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
|
/*-
* Copyright (C) 2007-2009 Semihalf, Rafal Jaworowski <raj@semihalf.com>
* Copyright (C) 2006 Semihalf, Marian Balakowicz <m8@semihalf.com>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN
* NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
* TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Some hw specific parts of this pmap were derived or influenced
* by NetBSD's ibm4xx pmap module. More generic code is shared with
* a few other pmap modules from the FreeBSD tree.
*/
/*
* VM layout notes:
*
* Kernel and user threads run within one common virtual address space
* defined by AS=0.
*
* Virtual address space layout:
* -----------------------------
* 0x0000_0000 - 0xafff_ffff : user process
* 0xb000_0000 - 0xbfff_ffff : pmap_mapdev()-ed area (PCI/PCIE etc.)
* 0xc000_0000 - 0xc0ff_ffff : kernel reserved
* 0xc000_0000 - data_end : kernel code+data, env, metadata etc.
* 0xc100_0000 - 0xfeef_ffff : KVA
* 0xc100_0000 - 0xc100_3fff : reserved for page zero/copy
* 0xc100_4000 - 0xc200_3fff : reserved for ptbl bufs
* 0xc200_4000 - 0xc200_8fff : guard page + kstack0
* 0xc200_9000 - 0xfeef_ffff : actual free KVA space
* 0xfef0_0000 - 0xffff_ffff : I/O devices region
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/conf.h>
#include <sys/malloc.h>
#include <sys/ktr.h>
#include <sys/proc.h>
#include <sys/user.h>
#include <sys/queue.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/kerneldump.h>
#include <sys/linker.h>
#include <sys/msgbuf.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/rwlock.h>
#include <sys/sched.h>
#include <sys/smp.h>
#include <sys/vmmeter.h>
#include <vm/vm.h>
#include <vm/vm_page.h>
#include <vm/vm_kern.h>
#include <vm/vm_pageout.h>
#include <vm/vm_extern.h>
#include <vm/vm_object.h>
#include <vm/vm_param.h>
#include <vm/vm_map.h>
#include <vm/vm_pager.h>
#include <vm/uma.h>
#include <machine/cpu.h>
#include <machine/pcb.h>
#include <machine/platform.h>
#include <machine/tlb.h>
#include <machine/spr.h>
#include <machine/md_var.h>
#include <machine/mmuvar.h>
#include <machine/pmap.h>
#include <machine/pte.h>
#include "mmu_if.h"
#ifdef DEBUG
#define debugf(fmt, args...) printf(fmt, ##args)
#else
#define debugf(fmt, args...)
#endif
#define TODO panic("%s: not implemented", __func__);
extern unsigned char _etext[];
extern unsigned char _end[];
extern uint32_t *bootinfo;
#ifdef SMP
extern uint32_t bp_ntlb1s;
#endif
vm_paddr_t kernload;
vm_offset_t kernstart;
vm_size_t kernsize;
/* Message buffer and tables. */
static vm_offset_t data_start;
static vm_size_t data_end;
/* Phys/avail memory regions. */
static struct mem_region *availmem_regions;
static int availmem_regions_sz;
static struct mem_region *physmem_regions;
static int physmem_regions_sz;
/* Reserved KVA space and mutex for mmu_booke_zero_page. */
static vm_offset_t zero_page_va;
static struct mtx zero_page_mutex;
static struct mtx tlbivax_mutex;
/*
* Reserved KVA space for mmu_booke_zero_page_idle. This is used
* by idle thred only, no lock required.
*/
static vm_offset_t zero_page_idle_va;
/* Reserved KVA space and mutex for mmu_booke_copy_page. */
static vm_offset_t copy_page_src_va;
static vm_offset_t copy_page_dst_va;
static struct mtx copy_page_mutex;
/**************************************************************************/
/* PMAP */
/**************************************************************************/
static int mmu_booke_enter_locked(mmu_t, pmap_t, vm_offset_t, vm_page_t,
vm_prot_t, u_int flags, int8_t psind);
unsigned int kptbl_min; /* Index of the first kernel ptbl. */
unsigned int kernel_ptbls; /* Number of KVA ptbls. */
/*
* If user pmap is processed with mmu_booke_remove and the resident count
* drops to 0, there are no more pages to remove, so we need not continue.
*/
#define PMAP_REMOVE_DONE(pmap) \
((pmap) != kernel_pmap && (pmap)->pm_stats.resident_count == 0)
extern void tid_flush(tlbtid_t);
extern int elf32_nxstack;
/**************************************************************************/
/* TLB and TID handling */
/**************************************************************************/
/* Translation ID busy table */
static volatile pmap_t tidbusy[MAXCPU][TID_MAX + 1];
/*
* TLB0 capabilities (entry, way numbers etc.). These can vary between e500
* core revisions and should be read from h/w registers during early config.
*/
uint32_t tlb0_entries;
uint32_t tlb0_ways;
uint32_t tlb0_entries_per_way;
#define TLB0_ENTRIES (tlb0_entries)
#define TLB0_WAYS (tlb0_ways)
#define TLB0_ENTRIES_PER_WAY (tlb0_entries_per_way)
#define TLB1_ENTRIES 16
/* In-ram copy of the TLB1 */
static tlb_entry_t tlb1[TLB1_ENTRIES];
/* Next free entry in the TLB1 */
static unsigned int tlb1_idx;
static vm_offset_t tlb1_map_base = VM_MAX_KERNEL_ADDRESS;
static tlbtid_t tid_alloc(struct pmap *);
static void tlb_print_entry(int, uint32_t, uint32_t, uint32_t, uint32_t);
static int tlb1_set_entry(vm_offset_t, vm_offset_t, vm_size_t, uint32_t);
static void tlb1_write_entry(unsigned int);
static int tlb1_iomapped(int, vm_paddr_t, vm_size_t, vm_offset_t *);
static vm_size_t tlb1_mapin_region(vm_offset_t, vm_paddr_t, vm_size_t);
static vm_size_t tsize2size(unsigned int);
static unsigned int size2tsize(vm_size_t);
static unsigned int ilog2(unsigned int);
static void set_mas4_defaults(void);
static inline void tlb0_flush_entry(vm_offset_t);
static inline unsigned int tlb0_tableidx(vm_offset_t, unsigned int);
/**************************************************************************/
/* Page table management */
/**************************************************************************/
static struct rwlock_padalign pvh_global_lock;
/* Data for the pv entry allocation mechanism */
static uma_zone_t pvzone;
static int pv_entry_count = 0, pv_entry_max = 0, pv_entry_high_water = 0;
#define PV_ENTRY_ZONE_MIN 2048 /* min pv entries in uma zone */
#ifndef PMAP_SHPGPERPROC
#define PMAP_SHPGPERPROC 200
#endif
static void ptbl_init(void);
static struct ptbl_buf *ptbl_buf_alloc(void);
static void ptbl_buf_free(struct ptbl_buf *);
static void ptbl_free_pmap_ptbl(pmap_t, pte_t *);
static pte_t *ptbl_alloc(mmu_t, pmap_t, unsigned int, boolean_t);
static void ptbl_free(mmu_t, pmap_t, unsigned int);
static void ptbl_hold(mmu_t, pmap_t, unsigned int);
static int ptbl_unhold(mmu_t, pmap_t, unsigned int);
static vm_paddr_t pte_vatopa(mmu_t, pmap_t, vm_offset_t);
static pte_t *pte_find(mmu_t, pmap_t, vm_offset_t);
static int pte_enter(mmu_t, pmap_t, vm_page_t, vm_offset_t, uint32_t, boolean_t);
static int pte_remove(mmu_t, pmap_t, vm_offset_t, uint8_t);
static pv_entry_t pv_alloc(void);
static void pv_free(pv_entry_t);
static void pv_insert(pmap_t, vm_offset_t, vm_page_t);
static void pv_remove(pmap_t, vm_offset_t, vm_page_t);
/* Number of kva ptbl buffers, each covering one ptbl (PTBL_PAGES). */
#define PTBL_BUFS (128 * 16)
struct ptbl_buf {
TAILQ_ENTRY(ptbl_buf) link; /* list link */
vm_offset_t kva; /* va of mapping */
};
/* ptbl free list and a lock used for access synchronization. */
static TAILQ_HEAD(, ptbl_buf) ptbl_buf_freelist;
static struct mtx ptbl_buf_freelist_lock;
/* Base address of kva space allocated fot ptbl bufs. */
static vm_offset_t ptbl_buf_pool_vabase;
/* Pointer to ptbl_buf structures. */
static struct ptbl_buf *ptbl_bufs;
void pmap_bootstrap_ap(volatile uint32_t *);
/*
* Kernel MMU interface
*/
static void mmu_booke_clear_modify(mmu_t, vm_page_t);
static void mmu_booke_copy(mmu_t, pmap_t, pmap_t, vm_offset_t,
vm_size_t, vm_offset_t);
static void mmu_booke_copy_page(mmu_t, vm_page_t, vm_page_t);
static void mmu_booke_copy_pages(mmu_t, vm_page_t *,
vm_offset_t, vm_page_t *, vm_offset_t, int);
static int mmu_booke_enter(mmu_t, pmap_t, vm_offset_t, vm_page_t,
vm_prot_t, u_int flags, int8_t psind);
static void mmu_booke_enter_object(mmu_t, pmap_t, vm_offset_t, vm_offset_t,
vm_page_t, vm_prot_t);
static void mmu_booke_enter_quick(mmu_t, pmap_t, vm_offset_t, vm_page_t,
vm_prot_t);
static vm_paddr_t mmu_booke_extract(mmu_t, pmap_t, vm_offset_t);
static vm_page_t mmu_booke_extract_and_hold(mmu_t, pmap_t, vm_offset_t,
vm_prot_t);
static void mmu_booke_init(mmu_t);
static boolean_t mmu_booke_is_modified(mmu_t, vm_page_t);
static boolean_t mmu_booke_is_prefaultable(mmu_t, pmap_t, vm_offset_t);
static boolean_t mmu_booke_is_referenced(mmu_t, vm_page_t);
static int mmu_booke_ts_referenced(mmu_t, vm_page_t);
static vm_offset_t mmu_booke_map(mmu_t, vm_offset_t *, vm_paddr_t, vm_paddr_t,
int);
static int mmu_booke_mincore(mmu_t, pmap_t, vm_offset_t,
vm_paddr_t *);
static void mmu_booke_object_init_pt(mmu_t, pmap_t, vm_offset_t,
vm_object_t, vm_pindex_t, vm_size_t);
static boolean_t mmu_booke_page_exists_quick(mmu_t, pmap_t, vm_page_t);
static void mmu_booke_page_init(mmu_t, vm_page_t);
static int mmu_booke_page_wired_mappings(mmu_t, vm_page_t);
static void mmu_booke_pinit(mmu_t, pmap_t);
static void mmu_booke_pinit0(mmu_t, pmap_t);
static void mmu_booke_protect(mmu_t, pmap_t, vm_offset_t, vm_offset_t,
vm_prot_t);
static void mmu_booke_qenter(mmu_t, vm_offset_t, vm_page_t *, int);
static void mmu_booke_qremove(mmu_t, vm_offset_t, int);
static void mmu_booke_release(mmu_t, pmap_t);
static void mmu_booke_remove(mmu_t, pmap_t, vm_offset_t, vm_offset_t);
static void mmu_booke_remove_all(mmu_t, vm_page_t);
static void mmu_booke_remove_write(mmu_t, vm_page_t);
static void mmu_booke_unwire(mmu_t, pmap_t, vm_offset_t, vm_offset_t);
static void mmu_booke_zero_page(mmu_t, vm_page_t);
static void mmu_booke_zero_page_area(mmu_t, vm_page_t, int, int);
static void mmu_booke_zero_page_idle(mmu_t, vm_page_t);
static void mmu_booke_activate(mmu_t, struct thread *);
static void mmu_booke_deactivate(mmu_t, struct thread *);
static void mmu_booke_bootstrap(mmu_t, vm_offset_t, vm_offset_t);
static void *mmu_booke_mapdev(mmu_t, vm_paddr_t, vm_size_t);
static void *mmu_booke_mapdev_attr(mmu_t, vm_paddr_t, vm_size_t, vm_memattr_t);
static void mmu_booke_unmapdev(mmu_t, vm_offset_t, vm_size_t);
static vm_paddr_t mmu_booke_kextract(mmu_t, vm_offset_t);
static void mmu_booke_kenter(mmu_t, vm_offset_t, vm_paddr_t);
static void mmu_booke_kenter_attr(mmu_t, vm_offset_t, vm_paddr_t, vm_memattr_t);
static void mmu_booke_kremove(mmu_t, vm_offset_t);
static boolean_t mmu_booke_dev_direct_mapped(mmu_t, vm_paddr_t, vm_size_t);
static void mmu_booke_sync_icache(mmu_t, pmap_t, vm_offset_t,
vm_size_t);
static void mmu_booke_dumpsys_map(mmu_t, vm_paddr_t pa, size_t,
void **);
static void mmu_booke_dumpsys_unmap(mmu_t, vm_paddr_t pa, size_t,
void *);
static void mmu_booke_scan_init(mmu_t);
static mmu_method_t mmu_booke_methods[] = {
/* pmap dispatcher interface */
MMUMETHOD(mmu_clear_modify, mmu_booke_clear_modify),
MMUMETHOD(mmu_copy, mmu_booke_copy),
MMUMETHOD(mmu_copy_page, mmu_booke_copy_page),
MMUMETHOD(mmu_copy_pages, mmu_booke_copy_pages),
MMUMETHOD(mmu_enter, mmu_booke_enter),
MMUMETHOD(mmu_enter_object, mmu_booke_enter_object),
MMUMETHOD(mmu_enter_quick, mmu_booke_enter_quick),
MMUMETHOD(mmu_extract, mmu_booke_extract),
MMUMETHOD(mmu_extract_and_hold, mmu_booke_extract_and_hold),
MMUMETHOD(mmu_init, mmu_booke_init),
MMUMETHOD(mmu_is_modified, mmu_booke_is_modified),
MMUMETHOD(mmu_is_prefaultable, mmu_booke_is_prefaultable),
MMUMETHOD(mmu_is_referenced, mmu_booke_is_referenced),
MMUMETHOD(mmu_ts_referenced, mmu_booke_ts_referenced),
MMUMETHOD(mmu_map, mmu_booke_map),
MMUMETHOD(mmu_mincore, mmu_booke_mincore),
MMUMETHOD(mmu_object_init_pt, mmu_booke_object_init_pt),
MMUMETHOD(mmu_page_exists_quick,mmu_booke_page_exists_quick),
MMUMETHOD(mmu_page_init, mmu_booke_page_init),
MMUMETHOD(mmu_page_wired_mappings, mmu_booke_page_wired_mappings),
MMUMETHOD(mmu_pinit, mmu_booke_pinit),
MMUMETHOD(mmu_pinit0, mmu_booke_pinit0),
MMUMETHOD(mmu_protect, mmu_booke_protect),
MMUMETHOD(mmu_qenter, mmu_booke_qenter),
MMUMETHOD(mmu_qremove, mmu_booke_qremove),
MMUMETHOD(mmu_release, mmu_booke_release),
MMUMETHOD(mmu_remove, mmu_booke_remove),
MMUMETHOD(mmu_remove_all, mmu_booke_remove_all),
MMUMETHOD(mmu_remove_write, mmu_booke_remove_write),
MMUMETHOD(mmu_sync_icache, mmu_booke_sync_icache),
MMUMETHOD(mmu_unwire, mmu_booke_unwire),
MMUMETHOD(mmu_zero_page, mmu_booke_zero_page),
MMUMETHOD(mmu_zero_page_area, mmu_booke_zero_page_area),
MMUMETHOD(mmu_zero_page_idle, mmu_booke_zero_page_idle),
MMUMETHOD(mmu_activate, mmu_booke_activate),
MMUMETHOD(mmu_deactivate, mmu_booke_deactivate),
/* Internal interfaces */
MMUMETHOD(mmu_bootstrap, mmu_booke_bootstrap),
MMUMETHOD(mmu_dev_direct_mapped,mmu_booke_dev_direct_mapped),
MMUMETHOD(mmu_mapdev, mmu_booke_mapdev),
MMUMETHOD(mmu_mapdev_attr, mmu_booke_mapdev_attr),
MMUMETHOD(mmu_kenter, mmu_booke_kenter),
MMUMETHOD(mmu_kenter_attr, mmu_booke_kenter_attr),
MMUMETHOD(mmu_kextract, mmu_booke_kextract),
/* MMUMETHOD(mmu_kremove, mmu_booke_kremove), */
MMUMETHOD(mmu_unmapdev, mmu_booke_unmapdev),
/* dumpsys() support */
MMUMETHOD(mmu_dumpsys_map, mmu_booke_dumpsys_map),
MMUMETHOD(mmu_dumpsys_unmap, mmu_booke_dumpsys_unmap),
MMUMETHOD(mmu_scan_init, mmu_booke_scan_init),
{ 0, 0 }
};
MMU_DEF(booke_mmu, MMU_TYPE_BOOKE, mmu_booke_methods, 0);
static __inline uint32_t
tlb_calc_wimg(vm_offset_t pa, vm_memattr_t ma)
{
uint32_t attrib;
int i;
if (ma != VM_MEMATTR_DEFAULT) {
switch (ma) {
case VM_MEMATTR_UNCACHEABLE:
return (PTE_I | PTE_G);
case VM_MEMATTR_WRITE_COMBINING:
case VM_MEMATTR_WRITE_BACK:
case VM_MEMATTR_PREFETCHABLE:
return (PTE_I);
case VM_MEMATTR_WRITE_THROUGH:
return (PTE_W | PTE_M);
}
}
/*
* Assume the page is cache inhibited and access is guarded unless
* it's in our available memory array.
*/
attrib = _TLB_ENTRY_IO;
for (i = 0; i < physmem_regions_sz; i++) {
if ((pa >= physmem_regions[i].mr_start) &&
(pa < (physmem_regions[i].mr_start +
physmem_regions[i].mr_size))) {
attrib = _TLB_ENTRY_MEM;
break;
}
}
return (attrib);
}
static inline void
tlb_miss_lock(void)
{
#ifdef SMP
struct pcpu *pc;
if (!smp_started)
return;
STAILQ_FOREACH(pc, &cpuhead, pc_allcpu) {
if (pc != pcpup) {
CTR3(KTR_PMAP, "%s: tlb miss LOCK of CPU=%d, "
"tlb_lock=%p", __func__, pc->pc_cpuid, pc->pc_booke_tlb_lock);
KASSERT((pc->pc_cpuid != PCPU_GET(cpuid)),
("tlb_miss_lock: tried to lock self"));
tlb_lock(pc->pc_booke_tlb_lock);
CTR1(KTR_PMAP, "%s: locked", __func__);
}
}
#endif
}
static inline void
tlb_miss_unlock(void)
{
#ifdef SMP
struct pcpu *pc;
if (!smp_started)
return;
STAILQ_FOREACH(pc, &cpuhead, pc_allcpu) {
if (pc != pcpup) {
CTR2(KTR_PMAP, "%s: tlb miss UNLOCK of CPU=%d",
__func__, pc->pc_cpuid);
tlb_unlock(pc->pc_booke_tlb_lock);
CTR1(KTR_PMAP, "%s: unlocked", __func__);
}
}
#endif
}
/* Return number of entries in TLB0. */
static __inline void
tlb0_get_tlbconf(void)
{
uint32_t tlb0_cfg;
tlb0_cfg = mfspr(SPR_TLB0CFG);
tlb0_entries = tlb0_cfg & TLBCFG_NENTRY_MASK;
tlb0_ways = (tlb0_cfg & TLBCFG_ASSOC_MASK) >> TLBCFG_ASSOC_SHIFT;
tlb0_entries_per_way = tlb0_entries / tlb0_ways;
}
/* Initialize pool of kva ptbl buffers. */
static void
ptbl_init(void)
{
int i;
CTR3(KTR_PMAP, "%s: s (ptbl_bufs = 0x%08x size 0x%08x)", __func__,
(uint32_t)ptbl_bufs, sizeof(struct ptbl_buf) * PTBL_BUFS);
CTR3(KTR_PMAP, "%s: s (ptbl_buf_pool_vabase = 0x%08x size = 0x%08x)",
__func__, ptbl_buf_pool_vabase, PTBL_BUFS * PTBL_PAGES * PAGE_SIZE);
mtx_init(&ptbl_buf_freelist_lock, "ptbl bufs lock", NULL, MTX_DEF);
TAILQ_INIT(&ptbl_buf_freelist);
for (i = 0; i < PTBL_BUFS; i++) {
ptbl_bufs[i].kva = ptbl_buf_pool_vabase + i * PTBL_PAGES * PAGE_SIZE;
TAILQ_INSERT_TAIL(&ptbl_buf_freelist, &ptbl_bufs[i], link);
}
}
/* Get a ptbl_buf from the freelist. */
static struct ptbl_buf *
ptbl_buf_alloc(void)
{
struct ptbl_buf *buf;
mtx_lock(&ptbl_buf_freelist_lock);
buf = TAILQ_FIRST(&ptbl_buf_freelist);
if (buf != NULL)
TAILQ_REMOVE(&ptbl_buf_freelist, buf, link);
mtx_unlock(&ptbl_buf_freelist_lock);
CTR2(KTR_PMAP, "%s: buf = %p", __func__, buf);
return (buf);
}
/* Return ptbl buff to free pool. */
static void
ptbl_buf_free(struct ptbl_buf *buf)
{
CTR2(KTR_PMAP, "%s: buf = %p", __func__, buf);
mtx_lock(&ptbl_buf_freelist_lock);
TAILQ_INSERT_TAIL(&ptbl_buf_freelist, buf, link);
mtx_unlock(&ptbl_buf_freelist_lock);
}
/*
* Search the list of allocated ptbl bufs and find on list of allocated ptbls
*/
static void
ptbl_free_pmap_ptbl(pmap_t pmap, pte_t *ptbl)
{
struct ptbl_buf *pbuf;
CTR2(KTR_PMAP, "%s: ptbl = %p", __func__, ptbl);
PMAP_LOCK_ASSERT(pmap, MA_OWNED);
TAILQ_FOREACH(pbuf, &pmap->pm_ptbl_list, link)
if (pbuf->kva == (vm_offset_t)ptbl) {
/* Remove from pmap ptbl buf list. */
TAILQ_REMOVE(&pmap->pm_ptbl_list, pbuf, link);
/* Free corresponding ptbl buf. */
ptbl_buf_free(pbuf);
break;
}
}
/* Allocate page table. */
static pte_t *
ptbl_alloc(mmu_t mmu, pmap_t pmap, unsigned int pdir_idx, boolean_t nosleep)
{
vm_page_t mtbl[PTBL_PAGES];
vm_page_t m;
struct ptbl_buf *pbuf;
unsigned int pidx;
pte_t *ptbl;
int i, j;
CTR4(KTR_PMAP, "%s: pmap = %p su = %d pdir_idx = %d", __func__, pmap,
(pmap == kernel_pmap), pdir_idx);
KASSERT((pdir_idx <= (VM_MAXUSER_ADDRESS / PDIR_SIZE)),
("ptbl_alloc: invalid pdir_idx"));
KASSERT((pmap->pm_pdir[pdir_idx] == NULL),
("pte_alloc: valid ptbl entry exists!"));
pbuf = ptbl_buf_alloc();
if (pbuf == NULL)
panic("pte_alloc: couldn't alloc kernel virtual memory");
ptbl = (pte_t *)pbuf->kva;
CTR2(KTR_PMAP, "%s: ptbl kva = %p", __func__, ptbl);
/* Allocate ptbl pages, this will sleep! */
for (i = 0; i < PTBL_PAGES; i++) {
pidx = (PTBL_PAGES * pdir_idx) + i;
while ((m = vm_page_alloc(NULL, pidx,
VM_ALLOC_NOOBJ | VM_ALLOC_WIRED)) == NULL) {
PMAP_UNLOCK(pmap);
rw_wunlock(&pvh_global_lock);
if (nosleep) {
ptbl_free_pmap_ptbl(pmap, ptbl);
for (j = 0; j < i; j++)
vm_page_free(mtbl[j]);
atomic_subtract_int(&vm_cnt.v_wire_count, i);
return (NULL);
}
VM_WAIT;
rw_wlock(&pvh_global_lock);
PMAP_LOCK(pmap);
}
mtbl[i] = m;
}
/* Map allocated pages into kernel_pmap. */
mmu_booke_qenter(mmu, (vm_offset_t)ptbl, mtbl, PTBL_PAGES);
/* Zero whole ptbl. */
bzero((caddr_t)ptbl, PTBL_PAGES * PAGE_SIZE);
/* Add pbuf to the pmap ptbl bufs list. */
TAILQ_INSERT_TAIL(&pmap->pm_ptbl_list, pbuf, link);
return (ptbl);
}
/* Free ptbl pages and invalidate pdir entry. */
static void
ptbl_free(mmu_t mmu, pmap_t pmap, unsigned int pdir_idx)
{
pte_t *ptbl;
vm_paddr_t pa;
vm_offset_t va;
vm_page_t m;
int i;
CTR4(KTR_PMAP, "%s: pmap = %p su = %d pdir_idx = %d", __func__, pmap,
(pmap == kernel_pmap), pdir_idx);
KASSERT((pdir_idx <= (VM_MAXUSER_ADDRESS / PDIR_SIZE)),
("ptbl_free: invalid pdir_idx"));
ptbl = pmap->pm_pdir[pdir_idx];
CTR2(KTR_PMAP, "%s: ptbl = %p", __func__, ptbl);
KASSERT((ptbl != NULL), ("ptbl_free: null ptbl"));
/*
* Invalidate the pdir entry as soon as possible, so that other CPUs
* don't attempt to look up the page tables we are releasing.
*/
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
pmap->pm_pdir[pdir_idx] = NULL;
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
for (i = 0; i < PTBL_PAGES; i++) {
va = ((vm_offset_t)ptbl + (i * PAGE_SIZE));
pa = pte_vatopa(mmu, kernel_pmap, va);
m = PHYS_TO_VM_PAGE(pa);
vm_page_free_zero(m);
atomic_subtract_int(&vm_cnt.v_wire_count, 1);
mmu_booke_kremove(mmu, va);
}
ptbl_free_pmap_ptbl(pmap, ptbl);
}
/*
* Decrement ptbl pages hold count and attempt to free ptbl pages.
* Called when removing pte entry from ptbl.
*
* Return 1 if ptbl pages were freed.
*/
static int
ptbl_unhold(mmu_t mmu, pmap_t pmap, unsigned int pdir_idx)
{
pte_t *ptbl;
vm_paddr_t pa;
vm_page_t m;
int i;
CTR4(KTR_PMAP, "%s: pmap = %p su = %d pdir_idx = %d", __func__, pmap,
(pmap == kernel_pmap), pdir_idx);
KASSERT((pdir_idx <= (VM_MAXUSER_ADDRESS / PDIR_SIZE)),
("ptbl_unhold: invalid pdir_idx"));
KASSERT((pmap != kernel_pmap),
("ptbl_unhold: unholding kernel ptbl!"));
ptbl = pmap->pm_pdir[pdir_idx];
//debugf("ptbl_unhold: ptbl = 0x%08x\n", (u_int32_t)ptbl);
KASSERT(((vm_offset_t)ptbl >= VM_MIN_KERNEL_ADDRESS),
("ptbl_unhold: non kva ptbl"));
/* decrement hold count */
for (i = 0; i < PTBL_PAGES; i++) {
pa = pte_vatopa(mmu, kernel_pmap,
(vm_offset_t)ptbl + (i * PAGE_SIZE));
m = PHYS_TO_VM_PAGE(pa);
m->wire_count--;
}
/*
* Free ptbl pages if there are no pte etries in this ptbl.
* wire_count has the same value for all ptbl pages, so check the last
* page.
*/
if (m->wire_count == 0) {
ptbl_free(mmu, pmap, pdir_idx);
//debugf("ptbl_unhold: e (freed ptbl)\n");
return (1);
}
return (0);
}
/*
* Increment hold count for ptbl pages. This routine is used when a new pte
* entry is being inserted into the ptbl.
*/
static void
ptbl_hold(mmu_t mmu, pmap_t pmap, unsigned int pdir_idx)
{
vm_paddr_t pa;
pte_t *ptbl;
vm_page_t m;
int i;
CTR3(KTR_PMAP, "%s: pmap = %p pdir_idx = %d", __func__, pmap,
pdir_idx);
KASSERT((pdir_idx <= (VM_MAXUSER_ADDRESS / PDIR_SIZE)),
("ptbl_hold: invalid pdir_idx"));
KASSERT((pmap != kernel_pmap),
("ptbl_hold: holding kernel ptbl!"));
ptbl = pmap->pm_pdir[pdir_idx];
KASSERT((ptbl != NULL), ("ptbl_hold: null ptbl"));
for (i = 0; i < PTBL_PAGES; i++) {
pa = pte_vatopa(mmu, kernel_pmap,
(vm_offset_t)ptbl + (i * PAGE_SIZE));
m = PHYS_TO_VM_PAGE(pa);
m->wire_count++;
}
}
/* Allocate pv_entry structure. */
pv_entry_t
pv_alloc(void)
{
pv_entry_t pv;
pv_entry_count++;
if (pv_entry_count > pv_entry_high_water)
pagedaemon_wakeup();
pv = uma_zalloc(pvzone, M_NOWAIT);
return (pv);
}
/* Free pv_entry structure. */
static __inline void
pv_free(pv_entry_t pve)
{
pv_entry_count--;
uma_zfree(pvzone, pve);
}
/* Allocate and initialize pv_entry structure. */
static void
pv_insert(pmap_t pmap, vm_offset_t va, vm_page_t m)
{
pv_entry_t pve;
//int su = (pmap == kernel_pmap);
//debugf("pv_insert: s (su = %d pmap = 0x%08x va = 0x%08x m = 0x%08x)\n", su,
// (u_int32_t)pmap, va, (u_int32_t)m);
pve = pv_alloc();
if (pve == NULL)
panic("pv_insert: no pv entries!");
pve->pv_pmap = pmap;
pve->pv_va = va;
/* add to pv_list */
PMAP_LOCK_ASSERT(pmap, MA_OWNED);
rw_assert(&pvh_global_lock, RA_WLOCKED);
TAILQ_INSERT_TAIL(&m->md.pv_list, pve, pv_link);
//debugf("pv_insert: e\n");
}
/* Destroy pv entry. */
static void
pv_remove(pmap_t pmap, vm_offset_t va, vm_page_t m)
{
pv_entry_t pve;
//int su = (pmap == kernel_pmap);
//debugf("pv_remove: s (su = %d pmap = 0x%08x va = 0x%08x)\n", su, (u_int32_t)pmap, va);
PMAP_LOCK_ASSERT(pmap, MA_OWNED);
rw_assert(&pvh_global_lock, RA_WLOCKED);
/* find pv entry */
TAILQ_FOREACH(pve, &m->md.pv_list, pv_link) {
if ((pmap == pve->pv_pmap) && (va == pve->pv_va)) {
/* remove from pv_list */
TAILQ_REMOVE(&m->md.pv_list, pve, pv_link);
if (TAILQ_EMPTY(&m->md.pv_list))
vm_page_aflag_clear(m, PGA_WRITEABLE);
/* free pv entry struct */
pv_free(pve);
break;
}
}
//debugf("pv_remove: e\n");
}
/*
* Clean pte entry, try to free page table page if requested.
*
* Return 1 if ptbl pages were freed, otherwise return 0.
*/
static int
pte_remove(mmu_t mmu, pmap_t pmap, vm_offset_t va, uint8_t flags)
{
unsigned int pdir_idx = PDIR_IDX(va);
unsigned int ptbl_idx = PTBL_IDX(va);
vm_page_t m;
pte_t *ptbl;
pte_t *pte;
//int su = (pmap == kernel_pmap);
//debugf("pte_remove: s (su = %d pmap = 0x%08x va = 0x%08x flags = %d)\n",
// su, (u_int32_t)pmap, va, flags);
ptbl = pmap->pm_pdir[pdir_idx];
KASSERT(ptbl, ("pte_remove: null ptbl"));
pte = &ptbl[ptbl_idx];
if (pte == NULL || !PTE_ISVALID(pte))
return (0);
if (PTE_ISWIRED(pte))
pmap->pm_stats.wired_count--;
/* Handle managed entry. */
if (PTE_ISMANAGED(pte)) {
/* Get vm_page_t for mapped pte. */
m = PHYS_TO_VM_PAGE(PTE_PA(pte));
if (PTE_ISMODIFIED(pte))
vm_page_dirty(m);
if (PTE_ISREFERENCED(pte))
vm_page_aflag_set(m, PGA_REFERENCED);
pv_remove(pmap, va, m);
}
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
tlb0_flush_entry(va);
pte->flags = 0;
pte->rpn = 0;
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
pmap->pm_stats.resident_count--;
if (flags & PTBL_UNHOLD) {
//debugf("pte_remove: e (unhold)\n");
return (ptbl_unhold(mmu, pmap, pdir_idx));
}
//debugf("pte_remove: e\n");
return (0);
}
/*
* Insert PTE for a given page and virtual address.
*/
static int
pte_enter(mmu_t mmu, pmap_t pmap, vm_page_t m, vm_offset_t va, uint32_t flags,
boolean_t nosleep)
{
unsigned int pdir_idx = PDIR_IDX(va);
unsigned int ptbl_idx = PTBL_IDX(va);
pte_t *ptbl, *pte;
CTR4(KTR_PMAP, "%s: su = %d pmap = %p va = %p", __func__,
pmap == kernel_pmap, pmap, va);
/* Get the page table pointer. */
ptbl = pmap->pm_pdir[pdir_idx];
if (ptbl == NULL) {
/* Allocate page table pages. */
ptbl = ptbl_alloc(mmu, pmap, pdir_idx, nosleep);
if (ptbl == NULL) {
KASSERT(nosleep, ("nosleep and NULL ptbl"));
return (ENOMEM);
}
} else {
/*
* Check if there is valid mapping for requested
* va, if there is, remove it.
*/
pte = &pmap->pm_pdir[pdir_idx][ptbl_idx];
if (PTE_ISVALID(pte)) {
pte_remove(mmu, pmap, va, PTBL_HOLD);
} else {
/*
* pte is not used, increment hold count
* for ptbl pages.
*/
if (pmap != kernel_pmap)
ptbl_hold(mmu, pmap, pdir_idx);
}
}
/*
* Insert pv_entry into pv_list for mapped page if part of managed
* memory.
*/
if ((m->oflags & VPO_UNMANAGED) == 0) {
flags |= PTE_MANAGED;
/* Create and insert pv entry. */
pv_insert(pmap, va, m);
}
pmap->pm_stats.resident_count++;
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
tlb0_flush_entry(va);
if (pmap->pm_pdir[pdir_idx] == NULL) {
/*
* If we just allocated a new page table, hook it in
* the pdir.
*/
pmap->pm_pdir[pdir_idx] = ptbl;
}
pte = &(pmap->pm_pdir[pdir_idx][ptbl_idx]);
pte->rpn = VM_PAGE_TO_PHYS(m) & ~PTE_PA_MASK;
pte->flags |= (PTE_VALID | flags);
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
return (0);
}
/* Return the pa for the given pmap/va. */
static vm_paddr_t
pte_vatopa(mmu_t mmu, pmap_t pmap, vm_offset_t va)
{
vm_paddr_t pa = 0;
pte_t *pte;
pte = pte_find(mmu, pmap, va);
if ((pte != NULL) && PTE_ISVALID(pte))
pa = (PTE_PA(pte) | (va & PTE_PA_MASK));
return (pa);
}
/* Get a pointer to a PTE in a page table. */
static pte_t *
pte_find(mmu_t mmu, pmap_t pmap, vm_offset_t va)
{
unsigned int pdir_idx = PDIR_IDX(va);
unsigned int ptbl_idx = PTBL_IDX(va);
KASSERT((pmap != NULL), ("pte_find: invalid pmap"));
if (pmap->pm_pdir[pdir_idx])
return (&(pmap->pm_pdir[pdir_idx][ptbl_idx]));
return (NULL);
}
/**************************************************************************/
/* PMAP related */
/**************************************************************************/
/*
* This is called during booke_init, before the system is really initialized.
*/
static void
mmu_booke_bootstrap(mmu_t mmu, vm_offset_t start, vm_offset_t kernelend)
{
vm_offset_t phys_kernelend;
struct mem_region *mp, *mp1;
int cnt, i, j;
u_int s, e, sz;
u_int phys_avail_count;
vm_size_t physsz, hwphyssz, kstack0_sz;
vm_offset_t kernel_pdir, kstack0, va;
vm_paddr_t kstack0_phys;
void *dpcpu;
pte_t *pte;
debugf("mmu_booke_bootstrap: entered\n");
/* Set interesting system properties */
hw_direct_map = 0;
elf32_nxstack = 1;
/* Initialize invalidation mutex */
mtx_init(&tlbivax_mutex, "tlbivax", NULL, MTX_SPIN);
/* Read TLB0 size and associativity. */
tlb0_get_tlbconf();
/*
* Align kernel start and end address (kernel image).
* Note that kernel end does not necessarily relate to kernsize.
* kernsize is the size of the kernel that is actually mapped.
* Also note that "start - 1" is deliberate. With SMP, the
* entry point is exactly a page from the actual load address.
* As such, trunc_page() has no effect and we're off by a page.
* Since we always have the ELF header between the load address
* and the entry point, we can safely subtract 1 to compensate.
*/
kernstart = trunc_page(start - 1);
data_start = round_page(kernelend);
data_end = data_start;
/*
* Addresses of preloaded modules (like file systems) use
* physical addresses. Make sure we relocate those into
* virtual addresses.
*/
preload_addr_relocate = kernstart - kernload;
/* Allocate the dynamic per-cpu area. */
dpcpu = (void *)data_end;
data_end += DPCPU_SIZE;
/* Allocate space for the message buffer. */
msgbufp = (struct msgbuf *)data_end;
data_end += msgbufsize;
debugf(" msgbufp at 0x%08x end = 0x%08x\n", (uint32_t)msgbufp,
data_end);
data_end = round_page(data_end);
/* Allocate space for ptbl_bufs. */
ptbl_bufs = (struct ptbl_buf *)data_end;
data_end += sizeof(struct ptbl_buf) * PTBL_BUFS;
debugf(" ptbl_bufs at 0x%08x end = 0x%08x\n", (uint32_t)ptbl_bufs,
data_end);
data_end = round_page(data_end);
/* Allocate PTE tables for kernel KVA. */
kernel_pdir = data_end;
kernel_ptbls = (VM_MAX_KERNEL_ADDRESS - VM_MIN_KERNEL_ADDRESS +
PDIR_SIZE - 1) / PDIR_SIZE;
data_end += kernel_ptbls * PTBL_PAGES * PAGE_SIZE;
debugf(" kernel ptbls: %d\n", kernel_ptbls);
debugf(" kernel pdir at 0x%08x end = 0x%08x\n", kernel_pdir, data_end);
debugf(" data_end: 0x%08x\n", data_end);
if (data_end - kernstart > kernsize) {
kernsize += tlb1_mapin_region(kernstart + kernsize,
kernload + kernsize, (data_end - kernstart) - kernsize);
}
data_end = kernstart + kernsize;
debugf(" updated data_end: 0x%08x\n", data_end);
/*
* Clear the structures - note we can only do it safely after the
* possible additional TLB1 translations are in place (above) so that
* all range up to the currently calculated 'data_end' is covered.
*/
dpcpu_init(dpcpu, 0);
memset((void *)ptbl_bufs, 0, sizeof(struct ptbl_buf) * PTBL_SIZE);
memset((void *)kernel_pdir, 0, kernel_ptbls * PTBL_PAGES * PAGE_SIZE);
/*******************************************************/
/* Set the start and end of kva. */
/*******************************************************/
virtual_avail = round_page(data_end);
virtual_end = VM_MAX_KERNEL_ADDRESS;
/* Allocate KVA space for page zero/copy operations. */
zero_page_va = virtual_avail;
virtual_avail += PAGE_SIZE;
zero_page_idle_va = virtual_avail;
virtual_avail += PAGE_SIZE;
copy_page_src_va = virtual_avail;
virtual_avail += PAGE_SIZE;
copy_page_dst_va = virtual_avail;
virtual_avail += PAGE_SIZE;
debugf("zero_page_va = 0x%08x\n", zero_page_va);
debugf("zero_page_idle_va = 0x%08x\n", zero_page_idle_va);
debugf("copy_page_src_va = 0x%08x\n", copy_page_src_va);
debugf("copy_page_dst_va = 0x%08x\n", copy_page_dst_va);
/* Initialize page zero/copy mutexes. */
mtx_init(&zero_page_mutex, "mmu_booke_zero_page", NULL, MTX_DEF);
mtx_init(©_page_mutex, "mmu_booke_copy_page", NULL, MTX_DEF);
/* Allocate KVA space for ptbl bufs. */
ptbl_buf_pool_vabase = virtual_avail;
virtual_avail += PTBL_BUFS * PTBL_PAGES * PAGE_SIZE;
debugf("ptbl_buf_pool_vabase = 0x%08x end = 0x%08x\n",
ptbl_buf_pool_vabase, virtual_avail);
/* Calculate corresponding physical addresses for the kernel region. */
phys_kernelend = kernload + kernsize;
debugf("kernel image and allocated data:\n");
debugf(" kernload = 0x%08x\n", kernload);
debugf(" kernstart = 0x%08x\n", kernstart);
debugf(" kernsize = 0x%08x\n", kernsize);
if (sizeof(phys_avail) / sizeof(phys_avail[0]) < availmem_regions_sz)
panic("mmu_booke_bootstrap: phys_avail too small");
/*
* Remove kernel physical address range from avail regions list. Page
* align all regions. Non-page aligned memory isn't very interesting
* to us. Also, sort the entries for ascending addresses.
*/
/* Retrieve phys/avail mem regions */
mem_regions(&physmem_regions, &physmem_regions_sz,
&availmem_regions, &availmem_regions_sz);
sz = 0;
cnt = availmem_regions_sz;
debugf("processing avail regions:\n");
for (mp = availmem_regions; mp->mr_size; mp++) {
s = mp->mr_start;
e = mp->mr_start + mp->mr_size;
debugf(" %08x-%08x -> ", s, e);
/* Check whether this region holds all of the kernel. */
if (s < kernload && e > phys_kernelend) {
availmem_regions[cnt].mr_start = phys_kernelend;
availmem_regions[cnt++].mr_size = e - phys_kernelend;
e = kernload;
}
/* Look whether this regions starts within the kernel. */
if (s >= kernload && s < phys_kernelend) {
if (e <= phys_kernelend)
goto empty;
s = phys_kernelend;
}
/* Now look whether this region ends within the kernel. */
if (e > kernload && e <= phys_kernelend) {
if (s >= kernload)
goto empty;
e = kernload;
}
/* Now page align the start and size of the region. */
s = round_page(s);
e = trunc_page(e);
if (e < s)
e = s;
sz = e - s;
debugf("%08x-%08x = %x\n", s, e, sz);
/* Check whether some memory is left here. */
if (sz == 0) {
empty:
memmove(mp, mp + 1,
(cnt - (mp - availmem_regions)) * sizeof(*mp));
cnt--;
mp--;
continue;
}
/* Do an insertion sort. */
for (mp1 = availmem_regions; mp1 < mp; mp1++)
if (s < mp1->mr_start)
break;
if (mp1 < mp) {
memmove(mp1 + 1, mp1, (char *)mp - (char *)mp1);
mp1->mr_start = s;
mp1->mr_size = sz;
} else {
mp->mr_start = s;
mp->mr_size = sz;
}
}
availmem_regions_sz = cnt;
/*******************************************************/
/* Steal physical memory for kernel stack from the end */
/* of the first avail region */
/*******************************************************/
kstack0_sz = KSTACK_PAGES * PAGE_SIZE;
kstack0_phys = availmem_regions[0].mr_start +
availmem_regions[0].mr_size;
kstack0_phys -= kstack0_sz;
availmem_regions[0].mr_size -= kstack0_sz;
/*******************************************************/
/* Fill in phys_avail table, based on availmem_regions */
/*******************************************************/
phys_avail_count = 0;
physsz = 0;
hwphyssz = 0;
TUNABLE_ULONG_FETCH("hw.physmem", (u_long *) &hwphyssz);
debugf("fill in phys_avail:\n");
for (i = 0, j = 0; i < availmem_regions_sz; i++, j += 2) {
debugf(" region: 0x%08x - 0x%08x (0x%08x)\n",
availmem_regions[i].mr_start,
availmem_regions[i].mr_start +
availmem_regions[i].mr_size,
availmem_regions[i].mr_size);
if (hwphyssz != 0 &&
(physsz + availmem_regions[i].mr_size) >= hwphyssz) {
debugf(" hw.physmem adjust\n");
if (physsz < hwphyssz) {
phys_avail[j] = availmem_regions[i].mr_start;
phys_avail[j + 1] =
availmem_regions[i].mr_start +
hwphyssz - physsz;
physsz = hwphyssz;
phys_avail_count++;
}
break;
}
phys_avail[j] = availmem_regions[i].mr_start;
phys_avail[j + 1] = availmem_regions[i].mr_start +
availmem_regions[i].mr_size;
phys_avail_count++;
physsz += availmem_regions[i].mr_size;
}
physmem = btoc(physsz);
/* Calculate the last available physical address. */
for (i = 0; phys_avail[i + 2] != 0; i += 2)
;
Maxmem = powerpc_btop(phys_avail[i + 1]);
debugf("Maxmem = 0x%08lx\n", Maxmem);
debugf("phys_avail_count = %d\n", phys_avail_count);
debugf("physsz = 0x%08x physmem = %ld (0x%08lx)\n", physsz, physmem,
physmem);
/*******************************************************/
/* Initialize (statically allocated) kernel pmap. */
/*******************************************************/
PMAP_LOCK_INIT(kernel_pmap);
kptbl_min = VM_MIN_KERNEL_ADDRESS / PDIR_SIZE;
debugf("kernel_pmap = 0x%08x\n", (uint32_t)kernel_pmap);
debugf("kptbl_min = %d, kernel_ptbls = %d\n", kptbl_min, kernel_ptbls);
debugf("kernel pdir range: 0x%08x - 0x%08x\n",
kptbl_min * PDIR_SIZE, (kptbl_min + kernel_ptbls) * PDIR_SIZE - 1);
/* Initialize kernel pdir */
for (i = 0; i < kernel_ptbls; i++)
kernel_pmap->pm_pdir[kptbl_min + i] =
(pte_t *)(kernel_pdir + (i * PAGE_SIZE * PTBL_PAGES));
for (i = 0; i < MAXCPU; i++) {
kernel_pmap->pm_tid[i] = TID_KERNEL;
/* Initialize each CPU's tidbusy entry 0 with kernel_pmap */
tidbusy[i][0] = kernel_pmap;
}
/*
* Fill in PTEs covering kernel code and data. They are not required
* for address translation, as this area is covered by static TLB1
* entries, but for pte_vatopa() to work correctly with kernel area
* addresses.
*/
for (va = kernstart; va < data_end; va += PAGE_SIZE) {
pte = &(kernel_pmap->pm_pdir[PDIR_IDX(va)][PTBL_IDX(va)]);
pte->rpn = kernload + (va - kernstart);
pte->flags = PTE_M | PTE_SR | PTE_SW | PTE_SX | PTE_WIRED |
PTE_VALID;
}
/* Mark kernel_pmap active on all CPUs */
CPU_FILL(&kernel_pmap->pm_active);
/*
* Initialize the global pv list lock.
*/
rw_init(&pvh_global_lock, "pmap pv global");
/*******************************************************/
/* Final setup */
/*******************************************************/
/* Enter kstack0 into kernel map, provide guard page */
kstack0 = virtual_avail + KSTACK_GUARD_PAGES * PAGE_SIZE;
thread0.td_kstack = kstack0;
thread0.td_kstack_pages = KSTACK_PAGES;
debugf("kstack_sz = 0x%08x\n", kstack0_sz);
debugf("kstack0_phys at 0x%08x - 0x%08x\n",
kstack0_phys, kstack0_phys + kstack0_sz);
debugf("kstack0 at 0x%08x - 0x%08x\n", kstack0, kstack0 + kstack0_sz);
virtual_avail += KSTACK_GUARD_PAGES * PAGE_SIZE + kstack0_sz;
for (i = 0; i < KSTACK_PAGES; i++) {
mmu_booke_kenter(mmu, kstack0, kstack0_phys);
kstack0 += PAGE_SIZE;
kstack0_phys += PAGE_SIZE;
}
pmap_bootstrapped = 1;
debugf("virtual_avail = %08x\n", virtual_avail);
debugf("virtual_end = %08x\n", virtual_end);
debugf("mmu_booke_bootstrap: exit\n");
}
void
pmap_bootstrap_ap(volatile uint32_t *trcp __unused)
{
int i;
/*
* Finish TLB1 configuration: the BSP already set up its TLB1 and we
* have the snapshot of its contents in the s/w tlb1[] table, so use
* these values directly to (re)program AP's TLB1 hardware.
*/
for (i = bp_ntlb1s; i < tlb1_idx; i++) {
/* Skip invalid entries */
if (!(tlb1[i].mas1 & MAS1_VALID))
continue;
tlb1_write_entry(i);
}
set_mas4_defaults();
}
/*
* Get the physical page address for the given pmap/virtual address.
*/
static vm_paddr_t
mmu_booke_extract(mmu_t mmu, pmap_t pmap, vm_offset_t va)
{
vm_paddr_t pa;
PMAP_LOCK(pmap);
pa = pte_vatopa(mmu, pmap, va);
PMAP_UNLOCK(pmap);
return (pa);
}
/*
* Extract the physical page address associated with the given
* kernel virtual address.
*/
static vm_paddr_t
mmu_booke_kextract(mmu_t mmu, vm_offset_t va)
{
int i;
/* Check TLB1 mappings */
for (i = 0; i < tlb1_idx; i++) {
if (!(tlb1[i].mas1 & MAS1_VALID))
continue;
if (va >= tlb1[i].virt && va < tlb1[i].virt + tlb1[i].size)
return (tlb1[i].phys + (va - tlb1[i].virt));
}
return (pte_vatopa(mmu, kernel_pmap, va));
}
/*
* Initialize the pmap module.
* Called by vm_init, to initialize any structures that the pmap
* system needs to map virtual memory.
*/
static void
mmu_booke_init(mmu_t mmu)
{
int shpgperproc = PMAP_SHPGPERPROC;
/*
* Initialize the address space (zone) for the pv entries. Set a
* high water mark so that the system can recover from excessive
* numbers of pv entries.
*/
pvzone = uma_zcreate("PV ENTRY", sizeof(struct pv_entry), NULL, NULL,
NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_VM | UMA_ZONE_NOFREE);
TUNABLE_INT_FETCH("vm.pmap.shpgperproc", &shpgperproc);
pv_entry_max = shpgperproc * maxproc + vm_cnt.v_page_count;
TUNABLE_INT_FETCH("vm.pmap.pv_entries", &pv_entry_max);
pv_entry_high_water = 9 * (pv_entry_max / 10);
uma_zone_reserve_kva(pvzone, pv_entry_max);
/* Pre-fill pvzone with initial number of pv entries. */
uma_prealloc(pvzone, PV_ENTRY_ZONE_MIN);
/* Initialize ptbl allocation. */
ptbl_init();
}
/*
* Map a list of wired pages into kernel virtual address space. This is
* intended for temporary mappings which do not need page modification or
* references recorded. Existing mappings in the region are overwritten.
*/
static void
mmu_booke_qenter(mmu_t mmu, vm_offset_t sva, vm_page_t *m, int count)
{
vm_offset_t va;
va = sva;
while (count-- > 0) {
mmu_booke_kenter(mmu, va, VM_PAGE_TO_PHYS(*m));
va += PAGE_SIZE;
m++;
}
}
/*
* Remove page mappings from kernel virtual address space. Intended for
* temporary mappings entered by mmu_booke_qenter.
*/
static void
mmu_booke_qremove(mmu_t mmu, vm_offset_t sva, int count)
{
vm_offset_t va;
va = sva;
while (count-- > 0) {
mmu_booke_kremove(mmu, va);
va += PAGE_SIZE;
}
}
/*
* Map a wired page into kernel virtual address space.
*/
static void
mmu_booke_kenter(mmu_t mmu, vm_offset_t va, vm_paddr_t pa)
{
mmu_booke_kenter_attr(mmu, va, pa, VM_MEMATTR_DEFAULT);
}
static void
mmu_booke_kenter_attr(mmu_t mmu, vm_offset_t va, vm_paddr_t pa, vm_memattr_t ma)
{
unsigned int pdir_idx = PDIR_IDX(va);
unsigned int ptbl_idx = PTBL_IDX(va);
uint32_t flags;
pte_t *pte;
KASSERT(((va >= VM_MIN_KERNEL_ADDRESS) &&
(va <= VM_MAX_KERNEL_ADDRESS)), ("mmu_booke_kenter: invalid va"));
flags = PTE_SR | PTE_SW | PTE_SX | PTE_WIRED | PTE_VALID;
flags |= tlb_calc_wimg(pa, ma);
pte = &(kernel_pmap->pm_pdir[pdir_idx][ptbl_idx]);
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
if (PTE_ISVALID(pte)) {
CTR1(KTR_PMAP, "%s: replacing entry!", __func__);
/* Flush entry from TLB0 */
tlb0_flush_entry(va);
}
pte->rpn = pa & ~PTE_PA_MASK;
pte->flags = flags;
//debugf("mmu_booke_kenter: pdir_idx = %d ptbl_idx = %d va=0x%08x "
// "pa=0x%08x rpn=0x%08x flags=0x%08x\n",
// pdir_idx, ptbl_idx, va, pa, pte->rpn, pte->flags);
/* Flush the real memory from the instruction cache. */
if ((flags & (PTE_I | PTE_G)) == 0) {
__syncicache((void *)va, PAGE_SIZE);
}
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
}
/*
* Remove a page from kernel page table.
*/
static void
mmu_booke_kremove(mmu_t mmu, vm_offset_t va)
{
unsigned int pdir_idx = PDIR_IDX(va);
unsigned int ptbl_idx = PTBL_IDX(va);
pte_t *pte;
// CTR2(KTR_PMAP,("%s: s (va = 0x%08x)\n", __func__, va));
KASSERT(((va >= VM_MIN_KERNEL_ADDRESS) &&
(va <= VM_MAX_KERNEL_ADDRESS)),
("mmu_booke_kremove: invalid va"));
pte = &(kernel_pmap->pm_pdir[pdir_idx][ptbl_idx]);
if (!PTE_ISVALID(pte)) {
CTR1(KTR_PMAP, "%s: invalid pte", __func__);
return;
}
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
/* Invalidate entry in TLB0, update PTE. */
tlb0_flush_entry(va);
pte->flags = 0;
pte->rpn = 0;
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
}
/*
* Initialize pmap associated with process 0.
*/
static void
mmu_booke_pinit0(mmu_t mmu, pmap_t pmap)
{
PMAP_LOCK_INIT(pmap);
mmu_booke_pinit(mmu, pmap);
PCPU_SET(curpmap, pmap);
}
/*
* Initialize a preallocated and zeroed pmap structure,
* such as one in a vmspace structure.
*/
static void
mmu_booke_pinit(mmu_t mmu, pmap_t pmap)
{
int i;
CTR4(KTR_PMAP, "%s: pmap = %p, proc %d '%s'", __func__, pmap,
curthread->td_proc->p_pid, curthread->td_proc->p_comm);
KASSERT((pmap != kernel_pmap), ("pmap_pinit: initializing kernel_pmap"));
for (i = 0; i < MAXCPU; i++)
pmap->pm_tid[i] = TID_NONE;
CPU_ZERO(&kernel_pmap->pm_active);
bzero(&pmap->pm_stats, sizeof(pmap->pm_stats));
bzero(&pmap->pm_pdir, sizeof(pte_t *) * PDIR_NENTRIES);
TAILQ_INIT(&pmap->pm_ptbl_list);
}
/*
* Release any resources held by the given physical map.
* Called when a pmap initialized by mmu_booke_pinit is being released.
* Should only be called if the map contains no valid mappings.
*/
static void
mmu_booke_release(mmu_t mmu, pmap_t pmap)
{
KASSERT(pmap->pm_stats.resident_count == 0,
("pmap_release: pmap resident count %ld != 0",
pmap->pm_stats.resident_count));
}
/*
* Insert the given physical page at the specified virtual address in the
* target physical map with the protection requested. If specified the page
* will be wired down.
*/
static int
mmu_booke_enter(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_page_t m,
vm_prot_t prot, u_int flags, int8_t psind)
{
int error;
rw_wlock(&pvh_global_lock);
PMAP_LOCK(pmap);
error = mmu_booke_enter_locked(mmu, pmap, va, m, prot, flags, psind);
rw_wunlock(&pvh_global_lock);
PMAP_UNLOCK(pmap);
return (error);
}
static int
mmu_booke_enter_locked(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_page_t m,
vm_prot_t prot, u_int pmap_flags, int8_t psind __unused)
{
pte_t *pte;
vm_paddr_t pa;
uint32_t flags;
int error, su, sync;
pa = VM_PAGE_TO_PHYS(m);
su = (pmap == kernel_pmap);
sync = 0;
//debugf("mmu_booke_enter_locked: s (pmap=0x%08x su=%d tid=%d m=0x%08x va=0x%08x "
// "pa=0x%08x prot=0x%08x flags=%#x)\n",
// (u_int32_t)pmap, su, pmap->pm_tid,
// (u_int32_t)m, va, pa, prot, flags);
if (su) {
KASSERT(((va >= virtual_avail) &&
(va <= VM_MAX_KERNEL_ADDRESS)),
("mmu_booke_enter_locked: kernel pmap, non kernel va"));
} else {
KASSERT((va <= VM_MAXUSER_ADDRESS),
("mmu_booke_enter_locked: user pmap, non user va"));
}
if ((m->oflags & VPO_UNMANAGED) == 0 && !vm_page_xbusied(m))
VM_OBJECT_ASSERT_LOCKED(m->object);
PMAP_LOCK_ASSERT(pmap, MA_OWNED);
/*
* If there is an existing mapping, and the physical address has not
* changed, must be protection or wiring change.
*/
if (((pte = pte_find(mmu, pmap, va)) != NULL) &&
(PTE_ISVALID(pte)) && (PTE_PA(pte) == pa)) {
/*
* Before actually updating pte->flags we calculate and
* prepare its new value in a helper var.
*/
flags = pte->flags;
flags &= ~(PTE_UW | PTE_UX | PTE_SW | PTE_SX | PTE_MODIFIED);
/* Wiring change, just update stats. */
if ((pmap_flags & PMAP_ENTER_WIRED) != 0) {
if (!PTE_ISWIRED(pte)) {
flags |= PTE_WIRED;
pmap->pm_stats.wired_count++;
}
} else {
if (PTE_ISWIRED(pte)) {
flags &= ~PTE_WIRED;
pmap->pm_stats.wired_count--;
}
}
if (prot & VM_PROT_WRITE) {
/* Add write permissions. */
flags |= PTE_SW;
if (!su)
flags |= PTE_UW;
if ((flags & PTE_MANAGED) != 0)
vm_page_aflag_set(m, PGA_WRITEABLE);
} else {
/* Handle modified pages, sense modify status. */
/*
* The PTE_MODIFIED flag could be set by underlying
* TLB misses since we last read it (above), possibly
* other CPUs could update it so we check in the PTE
* directly rather than rely on that saved local flags
* copy.
*/
if (PTE_ISMODIFIED(pte))
vm_page_dirty(m);
}
if (prot & VM_PROT_EXECUTE) {
flags |= PTE_SX;
if (!su)
flags |= PTE_UX;
/*
* Check existing flags for execute permissions: if we
* are turning execute permissions on, icache should
* be flushed.
*/
if ((pte->flags & (PTE_UX | PTE_SX)) == 0)
sync++;
}
flags &= ~PTE_REFERENCED;
/*
* The new flags value is all calculated -- only now actually
* update the PTE.
*/
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
tlb0_flush_entry(va);
pte->flags = flags;
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
} else {
/*
* If there is an existing mapping, but it's for a different
* physical address, pte_enter() will delete the old mapping.
*/
//if ((pte != NULL) && PTE_ISVALID(pte))
// debugf("mmu_booke_enter_locked: replace\n");
//else
// debugf("mmu_booke_enter_locked: new\n");
/* Now set up the flags and install the new mapping. */
flags = (PTE_SR | PTE_VALID);
flags |= PTE_M;
if (!su)
flags |= PTE_UR;
if (prot & VM_PROT_WRITE) {
flags |= PTE_SW;
if (!su)
flags |= PTE_UW;
if ((m->oflags & VPO_UNMANAGED) == 0)
vm_page_aflag_set(m, PGA_WRITEABLE);
}
if (prot & VM_PROT_EXECUTE) {
flags |= PTE_SX;
if (!su)
flags |= PTE_UX;
}
/* If its wired update stats. */
if ((pmap_flags & PMAP_ENTER_WIRED) != 0)
flags |= PTE_WIRED;
error = pte_enter(mmu, pmap, m, va, flags,
(pmap_flags & PMAP_ENTER_NOSLEEP) != 0);
if (error != 0)
return (KERN_RESOURCE_SHORTAGE);
if ((flags & PMAP_ENTER_WIRED) != 0)
pmap->pm_stats.wired_count++;
/* Flush the real memory from the instruction cache. */
if (prot & VM_PROT_EXECUTE)
sync++;
}
if (sync && (su || pmap == PCPU_GET(curpmap))) {
__syncicache((void *)va, PAGE_SIZE);
sync = 0;
}
return (KERN_SUCCESS);
}
/*
* Maps a sequence of resident pages belonging to the same object.
* The sequence begins with the given page m_start. This page is
* mapped at the given virtual address start. Each subsequent page is
* mapped at a virtual address that is offset from start by the same
* amount as the page is offset from m_start within the object. The
* last page in the sequence is the page with the largest offset from
* m_start that can be mapped at a virtual address less than the given
* virtual address end. Not every virtual page between start and end
* is mapped; only those for which a resident page exists with the
* corresponding offset from m_start are mapped.
*/
static void
mmu_booke_enter_object(mmu_t mmu, pmap_t pmap, vm_offset_t start,
vm_offset_t end, vm_page_t m_start, vm_prot_t prot)
{
vm_page_t m;
vm_pindex_t diff, psize;
VM_OBJECT_ASSERT_LOCKED(m_start->object);
psize = atop(end - start);
m = m_start;
rw_wlock(&pvh_global_lock);
PMAP_LOCK(pmap);
while (m != NULL && (diff = m->pindex - m_start->pindex) < psize) {
mmu_booke_enter_locked(mmu, pmap, start + ptoa(diff), m,
prot & (VM_PROT_READ | VM_PROT_EXECUTE),
PMAP_ENTER_NOSLEEP, 0);
m = TAILQ_NEXT(m, listq);
}
rw_wunlock(&pvh_global_lock);
PMAP_UNLOCK(pmap);
}
static void
mmu_booke_enter_quick(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_page_t m,
vm_prot_t prot)
{
rw_wlock(&pvh_global_lock);
PMAP_LOCK(pmap);
mmu_booke_enter_locked(mmu, pmap, va, m,
prot & (VM_PROT_READ | VM_PROT_EXECUTE), PMAP_ENTER_NOSLEEP,
0);
rw_wunlock(&pvh_global_lock);
PMAP_UNLOCK(pmap);
}
/*
* Remove the given range of addresses from the specified map.
*
* It is assumed that the start and end are properly rounded to the page size.
*/
static void
mmu_booke_remove(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_offset_t endva)
{
pte_t *pte;
uint8_t hold_flag;
int su = (pmap == kernel_pmap);
//debugf("mmu_booke_remove: s (su = %d pmap=0x%08x tid=%d va=0x%08x endva=0x%08x)\n",
// su, (u_int32_t)pmap, pmap->pm_tid, va, endva);
if (su) {
KASSERT(((va >= virtual_avail) &&
(va <= VM_MAX_KERNEL_ADDRESS)),
("mmu_booke_remove: kernel pmap, non kernel va"));
} else {
KASSERT((va <= VM_MAXUSER_ADDRESS),
("mmu_booke_remove: user pmap, non user va"));
}
if (PMAP_REMOVE_DONE(pmap)) {
//debugf("mmu_booke_remove: e (empty)\n");
return;
}
hold_flag = PTBL_HOLD_FLAG(pmap);
//debugf("mmu_booke_remove: hold_flag = %d\n", hold_flag);
rw_wlock(&pvh_global_lock);
PMAP_LOCK(pmap);
for (; va < endva; va += PAGE_SIZE) {
pte = pte_find(mmu, pmap, va);
if ((pte != NULL) && PTE_ISVALID(pte))
pte_remove(mmu, pmap, va, hold_flag);
}
PMAP_UNLOCK(pmap);
rw_wunlock(&pvh_global_lock);
//debugf("mmu_booke_remove: e\n");
}
/*
* Remove physical page from all pmaps in which it resides.
*/
static void
mmu_booke_remove_all(mmu_t mmu, vm_page_t m)
{
pv_entry_t pv, pvn;
uint8_t hold_flag;
rw_wlock(&pvh_global_lock);
for (pv = TAILQ_FIRST(&m->md.pv_list); pv != NULL; pv = pvn) {
pvn = TAILQ_NEXT(pv, pv_link);
PMAP_LOCK(pv->pv_pmap);
hold_flag = PTBL_HOLD_FLAG(pv->pv_pmap);
pte_remove(mmu, pv->pv_pmap, pv->pv_va, hold_flag);
PMAP_UNLOCK(pv->pv_pmap);
}
vm_page_aflag_clear(m, PGA_WRITEABLE);
rw_wunlock(&pvh_global_lock);
}
/*
* Map a range of physical addresses into kernel virtual address space.
*/
static vm_offset_t
mmu_booke_map(mmu_t mmu, vm_offset_t *virt, vm_paddr_t pa_start,
vm_paddr_t pa_end, int prot)
{
vm_offset_t sva = *virt;
vm_offset_t va = sva;
//debugf("mmu_booke_map: s (sva = 0x%08x pa_start = 0x%08x pa_end = 0x%08x)\n",
// sva, pa_start, pa_end);
while (pa_start < pa_end) {
mmu_booke_kenter(mmu, va, pa_start);
va += PAGE_SIZE;
pa_start += PAGE_SIZE;
}
*virt = va;
//debugf("mmu_booke_map: e (va = 0x%08x)\n", va);
return (sva);
}
/*
* The pmap must be activated before it's address space can be accessed in any
* way.
*/
static void
mmu_booke_activate(mmu_t mmu, struct thread *td)
{
pmap_t pmap;
u_int cpuid;
pmap = &td->td_proc->p_vmspace->vm_pmap;
CTR5(KTR_PMAP, "%s: s (td = %p, proc = '%s', id = %d, pmap = 0x%08x)",
__func__, td, td->td_proc->p_comm, td->td_proc->p_pid, pmap);
KASSERT((pmap != kernel_pmap), ("mmu_booke_activate: kernel_pmap!"));
sched_pin();
cpuid = PCPU_GET(cpuid);
CPU_SET_ATOMIC(cpuid, &pmap->pm_active);
PCPU_SET(curpmap, pmap);
if (pmap->pm_tid[cpuid] == TID_NONE)
tid_alloc(pmap);
/* Load PID0 register with pmap tid value. */
mtspr(SPR_PID0, pmap->pm_tid[cpuid]);
__asm __volatile("isync");
sched_unpin();
CTR3(KTR_PMAP, "%s: e (tid = %d for '%s')", __func__,
pmap->pm_tid[PCPU_GET(cpuid)], td->td_proc->p_comm);
}
/*
* Deactivate the specified process's address space.
*/
static void
mmu_booke_deactivate(mmu_t mmu, struct thread *td)
{
pmap_t pmap;
pmap = &td->td_proc->p_vmspace->vm_pmap;
CTR5(KTR_PMAP, "%s: td=%p, proc = '%s', id = %d, pmap = 0x%08x",
__func__, td, td->td_proc->p_comm, td->td_proc->p_pid, pmap);
CPU_CLR_ATOMIC(PCPU_GET(cpuid), &pmap->pm_active);
PCPU_SET(curpmap, NULL);
}
/*
* Copy the range specified by src_addr/len
* from the source map to the range dst_addr/len
* in the destination map.
*
* This routine is only advisory and need not do anything.
*/
static void
mmu_booke_copy(mmu_t mmu, pmap_t dst_pmap, pmap_t src_pmap,
vm_offset_t dst_addr, vm_size_t len, vm_offset_t src_addr)
{
}
/*
* Set the physical protection on the specified range of this map as requested.
*/
static void
mmu_booke_protect(mmu_t mmu, pmap_t pmap, vm_offset_t sva, vm_offset_t eva,
vm_prot_t prot)
{
vm_offset_t va;
vm_page_t m;
pte_t *pte;
if ((prot & VM_PROT_READ) == VM_PROT_NONE) {
mmu_booke_remove(mmu, pmap, sva, eva);
return;
}
if (prot & VM_PROT_WRITE)
return;
PMAP_LOCK(pmap);
for (va = sva; va < eva; va += PAGE_SIZE) {
if ((pte = pte_find(mmu, pmap, va)) != NULL) {
if (PTE_ISVALID(pte)) {
m = PHYS_TO_VM_PAGE(PTE_PA(pte));
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
/* Handle modified pages. */
if (PTE_ISMODIFIED(pte) && PTE_ISMANAGED(pte))
vm_page_dirty(m);
tlb0_flush_entry(va);
pte->flags &= ~(PTE_UW | PTE_SW | PTE_MODIFIED);
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
}
}
}
PMAP_UNLOCK(pmap);
}
/*
* Clear the write and modified bits in each of the given page's mappings.
*/
static void
mmu_booke_remove_write(mmu_t mmu, vm_page_t m)
{
pv_entry_t pv;
pte_t *pte;
KASSERT((m->oflags & VPO_UNMANAGED) == 0,
("mmu_booke_remove_write: page %p is not managed", m));
/*
* If the page is not exclusive busied, then PGA_WRITEABLE cannot be
* set by another thread while the object is locked. Thus,
* if PGA_WRITEABLE is clear, no page table entries need updating.
*/
VM_OBJECT_ASSERT_WLOCKED(m->object);
if (!vm_page_xbusied(m) && (m->aflags & PGA_WRITEABLE) == 0)
return;
rw_wlock(&pvh_global_lock);
TAILQ_FOREACH(pv, &m->md.pv_list, pv_link) {
PMAP_LOCK(pv->pv_pmap);
if ((pte = pte_find(mmu, pv->pv_pmap, pv->pv_va)) != NULL) {
if (PTE_ISVALID(pte)) {
m = PHYS_TO_VM_PAGE(PTE_PA(pte));
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
/* Handle modified pages. */
if (PTE_ISMODIFIED(pte))
vm_page_dirty(m);
/* Flush mapping from TLB0. */
pte->flags &= ~(PTE_UW | PTE_SW | PTE_MODIFIED);
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
}
}
PMAP_UNLOCK(pv->pv_pmap);
}
vm_page_aflag_clear(m, PGA_WRITEABLE);
rw_wunlock(&pvh_global_lock);
}
static void
mmu_booke_sync_icache(mmu_t mmu, pmap_t pm, vm_offset_t va, vm_size_t sz)
{
pte_t *pte;
pmap_t pmap;
vm_page_t m;
vm_offset_t addr;
vm_paddr_t pa = 0;
int active, valid;
va = trunc_page(va);
sz = round_page(sz);
rw_wlock(&pvh_global_lock);
pmap = PCPU_GET(curpmap);
active = (pm == kernel_pmap || pm == pmap) ? 1 : 0;
while (sz > 0) {
PMAP_LOCK(pm);
pte = pte_find(mmu, pm, va);
valid = (pte != NULL && PTE_ISVALID(pte)) ? 1 : 0;
if (valid)
pa = PTE_PA(pte);
PMAP_UNLOCK(pm);
if (valid) {
if (!active) {
/* Create a mapping in the active pmap. */
addr = 0;
m = PHYS_TO_VM_PAGE(pa);
PMAP_LOCK(pmap);
pte_enter(mmu, pmap, m, addr,
PTE_SR | PTE_VALID | PTE_UR, FALSE);
__syncicache((void *)addr, PAGE_SIZE);
pte_remove(mmu, pmap, addr, PTBL_UNHOLD);
PMAP_UNLOCK(pmap);
} else
__syncicache((void *)va, PAGE_SIZE);
}
va += PAGE_SIZE;
sz -= PAGE_SIZE;
}
rw_wunlock(&pvh_global_lock);
}
/*
* Atomically extract and hold the physical page with the given
* pmap and virtual address pair if that mapping permits the given
* protection.
*/
static vm_page_t
mmu_booke_extract_and_hold(mmu_t mmu, pmap_t pmap, vm_offset_t va,
vm_prot_t prot)
{
pte_t *pte;
vm_page_t m;
uint32_t pte_wbit;
vm_paddr_t pa;
m = NULL;
pa = 0;
PMAP_LOCK(pmap);
retry:
pte = pte_find(mmu, pmap, va);
if ((pte != NULL) && PTE_ISVALID(pte)) {
if (pmap == kernel_pmap)
pte_wbit = PTE_SW;
else
pte_wbit = PTE_UW;
if ((pte->flags & pte_wbit) || ((prot & VM_PROT_WRITE) == 0)) {
if (vm_page_pa_tryrelock(pmap, PTE_PA(pte), &pa))
goto retry;
m = PHYS_TO_VM_PAGE(PTE_PA(pte));
vm_page_hold(m);
}
}
PA_UNLOCK_COND(pa);
PMAP_UNLOCK(pmap);
return (m);
}
/*
* Initialize a vm_page's machine-dependent fields.
*/
static void
mmu_booke_page_init(mmu_t mmu, vm_page_t m)
{
TAILQ_INIT(&m->md.pv_list);
}
/*
* mmu_booke_zero_page_area zeros the specified hardware page by
* mapping it into virtual memory and using bzero to clear
* its contents.
*
* off and size must reside within a single page.
*/
static void
mmu_booke_zero_page_area(mmu_t mmu, vm_page_t m, int off, int size)
{
vm_offset_t va;
/* XXX KASSERT off and size are within a single page? */
mtx_lock(&zero_page_mutex);
va = zero_page_va;
mmu_booke_kenter(mmu, va, VM_PAGE_TO_PHYS(m));
bzero((caddr_t)va + off, size);
mmu_booke_kremove(mmu, va);
mtx_unlock(&zero_page_mutex);
}
/*
* mmu_booke_zero_page zeros the specified hardware page.
*/
static void
mmu_booke_zero_page(mmu_t mmu, vm_page_t m)
{
mmu_booke_zero_page_area(mmu, m, 0, PAGE_SIZE);
}
/*
* mmu_booke_copy_page copies the specified (machine independent) page by
* mapping the page into virtual memory and using memcopy to copy the page,
* one machine dependent page at a time.
*/
static void
mmu_booke_copy_page(mmu_t mmu, vm_page_t sm, vm_page_t dm)
{
vm_offset_t sva, dva;
sva = copy_page_src_va;
dva = copy_page_dst_va;
mtx_lock(©_page_mutex);
mmu_booke_kenter(mmu, sva, VM_PAGE_TO_PHYS(sm));
mmu_booke_kenter(mmu, dva, VM_PAGE_TO_PHYS(dm));
memcpy((caddr_t)dva, (caddr_t)sva, PAGE_SIZE);
mmu_booke_kremove(mmu, dva);
mmu_booke_kremove(mmu, sva);
mtx_unlock(©_page_mutex);
}
static inline void
mmu_booke_copy_pages(mmu_t mmu, vm_page_t *ma, vm_offset_t a_offset,
vm_page_t *mb, vm_offset_t b_offset, int xfersize)
{
void *a_cp, *b_cp;
vm_offset_t a_pg_offset, b_pg_offset;
int cnt;
mtx_lock(©_page_mutex);
while (xfersize > 0) {
a_pg_offset = a_offset & PAGE_MASK;
cnt = min(xfersize, PAGE_SIZE - a_pg_offset);
mmu_booke_kenter(mmu, copy_page_src_va,
VM_PAGE_TO_PHYS(ma[a_offset >> PAGE_SHIFT]));
a_cp = (char *)copy_page_src_va + a_pg_offset;
b_pg_offset = b_offset & PAGE_MASK;
cnt = min(cnt, PAGE_SIZE - b_pg_offset);
mmu_booke_kenter(mmu, copy_page_dst_va,
VM_PAGE_TO_PHYS(mb[b_offset >> PAGE_SHIFT]));
b_cp = (char *)copy_page_dst_va + b_pg_offset;
bcopy(a_cp, b_cp, cnt);
mmu_booke_kremove(mmu, copy_page_dst_va);
mmu_booke_kremove(mmu, copy_page_src_va);
a_offset += cnt;
b_offset += cnt;
xfersize -= cnt;
}
mtx_unlock(©_page_mutex);
}
/*
* mmu_booke_zero_page_idle zeros the specified hardware page by mapping it
* into virtual memory and using bzero to clear its contents. This is intended
* to be called from the vm_pagezero process only and outside of Giant. No
* lock is required.
*/
static void
mmu_booke_zero_page_idle(mmu_t mmu, vm_page_t m)
{
vm_offset_t va;
va = zero_page_idle_va;
mmu_booke_kenter(mmu, va, VM_PAGE_TO_PHYS(m));
bzero((caddr_t)va, PAGE_SIZE);
mmu_booke_kremove(mmu, va);
}
/*
* Return whether or not the specified physical page was modified
* in any of physical maps.
*/
static boolean_t
mmu_booke_is_modified(mmu_t mmu, vm_page_t m)
{
pte_t *pte;
pv_entry_t pv;
boolean_t rv;
KASSERT((m->oflags & VPO_UNMANAGED) == 0,
("mmu_booke_is_modified: page %p is not managed", m));
rv = FALSE;
/*
* If the page is not exclusive busied, then PGA_WRITEABLE cannot be
* concurrently set while the object is locked. Thus, if PGA_WRITEABLE
* is clear, no PTEs can be modified.
*/
VM_OBJECT_ASSERT_WLOCKED(m->object);
if (!vm_page_xbusied(m) && (m->aflags & PGA_WRITEABLE) == 0)
return (rv);
rw_wlock(&pvh_global_lock);
TAILQ_FOREACH(pv, &m->md.pv_list, pv_link) {
PMAP_LOCK(pv->pv_pmap);
if ((pte = pte_find(mmu, pv->pv_pmap, pv->pv_va)) != NULL &&
PTE_ISVALID(pte)) {
if (PTE_ISMODIFIED(pte))
rv = TRUE;
}
PMAP_UNLOCK(pv->pv_pmap);
if (rv)
break;
}
rw_wunlock(&pvh_global_lock);
return (rv);
}
/*
* Return whether or not the specified virtual address is eligible
* for prefault.
*/
static boolean_t
mmu_booke_is_prefaultable(mmu_t mmu, pmap_t pmap, vm_offset_t addr)
{
return (FALSE);
}
/*
* Return whether or not the specified physical page was referenced
* in any physical maps.
*/
static boolean_t
mmu_booke_is_referenced(mmu_t mmu, vm_page_t m)
{
pte_t *pte;
pv_entry_t pv;
boolean_t rv;
KASSERT((m->oflags & VPO_UNMANAGED) == 0,
("mmu_booke_is_referenced: page %p is not managed", m));
rv = FALSE;
rw_wlock(&pvh_global_lock);
TAILQ_FOREACH(pv, &m->md.pv_list, pv_link) {
PMAP_LOCK(pv->pv_pmap);
if ((pte = pte_find(mmu, pv->pv_pmap, pv->pv_va)) != NULL &&
PTE_ISVALID(pte)) {
if (PTE_ISREFERENCED(pte))
rv = TRUE;
}
PMAP_UNLOCK(pv->pv_pmap);
if (rv)
break;
}
rw_wunlock(&pvh_global_lock);
return (rv);
}
/*
* Clear the modify bits on the specified physical page.
*/
static void
mmu_booke_clear_modify(mmu_t mmu, vm_page_t m)
{
pte_t *pte;
pv_entry_t pv;
KASSERT((m->oflags & VPO_UNMANAGED) == 0,
("mmu_booke_clear_modify: page %p is not managed", m));
VM_OBJECT_ASSERT_WLOCKED(m->object);
KASSERT(!vm_page_xbusied(m),
("mmu_booke_clear_modify: page %p is exclusive busied", m));
/*
* If the page is not PG_AWRITEABLE, then no PTEs can be modified.
* If the object containing the page is locked and the page is not
* exclusive busied, then PG_AWRITEABLE cannot be concurrently set.
*/
if ((m->aflags & PGA_WRITEABLE) == 0)
return;
rw_wlock(&pvh_global_lock);
TAILQ_FOREACH(pv, &m->md.pv_list, pv_link) {
PMAP_LOCK(pv->pv_pmap);
if ((pte = pte_find(mmu, pv->pv_pmap, pv->pv_va)) != NULL &&
PTE_ISVALID(pte)) {
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
if (pte->flags & (PTE_SW | PTE_UW | PTE_MODIFIED)) {
tlb0_flush_entry(pv->pv_va);
pte->flags &= ~(PTE_SW | PTE_UW | PTE_MODIFIED |
PTE_REFERENCED);
}
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
}
PMAP_UNLOCK(pv->pv_pmap);
}
rw_wunlock(&pvh_global_lock);
}
/*
* Return a count of reference bits for a page, clearing those bits.
* It is not necessary for every reference bit to be cleared, but it
* is necessary that 0 only be returned when there are truly no
* reference bits set.
*
* XXX: The exact number of bits to check and clear is a matter that
* should be tested and standardized at some point in the future for
* optimal aging of shared pages.
*/
static int
mmu_booke_ts_referenced(mmu_t mmu, vm_page_t m)
{
pte_t *pte;
pv_entry_t pv;
int count;
KASSERT((m->oflags & VPO_UNMANAGED) == 0,
("mmu_booke_ts_referenced: page %p is not managed", m));
count = 0;
rw_wlock(&pvh_global_lock);
TAILQ_FOREACH(pv, &m->md.pv_list, pv_link) {
PMAP_LOCK(pv->pv_pmap);
if ((pte = pte_find(mmu, pv->pv_pmap, pv->pv_va)) != NULL &&
PTE_ISVALID(pte)) {
if (PTE_ISREFERENCED(pte)) {
mtx_lock_spin(&tlbivax_mutex);
tlb_miss_lock();
tlb0_flush_entry(pv->pv_va);
pte->flags &= ~PTE_REFERENCED;
tlb_miss_unlock();
mtx_unlock_spin(&tlbivax_mutex);
if (++count > 4) {
PMAP_UNLOCK(pv->pv_pmap);
break;
}
}
}
PMAP_UNLOCK(pv->pv_pmap);
}
rw_wunlock(&pvh_global_lock);
return (count);
}
/*
* Clear the wired attribute from the mappings for the specified range of
* addresses in the given pmap. Every valid mapping within that range must
* have the wired attribute set. In contrast, invalid mappings cannot have
* the wired attribute set, so they are ignored.
*
* The wired attribute of the page table entry is not a hardware feature, so
* there is no need to invalidate any TLB entries.
*/
static void
mmu_booke_unwire(mmu_t mmu, pmap_t pmap, vm_offset_t sva, vm_offset_t eva)
{
vm_offset_t va;
pte_t *pte;
PMAP_LOCK(pmap);
for (va = sva; va < eva; va += PAGE_SIZE) {
if ((pte = pte_find(mmu, pmap, va)) != NULL &&
PTE_ISVALID(pte)) {
if (!PTE_ISWIRED(pte))
panic("mmu_booke_unwire: pte %p isn't wired",
pte);
pte->flags &= ~PTE_WIRED;
pmap->pm_stats.wired_count--;
}
}
PMAP_UNLOCK(pmap);
}
/*
* Return true if the pmap's pv is one of the first 16 pvs linked to from this
* page. This count may be changed upwards or downwards in the future; it is
* only necessary that true be returned for a small subset of pmaps for proper
* page aging.
*/
static boolean_t
mmu_booke_page_exists_quick(mmu_t mmu, pmap_t pmap, vm_page_t m)
{
pv_entry_t pv;
int loops;
boolean_t rv;
KASSERT((m->oflags & VPO_UNMANAGED) == 0,
("mmu_booke_page_exists_quick: page %p is not managed", m));
loops = 0;
rv = FALSE;
rw_wlock(&pvh_global_lock);
TAILQ_FOREACH(pv, &m->md.pv_list, pv_link) {
if (pv->pv_pmap == pmap) {
rv = TRUE;
break;
}
if (++loops >= 16)
break;
}
rw_wunlock(&pvh_global_lock);
return (rv);
}
/*
* Return the number of managed mappings to the given physical page that are
* wired.
*/
static int
mmu_booke_page_wired_mappings(mmu_t mmu, vm_page_t m)
{
pv_entry_t pv;
pte_t *pte;
int count = 0;
if ((m->oflags & VPO_UNMANAGED) != 0)
return (count);
rw_wlock(&pvh_global_lock);
TAILQ_FOREACH(pv, &m->md.pv_list, pv_link) {
PMAP_LOCK(pv->pv_pmap);
if ((pte = pte_find(mmu, pv->pv_pmap, pv->pv_va)) != NULL)
if (PTE_ISVALID(pte) && PTE_ISWIRED(pte))
count++;
PMAP_UNLOCK(pv->pv_pmap);
}
rw_wunlock(&pvh_global_lock);
return (count);
}
static int
mmu_booke_dev_direct_mapped(mmu_t mmu, vm_paddr_t pa, vm_size_t size)
{
int i;
vm_offset_t va;
/*
* This currently does not work for entries that
* overlap TLB1 entries.
*/
for (i = 0; i < tlb1_idx; i ++) {
if (tlb1_iomapped(i, pa, size, &va) == 0)
return (0);
}
return (EFAULT);
}
void
mmu_booke_dumpsys_map(mmu_t mmu, vm_paddr_t pa, size_t sz, void **va)
{
vm_paddr_t ppa;
vm_offset_t ofs;
vm_size_t gran;
/* Minidumps are based on virtual memory addresses. */
if (do_minidump) {
*va = (void *)pa;
return;
}
/* Raw physical memory dumps don't have a virtual address. */
/* We always map a 256MB page at 256M. */
gran = 256 * 1024 * 1024;
ppa = pa & ~(gran - 1);
ofs = pa - ppa;
*va = (void *)gran;
tlb1_set_entry((vm_offset_t)va, ppa, gran, _TLB_ENTRY_IO);
if (sz > (gran - ofs))
tlb1_set_entry((vm_offset_t)(va + gran), ppa + gran, gran,
_TLB_ENTRY_IO);
}
void
mmu_booke_dumpsys_unmap(mmu_t mmu, vm_paddr_t pa, size_t sz, void *va)
{
vm_paddr_t ppa;
vm_offset_t ofs;
vm_size_t gran;
/* Minidumps are based on virtual memory addresses. */
/* Nothing to do... */
if (do_minidump)
return;
/* Raw physical memory dumps don't have a virtual address. */
tlb1_idx--;
tlb1[tlb1_idx].mas1 = 0;
tlb1[tlb1_idx].mas2 = 0;
tlb1[tlb1_idx].mas3 = 0;
tlb1_write_entry(tlb1_idx);
gran = 256 * 1024 * 1024;
ppa = pa & ~(gran - 1);
ofs = pa - ppa;
if (sz > (gran - ofs)) {
tlb1_idx--;
tlb1[tlb1_idx].mas1 = 0;
tlb1[tlb1_idx].mas2 = 0;
tlb1[tlb1_idx].mas3 = 0;
tlb1_write_entry(tlb1_idx);
}
}
extern struct dump_pa dump_map[PHYS_AVAIL_SZ + 1];
void
mmu_booke_scan_init(mmu_t mmu)
{
vm_offset_t va;
pte_t *pte;
int i;
if (!do_minidump) {
/* Initialize phys. segments for dumpsys(). */
memset(&dump_map, 0, sizeof(dump_map));
mem_regions(&physmem_regions, &physmem_regions_sz, &availmem_regions,
&availmem_regions_sz);
for (i = 0; i < physmem_regions_sz; i++) {
dump_map[i].pa_start = physmem_regions[i].mr_start;
dump_map[i].pa_size = physmem_regions[i].mr_size;
}
return;
}
/* Virtual segments for minidumps: */
memset(&dump_map, 0, sizeof(dump_map));
/* 1st: kernel .data and .bss. */
dump_map[0].pa_start = trunc_page((uintptr_t)_etext);
dump_map[0].pa_size =
round_page((uintptr_t)_end) - dump_map[0].pa_start;
/* 2nd: msgbuf and tables (see pmap_bootstrap()). */
dump_map[1].pa_start = data_start;
dump_map[1].pa_size = data_end - data_start;
/* 3rd: kernel VM. */
va = dump_map[1].pa_start + dump_map[1].pa_size;
/* Find start of next chunk (from va). */
while (va < virtual_end) {
/* Don't dump the buffer cache. */
if (va >= kmi.buffer_sva && va < kmi.buffer_eva) {
va = kmi.buffer_eva;
continue;
}
pte = pte_find(mmu, kernel_pmap, va);
if (pte != NULL && PTE_ISVALID(pte))
break;
va += PAGE_SIZE;
}
if (va < virtual_end) {
dump_map[2].pa_start = va;
va += PAGE_SIZE;
/* Find last page in chunk. */
while (va < virtual_end) {
/* Don't run into the buffer cache. */
if (va == kmi.buffer_sva)
break;
pte = pte_find(mmu, kernel_pmap, va);
if (pte == NULL || !PTE_ISVALID(pte))
break;
va += PAGE_SIZE;
}
dump_map[2].pa_size = va - dump_map[2].pa_start;
}
}
/*
* Map a set of physical memory pages into the kernel virtual address space.
* Return a pointer to where it is mapped. This routine is intended to be used
* for mapping device memory, NOT real memory.
*/
static void *
mmu_booke_mapdev(mmu_t mmu, vm_paddr_t pa, vm_size_t size)
{
return (mmu_booke_mapdev_attr(mmu, pa, size, VM_MEMATTR_DEFAULT));
}
static void *
mmu_booke_mapdev_attr(mmu_t mmu, vm_paddr_t pa, vm_size_t size, vm_memattr_t ma)
{
void *res;
uintptr_t va;
vm_size_t sz;
int i;
/*
* Check if this is premapped in TLB1. Note: this should probably also
* check whether a sequence of TLB1 entries exist that match the
* requirement, but now only checks the easy case.
*/
if (ma == VM_MEMATTR_DEFAULT) {
for (i = 0; i < tlb1_idx; i++) {
if (!(tlb1[i].mas1 & MAS1_VALID))
continue;
if (pa >= tlb1[i].phys &&
(pa + size) <= (tlb1[i].phys + tlb1[i].size))
return (void *)(tlb1[i].virt +
(pa - tlb1[i].phys));
}
}
size = roundup(size, PAGE_SIZE);
/*
* We leave a hole for device direct mapping between the maximum user
* address (0x8000000) and the minimum KVA address (0xc0000000). If
* devices are in there, just map them 1:1. If not, map them to the
* device mapping area about VM_MAX_KERNEL_ADDRESS. These mapped
* addresses should be pulled from an allocator, but since we do not
* ever free TLB1 entries, it is safe just to increment a counter.
* Note that there isn't a lot of address space here (128 MB) and it
* is not at all difficult to imagine running out, since that is a 4:1
* compression from the 0xc0000000 - 0xf0000000 address space that gets
* mapped there.
*/
if (pa >= (VM_MAXUSER_ADDRESS + PAGE_SIZE) &&
(pa + size - 1) < VM_MIN_KERNEL_ADDRESS)
va = pa;
else
va = atomic_fetchadd_int(&tlb1_map_base, size);
res = (void *)va;
do {
sz = 1 << (ilog2(size) & ~1);
if (bootverbose)
printf("Wiring VA=%x to PA=%x (size=%x), "
"using TLB1[%d]\n", va, pa, sz, tlb1_idx);
tlb1_set_entry(va, pa, sz, tlb_calc_wimg(pa, ma));
size -= sz;
pa += sz;
va += sz;
} while (size > 0);
return (res);
}
/*
* 'Unmap' a range mapped by mmu_booke_mapdev().
*/
static void
mmu_booke_unmapdev(mmu_t mmu, vm_offset_t va, vm_size_t size)
{
#ifdef SUPPORTS_SHRINKING_TLB1
vm_offset_t base, offset;
/*
* Unmap only if this is inside kernel virtual space.
*/
if ((va >= VM_MIN_KERNEL_ADDRESS) && (va <= VM_MAX_KERNEL_ADDRESS)) {
base = trunc_page(va);
offset = va & PAGE_MASK;
size = roundup(offset + size, PAGE_SIZE);
kva_free(base, size);
}
#endif
}
/*
* mmu_booke_object_init_pt preloads the ptes for a given object into the
* specified pmap. This eliminates the blast of soft faults on process startup
* and immediately after an mmap.
*/
static void
mmu_booke_object_init_pt(mmu_t mmu, pmap_t pmap, vm_offset_t addr,
vm_object_t object, vm_pindex_t pindex, vm_size_t size)
{
VM_OBJECT_ASSERT_WLOCKED(object);
KASSERT(object->type == OBJT_DEVICE || object->type == OBJT_SG,
("mmu_booke_object_init_pt: non-device object"));
}
/*
* Perform the pmap work for mincore.
*/
static int
mmu_booke_mincore(mmu_t mmu, pmap_t pmap, vm_offset_t addr,
vm_paddr_t *locked_pa)
{
/* XXX: this should be implemented at some point */
return (0);
}
/**************************************************************************/
/* TID handling */
/**************************************************************************/
/*
* Allocate a TID. If necessary, steal one from someone else.
* The new TID is flushed from the TLB before returning.
*/
static tlbtid_t
tid_alloc(pmap_t pmap)
{
tlbtid_t tid;
int thiscpu;
KASSERT((pmap != kernel_pmap), ("tid_alloc: kernel pmap"));
CTR2(KTR_PMAP, "%s: s (pmap = %p)", __func__, pmap);
thiscpu = PCPU_GET(cpuid);
tid = PCPU_GET(tid_next);
if (tid > TID_MAX)
tid = TID_MIN;
PCPU_SET(tid_next, tid + 1);
/* If we are stealing TID then clear the relevant pmap's field */
if (tidbusy[thiscpu][tid] != NULL) {
CTR2(KTR_PMAP, "%s: warning: stealing tid %d", __func__, tid);
tidbusy[thiscpu][tid]->pm_tid[thiscpu] = TID_NONE;
/* Flush all entries from TLB0 matching this TID. */
tid_flush(tid);
}
tidbusy[thiscpu][tid] = pmap;
pmap->pm_tid[thiscpu] = tid;
__asm __volatile("msync; isync");
CTR3(KTR_PMAP, "%s: e (%02d next = %02d)", __func__, tid,
PCPU_GET(tid_next));
return (tid);
}
/**************************************************************************/
/* TLB0 handling */
/**************************************************************************/
static void
tlb_print_entry(int i, uint32_t mas1, uint32_t mas2, uint32_t mas3,
uint32_t mas7)
{
int as;
char desc[3];
tlbtid_t tid;
vm_size_t size;
unsigned int tsize;
desc[2] = '\0';
if (mas1 & MAS1_VALID)
desc[0] = 'V';
else
desc[0] = ' ';
if (mas1 & MAS1_IPROT)
desc[1] = 'P';
else
desc[1] = ' ';
as = (mas1 & MAS1_TS_MASK) ? 1 : 0;
tid = MAS1_GETTID(mas1);
tsize = (mas1 & MAS1_TSIZE_MASK) >> MAS1_TSIZE_SHIFT;
size = 0;
if (tsize)
size = tsize2size(tsize);
debugf("%3d: (%s) [AS=%d] "
"sz = 0x%08x tsz = %d tid = %d mas1 = 0x%08x "
"mas2(va) = 0x%08x mas3(pa) = 0x%08x mas7 = 0x%08x\n",
i, desc, as, size, tsize, tid, mas1, mas2, mas3, mas7);
}
/* Convert TLB0 va and way number to tlb0[] table index. */
static inline unsigned int
tlb0_tableidx(vm_offset_t va, unsigned int way)
{
unsigned int idx;
idx = (way * TLB0_ENTRIES_PER_WAY);
idx += (va & MAS2_TLB0_ENTRY_IDX_MASK) >> MAS2_TLB0_ENTRY_IDX_SHIFT;
return (idx);
}
/*
* Invalidate TLB0 entry.
*/
static inline void
tlb0_flush_entry(vm_offset_t va)
{
CTR2(KTR_PMAP, "%s: s va=0x%08x", __func__, va);
mtx_assert(&tlbivax_mutex, MA_OWNED);
__asm __volatile("tlbivax 0, %0" :: "r"(va & MAS2_EPN_MASK));
__asm __volatile("isync; msync");
__asm __volatile("tlbsync; msync");
CTR1(KTR_PMAP, "%s: e", __func__);
}
/* Print out contents of the MAS registers for each TLB0 entry */
void
tlb0_print_tlbentries(void)
{
uint32_t mas0, mas1, mas2, mas3, mas7;
int entryidx, way, idx;
debugf("TLB0 entries:\n");
for (way = 0; way < TLB0_WAYS; way ++)
for (entryidx = 0; entryidx < TLB0_ENTRIES_PER_WAY; entryidx++) {
mas0 = MAS0_TLBSEL(0) | MAS0_ESEL(way);
mtspr(SPR_MAS0, mas0);
__asm __volatile("isync");
mas2 = entryidx << MAS2_TLB0_ENTRY_IDX_SHIFT;
mtspr(SPR_MAS2, mas2);
__asm __volatile("isync; tlbre");
mas1 = mfspr(SPR_MAS1);
mas2 = mfspr(SPR_MAS2);
mas3 = mfspr(SPR_MAS3);
mas7 = mfspr(SPR_MAS7);
idx = tlb0_tableidx(mas2, way);
tlb_print_entry(idx, mas1, mas2, mas3, mas7);
}
}
/**************************************************************************/
/* TLB1 handling */
/**************************************************************************/
/*
* TLB1 mapping notes:
*
* TLB1[0] Kernel text and data.
* TLB1[1-15] Additional kernel text and data mappings (if required), PCI
* windows, other devices mappings.
*/
/*
* Write given entry to TLB1 hardware.
* Use 32 bit pa, clear 4 high-order bits of RPN (mas7).
*/
static void
tlb1_write_entry(unsigned int idx)
{
uint32_t mas0, mas7;
//debugf("tlb1_write_entry: s\n");
/* Clear high order RPN bits */
mas7 = 0;
/* Select entry */
mas0 = MAS0_TLBSEL(1) | MAS0_ESEL(idx);
//debugf("tlb1_write_entry: mas0 = 0x%08x\n", mas0);
mtspr(SPR_MAS0, mas0);
__asm __volatile("isync");
mtspr(SPR_MAS1, tlb1[idx].mas1);
__asm __volatile("isync");
mtspr(SPR_MAS2, tlb1[idx].mas2);
__asm __volatile("isync");
mtspr(SPR_MAS3, tlb1[idx].mas3);
__asm __volatile("isync");
mtspr(SPR_MAS7, mas7);
__asm __volatile("isync; tlbwe; isync; msync");
//debugf("tlb1_write_entry: e\n");
}
/*
* Return the largest uint value log such that 2^log <= num.
*/
static unsigned int
ilog2(unsigned int num)
{
int lz;
__asm ("cntlzw %0, %1" : "=r" (lz) : "r" (num));
return (31 - lz);
}
/*
* Convert TLB TSIZE value to mapped region size.
*/
static vm_size_t
tsize2size(unsigned int tsize)
{
/*
* size = 4^tsize KB
* size = 4^tsize * 2^10 = 2^(2 * tsize - 10)
*/
return ((1 << (2 * tsize)) * 1024);
}
/*
* Convert region size (must be power of 4) to TLB TSIZE value.
*/
static unsigned int
size2tsize(vm_size_t size)
{
return (ilog2(size) / 2 - 5);
}
/*
* Register permanent kernel mapping in TLB1.
*
* Entries are created starting from index 0 (current free entry is
* kept in tlb1_idx) and are not supposed to be invalidated.
*/
static int
tlb1_set_entry(vm_offset_t va, vm_offset_t pa, vm_size_t size,
uint32_t flags)
{
uint32_t ts, tid;
int tsize, index;
index = atomic_fetchadd_int(&tlb1_idx, 1);
if (index >= TLB1_ENTRIES) {
printf("tlb1_set_entry: TLB1 full!\n");
return (-1);
}
/* Convert size to TSIZE */
tsize = size2tsize(size);
tid = (TID_KERNEL << MAS1_TID_SHIFT) & MAS1_TID_MASK;
/* XXX TS is hard coded to 0 for now as we only use single address space */
ts = (0 << MAS1_TS_SHIFT) & MAS1_TS_MASK;
/*
* Atomicity is preserved by the atomic increment above since nothing
* is ever removed from tlb1.
*/
tlb1[index].phys = pa;
tlb1[index].virt = va;
tlb1[index].size = size;
tlb1[index].mas1 = MAS1_VALID | MAS1_IPROT | ts | tid;
tlb1[index].mas1 |= ((tsize << MAS1_TSIZE_SHIFT) & MAS1_TSIZE_MASK);
tlb1[index].mas2 = (va & MAS2_EPN_MASK) | flags;
/* Set supervisor RWX permission bits */
tlb1[index].mas3 = (pa & MAS3_RPN) | MAS3_SR | MAS3_SW | MAS3_SX;
tlb1_write_entry(index);
/*
* XXX in general TLB1 updates should be propagated between CPUs,
* since current design assumes to have the same TLB1 set-up on all
* cores.
*/
return (0);
}
/*
* Map in contiguous RAM region into the TLB1 using maximum of
* KERNEL_REGION_MAX_TLB_ENTRIES entries.
*
* If necessary round up last entry size and return total size
* used by all allocated entries.
*/
vm_size_t
tlb1_mapin_region(vm_offset_t va, vm_paddr_t pa, vm_size_t size)
{
vm_size_t pgs[KERNEL_REGION_MAX_TLB_ENTRIES];
vm_size_t mapped, pgsz, base, mask;
int idx, nents;
/* Round up to the next 1M */
size = (size + (1 << 20) - 1) & ~((1 << 20) - 1);
mapped = 0;
idx = 0;
base = va;
pgsz = 64*1024*1024;
while (mapped < size) {
while (mapped < size && idx < KERNEL_REGION_MAX_TLB_ENTRIES) {
while (pgsz > (size - mapped))
pgsz >>= 2;
pgs[idx++] = pgsz;
mapped += pgsz;
}
/* We under-map. Correct for this. */
if (mapped < size) {
while (pgs[idx - 1] == pgsz) {
idx--;
mapped -= pgsz;
}
/* XXX We may increase beyond out starting point. */
pgsz <<= 2;
pgs[idx++] = pgsz;
mapped += pgsz;
}
}
nents = idx;
mask = pgs[0] - 1;
/* Align address to the boundary */
if (va & mask) {
va = (va + mask) & ~mask;
pa = (pa + mask) & ~mask;
}
for (idx = 0; idx < nents; idx++) {
pgsz = pgs[idx];
debugf("%u: %x -> %x, size=%x\n", idx, pa, va, pgsz);
tlb1_set_entry(va, pa, pgsz, _TLB_ENTRY_MEM);
pa += pgsz;
va += pgsz;
}
mapped = (va - base);
printf("mapped size 0x%08x (wasted space 0x%08x)\n",
mapped, mapped - size);
return (mapped);
}
/*
* TLB1 initialization routine, to be called after the very first
* assembler level setup done in locore.S.
*/
void
tlb1_init()
{
uint32_t mas0, mas1, mas2, mas3;
uint32_t tsz;
u_int i;
if (bootinfo != NULL && bootinfo[0] != 1) {
tlb1_idx = *((uint16_t *)(bootinfo + 8));
} else
tlb1_idx = 1;
/* The first entry/entries are used to map the kernel. */
for (i = 0; i < tlb1_idx; i++) {
mas0 = MAS0_TLBSEL(1) | MAS0_ESEL(i);
mtspr(SPR_MAS0, mas0);
__asm __volatile("isync; tlbre");
mas1 = mfspr(SPR_MAS1);
if ((mas1 & MAS1_VALID) == 0)
continue;
mas2 = mfspr(SPR_MAS2);
mas3 = mfspr(SPR_MAS3);
tlb1[i].mas1 = mas1;
tlb1[i].mas2 = mfspr(SPR_MAS2);
tlb1[i].mas3 = mas3;
tlb1[i].virt = mas2 & MAS2_EPN_MASK;
tlb1[i].phys = mas3 & MAS3_RPN;
if (i == 0)
kernload = mas3 & MAS3_RPN;
tsz = (mas1 & MAS1_TSIZE_MASK) >> MAS1_TSIZE_SHIFT;
tlb1[i].size = (tsz > 0) ? tsize2size(tsz) : 0;
kernsize += tlb1[i].size;
}
#ifdef SMP
bp_ntlb1s = tlb1_idx;
#endif
/* Purge the remaining entries */
for (i = tlb1_idx; i < TLB1_ENTRIES; i++)
tlb1_write_entry(i);
/* Setup TLB miss defaults */
set_mas4_defaults();
}
vm_offset_t
pmap_early_io_map(vm_paddr_t pa, vm_size_t size)
{
vm_paddr_t pa_base;
vm_offset_t va, sz;
int i;
KASSERT(!pmap_bootstrapped, ("Do not use after PMAP is up!"));
for (i = 0; i < tlb1_idx; i++) {
if (!(tlb1[i].mas1 & MAS1_VALID))
continue;
if (pa >= tlb1[i].phys && (pa + size) <=
(tlb1[i].phys + tlb1[i].size))
return (tlb1[i].virt + (pa - tlb1[i].phys));
}
pa_base = trunc_page(pa);
size = roundup(size + (pa - pa_base), PAGE_SIZE);
tlb1_map_base = roundup2(tlb1_map_base, 1 << (ilog2(size) & ~1));
va = tlb1_map_base + (pa - pa_base);
do {
sz = 1 << (ilog2(size) & ~1);
tlb1_set_entry(tlb1_map_base, pa_base, sz, _TLB_ENTRY_IO);
size -= sz;
pa_base += sz;
tlb1_map_base += sz;
} while (size > 0);
#ifdef SMP
bp_ntlb1s = tlb1_idx;
#endif
return (va);
}
/*
* Setup MAS4 defaults.
* These values are loaded to MAS0-2 on a TLB miss.
*/
static void
set_mas4_defaults(void)
{
uint32_t mas4;
/* Defaults: TLB0, PID0, TSIZED=4K */
mas4 = MAS4_TLBSELD0;
mas4 |= (TLB_SIZE_4K << MAS4_TSIZED_SHIFT) & MAS4_TSIZED_MASK;
#ifdef SMP
mas4 |= MAS4_MD;
#endif
mtspr(SPR_MAS4, mas4);
__asm __volatile("isync");
}
/*
* Print out contents of the MAS registers for each TLB1 entry
*/
void
tlb1_print_tlbentries(void)
{
uint32_t mas0, mas1, mas2, mas3, mas7;
int i;
debugf("TLB1 entries:\n");
for (i = 0; i < TLB1_ENTRIES; i++) {
mas0 = MAS0_TLBSEL(1) | MAS0_ESEL(i);
mtspr(SPR_MAS0, mas0);
__asm __volatile("isync; tlbre");
mas1 = mfspr(SPR_MAS1);
mas2 = mfspr(SPR_MAS2);
mas3 = mfspr(SPR_MAS3);
mas7 = mfspr(SPR_MAS7);
tlb_print_entry(i, mas1, mas2, mas3, mas7);
}
}
/*
* Print out contents of the in-ram tlb1 table.
*/
void
tlb1_print_entries(void)
{
int i;
debugf("tlb1[] table entries:\n");
for (i = 0; i < TLB1_ENTRIES; i++)
tlb_print_entry(i, tlb1[i].mas1, tlb1[i].mas2, tlb1[i].mas3, 0);
}
/*
* Return 0 if the physical IO range is encompassed by one of the
* the TLB1 entries, otherwise return related error code.
*/
static int
tlb1_iomapped(int i, vm_paddr_t pa, vm_size_t size, vm_offset_t *va)
{
uint32_t prot;
vm_paddr_t pa_start;
vm_paddr_t pa_end;
unsigned int entry_tsize;
vm_size_t entry_size;
*va = (vm_offset_t)NULL;
/* Skip invalid entries */
if (!(tlb1[i].mas1 & MAS1_VALID))
return (EINVAL);
/*
* The entry must be cache-inhibited, guarded, and r/w
* so it can function as an i/o page
*/
prot = tlb1[i].mas2 & (MAS2_I | MAS2_G);
if (prot != (MAS2_I | MAS2_G))
return (EPERM);
prot = tlb1[i].mas3 & (MAS3_SR | MAS3_SW);
if (prot != (MAS3_SR | MAS3_SW))
return (EPERM);
/* The address should be within the entry range. */
entry_tsize = (tlb1[i].mas1 & MAS1_TSIZE_MASK) >> MAS1_TSIZE_SHIFT;
KASSERT((entry_tsize), ("tlb1_iomapped: invalid entry tsize"));
entry_size = tsize2size(entry_tsize);
pa_start = tlb1[i].mas3 & MAS3_RPN;
pa_end = pa_start + entry_size - 1;
if ((pa < pa_start) || ((pa + size) > pa_end))
return (ERANGE);
/* Return virtual address of this mapping. */
*va = (tlb1[i].mas2 & MAS2_EPN_MASK) + (pa - pa_start);
return (0);
}
|