1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
|
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or http://www.opensolaris.org/os/licensing.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 1993, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2012 DEY Storage Systems, Inc. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc.
* Copyright (c) 2018 Joyent, Inc.
* Copyright (c) 2015 by Delphix. All rights reserved.
* Copyright 2020 Oxide Computer Company
*/
/*
* Copyright (c) 2010, Intel Corporation.
* All rights reserved.
*/
#include <sys/types.h>
#include <sys/t_lock.h>
#include <sys/param.h>
#include <sys/sysmacros.h>
#include <sys/signal.h>
#include <sys/systm.h>
#include <sys/user.h>
#include <sys/mman.h>
#include <sys/vm.h>
#include <sys/conf.h>
#include <sys/avintr.h>
#include <sys/autoconf.h>
#include <sys/disp.h>
#include <sys/class.h>
#include <sys/bitmap.h>
#include <sys/privregs.h>
#include <sys/proc.h>
#include <sys/buf.h>
#include <sys/kmem.h>
#include <sys/mem.h>
#include <sys/kstat.h>
#include <sys/reboot.h>
#include <sys/cred.h>
#include <sys/vnode.h>
#include <sys/file.h>
#include <sys/procfs.h>
#include <sys/vfs.h>
#include <sys/cmn_err.h>
#include <sys/utsname.h>
#include <sys/debug.h>
#include <sys/kdi.h>
#include <sys/dumphdr.h>
#include <sys/bootconf.h>
#include <sys/memlist_plat.h>
#include <sys/varargs.h>
#include <sys/promif.h>
#include <sys/prom_debug.h>
#include <sys/modctl.h>
#include <sys/sunddi.h>
#include <sys/sunndi.h>
#include <sys/ndi_impldefs.h>
#include <sys/ddidmareq.h>
#include <sys/psw.h>
#include <sys/regset.h>
#include <sys/clock.h>
#include <sys/pte.h>
#include <sys/tss.h>
#include <sys/stack.h>
#include <sys/trap.h>
#include <sys/fp.h>
#include <vm/kboot_mmu.h>
#include <vm/anon.h>
#include <vm/as.h>
#include <vm/page.h>
#include <vm/seg.h>
#include <vm/seg_dev.h>
#include <vm/seg_kmem.h>
#include <vm/seg_kpm.h>
#include <vm/seg_map.h>
#include <vm/seg_vn.h>
#include <vm/seg_kp.h>
#include <sys/memnode.h>
#include <vm/vm_dep.h>
#include <sys/thread.h>
#include <sys/sysconf.h>
#include <sys/vm_machparam.h>
#include <sys/archsystm.h>
#include <sys/machsystm.h>
#include <vm/hat.h>
#include <vm/hat_i86.h>
#include <sys/pmem.h>
#include <sys/smp_impldefs.h>
#include <sys/x86_archext.h>
#include <sys/cpuvar.h>
#include <sys/segments.h>
#include <sys/clconf.h>
#include <sys/kobj.h>
#include <sys/kobj_lex.h>
#include <sys/cpc_impl.h>
#include <sys/cpu_module.h>
#include <sys/smbios.h>
#include <sys/debug_info.h>
#include <sys/bootinfo.h>
#include <sys/ddi_periodic.h>
#include <sys/systeminfo.h>
#include <sys/multiboot.h>
#include <sys/ramdisk.h>
#ifdef __xpv
#include <sys/hypervisor.h>
#include <sys/xen_mmu.h>
#include <sys/evtchn_impl.h>
#include <sys/gnttab.h>
#include <sys/xpv_panic.h>
#include <xen/sys/xenbus_comms.h>
#include <xen/public/physdev.h>
extern void xen_late_startup(void);
struct xen_evt_data cpu0_evt_data;
#else /* __xpv */
#include <sys/memlist_impl.h>
extern void mem_config_init(void);
#endif /* __xpv */
extern void progressbar_init(void);
extern void brand_init(void);
extern void pcf_init(void);
extern void pg_init(void);
extern void ssp_init(void);
extern int size_pse_array(pgcnt_t, int);
#if defined(_SOFT_HOSTID)
#include <sys/rtc.h>
static int32_t set_soft_hostid(void);
static char hostid_file[] = "/etc/hostid";
#endif
void *gfx_devinfo_list;
#if defined(__amd64) && !defined(__xpv)
extern void immu_startup(void);
#endif
/*
* XXX make declaration below "static" when drivers no longer use this
* interface.
*/
extern caddr_t p0_va; /* Virtual address for accessing physical page 0 */
/*
* segkp
*/
extern int segkp_fromheap;
static void kvm_init(void);
static void startup_init(void);
static void startup_memlist(void);
static void startup_kmem(void);
static void startup_modules(void);
static void startup_vm(void);
static void startup_end(void);
static void layout_kernel_va(void);
/*
* Declare these as initialized data so we can patch them.
*/
#ifdef __i386
/*
* Due to virtual address space limitations running in 32 bit mode, restrict
* the amount of physical memory configured to a max of PHYSMEM pages (16g).
*
* If the physical max memory size of 64g were allowed to be configured, the
* size of user virtual address space will be less than 1g. A limited user
* address space greatly reduces the range of applications that can run.
*
* If more physical memory than PHYSMEM is required, users should preferably
* run in 64 bit mode which has far looser virtual address space limitations.
*
* If 64 bit mode is not available (as in IA32) and/or more physical memory
* than PHYSMEM is required in 32 bit mode, physmem can be set to the desired
* value or to 0 (to configure all available memory) via eeprom(1M). kernelbase
* should also be carefully tuned to balance out the need of the user
* application while minimizing the risk of kernel heap exhaustion due to
* kernelbase being set too high.
*/
#define PHYSMEM 0x400000
#else /* __amd64 */
/*
* For now we can handle memory with physical addresses up to about
* 64 Terabytes. This keeps the kernel above the VA hole, leaving roughly
* half the VA space for seg_kpm. When systems get bigger than 64TB this
* code will need revisiting. There is an implicit assumption that there
* are no *huge* holes in the physical address space too.
*/
#define TERABYTE (1ul << 40)
#define PHYSMEM_MAX64 mmu_btop(64 * TERABYTE)
#define PHYSMEM PHYSMEM_MAX64
#define AMD64_VA_HOLE_END 0xFFFF800000000000ul
#endif /* __amd64 */
pgcnt_t physmem = PHYSMEM;
pgcnt_t obp_pages; /* Memory used by PROM for its text and data */
char *kobj_file_buf;
int kobj_file_bufsize; /* set in /etc/system */
/* Global variables for MP support. Used in mp_startup */
caddr_t rm_platter_va = 0;
uint32_t rm_platter_pa;
int auto_lpg_disable = 1;
/*
* Some CPUs have holes in the middle of the 64-bit virtual address range.
*/
uintptr_t hole_start, hole_end;
/*
* kpm mapping window
*/
caddr_t kpm_vbase;
size_t kpm_size;
static int kpm_desired;
#ifdef __amd64
static uintptr_t segkpm_base = (uintptr_t)SEGKPM_BASE;
#endif
/*
* Configuration parameters set at boot time.
*/
caddr_t econtig; /* end of first block of contiguous kernel */
struct bootops *bootops = 0; /* passed in from boot */
struct bootops **bootopsp;
struct boot_syscalls *sysp; /* passed in from boot */
char bootblock_fstype[16];
char kern_bootargs[OBP_MAXPATHLEN];
char kern_bootfile[OBP_MAXPATHLEN];
/*
* ZFS zio segment. This allows us to exclude large portions of ZFS data that
* gets cached in kmem caches on the heap. If this is set to zero, we allocate
* zio buffers from their own segment, otherwise they are allocated from the
* heap. The optimization of allocating zio buffers from their own segment is
* only valid on 64-bit kernels.
*/
#if defined(__amd64)
int segzio_fromheap = 0;
#else
int segzio_fromheap = 1;
#endif
/*
* Give folks an escape hatch for disabling SMAP via kmdb. Doesn't work
* post-boot.
*/
int disable_smap = 0;
/*
* new memory fragmentations are possible in startup() due to BOP_ALLOCs. this
* depends on number of BOP_ALLOC calls made and requested size, memory size
* combination and whether boot.bin memory needs to be freed.
*/
#define POSS_NEW_FRAGMENTS 12
/*
* VM data structures
*/
long page_hashsz; /* Size of page hash table (power of two) */
unsigned int page_hashsz_shift; /* log2(page_hashsz) */
struct page *pp_base; /* Base of initial system page struct array */
struct page **page_hash; /* Page hash table */
pad_mutex_t *pse_mutex; /* Locks protecting pp->p_selock */
size_t pse_table_size; /* Number of mutexes in pse_mutex[] */
int pse_shift; /* log2(pse_table_size) */
struct seg ktextseg; /* Segment used for kernel executable image */
struct seg kvalloc; /* Segment used for "valloc" mapping */
struct seg kpseg; /* Segment used for pageable kernel virt mem */
struct seg kmapseg; /* Segment used for generic kernel mappings */
struct seg kdebugseg; /* Segment used for the kernel debugger */
struct seg *segkmap = &kmapseg; /* Kernel generic mapping segment */
static struct seg *segmap = &kmapseg; /* easier to use name for in here */
struct seg *segkp = &kpseg; /* Pageable kernel virtual memory segment */
struct seg kvseg_core; /* Segment used for the core heap */
struct seg kpmseg; /* Segment used for physical mapping */
struct seg *segkpm = &kpmseg; /* 64bit kernel physical mapping segment */
caddr_t segkp_base; /* Base address of segkp */
caddr_t segzio_base; /* Base address of segzio */
pgcnt_t segkpsize = btop(SEGKPDEFSIZE); /* size of segkp segment in pages */
caddr_t segkvmm_base;
pgcnt_t segkvmmsize;
pgcnt_t segziosize;
/*
* A static DR page_t VA map is reserved that can map the page structures
* for a domain's entire RA space. The pages that back this space are
* dynamically allocated and need not be physically contiguous. The DR
* map size is derived from KPM size.
* This mechanism isn't used by x86 yet, so just stubs here.
*/
int ppvm_enable = 0; /* Static virtual map for page structs */
page_t *ppvm_base = NULL; /* Base of page struct map */
pgcnt_t ppvm_size = 0; /* Size of page struct map */
/*
* VA range available to the debugger
*/
const caddr_t kdi_segdebugbase = (const caddr_t)SEGDEBUGBASE;
const size_t kdi_segdebugsize = SEGDEBUGSIZE;
struct memseg *memseg_base;
struct vnode unused_pages_vp;
#define FOURGB 0x100000000LL
struct memlist *memlist;
caddr_t s_text; /* start of kernel text segment */
caddr_t e_text; /* end of kernel text segment */
caddr_t s_data; /* start of kernel data segment */
caddr_t e_data; /* end of kernel data segment */
caddr_t modtext; /* start of loadable module text reserved */
caddr_t e_modtext; /* end of loadable module text reserved */
caddr_t moddata; /* start of loadable module data reserved */
caddr_t e_moddata; /* end of loadable module data reserved */
struct memlist *phys_install; /* Total installed physical memory */
struct memlist *phys_avail; /* Total available physical memory */
struct memlist *bios_rsvd; /* Bios reserved memory */
/*
* kphysm_init returns the number of pages that were processed
*/
static pgcnt_t kphysm_init(page_t *, pgcnt_t);
#define IO_PROP_SIZE 64 /* device property size */
/*
* a couple useful roundup macros
*/
#define ROUND_UP_PAGE(x) \
((uintptr_t)P2ROUNDUP((uintptr_t)(x), (uintptr_t)MMU_PAGESIZE))
#define ROUND_UP_LPAGE(x) \
((uintptr_t)P2ROUNDUP((uintptr_t)(x), mmu.level_size[1]))
#define ROUND_UP_4MEG(x) \
((uintptr_t)P2ROUNDUP((uintptr_t)(x), (uintptr_t)FOUR_MEG))
#define ROUND_UP_TOPLEVEL(x) \
((uintptr_t)P2ROUNDUP((uintptr_t)(x), mmu.level_size[mmu.max_level]))
/*
* 32-bit Kernel's Virtual memory layout.
* +-----------------------+
* | |
* 0xFFC00000 -|-----------------------|- ARGSBASE
* | debugger |
* 0xFF800000 -|-----------------------|- SEGDEBUGBASE
* | Kernel Data |
* 0xFEC00000 -|-----------------------|
* | Kernel Text |
* 0xFE800000 -|-----------------------|- KERNEL_TEXT (0xFB400000 on Xen)
* |--- GDT ---|- GDT page (GDT_VA)
* |--- debug info ---|- debug info (DEBUG_INFO_VA)
* | |
* | page_t structures |
* | memsegs, memlists, |
* | page hash, etc. |
* --- -|-----------------------|- ekernelheap, valloc_base (floating)
* | | (segkp is just an arena in the heap)
* | |
* | kvseg |
* | |
* | |
* --- -|-----------------------|- kernelheap (floating)
* | Segkmap |
* 0xC3002000 -|-----------------------|- segmap_start (floating)
* | Red Zone |
* 0xC3000000 -|-----------------------|- kernelbase / userlimit (floating)
* | | ||
* | Shared objects | \/
* | |
* : :
* | user data |
* |-----------------------|
* | user text |
* 0x08048000 -|-----------------------|
* | user stack |
* : :
* | invalid |
* 0x00000000 +-----------------------+
*
*
* 64-bit Kernel's Virtual memory layout. (assuming 64 bit app)
* +-----------------------+
* | |
* 0xFFFFFFFF.FFC00000 |-----------------------|- ARGSBASE
* | debugger (?) |
* 0xFFFFFFFF.FF800000 |-----------------------|- SEGDEBUGBASE
* | unused |
* +-----------------------+
* | Kernel Data |
* 0xFFFFFFFF.FBC00000 |-----------------------|
* | Kernel Text |
* 0xFFFFFFFF.FB800000 |-----------------------|- KERNEL_TEXT
* |--- debug info ---|- debug info (DEBUG_INFO_VA)
* |--- GDT ---|- GDT page (GDT_VA)
* |--- IDT ---|- IDT page (IDT_VA)
* |--- LDT ---|- LDT pages (LDT_VA)
* | |
* | Core heap | (used for loadable modules)
* 0xFFFFFFFF.C0000000 |-----------------------|- core_base / ekernelheap
* | Kernel |
* | heap |
* | |
* | |
* 0xFFFFFXXX.XXX00000 |-----------------------|- kernelheap (floating)
* | segmap |
* 0xFFFFFXXX.XXX00000 |-----------------------|- segmap_start (floating)
* | device mappings |
* 0xFFFFFXXX.XXX00000 |-----------------------|- toxic_addr (floating)
* | segzio |
* 0xFFFFFXXX.XXX00000 |-----------------------|- segzio_base (floating)
* | segkvmm |
* | |
* | |
* | |
* 0xFFFFFXXX.XXX00000 |-----------------------|- segkvmm_base (floating)
* | segkp |
* |-----------------------|- segkp_base (floating)
* | page_t structures | valloc_base + valloc_sz
* | memsegs, memlists, |
* | page hash, etc. |
* 0xFFFFFE00.00000000 |-----------------------|- valloc_base (lower if >256GB)
* | segkpm |
* | |
* 0xFFFFFD00.00000000 |-----------------------|- SEGKPM_BASE (lower if >256GB)
* | Red Zone |
* 0xFFFFFC80.00000000 |-----------------------|- KERNELBASE (lower if >256GB)
* 0xFFFFFC7F.FFE00000 |-----------------------|- USERLIMIT (lower if >256GB)
* | User stack |- User space memory
* | |
* | shared objects, etc | (grows downwards)
* : :
* | |
* 0xFFFF8000.00000000 |-----------------------|
* | |
* | VA Hole / unused |
* | |
* 0x00008000.00000000 |-----------------------|
* | |
* | |
* : :
* | user heap | (grows upwards)
* | |
* | user data |
* |-----------------------|
* | user text |
* 0x00000000.04000000 |-----------------------|
* | invalid |
* 0x00000000.00000000 +-----------------------+
*
* A 32 bit app on the 64 bit kernel sees the same layout as on the 32 bit
* kernel, except that userlimit is raised to 0xfe000000
*
* Floating values:
*
* valloc_base: start of the kernel's memory management/tracking data
* structures. This region contains page_t structures for
* physical memory, memsegs, memlists, and the page hash.
*
* core_base: start of the kernel's "core" heap area on 64-bit systems.
* This area is intended to be used for global data as well as for module
* text/data that does not fit into the nucleus pages. The core heap is
* restricted to a 2GB range, allowing every address within it to be
* accessed using rip-relative addressing
*
* ekernelheap: end of kernelheap and start of segmap.
*
* kernelheap: start of kernel heap. On 32-bit systems, this starts right
* above a red zone that separates the user's address space from the
* kernel's. On 64-bit systems, it sits above segkp and segkpm.
*
* segmap_start: start of segmap. The length of segmap can be modified
* through eeprom. The default length is 16MB on 32-bit systems and 64MB
* on 64-bit systems.
*
* kernelbase: On a 32-bit kernel the default value of 0xd4000000 will be
* decreased by 2X the size required for page_t. This allows the kernel
* heap to grow in size with physical memory. With sizeof(page_t) == 80
* bytes, the following shows the values of kernelbase and kernel heap
* sizes for different memory configurations (assuming default segmap and
* segkp sizes).
*
* mem size for kernelbase kernel heap
* size page_t's size
* ---- --------- ---------- -----------
* 1gb 0x01400000 0xd1800000 684MB
* 2gb 0x02800000 0xcf000000 704MB
* 4gb 0x05000000 0xca000000 744MB
* 6gb 0x07800000 0xc5000000 784MB
* 8gb 0x0a000000 0xc0000000 824MB
* 16gb 0x14000000 0xac000000 984MB
* 32gb 0x28000000 0x84000000 1304MB
* 64gb 0x50000000 0x34000000 1944MB (*)
*
* kernelbase is less than the abi minimum of 0xc0000000 for memory
* configurations above 8gb.
*
* (*) support for memory configurations above 32gb will require manual tuning
* of kernelbase to balance out the need of user applications.
*/
/* real-time-clock initialization parameters */
extern time_t process_rtc_config_file(void);
uintptr_t kernelbase;
uintptr_t postbootkernelbase; /* not set till boot loader is gone */
uintptr_t eprom_kernelbase;
size_t segmapsize;
uintptr_t segmap_start;
int segmapfreelists;
pgcnt_t npages;
pgcnt_t orig_npages;
size_t core_size; /* size of "core" heap */
uintptr_t core_base; /* base address of "core" heap */
/*
* List of bootstrap pages. We mark these as allocated in startup.
* release_bootstrap() will free them when we're completely done with
* the bootstrap.
*/
static page_t *bootpages;
/*
* boot time pages that have a vnode from the ramdisk will keep that forever.
*/
static page_t *rd_pages;
/*
* Lower 64K
*/
static page_t *lower_pages = NULL;
static int lower_pages_count = 0;
struct system_hardware system_hardware;
/*
* Enable some debugging messages concerning memory usage...
*/
static void
print_memlist(char *title, struct memlist *mp)
{
prom_printf("MEMLIST: %s:\n", title);
while (mp != NULL) {
prom_printf("\tAddress 0x%" PRIx64 ", size 0x%" PRIx64 "\n",
mp->ml_address, mp->ml_size);
mp = mp->ml_next;
}
}
/*
* XX64 need a comment here.. are these just default values, surely
* we read the "cpuid" type information to figure this out.
*/
int l2cache_sz = 0x80000;
int l2cache_linesz = 0x40;
int l2cache_assoc = 1;
static size_t textrepl_min_gb = 10;
/*
* on 64 bit we use a predifined VA range for mapping devices in the kernel
* on 32 bit the mappings are intermixed in the heap, so we use a bit map
*/
#ifdef __amd64
vmem_t *device_arena;
uintptr_t toxic_addr = (uintptr_t)NULL;
size_t toxic_size = 1024 * 1024 * 1024; /* Sparc uses 1 gig too */
#else /* __i386 */
ulong_t *toxic_bit_map; /* one bit for each 4k of VA in heap_arena */
size_t toxic_bit_map_len = 0; /* in bits */
#endif /* __i386 */
int prom_debug;
/*
* This structure is used to keep track of the intial allocations
* done in startup_memlist(). The value of NUM_ALLOCATIONS needs to
* be >= the number of ADD_TO_ALLOCATIONS() executed in the code.
*/
#define NUM_ALLOCATIONS 8
int num_allocations = 0;
struct {
void **al_ptr;
size_t al_size;
} allocations[NUM_ALLOCATIONS];
size_t valloc_sz = 0;
uintptr_t valloc_base;
#define ADD_TO_ALLOCATIONS(ptr, size) { \
size = ROUND_UP_PAGE(size); \
if (num_allocations == NUM_ALLOCATIONS) \
panic("too many ADD_TO_ALLOCATIONS()"); \
allocations[num_allocations].al_ptr = (void**)&ptr; \
allocations[num_allocations].al_size = size; \
valloc_sz += size; \
++num_allocations; \
}
/*
* Allocate all the initial memory needed by the page allocator.
*/
static void
perform_allocations(void)
{
caddr_t mem;
int i;
int valloc_align;
PRM_DEBUG(valloc_base);
PRM_DEBUG(valloc_sz);
valloc_align = mmu.level_size[mmu.max_page_level > 0];
mem = BOP_ALLOC(bootops, (caddr_t)valloc_base, valloc_sz, valloc_align);
if (mem != (caddr_t)valloc_base)
panic("BOP_ALLOC() failed");
bzero(mem, valloc_sz);
for (i = 0; i < num_allocations; ++i) {
*allocations[i].al_ptr = (void *)mem;
mem += allocations[i].al_size;
}
}
/*
* Set up and enable SMAP now before we start other CPUs, but after the kernel's
* VM has been set up so we can use hot_patch_kernel_text().
*
* We can only patch 1, 2, or 4 bytes, but not three bytes. So instead, we
* replace the four byte word at the patch point. See uts/intel/ia32/ml/copy.s
* for more information on what's going on here.
*/
static void
startup_smap(void)
{
int i;
uint32_t inst;
uint8_t *instp;
char sym[128];
struct modctl *modp;
extern int _smap_enable_patch_count;
extern int _smap_disable_patch_count;
if (disable_smap != 0)
remove_x86_feature(x86_featureset, X86FSET_SMAP);
if (is_x86_feature(x86_featureset, X86FSET_SMAP) == B_FALSE)
return;
for (i = 0; i < _smap_enable_patch_count; i++) {
int sizep;
VERIFY3U(i, <, _smap_enable_patch_count);
VERIFY(snprintf(sym, sizeof (sym), "_smap_enable_patch_%d", i) <
sizeof (sym));
instp = (uint8_t *)(void *)kobj_getelfsym(sym, NULL, &sizep);
VERIFY(instp != 0);
inst = (instp[3] << 24) | (SMAP_CLAC_INSTR & 0x00ffffff);
hot_patch_kernel_text((caddr_t)instp, inst, 4);
}
for (i = 0; i < _smap_disable_patch_count; i++) {
int sizep;
VERIFY(snprintf(sym, sizeof (sym), "_smap_disable_patch_%d",
i) < sizeof (sym));
instp = (uint8_t *)(void *)kobj_getelfsym(sym, NULL, &sizep);
VERIFY(instp != 0);
inst = (instp[3] << 24) | (SMAP_STAC_INSTR & 0x00ffffff);
hot_patch_kernel_text((caddr_t)instp, inst, 4);
}
/*
* Hotinline calls to smap_enable and smap_disable within
* unix module. Hotinlines in other modules are done on
* mod_load().
*/
modp = mod_hold_by_name("unix");
do_hotinlines(modp->mod_mp);
mod_release_mod(modp);
setcr4(getcr4() | CR4_SMAP);
smap_enable();
}
/*
* Our world looks like this at startup time.
*
* In a 32-bit OS, boot loads the kernel text at 0xfe800000 and kernel data
* at 0xfec00000. On a 64-bit OS, kernel text and data are loaded at
* 0xffffffff.fe800000 and 0xffffffff.fec00000 respectively. Those
* addresses are fixed in the binary at link time.
*
* On the text page:
* unix/genunix/krtld/module text loads.
*
* On the data page:
* unix/genunix/krtld/module data loads.
*
* Machine-dependent startup code
*/
void
startup(void)
{
#if !defined(__xpv)
extern void startup_pci_bios(void);
#endif
extern cpuset_t cpu_ready_set;
/*
* Make sure that nobody tries to use sekpm until we have
* initialized it properly.
*/
#if defined(__amd64)
kpm_desired = 1;
#endif
kpm_enable = 0;
CPUSET_ONLY(cpu_ready_set, 0); /* cpu 0 is boot cpu */
#if defined(__xpv) /* XXPV fix me! */
{
extern int segvn_use_regions;
segvn_use_regions = 0;
}
#endif
ssp_init();
progressbar_init();
startup_init();
#if defined(__xpv)
startup_xen_version();
#endif
startup_memlist();
startup_kmem();
startup_vm();
#if !defined(__xpv)
/*
* Note we need to do this even on fast reboot in order to access
* the irq routing table (used for pci labels).
*/
startup_pci_bios();
startup_smap();
#endif
#if defined(__xpv)
startup_xen_mca();
#endif
startup_modules();
startup_end();
}
static void
startup_init()
{
PRM_POINT("startup_init() starting...");
/*
* Complete the extraction of cpuid data
*/
cpuid_pass2(CPU);
(void) check_boot_version(BOP_GETVERSION(bootops));
/*
* Check for prom_debug in boot environment
*/
if (BOP_GETPROPLEN(bootops, "prom_debug") >= 0) {
++prom_debug;
PRM_POINT("prom_debug found in boot enviroment");
}
/*
* Collect node, cpu and memory configuration information.
*/
get_system_configuration();
/*
* Halt if this is an unsupported processor.
*/
if (x86_type == X86_TYPE_486 || x86_type == X86_TYPE_CYRIX_486) {
printf("\n486 processor (\"%s\") detected.\n",
CPU->cpu_brandstr);
halt("This processor is not supported by this release "
"of Solaris.");
}
PRM_POINT("startup_init() done");
}
/*
* Callback for copy_memlist_filter() to filter nucleus, kadb/kmdb, (ie.
* everything mapped above KERNEL_TEXT) pages from phys_avail. Note it
* also filters out physical page zero. There is some reliance on the
* boot loader allocating only a few contiguous physical memory chunks.
*/
static void
avail_filter(uint64_t *addr, uint64_t *size)
{
uintptr_t va;
uintptr_t next_va;
pfn_t pfn;
uint64_t pfn_addr;
uint64_t pfn_eaddr;
uint_t prot;
size_t len;
uint_t change;
if (prom_debug)
prom_printf("\tFilter: in: a=%" PRIx64 ", s=%" PRIx64 "\n",
*addr, *size);
/*
* page zero is required for BIOS.. never make it available
*/
if (*addr == 0) {
*addr += MMU_PAGESIZE;
*size -= MMU_PAGESIZE;
}
/*
* First we trim from the front of the range. Since kbm_probe()
* walks ranges in virtual order, but addr/size are physical, we need
* to the list until no changes are seen. This deals with the case
* where page "p" is mapped at v, page "p + PAGESIZE" is mapped at w
* but w < v.
*/
do {
change = 0;
for (va = KERNEL_TEXT;
*size > 0 && kbm_probe(&va, &len, &pfn, &prot) != 0;
va = next_va) {
next_va = va + len;
pfn_addr = pfn_to_pa(pfn);
pfn_eaddr = pfn_addr + len;
if (pfn_addr <= *addr && pfn_eaddr > *addr) {
change = 1;
while (*size > 0 && len > 0) {
*addr += MMU_PAGESIZE;
*size -= MMU_PAGESIZE;
len -= MMU_PAGESIZE;
}
}
}
if (change && prom_debug)
prom_printf("\t\ttrim: a=%" PRIx64 ", s=%" PRIx64 "\n",
*addr, *size);
} while (change);
/*
* Trim pages from the end of the range.
*/
for (va = KERNEL_TEXT;
*size > 0 && kbm_probe(&va, &len, &pfn, &prot) != 0;
va = next_va) {
next_va = va + len;
pfn_addr = pfn_to_pa(pfn);
if (pfn_addr >= *addr && pfn_addr < *addr + *size)
*size = pfn_addr - *addr;
}
if (prom_debug)
prom_printf("\tFilter out: a=%" PRIx64 ", s=%" PRIx64 "\n",
*addr, *size);
}
static void
kpm_init()
{
struct segkpm_crargs b;
/*
* These variables were all designed for sfmmu in which segkpm is
* mapped using a single pagesize - either 8KB or 4MB. On x86, we
* might use 2+ page sizes on a single machine, so none of these
* variables have a single correct value. They are set up as if we
* always use a 4KB pagesize, which should do no harm. In the long
* run, we should get rid of KPM's assumption that only a single
* pagesize is used.
*/
kpm_pgshft = MMU_PAGESHIFT;
kpm_pgsz = MMU_PAGESIZE;
kpm_pgoff = MMU_PAGEOFFSET;
kpmp2pshft = 0;
kpmpnpgs = 1;
ASSERT(((uintptr_t)kpm_vbase & (kpm_pgsz - 1)) == 0);
PRM_POINT("about to create segkpm");
rw_enter(&kas.a_lock, RW_WRITER);
if (seg_attach(&kas, kpm_vbase, kpm_size, segkpm) < 0)
panic("cannot attach segkpm");
b.prot = PROT_READ | PROT_WRITE;
b.nvcolors = 1;
if (segkpm_create(segkpm, (caddr_t)&b) != 0)
panic("segkpm_create segkpm");
rw_exit(&kas.a_lock);
kpm_enable = 1;
/*
* As the KPM was disabled while setting up the system, go back and fix
* CPU zero's access to its user page table. This is a bit gross, but
* we have a chicken and egg problem otherwise.
*/
ASSERT(CPU->cpu_hat_info->hci_user_l3ptes == NULL);
CPU->cpu_hat_info->hci_user_l3ptes =
(x86pte_t *)hat_kpm_mapin_pfn(CPU->cpu_hat_info->hci_user_l3pfn);
}
/*
* The debug info page provides enough information to allow external
* inspectors (e.g. when running under a hypervisor) to bootstrap
* themselves into allowing full-blown kernel debugging.
*/
static void
init_debug_info(void)
{
caddr_t mem;
debug_info_t *di;
#ifndef __lint
ASSERT(sizeof (debug_info_t) < MMU_PAGESIZE);
#endif
mem = BOP_ALLOC(bootops, (caddr_t)DEBUG_INFO_VA, MMU_PAGESIZE,
MMU_PAGESIZE);
if (mem != (caddr_t)DEBUG_INFO_VA)
panic("BOP_ALLOC() failed");
bzero(mem, MMU_PAGESIZE);
di = (debug_info_t *)mem;
di->di_magic = DEBUG_INFO_MAGIC;
di->di_version = DEBUG_INFO_VERSION;
di->di_modules = (uintptr_t)&modules;
di->di_s_text = (uintptr_t)s_text;
di->di_e_text = (uintptr_t)e_text;
di->di_s_data = (uintptr_t)s_data;
di->di_e_data = (uintptr_t)e_data;
di->di_hat_htable_off = offsetof(hat_t, hat_htable);
di->di_ht_pfn_off = offsetof(htable_t, ht_pfn);
}
/*
* Build the memlists and other kernel essential memory system data structures.
* This is everything at valloc_base.
*/
static void
startup_memlist(void)
{
size_t memlist_sz;
size_t memseg_sz;
size_t pagehash_sz;
size_t pp_sz;
uintptr_t va;
size_t len;
uint_t prot;
pfn_t pfn;
int memblocks;
pfn_t rsvd_high_pfn;
pgcnt_t rsvd_pgcnt;
size_t rsvdmemlist_sz;
int rsvdmemblocks;
caddr_t pagecolor_mem;
size_t pagecolor_memsz;
caddr_t page_ctrs_mem;
size_t page_ctrs_size;
size_t pse_table_alloc_size;
struct memlist *current;
extern void startup_build_mem_nodes(struct memlist *);
/* XX64 fix these - they should be in include files */
extern size_t page_coloring_init(uint_t, int, int);
extern void page_coloring_setup(caddr_t);
PRM_POINT("startup_memlist() starting...");
/*
* Use leftover large page nucleus text/data space for loadable modules.
* Use at most MODTEXT/MODDATA.
*/
len = kbm_nucleus_size;
ASSERT(len > MMU_PAGESIZE);
moddata = (caddr_t)ROUND_UP_PAGE(e_data);
e_moddata = (caddr_t)P2ROUNDUP((uintptr_t)e_data, (uintptr_t)len);
if (e_moddata - moddata > MODDATA)
e_moddata = moddata + MODDATA;
modtext = (caddr_t)ROUND_UP_PAGE(e_text);
e_modtext = (caddr_t)P2ROUNDUP((uintptr_t)e_text, (uintptr_t)len);
if (e_modtext - modtext > MODTEXT)
e_modtext = modtext + MODTEXT;
econtig = e_moddata;
PRM_DEBUG(modtext);
PRM_DEBUG(e_modtext);
PRM_DEBUG(moddata);
PRM_DEBUG(e_moddata);
PRM_DEBUG(econtig);
/*
* Examine the boot loader physical memory map to find out:
* - total memory in system - physinstalled
* - the max physical address - physmax
* - the number of discontiguous segments of memory.
*/
if (prom_debug)
print_memlist("boot physinstalled",
bootops->boot_mem->physinstalled);
installed_top_size_ex(bootops->boot_mem->physinstalled, &physmax,
&physinstalled, &memblocks);
PRM_DEBUG(physmax);
PRM_DEBUG(physinstalled);
PRM_DEBUG(memblocks);
/*
* We no longer support any form of memory DR.
*/
plat_dr_physmax = 0;
/*
* Examine the bios reserved memory to find out:
* - the number of discontiguous segments of memory.
*/
if (prom_debug)
print_memlist("boot reserved mem",
bootops->boot_mem->rsvdmem);
installed_top_size_ex(bootops->boot_mem->rsvdmem, &rsvd_high_pfn,
&rsvd_pgcnt, &rsvdmemblocks);
PRM_DEBUG(rsvd_high_pfn);
PRM_DEBUG(rsvd_pgcnt);
PRM_DEBUG(rsvdmemblocks);
/*
* Initialize hat's mmu parameters.
* Check for enforce-prot-exec in boot environment. It's used to
* enable/disable support for the page table entry NX bit.
* The default is to enforce PROT_EXEC on processors that support NX.
* Boot seems to round up the "len", but 8 seems to be big enough.
*/
mmu_init();
#ifdef __i386
/*
* physmax is lowered if there is more memory than can be
* physically addressed in 32 bit (PAE/non-PAE) modes.
*/
if (mmu.pae_hat) {
if (PFN_ABOVE64G(physmax)) {
physinstalled -= (physmax - (PFN_64G - 1));
physmax = PFN_64G - 1;
}
} else {
if (PFN_ABOVE4G(physmax)) {
physinstalled -= (physmax - (PFN_4G - 1));
physmax = PFN_4G - 1;
}
}
#endif
startup_build_mem_nodes(bootops->boot_mem->physinstalled);
if (BOP_GETPROPLEN(bootops, "enforce-prot-exec") >= 0) {
int len = BOP_GETPROPLEN(bootops, "enforce-prot-exec");
char value[8];
if (len < 8)
(void) BOP_GETPROP(bootops, "enforce-prot-exec", value);
else
(void) strcpy(value, "");
if (strcmp(value, "off") == 0)
mmu.pt_nx = 0;
}
PRM_DEBUG(mmu.pt_nx);
/*
* We will need page_t's for every page in the system, except for
* memory mapped at or above above the start of the kernel text segment.
*
* pages above e_modtext are attributed to kernel debugger (obp_pages)
*/
npages = physinstalled - 1; /* avail_filter() skips page 0, so "- 1" */
obp_pages = 0;
va = KERNEL_TEXT;
while (kbm_probe(&va, &len, &pfn, &prot) != 0) {
npages -= len >> MMU_PAGESHIFT;
if (va >= (uintptr_t)e_moddata)
obp_pages += len >> MMU_PAGESHIFT;
va += len;
}
PRM_DEBUG(npages);
PRM_DEBUG(obp_pages);
/*
* If physmem is patched to be non-zero, use it instead of the computed
* value unless it is larger than the actual amount of memory on hand.
*/
if (physmem == 0 || physmem > npages) {
physmem = npages;
} else if (physmem < npages) {
orig_npages = npages;
npages = physmem;
}
PRM_DEBUG(physmem);
/*
* We now compute the sizes of all the initial allocations for
* structures the kernel needs in order do kmem_alloc(). These
* include:
* memsegs
* memlists
* page hash table
* page_t's
* page coloring data structs
*/
memseg_sz = sizeof (struct memseg) * (memblocks + POSS_NEW_FRAGMENTS);
ADD_TO_ALLOCATIONS(memseg_base, memseg_sz);
PRM_DEBUG(memseg_sz);
/*
* Reserve space for memlists. There's no real good way to know exactly
* how much room we'll need, but this should be a good upper bound.
*/
memlist_sz = ROUND_UP_PAGE(2 * sizeof (struct memlist) *
(memblocks + POSS_NEW_FRAGMENTS));
ADD_TO_ALLOCATIONS(memlist, memlist_sz);
PRM_DEBUG(memlist_sz);
/*
* Reserve space for bios reserved memlists.
*/
rsvdmemlist_sz = ROUND_UP_PAGE(2 * sizeof (struct memlist) *
(rsvdmemblocks + POSS_NEW_FRAGMENTS));
ADD_TO_ALLOCATIONS(bios_rsvd, rsvdmemlist_sz);
PRM_DEBUG(rsvdmemlist_sz);
/* LINTED */
ASSERT(P2SAMEHIGHBIT((1 << PP_SHIFT), sizeof (struct page)));
/*
* The page structure hash table size is a power of 2
* such that the average hash chain length is PAGE_HASHAVELEN.
*/
page_hashsz = npages / PAGE_HASHAVELEN;
page_hashsz_shift = highbit(page_hashsz);
page_hashsz = 1 << page_hashsz_shift;
pagehash_sz = sizeof (struct page *) * page_hashsz;
ADD_TO_ALLOCATIONS(page_hash, pagehash_sz);
PRM_DEBUG(pagehash_sz);
/*
* Set aside room for the page structures themselves.
*/
PRM_DEBUG(npages);
pp_sz = sizeof (struct page) * npages;
ADD_TO_ALLOCATIONS(pp_base, pp_sz);
PRM_DEBUG(pp_sz);
/*
* determine l2 cache info and memory size for page coloring
*/
(void) getl2cacheinfo(CPU,
&l2cache_sz, &l2cache_linesz, &l2cache_assoc);
pagecolor_memsz =
page_coloring_init(l2cache_sz, l2cache_linesz, l2cache_assoc);
ADD_TO_ALLOCATIONS(pagecolor_mem, pagecolor_memsz);
PRM_DEBUG(pagecolor_memsz);
page_ctrs_size = page_ctrs_sz();
ADD_TO_ALLOCATIONS(page_ctrs_mem, page_ctrs_size);
PRM_DEBUG(page_ctrs_size);
/*
* Allocate the array that protects pp->p_selock.
*/
pse_shift = size_pse_array(physmem, max_ncpus);
pse_table_size = 1 << pse_shift;
pse_table_alloc_size = pse_table_size * sizeof (pad_mutex_t);
ADD_TO_ALLOCATIONS(pse_mutex, pse_table_alloc_size);
valloc_sz = ROUND_UP_LPAGE(valloc_sz);
valloc_base = VALLOC_BASE;
/*
* The signicant memory-sized regions are roughly sized as follows in
* the default layout with max physmem:
* segkpm: 1x physmem allocated (but 1Tb room, below VALLOC_BASE)
* segzio: 1.5x physmem
* segkvmm: 4x physmem
* heap: whatever's left up to COREHEAP_BASE, at least 1.5x physmem
*
* The idea is that we leave enough room to avoid fragmentation issues,
* so we would like the VA arenas to have some extra.
*
* Ignoring the loose change of segkp, valloc, and such, this means that
* as COREHEAP_BASE-VALLOC_BASE=2Tb, we can accommodate a physmem up to
* about (2Tb / 7.0), rounded down to 256Gb in the check below.
*
* Note that KPM lives below VALLOC_BASE, but we want to include it in
* adjustments, hence the 8 below.
*
* Beyond 256Gb, we push segkpm_base (and hence kernelbase and
* _userlimit) down to accommodate the VA requirements above.
*/
if (physmax + 1 > mmu_btop(TERABYTE / 4)) {
uint64_t physmem_bytes = mmu_ptob(physmax + 1);
uint64_t adjustment = 8 * (physmem_bytes - (TERABYTE / 4));
PRM_DEBUG(adjustment);
/*
* segkpm_base is always aligned on a L3 PTE boundary.
*/
segkpm_base -= P2ROUNDUP(adjustment, KERNEL_REDZONE_SIZE);
/*
* But make sure we leave some space for user apps above hole.
*/
segkpm_base = MAX(segkpm_base, AMD64_VA_HOLE_END + TERABYTE);
ASSERT(segkpm_base <= SEGKPM_BASE);
valloc_base = segkpm_base + P2ROUNDUP(physmem_bytes, ONE_GIG);
if (valloc_base < segkpm_base)
panic("not enough kernel VA to support memory size");
}
PRM_DEBUG(segkpm_base);
PRM_DEBUG(valloc_base);
/*
* do all the initial allocations
*/
perform_allocations();
/*
* Build phys_install and phys_avail in kernel memspace.
* - phys_install should be all memory in the system.
* - phys_avail is phys_install minus any memory mapped before this
* point above KERNEL_TEXT.
*/
current = phys_install = memlist;
copy_memlist_filter(bootops->boot_mem->physinstalled, ¤t, NULL);
if ((caddr_t)current > (caddr_t)memlist + memlist_sz)
panic("physinstalled was too big!");
if (prom_debug)
print_memlist("phys_install", phys_install);
phys_avail = current;
PRM_POINT("Building phys_avail:\n");
copy_memlist_filter(bootops->boot_mem->physinstalled, ¤t,
avail_filter);
if ((caddr_t)current > (caddr_t)memlist + memlist_sz)
panic("physavail was too big!");
if (prom_debug)
print_memlist("phys_avail", phys_avail);
#ifndef __xpv
/*
* Free unused memlist items, which may be used by memory DR driver
* at runtime.
*/
if ((caddr_t)current < (caddr_t)memlist + memlist_sz) {
memlist_free_block((caddr_t)current,
(caddr_t)memlist + memlist_sz - (caddr_t)current);
}
#endif
/*
* Build bios reserved memspace
*/
current = bios_rsvd;
copy_memlist_filter(bootops->boot_mem->rsvdmem, ¤t, NULL);
if ((caddr_t)current > (caddr_t)bios_rsvd + rsvdmemlist_sz)
panic("bios_rsvd was too big!");
if (prom_debug)
print_memlist("bios_rsvd", bios_rsvd);
#ifndef __xpv
/*
* Free unused memlist items, which may be used by memory DR driver
* at runtime.
*/
if ((caddr_t)current < (caddr_t)bios_rsvd + rsvdmemlist_sz) {
memlist_free_block((caddr_t)current,
(caddr_t)bios_rsvd + rsvdmemlist_sz - (caddr_t)current);
}
#endif
/*
* setup page coloring
*/
page_coloring_setup(pagecolor_mem);
page_lock_init(); /* currently a no-op */
/*
* free page list counters
*/
(void) page_ctrs_alloc(page_ctrs_mem);
/*
* Size the pcf array based on the number of cpus in the box at
* boot time.
*/
pcf_init();
/*
* Initialize the page structures from the memory lists.
*/
availrmem_initial = availrmem = freemem = 0;
PRM_POINT("Calling kphysm_init()...");
npages = kphysm_init(pp_base, npages);
PRM_POINT("kphysm_init() done");
PRM_DEBUG(npages);
init_debug_info();
/*
* Now that page_t's have been initialized, remove all the
* initial allocation pages from the kernel free page lists.
*/
boot_mapin((caddr_t)valloc_base, valloc_sz);
boot_mapin((caddr_t)MISC_VA_BASE, MISC_VA_SIZE);
PRM_POINT("startup_memlist() done");
PRM_DEBUG(valloc_sz);
#if defined(__amd64)
if ((availrmem >> (30 - MMU_PAGESHIFT)) >=
textrepl_min_gb && l2cache_sz <= 2 << 20) {
extern size_t textrepl_size_thresh;
textrepl_size_thresh = (16 << 20) - 1;
}
#endif
}
/*
* Layout the kernel's part of address space and initialize kmem allocator.
*/
static void
startup_kmem(void)
{
extern void page_set_colorequiv_arr(void);
#if !defined(__xpv)
extern uint64_t kpti_kbase;
#endif
PRM_POINT("startup_kmem() starting...");
#if defined(__amd64)
if (eprom_kernelbase && eprom_kernelbase != KERNELBASE)
cmn_err(CE_NOTE, "!kernelbase cannot be changed on 64-bit "
"systems.");
kernelbase = segkpm_base - KERNEL_REDZONE_SIZE;
core_base = (uintptr_t)COREHEAP_BASE;
core_size = (size_t)MISC_VA_BASE - COREHEAP_BASE;
#else /* __i386 */
/*
* We configure kernelbase based on:
*
* 1. user specified kernelbase via eeprom command. Value cannot exceed
* KERNELBASE_MAX. we large page align eprom_kernelbase
*
* 2. Default to KERNELBASE and adjust to 2X less the size for page_t.
* On large memory systems we must lower kernelbase to allow
* enough room for page_t's for all of memory.
*
* The value set here, might be changed a little later.
*/
if (eprom_kernelbase) {
kernelbase = eprom_kernelbase & mmu.level_mask[1];
if (kernelbase > KERNELBASE_MAX)
kernelbase = KERNELBASE_MAX;
} else {
kernelbase = (uintptr_t)KERNELBASE;
kernelbase -= ROUND_UP_4MEG(2 * valloc_sz);
}
ASSERT((kernelbase & mmu.level_offset[1]) == 0);
core_base = valloc_base;
core_size = 0;
#endif /* __i386 */
PRM_DEBUG(core_base);
PRM_DEBUG(core_size);
PRM_DEBUG(kernelbase);
#if defined(__i386)
segkp_fromheap = 1;
#endif /* __i386 */
ekernelheap = (char *)core_base;
PRM_DEBUG(ekernelheap);
/*
* Now that we know the real value of kernelbase,
* update variables that were initialized with a value of
* KERNELBASE (in common/conf/param.c).
*
* XXX The problem with this sort of hackery is that the
* compiler just may feel like putting the const declarations
* (in param.c) into the .text section. Perhaps they should
* just be declared as variables there?
*/
*(uintptr_t *)&_kernelbase = kernelbase;
*(uintptr_t *)&_userlimit = kernelbase;
#if defined(__amd64)
*(uintptr_t *)&_userlimit -= KERNELBASE - USERLIMIT;
#if !defined(__xpv)
kpti_kbase = kernelbase;
#endif
#else
*(uintptr_t *)&_userlimit32 = _userlimit;
#endif
PRM_DEBUG(_kernelbase);
PRM_DEBUG(_userlimit);
PRM_DEBUG(_userlimit32);
/* We have to re-do this now that we've modified _userlimit. */
mmu_calc_user_slots();
layout_kernel_va();
#if defined(__i386)
/*
* If segmap is too large we can push the bottom of the kernel heap
* higher than the base. Or worse, it could exceed the top of the
* VA space entirely, causing it to wrap around.
*/
if (kernelheap >= ekernelheap || (uintptr_t)kernelheap < kernelbase)
panic("too little address space available for kernelheap,"
" use eeprom for lower kernelbase or smaller segmapsize");
#endif /* __i386 */
/*
* Initialize the kernel heap. Note 3rd argument must be > 1st.
*/
kernelheap_init(kernelheap, ekernelheap,
kernelheap + MMU_PAGESIZE,
(void *)core_base, (void *)(core_base + core_size));
#if defined(__xpv)
/*
* Link pending events struct into cpu struct
*/
CPU->cpu_m.mcpu_evt_pend = &cpu0_evt_data;
#endif
/*
* Initialize kernel memory allocator.
*/
kmem_init();
/*
* Factor in colorequiv to check additional 'equivalent' bins
*/
page_set_colorequiv_arr();
/*
* print this out early so that we know what's going on
*/
print_x86_featureset(x86_featureset);
/*
* Initialize bp_mapin().
*/
bp_init(MMU_PAGESIZE, HAT_STORECACHING_OK);
/*
* orig_npages is non-zero if physmem has been configured for less
* than the available memory.
*/
if (orig_npages) {
cmn_err(CE_WARN, "!%slimiting physmem to 0x%lx of 0x%lx pages",
(npages == PHYSMEM ? "Due to virtual address space " : ""),
npages, orig_npages);
}
#if defined(__i386)
if (eprom_kernelbase && (eprom_kernelbase != kernelbase))
cmn_err(CE_WARN, "kernelbase value, User specified 0x%lx, "
"System using 0x%lx",
(uintptr_t)eprom_kernelbase, (uintptr_t)kernelbase);
#endif
#ifdef KERNELBASE_ABI_MIN
if (kernelbase < (uintptr_t)KERNELBASE_ABI_MIN) {
cmn_err(CE_NOTE, "!kernelbase set to 0x%lx, system is not "
"i386 ABI compliant.", (uintptr_t)kernelbase);
}
#endif
#ifndef __xpv
if (plat_dr_support_memory()) {
mem_config_init();
}
#else /* __xpv */
/*
* Some of the xen start information has to be relocated up
* into the kernel's permanent address space.
*/
PRM_POINT("calling xen_relocate_start_info()");
xen_relocate_start_info();
PRM_POINT("xen_relocate_start_info() done");
/*
* (Update the vcpu pointer in our cpu structure to point into
* the relocated shared info.)
*/
CPU->cpu_m.mcpu_vcpu_info =
&HYPERVISOR_shared_info->vcpu_info[CPU->cpu_id];
#endif /* __xpv */
PRM_POINT("startup_kmem() done");
}
#ifndef __xpv
/*
* If we have detected that we are running in an HVM environment, we need
* to prepend the PV driver directory to the module search path.
*/
#define HVM_MOD_DIR "/platform/i86hvm/kernel"
static void
update_default_path()
{
char *current, *newpath;
int newlen;
/*
* We are about to resync with krtld. krtld will reset its
* internal module search path iff Solaris has set default_path.
* We want to be sure we're prepending this new directory to the
* right search path.
*/
current = (default_path == NULL) ? kobj_module_path : default_path;
newlen = strlen(HVM_MOD_DIR) + strlen(current) + 2;
newpath = kmem_alloc(newlen, KM_SLEEP);
(void) strcpy(newpath, HVM_MOD_DIR);
(void) strcat(newpath, " ");
(void) strcat(newpath, current);
default_path = newpath;
}
#endif
static void
startup_modules(void)
{
int cnt;
extern void prom_setup(void);
int32_t v, h;
char d[11];
char *cp;
cmi_hdl_t hdl;
PRM_POINT("startup_modules() starting...");
#ifndef __xpv
/*
* Initialize ten-micro second timer so that drivers will
* not get short changed in their init phase. This was
* not getting called until clkinit which, on fast cpu's
* caused the drv_usecwait to be way too short.
*/
microfind();
if ((get_hwenv() & HW_XEN_HVM) != 0)
update_default_path();
#endif
/*
* Read the GMT lag from /etc/rtc_config.
*/
sgmtl(process_rtc_config_file());
/*
* Calculate default settings of system parameters based upon
* maxusers, yet allow to be overridden via the /etc/system file.
*/
param_calc(0);
mod_setup();
/*
* Initialize system parameters.
*/
param_init();
/*
* Initialize the default brands
*/
brand_init();
/*
* maxmem is the amount of physical memory we're playing with.
*/
maxmem = physmem;
/*
* Initialize segment management stuff.
*/
seg_init();
if (modload("fs", "specfs") == -1)
halt("Can't load specfs");
if (modload("fs", "devfs") == -1)
halt("Can't load devfs");
if (modload("fs", "dev") == -1)
halt("Can't load dev");
if (modload("fs", "procfs") == -1)
halt("Can't load procfs");
(void) modloadonly("sys", "lbl_edition");
dispinit();
/* Read cluster configuration data. */
clconf_init();
#if defined(__xpv)
(void) ec_init();
gnttab_init();
(void) xs_early_init();
#endif /* __xpv */
/*
* Create a kernel device tree. First, create rootnex and
* then invoke bus specific code to probe devices.
*/
setup_ddi();
#ifdef __xpv
if (DOMAIN_IS_INITDOMAIN(xen_info))
#endif
{
id_t smid;
smbios_system_t smsys;
smbios_info_t sminfo;
char *mfg;
/*
* Load the System Management BIOS into the global ksmbios
* handle, if an SMBIOS is present on this system.
* Also set "si-hw-provider" property, if not already set.
*/
ksmbios = smbios_open(NULL, SMB_VERSION, ksmbios_flags, NULL);
if (ksmbios != NULL &&
((smid = smbios_info_system(ksmbios, &smsys)) != SMB_ERR) &&
(smbios_info_common(ksmbios, smid, &sminfo)) != SMB_ERR) {
mfg = (char *)sminfo.smbi_manufacturer;
if (BOP_GETPROPLEN(bootops, "si-hw-provider") < 0) {
extern char hw_provider[];
int i;
for (i = 0; i < SYS_NMLN; i++) {
if (isprint(mfg[i]))
hw_provider[i] = mfg[i];
else {
hw_provider[i] = '\0';
break;
}
}
hw_provider[SYS_NMLN - 1] = '\0';
}
}
}
/*
* Originally clconf_init() apparently needed the hostid. But
* this no longer appears to be true - it uses its own nodeid.
* By placing the hostid logic here, we are able to make use of
* the SMBIOS UUID.
*/
if ((h = set_soft_hostid()) == HW_INVALID_HOSTID) {
cmn_err(CE_WARN, "Unable to set hostid");
} else {
for (v = h, cnt = 0; cnt < 10; cnt++) {
d[cnt] = (char)(v % 10);
v /= 10;
if (v == 0)
break;
}
for (cp = hw_serial; cnt >= 0; cnt--)
*cp++ = d[cnt] + '0';
*cp = 0;
}
/*
* Set up the CPU module subsystem for the boot cpu in the native
* case, and all physical cpu resource in the xpv dom0 case.
* Modifies the device tree, so this must be done after
* setup_ddi().
*/
#ifdef __xpv
/*
* If paravirtualized and on dom0 then we initialize all physical
* cpu handles now; if paravirtualized on a domU then do not
* initialize.
*/
if (DOMAIN_IS_INITDOMAIN(xen_info)) {
xen_mc_lcpu_cookie_t cpi;
for (cpi = xen_physcpu_next(NULL); cpi != NULL;
cpi = xen_physcpu_next(cpi)) {
if ((hdl = cmi_init(CMI_HDL_SOLARIS_xVM_MCA,
xen_physcpu_chipid(cpi), xen_physcpu_coreid(cpi),
xen_physcpu_strandid(cpi))) != NULL &&
is_x86_feature(x86_featureset, X86FSET_MCA))
cmi_mca_init(hdl);
}
}
#else
/*
* Initialize a handle for the boot cpu - others will initialize
* as they startup.
*/
if ((hdl = cmi_init(CMI_HDL_NATIVE, cmi_ntv_hwchipid(CPU),
cmi_ntv_hwcoreid(CPU), cmi_ntv_hwstrandid(CPU))) != NULL) {
if (is_x86_feature(x86_featureset, X86FSET_MCA))
cmi_mca_init(hdl);
CPU->cpu_m.mcpu_cmi_hdl = hdl;
}
#endif /* __xpv */
/*
* Fake a prom tree such that /dev/openprom continues to work
*/
PRM_POINT("startup_modules: calling prom_setup...");
prom_setup();
PRM_POINT("startup_modules: done");
/*
* Load all platform specific modules
*/
PRM_POINT("startup_modules: calling psm_modload...");
psm_modload();
PRM_POINT("startup_modules() done");
}
/*
* claim a "setaside" boot page for use in the kernel
*/
page_t *
boot_claim_page(pfn_t pfn)
{
page_t *pp;
pp = page_numtopp_nolock(pfn);
ASSERT(pp != NULL);
if (PP_ISBOOTPAGES(pp)) {
if (pp->p_next != NULL)
pp->p_next->p_prev = pp->p_prev;
if (pp->p_prev == NULL)
bootpages = pp->p_next;
else
pp->p_prev->p_next = pp->p_next;
} else {
/*
* htable_attach() expects a base pagesize page
*/
if (pp->p_szc != 0)
page_boot_demote(pp);
pp = page_numtopp(pfn, SE_EXCL);
}
return (pp);
}
/*
* Walk through the pagetables looking for pages mapped in by boot. If the
* setaside flag is set the pages are expected to be returned to the
* kernel later in boot, so we add them to the bootpages list.
*/
static void
protect_boot_range(uintptr_t low, uintptr_t high, int setaside)
{
uintptr_t va = low;
size_t len;
uint_t prot;
pfn_t pfn;
page_t *pp;
pgcnt_t boot_protect_cnt = 0;
while (kbm_probe(&va, &len, &pfn, &prot) != 0 && va < high) {
if (va + len >= high)
panic("0x%lx byte mapping at 0x%p exceeds boot's "
"legal range.", len, (void *)va);
while (len > 0) {
pp = page_numtopp_alloc(pfn);
if (pp != NULL) {
if (setaside == 0)
panic("Unexpected mapping by boot. "
"addr=%p pfn=%lx\n",
(void *)va, pfn);
pp->p_next = bootpages;
pp->p_prev = NULL;
PP_SETBOOTPAGES(pp);
if (bootpages != NULL) {
bootpages->p_prev = pp;
}
bootpages = pp;
++boot_protect_cnt;
}
++pfn;
len -= MMU_PAGESIZE;
va += MMU_PAGESIZE;
}
}
PRM_DEBUG(boot_protect_cnt);
}
/*
* Establish the final size of the kernel's heap, size of segmap, segkp, etc.
*/
static void
layout_kernel_va(void)
{
const size_t physmem_size = mmu_ptob(physmem);
size_t size;
PRM_POINT("layout_kernel_va() starting...");
kpm_vbase = (caddr_t)segkpm_base;
kpm_size = ROUND_UP_LPAGE(mmu_ptob(physmax + 1));
if ((uintptr_t)kpm_vbase + kpm_size > (uintptr_t)valloc_base)
panic("not enough room for kpm!");
PRM_DEBUG(kpm_size);
PRM_DEBUG(kpm_vbase);
segkp_base = (caddr_t)valloc_base + valloc_sz;
if (!segkp_fromheap) {
size = mmu_ptob(segkpsize);
/*
* determine size of segkp
*/
if (size < SEGKPMINSIZE || size > SEGKPMAXSIZE) {
size = SEGKPDEFSIZE;
cmn_err(CE_WARN, "!Illegal value for segkpsize. "
"segkpsize has been reset to %ld pages",
mmu_btop(size));
}
size = MIN(size, MAX(SEGKPMINSIZE, physmem_size));
segkpsize = mmu_btop(ROUND_UP_LPAGE(size));
}
PRM_DEBUG(segkp_base);
PRM_DEBUG(segkpsize);
/*
* segkvmm: backing for vmm guest memory. Like segzio, we have a
* separate segment for two reasons: it makes it easy to skip our pages
* on kernel crash dumps, and it helps avoid fragmentation. With this
* segment, we're expecting significantly-sized allocations only; we'll
* default to 4x the size of physmem.
*/
segkvmm_base = segkp_base + mmu_ptob(segkpsize);
size = segkvmmsize != 0 ? mmu_ptob(segkvmmsize) : (physmem_size * 4);
size = MAX(size, SEGVMMMINSIZE);
segkvmmsize = mmu_btop(ROUND_UP_LPAGE(size));
PRM_DEBUG(segkvmmsize);
PRM_DEBUG(segkvmm_base);
/*
* segzio is used for ZFS cached data. For segzio, we use 1.5x physmem.
*/
segzio_base = segkvmm_base + mmu_ptob(segkvmmsize);
if (segzio_fromheap) {
segziosize = 0;
} else {
size = (segziosize != 0) ? mmu_ptob(segziosize) :
(physmem_size * 3) / 2;
size = MAX(size, SEGZIOMINSIZE);
segziosize = mmu_btop(ROUND_UP_LPAGE(size));
}
PRM_DEBUG(segziosize);
PRM_DEBUG(segzio_base);
/*
* Put the range of VA for device mappings next, kmdb knows to not
* grep in this range of addresses.
*/
toxic_addr =
ROUND_UP_LPAGE((uintptr_t)segzio_base + mmu_ptob(segziosize));
PRM_DEBUG(toxic_addr);
segmap_start = ROUND_UP_LPAGE(toxic_addr + toxic_size);
/*
* Users can change segmapsize through eeprom. If the variable
* is tuned through eeprom, there is no upper bound on the
* size of segmap.
*/
segmapsize = MAX(ROUND_UP_LPAGE(segmapsize), SEGMAPDEFAULT);
PRM_DEBUG(segmap_start);
PRM_DEBUG(segmapsize);
kernelheap = (caddr_t)ROUND_UP_LPAGE(segmap_start + segmapsize);
PRM_DEBUG(kernelheap);
PRM_POINT("layout_kernel_va() done...");
}
/*
* Finish initializing the VM system, now that we are no longer
* relying on the boot time memory allocators.
*/
static void
startup_vm(void)
{
struct segmap_crargs a;
extern int use_brk_lpg, use_stk_lpg;
PRM_POINT("startup_vm() starting...");
/*
* Initialize the hat layer.
*/
hat_init();
/*
* Do final allocations of HAT data structures that need to
* be allocated before quiescing the boot loader.
*/
PRM_POINT("Calling hat_kern_alloc()...");
hat_kern_alloc((caddr_t)segmap_start, segmapsize, ekernelheap);
PRM_POINT("hat_kern_alloc() done");
#ifndef __xpv
/*
* Setup Page Attribute Table
*/
pat_sync();
#endif
/*
* The next two loops are done in distinct steps in order
* to be sure that any page that is doubly mapped (both above
* KERNEL_TEXT and below kernelbase) is dealt with correctly.
* Note this may never happen, but it might someday.
*/
bootpages = NULL;
PRM_POINT("Protecting boot pages");
/*
* Protect any pages mapped above KERNEL_TEXT that somehow have
* page_t's. This can only happen if something weird allocated
* in this range (like kadb/kmdb).
*/
protect_boot_range(KERNEL_TEXT, (uintptr_t)-1, 0);
/*
* Before we can take over memory allocation/mapping from the boot
* loader we must remove from our free page lists any boot allocated
* pages that stay mapped until release_bootstrap().
*/
protect_boot_range(0, kernelbase, 1);
/*
* Switch to running on regular HAT (not boot_mmu)
*/
PRM_POINT("Calling hat_kern_setup()...");
hat_kern_setup();
/*
* It is no longer safe to call BOP_ALLOC(), so make sure we don't.
*/
bop_no_more_mem();
PRM_POINT("hat_kern_setup() done");
hat_cpu_online(CPU);
/*
* Initialize VM system
*/
PRM_POINT("Calling kvm_init()...");
kvm_init();
PRM_POINT("kvm_init() done");
/*
* Tell kmdb that the VM system is now working
*/
if (boothowto & RB_DEBUG)
kdi_dvec_vmready();
#if defined(__xpv)
/*
* Populate the I/O pool on domain 0
*/
if (DOMAIN_IS_INITDOMAIN(xen_info)) {
extern long populate_io_pool(void);
long init_io_pool_cnt;
PRM_POINT("Populating reserve I/O page pool");
init_io_pool_cnt = populate_io_pool();
PRM_DEBUG(init_io_pool_cnt);
}
#endif
/*
* Mangle the brand string etc.
*/
cpuid_pass3(CPU);
#if defined(__amd64)
/*
* Create the device arena for toxic (to dtrace/kmdb) mappings.
*/
device_arena = vmem_create("device", (void *)toxic_addr,
toxic_size, MMU_PAGESIZE, NULL, NULL, NULL, 0, VM_SLEEP);
#else /* __i386 */
/*
* allocate the bit map that tracks toxic pages
*/
toxic_bit_map_len = btop((ulong_t)(valloc_base - kernelbase));
PRM_DEBUG(toxic_bit_map_len);
toxic_bit_map =
kmem_zalloc(BT_SIZEOFMAP(toxic_bit_map_len), KM_NOSLEEP);
ASSERT(toxic_bit_map != NULL);
PRM_DEBUG(toxic_bit_map);
#endif /* __i386 */
/*
* Now that we've got more VA, as well as the ability to allocate from
* it, tell the debugger.
*/
if (boothowto & RB_DEBUG)
kdi_dvec_memavail();
#if !defined(__xpv)
/*
* Map page pfn=0 for drivers, such as kd, that need to pick up
* parameters left there by controllers/BIOS.
*/
PRM_POINT("setup up p0_va");
p0_va = i86devmap(0, 1, PROT_READ);
PRM_DEBUG(p0_va);
#endif
cmn_err(CE_CONT, "?mem = %luK (0x%lx)\n",
physinstalled << (MMU_PAGESHIFT - 10), ptob(physinstalled));
/*
* disable automatic large pages for small memory systems or
* when the disable flag is set.
*
* Do not yet consider page sizes larger than 2m/4m.
*/
if (!auto_lpg_disable && mmu.max_page_level > 0) {
max_uheap_lpsize = LEVEL_SIZE(1);
max_ustack_lpsize = LEVEL_SIZE(1);
max_privmap_lpsize = LEVEL_SIZE(1);
max_uidata_lpsize = LEVEL_SIZE(1);
max_utext_lpsize = LEVEL_SIZE(1);
max_shm_lpsize = LEVEL_SIZE(1);
}
if (physmem < privm_lpg_min_physmem || mmu.max_page_level == 0 ||
auto_lpg_disable) {
use_brk_lpg = 0;
use_stk_lpg = 0;
}
mcntl0_lpsize = LEVEL_SIZE(mmu.umax_page_level);
PRM_POINT("Calling hat_init_finish()...");
hat_init_finish();
PRM_POINT("hat_init_finish() done");
/*
* Initialize the segkp segment type.
*/
rw_enter(&kas.a_lock, RW_WRITER);
PRM_POINT("Attaching segkp");
if (segkp_fromheap) {
segkp->s_as = &kas;
} else if (seg_attach(&kas, (caddr_t)segkp_base, mmu_ptob(segkpsize),
segkp) < 0) {
panic("startup: cannot attach segkp");
/*NOTREACHED*/
}
PRM_POINT("Doing segkp_create()");
if (segkp_create(segkp) != 0) {
panic("startup: segkp_create failed");
/*NOTREACHED*/
}
PRM_DEBUG(segkp);
rw_exit(&kas.a_lock);
/*
* kpm segment
*/
segmap_kpm = 0;
if (kpm_desired)
kpm_init();
/*
* Now create segmap segment.
*/
rw_enter(&kas.a_lock, RW_WRITER);
if (seg_attach(&kas, (caddr_t)segmap_start, segmapsize, segmap) < 0) {
panic("cannot attach segmap");
/*NOTREACHED*/
}
PRM_DEBUG(segmap);
a.prot = PROT_READ | PROT_WRITE;
a.shmsize = 0;
a.nfreelist = segmapfreelists;
if (segmap_create(segmap, (caddr_t)&a) != 0)
panic("segmap_create segmap");
rw_exit(&kas.a_lock);
setup_vaddr_for_ppcopy(CPU);
segdev_init();
#if defined(__xpv)
if (DOMAIN_IS_INITDOMAIN(xen_info))
#endif
pmem_init();
PRM_POINT("startup_vm() done");
}
/*
* Load a tod module for the non-standard tod part found on this system.
*/
static void
load_tod_module(char *todmod)
{
if (modload("tod", todmod) == -1)
halt("Can't load TOD module");
}
static void
startup_end(void)
{
int i;
extern void setx86isalist(void);
extern void cpu_event_init(void);
PRM_POINT("startup_end() starting...");
/*
* Perform tasks that get done after most of the VM
* initialization has been done but before the clock
* and other devices get started.
*/
kern_setup1();
/*
* Perform CPC initialization for this CPU.
*/
kcpc_hw_init(CPU);
/*
* Initialize cpu event framework.
*/
cpu_event_init();
#if defined(OPTERON_WORKAROUND_6323525)
if (opteron_workaround_6323525)
patch_workaround_6323525();
#endif
/*
* If needed, load TOD module now so that ddi_get_time(9F) etc. work
* (For now, "needed" is defined as set tod_module_name in /etc/system)
*/
if (tod_module_name != NULL) {
PRM_POINT("load_tod_module()");
load_tod_module(tod_module_name);
}
#if defined(__xpv)
/*
* Forceload interposing TOD module for the hypervisor.
*/
PRM_POINT("load_tod_module()");
load_tod_module("xpvtod");
#endif
/*
* Configure the system.
*/
PRM_POINT("Calling configure()...");
configure(); /* set up devices */
PRM_POINT("configure() done");
/*
* We can now setup for XSAVE because fpu_probe is done in configure().
*/
if (fp_save_mech == FP_XSAVE) {
PRM_POINT("xsave_setup_msr()");
xsave_setup_msr(CPU);
}
/*
* Set the isa_list string to the defined instruction sets we
* support.
*/
setx86isalist();
PRM_POINT("cpu_intr_alloc()");
cpu_intr_alloc(CPU, NINTR_THREADS);
PRM_POINT("psm_install()");
psm_install();
/*
* We're done with bootops. We don't unmap the bootstrap yet because
* we're still using bootsvcs.
*/
PRM_POINT("NULLing out bootops");
*bootopsp = (struct bootops *)NULL;
bootops = (struct bootops *)NULL;
#if defined(__xpv)
ec_init_debug_irq();
xs_domu_init();
#endif
#if !defined(__xpv)
/*
* Intel IOMMU has been setup/initialized in ddi_impl.c
* Start it up now.
*/
immu_startup();
/*
* Now that we're no longer going to drop into real mode for a BIOS call
* via bootops, we can enable PCID (which requires CR0.PG).
*/
enable_pcid();
#endif
PRM_POINT("Enabling interrupts");
(*picinitf)();
sti();
#if defined(__xpv)
ASSERT(CPU->cpu_m.mcpu_vcpu_info->evtchn_upcall_mask == 0);
xen_late_startup();
#endif
(void) add_avsoftintr((void *)&softlevel1_hdl, 1, softlevel1,
"softlevel1", NULL, NULL); /* XXX to be moved later */
/*
* Register software interrupt handlers for ddi_periodic_add(9F).
* Software interrupts up to the level 10 are supported.
*/
for (i = DDI_IPL_1; i <= DDI_IPL_10; i++) {
(void) add_avsoftintr((void *)&softlevel_hdl[i-1], i,
(avfunc)(uintptr_t)ddi_periodic_softintr, "ddi_periodic",
(caddr_t)(uintptr_t)i, NULL);
}
#if !defined(__xpv)
if (modload("drv", "amd_iommu") < 0) {
PRM_POINT("No AMD IOMMU present\n");
} else if (ddi_hold_installed_driver(ddi_name_to_major(
"amd_iommu")) == NULL) {
prom_printf("ERROR: failed to attach AMD IOMMU\n");
}
#endif
post_startup_cpu_fixups();
PRM_POINT("startup_end() done");
}
/*
* Don't remove the following 2 variables. They are necessary
* for reading the hostid from the legacy file (/kernel/misc/sysinit).
*/
char *_hs1107 = hw_serial;
ulong_t _bdhs34;
void
post_startup(void)
{
extern void cpupm_init(cpu_t *);
extern void cpu_event_init_cpu(cpu_t *);
/*
* Set the system wide, processor-specific flags to be passed
* to userland via the aux vector for performance hints and
* instruction set extensions.
*/
bind_hwcap();
#ifdef __xpv
if (DOMAIN_IS_INITDOMAIN(xen_info))
#endif
{
#if defined(__xpv)
xpv_panic_init();
#else
/*
* Startup the memory scrubber.
* XXPV This should be running somewhere ..
*/
if ((get_hwenv() & HW_VIRTUAL) == 0)
memscrub_init();
#endif
}
/*
* Complete CPU module initialization
*/
cmi_post_startup();
/*
* Perform forceloading tasks for /etc/system.
*/
(void) mod_sysctl(SYS_FORCELOAD, NULL);
/*
* ON4.0: Force /proc module in until clock interrupt handle fixed
* ON4.0: This must be fixed or restated in /etc/systems.
*/
(void) modload("fs", "procfs");
(void) i_ddi_attach_hw_nodes("pit_beep");
#if defined(__i386)
/*
* Check for required functional Floating Point hardware,
* unless FP hardware explicitly disabled.
*/
if (fpu_exists && (fpu_pentium_fdivbug || fp_kind == FP_NO))
halt("No working FP hardware found");
#endif
maxmem = freemem;
cpu_event_init_cpu(CPU);
cpupm_init(CPU);
(void) mach_cpu_create_device_node(CPU, NULL);
pg_init();
}
static int
pp_in_range(page_t *pp, uint64_t low_addr, uint64_t high_addr)
{
return ((pp->p_pagenum >= btop(low_addr)) &&
(pp->p_pagenum < btopr(high_addr)));
}
static int
pp_in_module(page_t *pp, const rd_existing_t *modranges)
{
uint_t i;
for (i = 0; modranges[i].phys != 0; i++) {
if (pp_in_range(pp, modranges[i].phys,
modranges[i].phys + modranges[i].size))
return (1);
}
return (0);
}
void
release_bootstrap(void)
{
int root_is_ramdisk;
page_t *pp;
extern void kobj_boot_unmountroot(void);
extern dev_t rootdev;
uint_t i;
char propname[32];
rd_existing_t *modranges;
#if !defined(__xpv)
pfn_t pfn;
#endif
/*
* Save the bootfs module ranges so that we can reserve them below
* for the real bootfs.
*/
modranges = kmem_alloc(sizeof (rd_existing_t) * MAX_BOOT_MODULES,
KM_SLEEP);
for (i = 0; ; i++) {
uint64_t start, size;
modranges[i].phys = 0;
(void) snprintf(propname, sizeof (propname),
"module-addr-%u", i);
if (do_bsys_getproplen(NULL, propname) <= 0)
break;
(void) do_bsys_getprop(NULL, propname, &start);
(void) snprintf(propname, sizeof (propname),
"module-size-%u", i);
if (do_bsys_getproplen(NULL, propname) <= 0)
break;
(void) do_bsys_getprop(NULL, propname, &size);
modranges[i].phys = start;
modranges[i].size = size;
}
/* unmount boot ramdisk and release kmem usage */
kobj_boot_unmountroot();
/*
* We're finished using the boot loader so free its pages.
*/
PRM_POINT("Unmapping lower boot pages");
clear_boot_mappings(0, _userlimit);
postbootkernelbase = kernelbase;
/*
* If root isn't on ramdisk, destroy the hardcoded
* ramdisk node now and release the memory. Else,
* ramdisk memory is kept in rd_pages.
*/
root_is_ramdisk = (getmajor(rootdev) == ddi_name_to_major("ramdisk"));
if (!root_is_ramdisk) {
dev_info_t *dip = ddi_find_devinfo("ramdisk", -1, 0);
ASSERT(dip && ddi_get_parent(dip) == ddi_root_node());
ndi_rele_devi(dip); /* held from ddi_find_devinfo */
(void) ddi_remove_child(dip, 0);
}
PRM_POINT("Releasing boot pages");
while (bootpages) {
extern uint64_t ramdisk_start, ramdisk_end;
pp = bootpages;
bootpages = pp->p_next;
/* Keep pages for the lower 64K */
if (pp_in_range(pp, 0, 0x40000)) {
pp->p_next = lower_pages;
lower_pages = pp;
lower_pages_count++;
continue;
}
if (root_is_ramdisk && pp_in_range(pp, ramdisk_start,
ramdisk_end) || pp_in_module(pp, modranges)) {
pp->p_next = rd_pages;
rd_pages = pp;
continue;
}
pp->p_next = (struct page *)0;
pp->p_prev = (struct page *)0;
PP_CLRBOOTPAGES(pp);
page_free(pp, 1);
}
PRM_POINT("Boot pages released");
kmem_free(modranges, sizeof (rd_existing_t) * 99);
#if !defined(__xpv)
/* XXPV -- note this following bunch of code needs to be revisited in Xen 3.0 */
/*
* Find 1 page below 1 MB so that other processors can boot up or
* so that any processor can resume.
* Make sure it has a kernel VA as well as a 1:1 mapping.
* We should have just free'd one up.
*/
/*
* 0x10 pages is 64K. Leave the bottom 64K alone
* for BIOS.
*/
for (pfn = 0x10; pfn < btop(1*1024*1024); pfn++) {
if (page_numtopp_alloc(pfn) == NULL)
continue;
rm_platter_va = i86devmap(pfn, 1,
PROT_READ | PROT_WRITE | PROT_EXEC);
rm_platter_pa = ptob(pfn);
break;
}
if (pfn == btop(1*1024*1024) && use_mp)
panic("No page below 1M available for starting "
"other processors or for resuming from system-suspend");
#endif /* !__xpv */
}
/*
* Initialize the platform-specific parts of a page_t.
*/
void
add_physmem_cb(page_t *pp, pfn_t pnum)
{
pp->p_pagenum = pnum;
pp->p_mapping = NULL;
pp->p_embed = 0;
pp->p_share = 0;
pp->p_zoneid = ALL_ZONES;
pp->p_mlentry = 0;
}
/*
* kphysm_init() initializes physical memory.
*/
static pgcnt_t
kphysm_init(page_t *pp, pgcnt_t npages)
{
struct memlist *pmem;
struct memseg *cur_memseg;
pfn_t base_pfn;
pfn_t end_pfn;
pgcnt_t num;
pgcnt_t pages_done = 0;
uint64_t addr;
uint64_t size;
extern pfn_t ddiphysmin;
extern int mnode_xwa;
int ms = 0, me = 0;
ASSERT(page_hash != NULL && page_hashsz != 0);
cur_memseg = memseg_base;
for (pmem = phys_avail; pmem && npages; pmem = pmem->ml_next) {
/*
* In a 32 bit kernel can't use higher memory if we're
* not booting in PAE mode. This check takes care of that.
*/
addr = pmem->ml_address;
size = pmem->ml_size;
if (btop(addr) > physmax)
continue;
/*
* align addr and size - they may not be at page boundaries
*/
if ((addr & MMU_PAGEOFFSET) != 0) {
addr += MMU_PAGEOFFSET;
addr &= ~(uint64_t)MMU_PAGEOFFSET;
size -= addr - pmem->ml_address;
}
/* only process pages below or equal to physmax */
if ((btop(addr + size) - 1) > physmax)
size = ptob(physmax - btop(addr) + 1);
num = btop(size);
if (num == 0)
continue;
if (num > npages)
num = npages;
npages -= num;
pages_done += num;
base_pfn = btop(addr);
if (prom_debug)
prom_printf("MEMSEG addr=0x%" PRIx64
" pgs=0x%lx pfn 0x%lx-0x%lx\n",
addr, num, base_pfn, base_pfn + num);
/*
* Ignore pages below ddiphysmin to simplify ddi memory
* allocation with non-zero addr_lo requests.
*/
if (base_pfn < ddiphysmin) {
if (base_pfn + num <= ddiphysmin)
continue;
pp += (ddiphysmin - base_pfn);
num -= (ddiphysmin - base_pfn);
base_pfn = ddiphysmin;
}
/*
* mnode_xwa is greater than 1 when large pages regions can
* cross memory node boundaries. To prevent the formation
* of these large pages, configure the memsegs based on the
* memory node ranges which had been made non-contiguous.
*/
end_pfn = base_pfn + num - 1;
if (mnode_xwa > 1) {
ms = PFN_2_MEM_NODE(base_pfn);
me = PFN_2_MEM_NODE(end_pfn);
if (ms != me) {
/*
* current range spans more than 1 memory node.
* Set num to only the pfn range in the start
* memory node.
*/
num = mem_node_config[ms].physmax - base_pfn
+ 1;
ASSERT(end_pfn > mem_node_config[ms].physmax);
}
}
for (;;) {
/*
* Build the memsegs entry
*/
cur_memseg->pages = pp;
cur_memseg->epages = pp + num;
cur_memseg->pages_base = base_pfn;
cur_memseg->pages_end = base_pfn + num;
/*
* Insert into memseg list in decreasing pfn range
* order. Low memory is typically more fragmented such
* that this ordering keeps the larger ranges at the
* front of the list for code that searches memseg.
* This ASSERTS that the memsegs coming in from boot
* are in increasing physical address order and not
* contiguous.
*/
if (memsegs != NULL) {
ASSERT(cur_memseg->pages_base >=
memsegs->pages_end);
cur_memseg->next = memsegs;
}
memsegs = cur_memseg;
/*
* add_physmem() initializes the PSM part of the page
* struct by calling the PSM back with add_physmem_cb().
* In addition it coalesces pages into larger pages as
* it initializes them.
*/
add_physmem(pp, num, base_pfn);
cur_memseg++;
availrmem_initial += num;
availrmem += num;
pp += num;
if (ms >= me)
break;
/* process next memory node range */
ms++;
base_pfn = mem_node_config[ms].physbase;
if (mnode_xwa > 1) {
num = MIN(mem_node_config[ms].physmax,
end_pfn) - base_pfn + 1;
} else {
num = mem_node_config[ms].physmax -
base_pfn + 1;
}
}
}
PRM_DEBUG(availrmem_initial);
PRM_DEBUG(availrmem);
PRM_DEBUG(freemem);
build_pfn_hash();
return (pages_done);
}
/*
* Kernel VM initialization.
*/
static void
kvm_init(void)
{
ASSERT((((uintptr_t)s_text) & MMU_PAGEOFFSET) == 0);
/*
* Put the kernel segments in kernel address space.
*/
rw_enter(&kas.a_lock, RW_WRITER);
as_avlinit(&kas);
(void) seg_attach(&kas, s_text, e_moddata - s_text, &ktextseg);
(void) segkmem_create(&ktextseg);
(void) seg_attach(&kas, (caddr_t)valloc_base, valloc_sz, &kvalloc);
(void) segkmem_create(&kvalloc);
(void) seg_attach(&kas, kernelheap,
ekernelheap - kernelheap, &kvseg);
(void) segkmem_create(&kvseg);
if (core_size > 0) {
PRM_POINT("attaching kvseg_core");
(void) seg_attach(&kas, (caddr_t)core_base, core_size,
&kvseg_core);
(void) segkmem_create(&kvseg_core);
}
PRM_POINT("attaching segkvmm");
(void) seg_attach(&kas, segkvmm_base, mmu_ptob(segkvmmsize), &kvmmseg);
(void) segkmem_create(&kvmmseg);
segkmem_kvmm_init(segkvmm_base, mmu_ptob(segkvmmsize));
if (segziosize > 0) {
PRM_POINT("attaching segzio");
(void) seg_attach(&kas, segzio_base, mmu_ptob(segziosize),
&kzioseg);
(void) segkmem_create(&kzioseg);
/* create zio area covering new segment */
segkmem_zio_init(segzio_base, mmu_ptob(segziosize));
}
(void) seg_attach(&kas, kdi_segdebugbase, kdi_segdebugsize, &kdebugseg);
(void) segkmem_create(&kdebugseg);
rw_exit(&kas.a_lock);
/*
* Ensure that the red zone at kernelbase is never accessible.
*/
PRM_POINT("protecting redzone");
(void) as_setprot(&kas, (caddr_t)kernelbase, KERNEL_REDZONE_SIZE, 0);
/*
* Make the text writable so that it can be hot patched by DTrace.
*/
(void) as_setprot(&kas, s_text, e_modtext - s_text,
PROT_READ | PROT_WRITE | PROT_EXEC);
/*
* Make data writable until end.
*/
(void) as_setprot(&kas, s_data, e_moddata - s_data,
PROT_READ | PROT_WRITE | PROT_EXEC);
}
#ifndef __xpv
/*
* Solaris adds an entry for Write Combining caching to the PAT
*/
static uint64_t pat_attr_reg = PAT_DEFAULT_ATTRIBUTE;
void
pat_sync(void)
{
ulong_t cr0, cr0_orig, cr4;
if (!is_x86_feature(x86_featureset, X86FSET_PAT))
return;
cr0_orig = cr0 = getcr0();
cr4 = getcr4();
/* disable caching and flush all caches and TLBs */
cr0 |= CR0_CD;
cr0 &= ~CR0_NW;
setcr0(cr0);
invalidate_cache();
if (cr4 & CR4_PGE) {
setcr4(cr4 & ~(ulong_t)CR4_PGE);
setcr4(cr4);
} else {
reload_cr3();
}
/* add our entry to the PAT */
wrmsr(REG_PAT, pat_attr_reg);
/* flush TLBs and cache again, then reenable cr0 caching */
if (cr4 & CR4_PGE) {
setcr4(cr4 & ~(ulong_t)CR4_PGE);
setcr4(cr4);
} else {
reload_cr3();
}
invalidate_cache();
setcr0(cr0_orig);
}
#endif /* !__xpv */
#if defined(_SOFT_HOSTID)
/*
* On platforms that do not have a hardware serial number, attempt
* to set one based on the contents of /etc/hostid. If this file does
* not exist, assume that we are to generate a new hostid and set
* it in the kernel, for subsequent saving by a userland process
* once the system is up and the root filesystem is mounted r/w.
*
* In order to gracefully support upgrade on OpenSolaris, if
* /etc/hostid does not exist, we will attempt to get a serial number
* using the legacy method (/kernel/misc/sysinit).
*
* If that isn't present, we attempt to use an SMBIOS UUID, which is
* a hardware serial number. Note that we don't automatically trust
* all SMBIOS UUIDs (some older platforms are defective and ship duplicate
* UUIDs in violation of the standard), we check against a blacklist.
*
* In an attempt to make the hostid less prone to abuse
* (for license circumvention, etc), we store it in /etc/hostid
* in rot47 format.
*/
extern volatile unsigned long tenmicrodata;
static int atoi(char *);
/*
* Set this to non-zero in /etc/system if you think your SMBIOS returns a
* UUID that is not unique. (Also report it so that the smbios_uuid_blacklist
* array can be updated.)
*/
int smbios_broken_uuid = 0;
/*
* List of known bad UUIDs. This is just the lower 32-bit values, since
* that's what we use for the host id. If your hostid falls here, you need
* to contact your hardware OEM for a fix for your BIOS.
*/
static unsigned char
smbios_uuid_blacklist[][16] = {
{ /* Reported bad UUID (Google search) */
0x00, 0x02, 0x00, 0x03, 0x00, 0x04, 0x00, 0x05,
0x00, 0x06, 0x00, 0x07, 0x00, 0x08, 0x00, 0x09,
},
{ /* Known bad DELL UUID */
0x4C, 0x4C, 0x45, 0x44, 0x00, 0x00, 0x20, 0x10,
0x80, 0x20, 0x80, 0xC0, 0x4F, 0x20, 0x20, 0x20,
},
{ /* Uninitialized flash */
0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff
},
{ /* All zeros */
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
},
};
static int32_t
uuid_to_hostid(const uint8_t *uuid)
{
/*
* Although the UUIDs are 128-bits, they may not distribute entropy
* evenly. We would like to use SHA or MD5, but those are located
* in loadable modules and not available this early in boot. As we
* don't need the values to be cryptographically strong, we just
* generate 32-bit vaue by xor'ing the various sequences together,
* which ensures that the entire UUID contributes to the hostid.
*/
uint32_t id = 0;
/* first check against the blacklist */
for (int i = 0; i < (sizeof (smbios_uuid_blacklist) / 16); i++) {
if (bcmp(smbios_uuid_blacklist[0], uuid, 16) == 0) {
cmn_err(CE_CONT, "?Broken SMBIOS UUID. "
"Contact BIOS manufacturer for repair.\n");
return ((int32_t)HW_INVALID_HOSTID);
}
}
for (int i = 0; i < 16; i++)
id ^= ((uuid[i]) << (8 * (i % sizeof (id))));
/* Make sure return value is positive */
return (id & 0x7fffffff);
}
static int32_t
set_soft_hostid(void)
{
struct _buf *file;
char tokbuf[MAXNAMELEN];
token_t token;
int done = 0;
u_longlong_t tmp;
int i;
int32_t hostid = (int32_t)HW_INVALID_HOSTID;
unsigned char *c;
hrtime_t tsc;
smbios_system_t smsys;
/*
* If /etc/hostid file not found, we'd like to get a pseudo
* random number to use at the hostid. A nice way to do this
* is to read the real time clock. To remain xen-compatible,
* we can't poke the real hardware, so we use tsc_read() to
* read the real time clock. However, there is an ominous
* warning in tsc_read that says it can return zero, so we
* deal with that possibility by falling back to using the
* (hopefully random enough) value in tenmicrodata.
*/
if ((file = kobj_open_file(hostid_file)) == (struct _buf *)-1) {
/*
* hostid file not found - try to load sysinit module
* and see if it has a nonzero hostid value...use that
* instead of generating a new hostid here if so.
*/
if ((i = modload("misc", "sysinit")) != -1) {
if (strlen(hw_serial) > 0)
hostid = (int32_t)atoi(hw_serial);
(void) modunload(i);
}
/*
* We try to use the SMBIOS UUID. But not if it is blacklisted
* in /etc/system.
*/
if ((hostid == HW_INVALID_HOSTID) &&
(smbios_broken_uuid == 0) &&
(ksmbios != NULL) &&
(smbios_info_system(ksmbios, &smsys) != SMB_ERR) &&
(smsys.smbs_uuidlen >= 16)) {
hostid = uuid_to_hostid(smsys.smbs_uuid);
}
/*
* Generate a "random" hostid using the clock. These
* hostids will change on each boot if the value is not
* saved to a persistent /etc/hostid file.
*/
if (hostid == HW_INVALID_HOSTID) {
tsc = tsc_read();
if (tsc == 0) /* tsc_read can return zero sometimes */
hostid = (int32_t)tenmicrodata & 0x0CFFFFF;
else
hostid = (int32_t)tsc & 0x0CFFFFF;
}
} else {
/* hostid file found */
while (!done) {
token = kobj_lex(file, tokbuf, sizeof (tokbuf));
switch (token) {
case POUND:
/*
* skip comments
*/
kobj_find_eol(file);
break;
case STRING:
/*
* un-rot47 - obviously this
* nonsense is ascii-specific
*/
for (c = (unsigned char *)tokbuf;
*c != '\0'; c++) {
*c += 47;
if (*c > '~')
*c -= 94;
else if (*c < '!')
*c += 94;
}
/*
* now we should have a real number
*/
if (kobj_getvalue(tokbuf, &tmp) != 0)
kobj_file_err(CE_WARN, file,
"Bad value %s for hostid",
tokbuf);
else
hostid = (int32_t)tmp;
break;
case EOF:
done = 1;
/* FALLTHROUGH */
case NEWLINE:
kobj_newline(file);
break;
default:
break;
}
}
if (hostid == HW_INVALID_HOSTID) /* didn't find a hostid */
kobj_file_err(CE_WARN, file,
"hostid missing or corrupt");
kobj_close_file(file);
}
/*
* hostid is now the value read from /etc/hostid, or the
* new hostid we generated in this routine or HW_INVALID_HOSTID if not
* set.
*/
return (hostid);
}
static int
atoi(char *p)
{
int i = 0;
while (*p != '\0')
i = 10 * i + (*p++ - '0');
return (i);
}
#endif /* _SOFT_HOSTID */
void
get_system_configuration(void)
{
char prop[32];
u_longlong_t nodes_ll, cpus_pernode_ll, lvalue;
if (BOP_GETPROPLEN(bootops, "nodes") > sizeof (prop) ||
BOP_GETPROP(bootops, "nodes", prop) < 0 ||
kobj_getvalue(prop, &nodes_ll) == -1 ||
nodes_ll > MAXNODES ||
BOP_GETPROPLEN(bootops, "cpus_pernode") > sizeof (prop) ||
BOP_GETPROP(bootops, "cpus_pernode", prop) < 0 ||
kobj_getvalue(prop, &cpus_pernode_ll) == -1) {
system_hardware.hd_nodes = 1;
system_hardware.hd_cpus_per_node = 0;
} else {
system_hardware.hd_nodes = (int)nodes_ll;
system_hardware.hd_cpus_per_node = (int)cpus_pernode_ll;
}
if (BOP_GETPROPLEN(bootops, "kernelbase") > sizeof (prop) ||
BOP_GETPROP(bootops, "kernelbase", prop) < 0 ||
kobj_getvalue(prop, &lvalue) == -1)
eprom_kernelbase = 0;
else
eprom_kernelbase = (uintptr_t)lvalue;
if (BOP_GETPROPLEN(bootops, "segmapsize") > sizeof (prop) ||
BOP_GETPROP(bootops, "segmapsize", prop) < 0 ||
kobj_getvalue(prop, &lvalue) == -1)
segmapsize = SEGMAPDEFAULT;
else
segmapsize = (uintptr_t)lvalue;
if (BOP_GETPROPLEN(bootops, "segmapfreelists") > sizeof (prop) ||
BOP_GETPROP(bootops, "segmapfreelists", prop) < 0 ||
kobj_getvalue(prop, &lvalue) == -1)
segmapfreelists = 0; /* use segmap driver default */
else
segmapfreelists = (int)lvalue;
/* physmem used to be here, but moved much earlier to fakebop.c */
}
/*
* Add to a memory list.
* start = start of new memory segment
* len = length of new memory segment in bytes
* new = pointer to a new struct memlist
* memlistp = memory list to which to add segment.
*/
void
memlist_add(
uint64_t start,
uint64_t len,
struct memlist *new,
struct memlist **memlistp)
{
struct memlist *cur;
uint64_t end = start + len;
new->ml_address = start;
new->ml_size = len;
cur = *memlistp;
while (cur) {
if (cur->ml_address >= end) {
new->ml_next = cur;
*memlistp = new;
new->ml_prev = cur->ml_prev;
cur->ml_prev = new;
return;
}
ASSERT(cur->ml_address + cur->ml_size <= start);
if (cur->ml_next == NULL) {
cur->ml_next = new;
new->ml_prev = cur;
new->ml_next = NULL;
return;
}
memlistp = &cur->ml_next;
cur = cur->ml_next;
}
}
void
kobj_vmem_init(vmem_t **text_arena, vmem_t **data_arena)
{
size_t tsize = e_modtext - modtext;
size_t dsize = e_moddata - moddata;
*text_arena = vmem_create("module_text", tsize ? modtext : NULL, tsize,
1, segkmem_alloc, segkmem_free, heaptext_arena, 0, VM_SLEEP);
*data_arena = vmem_create("module_data", dsize ? moddata : NULL, dsize,
1, segkmem_alloc, segkmem_free, heap32_arena, 0, VM_SLEEP);
}
caddr_t
kobj_text_alloc(vmem_t *arena, size_t size)
{
return (vmem_alloc(arena, size, VM_SLEEP | VM_BESTFIT));
}
/*ARGSUSED*/
caddr_t
kobj_texthole_alloc(caddr_t addr, size_t size)
{
panic("unexpected call to kobj_texthole_alloc()");
/*NOTREACHED*/
return (0);
}
/*ARGSUSED*/
void
kobj_texthole_free(caddr_t addr, size_t size)
{
panic("unexpected call to kobj_texthole_free()");
}
/*
* This is called just after configure() in startup().
*
* The ISALIST concept is a bit hopeless on Intel, because
* there's no guarantee of an ever-more-capable processor
* given that various parts of the instruction set may appear
* and disappear between different implementations.
*
* While it would be possible to correct it and even enhance
* it somewhat, the explicit hardware capability bitmask allows
* more flexibility.
*
* So, we just leave this alone.
*/
void
setx86isalist(void)
{
char *tp;
size_t len;
extern char *isa_list;
#define TBUFSIZE 1024
tp = kmem_alloc(TBUFSIZE, KM_SLEEP);
*tp = '\0';
#if defined(__amd64)
(void) strcpy(tp, "amd64 ");
#endif
switch (x86_vendor) {
case X86_VENDOR_Intel:
case X86_VENDOR_AMD:
case X86_VENDOR_HYGON:
case X86_VENDOR_TM:
if (is_x86_feature(x86_featureset, X86FSET_CMOV)) {
/*
* Pentium Pro or later
*/
(void) strcat(tp, "pentium_pro");
(void) strcat(tp,
is_x86_feature(x86_featureset, X86FSET_MMX) ?
"+mmx pentium_pro " : " ");
}
/*FALLTHROUGH*/
case X86_VENDOR_Cyrix:
/*
* The Cyrix 6x86 does not have any Pentium features
* accessible while not at privilege level 0.
*/
if (is_x86_feature(x86_featureset, X86FSET_CPUID)) {
(void) strcat(tp, "pentium");
(void) strcat(tp,
is_x86_feature(x86_featureset, X86FSET_MMX) ?
"+mmx pentium " : " ");
}
break;
default:
break;
}
(void) strcat(tp, "i486 i386 i86");
len = strlen(tp) + 1; /* account for NULL at end of string */
isa_list = strcpy(kmem_alloc(len, KM_SLEEP), tp);
kmem_free(tp, TBUFSIZE);
#undef TBUFSIZE
}
#ifdef __amd64
void *
device_arena_alloc(size_t size, int vm_flag)
{
return (vmem_alloc(device_arena, size, vm_flag));
}
void
device_arena_free(void *vaddr, size_t size)
{
vmem_free(device_arena, vaddr, size);
}
#else /* __i386 */
void *
device_arena_alloc(size_t size, int vm_flag)
{
caddr_t vaddr;
uintptr_t v;
size_t start;
size_t end;
vaddr = vmem_alloc(heap_arena, size, vm_flag);
if (vaddr == NULL)
return (NULL);
v = (uintptr_t)vaddr;
ASSERT(v >= kernelbase);
ASSERT(v + size <= valloc_base);
start = btop(v - kernelbase);
end = btop(v + size - 1 - kernelbase);
ASSERT(start < toxic_bit_map_len);
ASSERT(end < toxic_bit_map_len);
while (start <= end) {
BT_ATOMIC_SET(toxic_bit_map, start);
++start;
}
return (vaddr);
}
void
device_arena_free(void *vaddr, size_t size)
{
uintptr_t v = (uintptr_t)vaddr;
size_t start;
size_t end;
ASSERT(v >= kernelbase);
ASSERT(v + size <= valloc_base);
start = btop(v - kernelbase);
end = btop(v + size - 1 - kernelbase);
ASSERT(start < toxic_bit_map_len);
ASSERT(end < toxic_bit_map_len);
while (start <= end) {
ASSERT(BT_TEST(toxic_bit_map, start) != 0);
BT_ATOMIC_CLEAR(toxic_bit_map, start);
++start;
}
vmem_free(heap_arena, vaddr, size);
}
/*
* returns 1st address in range that is in device arena, or NULL
* if len is not NULL it returns the length of the toxic range
*/
void *
device_arena_contains(void *vaddr, size_t size, size_t *len)
{
uintptr_t v = (uintptr_t)vaddr;
uintptr_t eaddr = v + size;
size_t start;
size_t end;
/*
* if called very early by kmdb, just return NULL
*/
if (toxic_bit_map == NULL)
return (NULL);
/*
* First check if we're completely outside the bitmap range.
*/
if (v >= valloc_base || eaddr < kernelbase)
return (NULL);
/*
* Trim ends of search to look at only what the bitmap covers.
*/
if (v < kernelbase)
v = kernelbase;
start = btop(v - kernelbase);
end = btop(eaddr - kernelbase);
if (end >= toxic_bit_map_len)
end = toxic_bit_map_len;
if (bt_range(toxic_bit_map, &start, &end, end) == 0)
return (NULL);
v = kernelbase + ptob(start);
if (len != NULL)
*len = ptob(end - start);
return ((void *)v);
}
#endif /* __i386 */
|