summaryrefslogtreecommitdiffstats
path: root/sys/powerpc/aim/mmu_oea64.c
blob: d58e8dee1649bcc890e588bee70bece88db303b6 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
/*-
 * Copyright (c) 2001 The NetBSD Foundation, Inc.
 * All rights reserved.
 *
 * This code is derived from software contributed to The NetBSD Foundation
 * by Matt Thomas <matt@3am-software.com> of Allegro Networks, Inc.
 *
 * 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.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *        This product includes software developed by the NetBSD
 *        Foundation, Inc. and its contributors.
 * 4. Neither the name of The NetBSD Foundation nor the names of its
 *    contributors may be used to endorse or promote products derived
 *    from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
 * ``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 FOUNDATION OR CONTRIBUTORS
 * 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.
 */
/*-
 * Copyright (C) 1995, 1996 Wolfgang Solfrank.
 * Copyright (C) 1995, 1996 TooLs GmbH.
 * 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.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by TooLs GmbH.
 * 4. The name of TooLs GmbH may not be used to endorse or promote products
 *    derived from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY TOOLS GMBH ``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 TOOLS GMBH 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.
 *
 * $NetBSD: pmap.c,v 1.28 2000/03/26 20:42:36 kleink Exp $
 */
/*-
 * Copyright (C) 2001 Benno Rice.
 * 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 Benno Rice ``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 TOOLS GMBH BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

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

/*
 * Manages physical address maps.
 *
 * In addition to hardware address maps, this module is called upon to
 * provide software-use-only maps which may or may not be stored in the
 * same form as hardware maps.  These pseudo-maps are used to store
 * intermediate results from copy operations to and from address spaces.
 *
 * Since the information managed by this module is also stored by the
 * logical address mapping module, this module may throw away valid virtual
 * to physical mappings at almost any time.  However, invalidations of
 * mappings must be done as requested.
 *
 * In order to cope with hardware architectures which make virtual to
 * physical map invalidates expensive, this module may delay invalidate
 * reduced protection operations until such time as they are actually
 * necessary.  This module is given full information as to which processors
 * are currently using which maps, and to when physical maps must be made
 * correct.
 */

#include "opt_kstack_pages.h"

#include <sys/param.h>
#include <sys/kernel.h>
#include <sys/queue.h>
#include <sys/cpuset.h>
#include <sys/ktr.h>
#include <sys/lock.h>
#include <sys/msgbuf.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/sched.h>
#include <sys/sysctl.h>
#include <sys/systm.h>
#include <sys/vmmeter.h>

#include <sys/kdb.h>

#include <dev/ofw/openfirm.h>

#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/vm_kern.h>
#include <vm/vm_page.h>
#include <vm/vm_map.h>
#include <vm/vm_object.h>
#include <vm/vm_extern.h>
#include <vm/vm_pageout.h>
#include <vm/vm_pager.h>
#include <vm/uma.h>

#include <machine/_inttypes.h>
#include <machine/cpu.h>
#include <machine/platform.h>
#include <machine/frame.h>
#include <machine/md_var.h>
#include <machine/psl.h>
#include <machine/bat.h>
#include <machine/hid.h>
#include <machine/pte.h>
#include <machine/sr.h>
#include <machine/trap.h>
#include <machine/mmuvar.h>

#include "mmu_oea64.h"
#include "mmu_if.h"
#include "moea64_if.h"

void moea64_release_vsid(uint64_t vsid);
uintptr_t moea64_get_unique_vsid(void); 

#define DISABLE_TRANS(msr)	msr = mfmsr(); mtmsr(msr & ~PSL_DR)
#define ENABLE_TRANS(msr)	mtmsr(msr)

#define	VSID_MAKE(sr, hash)	((sr) | (((hash) & 0xfffff) << 4))
#define	VSID_TO_HASH(vsid)	(((vsid) >> 4) & 0xfffff)
#define	VSID_HASH_MASK		0x0000007fffffffffULL

#define LOCK_TABLE() mtx_lock(&moea64_table_mutex)
#define UNLOCK_TABLE() mtx_unlock(&moea64_table_mutex);
#define ASSERT_TABLE_LOCK() mtx_assert(&moea64_table_mutex, MA_OWNED)

struct ofw_map {
	cell_t	om_va;
	cell_t	om_len;
	cell_t	om_pa_hi;
	cell_t	om_pa_lo;
	cell_t	om_mode;
};

/*
 * Map of physical memory regions.
 */
static struct	mem_region *regions;
static struct	mem_region *pregions;
static u_int	phys_avail_count;
static int	regions_sz, pregions_sz;

extern void bs_remap_earlyboot(void);

/*
 * Lock for the pteg and pvo tables.
 */
struct mtx	moea64_table_mutex;
struct mtx	moea64_slb_mutex;

/*
 * PTEG data.
 */
u_int		moea64_pteg_count;
u_int		moea64_pteg_mask;

/*
 * PVO data.
 */
struct	pvo_head *moea64_pvo_table;		/* pvo entries by pteg index */
struct	pvo_head moea64_pvo_kunmanaged =	/* list of unmanaged pages */
    LIST_HEAD_INITIALIZER(moea64_pvo_kunmanaged);

uma_zone_t	moea64_upvo_zone; /* zone for pvo entries for unmanaged pages */
uma_zone_t	moea64_mpvo_zone; /* zone for pvo entries for managed pages */

#define	BPVO_POOL_SIZE	327680
static struct	pvo_entry *moea64_bpvo_pool;
static int	moea64_bpvo_pool_index = 0;

#define	VSID_NBPW	(sizeof(u_int32_t) * 8)
#ifdef __powerpc64__
#define	NVSIDS		(NPMAPS * 16)
#define VSID_HASHMASK	0xffffffffUL
#else
#define NVSIDS		NPMAPS
#define VSID_HASHMASK	0xfffffUL
#endif
static u_int	moea64_vsid_bitmap[NVSIDS / VSID_NBPW];

static boolean_t moea64_initialized = FALSE;

/*
 * Statistics.
 */
u_int	moea64_pte_valid = 0;
u_int	moea64_pte_overflow = 0;
u_int	moea64_pvo_entries = 0;
u_int	moea64_pvo_enter_calls = 0;
u_int	moea64_pvo_remove_calls = 0;
SYSCTL_INT(_machdep, OID_AUTO, moea64_pte_valid, CTLFLAG_RD, 
    &moea64_pte_valid, 0, "");
SYSCTL_INT(_machdep, OID_AUTO, moea64_pte_overflow, CTLFLAG_RD,
    &moea64_pte_overflow, 0, "");
SYSCTL_INT(_machdep, OID_AUTO, moea64_pvo_entries, CTLFLAG_RD, 
    &moea64_pvo_entries, 0, "");
SYSCTL_INT(_machdep, OID_AUTO, moea64_pvo_enter_calls, CTLFLAG_RD,
    &moea64_pvo_enter_calls, 0, "");
SYSCTL_INT(_machdep, OID_AUTO, moea64_pvo_remove_calls, CTLFLAG_RD,
    &moea64_pvo_remove_calls, 0, "");

vm_offset_t	moea64_scratchpage_va[2];
struct pvo_entry *moea64_scratchpage_pvo[2];
uintptr_t	moea64_scratchpage_pte[2];
struct	mtx	moea64_scratchpage_mtx;

uint64_t 	moea64_large_page_mask = 0;
int		moea64_large_page_size = 0;
int		moea64_large_page_shift = 0;

/*
 * PVO calls.
 */
static int	moea64_pvo_enter(mmu_t, pmap_t, uma_zone_t, struct pvo_head *,
		    vm_offset_t, vm_offset_t, uint64_t, int);
static void	moea64_pvo_remove(mmu_t, struct pvo_entry *);
static struct	pvo_entry *moea64_pvo_find_va(pmap_t, vm_offset_t);

/*
 * Utility routines.
 */
static void		moea64_enter_locked(mmu_t, pmap_t, vm_offset_t,
			    vm_page_t, vm_prot_t, boolean_t);
static boolean_t	moea64_query_bit(mmu_t, vm_page_t, u_int64_t);
static u_int		moea64_clear_bit(mmu_t, vm_page_t, u_int64_t);
static void		moea64_kremove(mmu_t, vm_offset_t);
static void		moea64_syncicache(mmu_t, pmap_t pmap, vm_offset_t va, 
			    vm_offset_t pa, vm_size_t sz);

/*
 * Kernel MMU interface
 */
void moea64_change_wiring(mmu_t, pmap_t, vm_offset_t, boolean_t);
void moea64_clear_modify(mmu_t, vm_page_t);
void moea64_clear_reference(mmu_t, vm_page_t);
void moea64_copy_page(mmu_t, vm_page_t, vm_page_t);
void moea64_enter(mmu_t, pmap_t, vm_offset_t, vm_page_t, vm_prot_t, boolean_t);
void moea64_enter_object(mmu_t, pmap_t, vm_offset_t, vm_offset_t, vm_page_t,
    vm_prot_t);
void moea64_enter_quick(mmu_t, pmap_t, vm_offset_t, vm_page_t, vm_prot_t);
vm_paddr_t moea64_extract(mmu_t, pmap_t, vm_offset_t);
vm_page_t moea64_extract_and_hold(mmu_t, pmap_t, vm_offset_t, vm_prot_t);
void moea64_init(mmu_t);
boolean_t moea64_is_modified(mmu_t, vm_page_t);
boolean_t moea64_is_prefaultable(mmu_t, pmap_t, vm_offset_t);
boolean_t moea64_is_referenced(mmu_t, vm_page_t);
boolean_t moea64_ts_referenced(mmu_t, vm_page_t);
vm_offset_t moea64_map(mmu_t, vm_offset_t *, vm_offset_t, vm_offset_t, int);
boolean_t moea64_page_exists_quick(mmu_t, pmap_t, vm_page_t);
int moea64_page_wired_mappings(mmu_t, vm_page_t);
void moea64_pinit(mmu_t, pmap_t);
void moea64_pinit0(mmu_t, pmap_t);
void moea64_protect(mmu_t, pmap_t, vm_offset_t, vm_offset_t, vm_prot_t);
void moea64_qenter(mmu_t, vm_offset_t, vm_page_t *, int);
void moea64_qremove(mmu_t, vm_offset_t, int);
void moea64_release(mmu_t, pmap_t);
void moea64_remove(mmu_t, pmap_t, vm_offset_t, vm_offset_t);
void moea64_remove_all(mmu_t, vm_page_t);
void moea64_remove_write(mmu_t, vm_page_t);
void moea64_zero_page(mmu_t, vm_page_t);
void moea64_zero_page_area(mmu_t, vm_page_t, int, int);
void moea64_zero_page_idle(mmu_t, vm_page_t);
void moea64_activate(mmu_t, struct thread *);
void moea64_deactivate(mmu_t, struct thread *);
void *moea64_mapdev(mmu_t, vm_offset_t, vm_size_t);
void *moea64_mapdev_attr(mmu_t, vm_offset_t, vm_size_t, vm_memattr_t);
void moea64_unmapdev(mmu_t, vm_offset_t, vm_size_t);
vm_offset_t moea64_kextract(mmu_t, vm_offset_t);
void moea64_page_set_memattr(mmu_t, vm_page_t m, vm_memattr_t ma);
void moea64_kenter_attr(mmu_t, vm_offset_t, vm_offset_t, vm_memattr_t ma);
void moea64_kenter(mmu_t, vm_offset_t, vm_offset_t);
boolean_t moea64_dev_direct_mapped(mmu_t, vm_offset_t, vm_size_t);
static void moea64_sync_icache(mmu_t, pmap_t, vm_offset_t, vm_size_t);

static mmu_method_t moea64_methods[] = {
	MMUMETHOD(mmu_change_wiring,	moea64_change_wiring),
	MMUMETHOD(mmu_clear_modify,	moea64_clear_modify),
	MMUMETHOD(mmu_clear_reference,	moea64_clear_reference),
	MMUMETHOD(mmu_copy_page,	moea64_copy_page),
	MMUMETHOD(mmu_enter,		moea64_enter),
	MMUMETHOD(mmu_enter_object,	moea64_enter_object),
	MMUMETHOD(mmu_enter_quick,	moea64_enter_quick),
	MMUMETHOD(mmu_extract,		moea64_extract),
	MMUMETHOD(mmu_extract_and_hold,	moea64_extract_and_hold),
	MMUMETHOD(mmu_init,		moea64_init),
	MMUMETHOD(mmu_is_modified,	moea64_is_modified),
	MMUMETHOD(mmu_is_prefaultable,	moea64_is_prefaultable),
	MMUMETHOD(mmu_is_referenced,	moea64_is_referenced),
	MMUMETHOD(mmu_ts_referenced,	moea64_ts_referenced),
	MMUMETHOD(mmu_map,     		moea64_map),
	MMUMETHOD(mmu_page_exists_quick,moea64_page_exists_quick),
	MMUMETHOD(mmu_page_wired_mappings,moea64_page_wired_mappings),
	MMUMETHOD(mmu_pinit,		moea64_pinit),
	MMUMETHOD(mmu_pinit0,		moea64_pinit0),
	MMUMETHOD(mmu_protect,		moea64_protect),
	MMUMETHOD(mmu_qenter,		moea64_qenter),
	MMUMETHOD(mmu_qremove,		moea64_qremove),
	MMUMETHOD(mmu_release,		moea64_release),
	MMUMETHOD(mmu_remove,		moea64_remove),
	MMUMETHOD(mmu_remove_all,      	moea64_remove_all),
	MMUMETHOD(mmu_remove_write,	moea64_remove_write),
	MMUMETHOD(mmu_sync_icache,	moea64_sync_icache),
	MMUMETHOD(mmu_zero_page,       	moea64_zero_page),
	MMUMETHOD(mmu_zero_page_area,	moea64_zero_page_area),
	MMUMETHOD(mmu_zero_page_idle,	moea64_zero_page_idle),
	MMUMETHOD(mmu_activate,		moea64_activate),
	MMUMETHOD(mmu_deactivate,      	moea64_deactivate),
	MMUMETHOD(mmu_page_set_memattr,	moea64_page_set_memattr),

	/* Internal interfaces */
	MMUMETHOD(mmu_mapdev,		moea64_mapdev),
	MMUMETHOD(mmu_mapdev_attr,	moea64_mapdev_attr),
	MMUMETHOD(mmu_unmapdev,		moea64_unmapdev),
	MMUMETHOD(mmu_kextract,		moea64_kextract),
	MMUMETHOD(mmu_kenter,		moea64_kenter),
	MMUMETHOD(mmu_kenter_attr,	moea64_kenter_attr),
	MMUMETHOD(mmu_dev_direct_mapped,moea64_dev_direct_mapped),

	{ 0, 0 }
};

MMU_DEF(oea64_mmu, "mmu_oea64_base", moea64_methods, 0);

static __inline u_int
va_to_pteg(uint64_t vsid, vm_offset_t addr, int large)
{
	uint64_t hash;
	int shift;

	shift = large ? moea64_large_page_shift : ADDR_PIDX_SHFT;
	hash = (vsid & VSID_HASH_MASK) ^ (((uint64_t)addr & ADDR_PIDX) >>
	    shift);
	return (hash & moea64_pteg_mask);
}

static __inline struct pvo_head *
vm_page_to_pvoh(vm_page_t m)
{

	return (&m->md.mdpg_pvoh);
}

static __inline void
moea64_attr_clear(vm_page_t m, u_int64_t ptebit)
{

	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
	m->md.mdpg_attrs &= ~ptebit;
}

static __inline u_int64_t
moea64_attr_fetch(vm_page_t m)
{

	return (m->md.mdpg_attrs);
}

static __inline void
moea64_attr_save(vm_page_t m, u_int64_t ptebit)
{

	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
	m->md.mdpg_attrs |= ptebit;
}

static __inline void
moea64_pte_create(struct lpte *pt, uint64_t vsid, vm_offset_t va, 
    uint64_t pte_lo, int flags)
{

	ASSERT_TABLE_LOCK();

	/*
	 * Construct a PTE.  Default to IMB initially.  Valid bit only gets
	 * set when the real pte is set in memory.
	 *
	 * Note: Don't set the valid bit for correct operation of tlb update.
	 */
	pt->pte_hi = (vsid << LPTE_VSID_SHIFT) |
	    (((uint64_t)(va & ADDR_PIDX) >> ADDR_API_SHFT64) & LPTE_API);

	if (flags & PVO_LARGE)
		pt->pte_hi |= LPTE_BIG;

	pt->pte_lo = pte_lo;
}

static __inline uint64_t
moea64_calc_wimg(vm_offset_t pa, vm_memattr_t ma)
{
	uint64_t pte_lo;
	int i;

	if (ma != VM_MEMATTR_DEFAULT) {
		switch (ma) {
		case VM_MEMATTR_UNCACHEABLE:
			return (LPTE_I | LPTE_G);
		case VM_MEMATTR_WRITE_COMBINING:
		case VM_MEMATTR_WRITE_BACK:
		case VM_MEMATTR_PREFETCHABLE:
			return (LPTE_I);
		case VM_MEMATTR_WRITE_THROUGH:
			return (LPTE_W | LPTE_M);
		}
	}

	/*
	 * Assume the page is cache inhibited and access is guarded unless
	 * it's in our available memory array.
	 */
	pte_lo = LPTE_I | LPTE_G;
	for (i = 0; i < pregions_sz; i++) {
		if ((pa >= pregions[i].mr_start) &&
		    (pa < (pregions[i].mr_start + pregions[i].mr_size))) {
			pte_lo &= ~(LPTE_I | LPTE_G);
			pte_lo |= LPTE_M;
			break;
		}
	}

	return pte_lo;
}

/*
 * Quick sort callout for comparing memory regions.
 */
static int	om_cmp(const void *a, const void *b);

static int
om_cmp(const void *a, const void *b)
{
	const struct	ofw_map *mapa;
	const struct	ofw_map *mapb;

	mapa = a;
	mapb = b;
	if (mapa->om_pa_hi < mapb->om_pa_hi)
		return (-1);
	else if (mapa->om_pa_hi > mapb->om_pa_hi)
		return (1);
	else if (mapa->om_pa_lo < mapb->om_pa_lo)
		return (-1);
	else if (mapa->om_pa_lo > mapb->om_pa_lo)
		return (1);
	else
		return (0);
}

static void
moea64_add_ofw_mappings(mmu_t mmup, phandle_t mmu, size_t sz)
{
	struct ofw_map	translations[sz/sizeof(struct ofw_map)];
	register_t	msr;
	vm_offset_t	off;
	vm_paddr_t	pa_base;
	int		i;

	bzero(translations, sz);
	if (OF_getprop(mmu, "translations", translations, sz) == -1)
		panic("moea64_bootstrap: can't get ofw translations");

	CTR0(KTR_PMAP, "moea64_add_ofw_mappings: translations");
	sz /= sizeof(*translations);
	qsort(translations, sz, sizeof (*translations), om_cmp);

	for (i = 0; i < sz; i++) {
		CTR3(KTR_PMAP, "translation: pa=%#x va=%#x len=%#x",
		    (uint32_t)(translations[i].om_pa_lo), translations[i].om_va,
		    translations[i].om_len);

		if (translations[i].om_pa_lo % PAGE_SIZE)
			panic("OFW translation not page-aligned!");

		pa_base = translations[i].om_pa_lo;

	      #ifdef __powerpc64__
		pa_base += (vm_offset_t)translations[i].om_pa_hi << 32;
	      #else
		if (translations[i].om_pa_hi)
			panic("OFW translations above 32-bit boundary!");
	      #endif

		/* Now enter the pages for this mapping */

		DISABLE_TRANS(msr);
		for (off = 0; off < translations[i].om_len; off += PAGE_SIZE) {
			if (moea64_pvo_find_va(kernel_pmap,
			    translations[i].om_va + off) != NULL)
				continue;

			moea64_kenter(mmup, translations[i].om_va + off,
			    pa_base + off);
		}
		ENABLE_TRANS(msr);
	}
}

#ifdef __powerpc64__
static void
moea64_probe_large_page(void)
{
	uint16_t pvr = mfpvr() >> 16;

	switch (pvr) {
	case IBM970:
	case IBM970FX:
	case IBM970MP:
		powerpc_sync(); isync();
		mtspr(SPR_HID4, mfspr(SPR_HID4) & ~HID4_970_DISABLE_LG_PG);
		powerpc_sync(); isync();
		
		/* FALLTHROUGH */
	case IBMCELLBE:
		moea64_large_page_size = 0x1000000; /* 16 MB */
		moea64_large_page_shift = 24;
		break;
	default:
		moea64_large_page_size = 0;
	}

	moea64_large_page_mask = moea64_large_page_size - 1;
}

static void
moea64_bootstrap_slb_prefault(vm_offset_t va, int large)
{
	struct slb *cache;
	struct slb entry;
	uint64_t esid, slbe;
	uint64_t i;

	cache = PCPU_GET(slb);
	esid = va >> ADDR_SR_SHFT;
	slbe = (esid << SLBE_ESID_SHIFT) | SLBE_VALID;

	for (i = 0; i < 64; i++) {
		if (cache[i].slbe == (slbe | i))
			return;
	}

	entry.slbe = slbe;
	entry.slbv = KERNEL_VSID(esid) << SLBV_VSID_SHIFT;
	if (large)
		entry.slbv |= SLBV_L;

	slb_insert_kernel(entry.slbe, entry.slbv);
}
#endif

static void
moea64_setup_direct_map(mmu_t mmup, vm_offset_t kernelstart,
    vm_offset_t kernelend)
{
	register_t msr;
	vm_paddr_t pa;
	vm_offset_t size, off;
	uint64_t pte_lo;
	int i;

	if (moea64_large_page_size == 0) 
		hw_direct_map = 0;

	DISABLE_TRANS(msr);
	if (hw_direct_map) {
		PMAP_LOCK(kernel_pmap);
		for (i = 0; i < pregions_sz; i++) {
		  for (pa = pregions[i].mr_start; pa < pregions[i].mr_start +
		     pregions[i].mr_size; pa += moea64_large_page_size) {
			pte_lo = LPTE_M;

			/*
			 * Set memory access as guarded if prefetch within
			 * the page could exit the available physmem area.
			 */
			if (pa & moea64_large_page_mask) {
				pa &= moea64_large_page_mask;
				pte_lo |= LPTE_G;
			}
			if (pa + moea64_large_page_size >
			    pregions[i].mr_start + pregions[i].mr_size)
				pte_lo |= LPTE_G;

			moea64_pvo_enter(mmup, kernel_pmap, moea64_upvo_zone,
				    &moea64_pvo_kunmanaged, pa, pa,
				    pte_lo, PVO_WIRED | PVO_LARGE);
		  }
		}
		PMAP_UNLOCK(kernel_pmap);
	} else {
		size = sizeof(struct pvo_head) * moea64_pteg_count;
		off = (vm_offset_t)(moea64_pvo_table);
		for (pa = off; pa < off + size; pa += PAGE_SIZE) 
			moea64_kenter(mmup, pa, pa);
		size = BPVO_POOL_SIZE*sizeof(struct pvo_entry);
		off = (vm_offset_t)(moea64_bpvo_pool);
		for (pa = off; pa < off + size; pa += PAGE_SIZE) 
		moea64_kenter(mmup, pa, pa);

		/*
		 * Map certain important things, like ourselves.
		 *
		 * NOTE: We do not map the exception vector space. That code is
		 * used only in real mode, and leaving it unmapped allows us to
		 * catch NULL pointer deferences, instead of making NULL a valid
		 * address.
		 */

		for (pa = kernelstart & ~PAGE_MASK; pa < kernelend;
		    pa += PAGE_SIZE) 
			moea64_kenter(mmup, pa, pa);
	}
	ENABLE_TRANS(msr);
}

void
moea64_early_bootstrap(mmu_t mmup, vm_offset_t kernelstart, vm_offset_t kernelend)
{
	int		i, j;
	vm_size_t	physsz, hwphyssz;

#ifndef __powerpc64__
	/* We don't have a direct map since there is no BAT */
	hw_direct_map = 0;

	/* Make sure battable is zero, since we have no BAT */
	for (i = 0; i < 16; i++) {
		battable[i].batu = 0;
		battable[i].batl = 0;
	}
#else
	moea64_probe_large_page();

	/* Use a direct map if we have large page support */
	if (moea64_large_page_size > 0)
		hw_direct_map = 1;
	else
		hw_direct_map = 0;
#endif

	/* Get physical memory regions from firmware */
	mem_regions(&pregions, &pregions_sz, &regions, &regions_sz);
	CTR0(KTR_PMAP, "moea64_bootstrap: physical memory");

	if (sizeof(phys_avail)/sizeof(phys_avail[0]) < regions_sz)
		panic("moea64_bootstrap: phys_avail too small");

	phys_avail_count = 0;
	physsz = 0;
	hwphyssz = 0;
	TUNABLE_ULONG_FETCH("hw.physmem", (u_long *) &hwphyssz);
	for (i = 0, j = 0; i < regions_sz; i++, j += 2) {
		CTR3(KTR_PMAP, "region: %#x - %#x (%#x)", regions[i].mr_start,
		    regions[i].mr_start + regions[i].mr_size,
		    regions[i].mr_size);
		if (hwphyssz != 0 &&
		    (physsz + regions[i].mr_size) >= hwphyssz) {
			if (physsz < hwphyssz) {
				phys_avail[j] = regions[i].mr_start;
				phys_avail[j + 1] = regions[i].mr_start +
				    hwphyssz - physsz;
				physsz = hwphyssz;
				phys_avail_count++;
			}
			break;
		}
		phys_avail[j] = regions[i].mr_start;
		phys_avail[j + 1] = regions[i].mr_start + regions[i].mr_size;
		phys_avail_count++;
		physsz += regions[i].mr_size;
	}

	/* Check for overlap with the kernel and exception vectors */
	for (j = 0; j < 2*phys_avail_count; j+=2) {
		if (phys_avail[j] < EXC_LAST)
			phys_avail[j] += EXC_LAST;

		if (kernelstart >= phys_avail[j] &&
		    kernelstart < phys_avail[j+1]) {
			if (kernelend < phys_avail[j+1]) {
				phys_avail[2*phys_avail_count] =
				    (kernelend & ~PAGE_MASK) + PAGE_SIZE;
				phys_avail[2*phys_avail_count + 1] =
				    phys_avail[j+1];
				phys_avail_count++;
			}

			phys_avail[j+1] = kernelstart & ~PAGE_MASK;
		}

		if (kernelend >= phys_avail[j] &&
		    kernelend < phys_avail[j+1]) {
			if (kernelstart > phys_avail[j]) {
				phys_avail[2*phys_avail_count] = phys_avail[j];
				phys_avail[2*phys_avail_count + 1] =
				    kernelstart & ~PAGE_MASK;
				phys_avail_count++;
			}

			phys_avail[j] = (kernelend & ~PAGE_MASK) + PAGE_SIZE;
		}
	}

	physmem = btoc(physsz);

#ifdef PTEGCOUNT
	moea64_pteg_count = PTEGCOUNT;
#else
	moea64_pteg_count = 0x1000;

	while (moea64_pteg_count < physmem)
		moea64_pteg_count <<= 1;

	moea64_pteg_count >>= 1;
#endif /* PTEGCOUNT */
}

void
moea64_mid_bootstrap(mmu_t mmup, vm_offset_t kernelstart, vm_offset_t kernelend)
{
	vm_size_t	size;
	register_t	msr;
	int		i;

	/*
	 * Set PTEG mask
	 */
	moea64_pteg_mask = moea64_pteg_count - 1;

	/*
	 * Allocate pv/overflow lists.
	 */
	size = sizeof(struct pvo_head) * moea64_pteg_count;

	moea64_pvo_table = (struct pvo_head *)moea64_bootstrap_alloc(size,
	    PAGE_SIZE);
	CTR1(KTR_PMAP, "moea64_bootstrap: PVO table at %p", moea64_pvo_table);

	DISABLE_TRANS(msr);
	for (i = 0; i < moea64_pteg_count; i++)
		LIST_INIT(&moea64_pvo_table[i]);
	ENABLE_TRANS(msr);

	/*
	 * Initialize the lock that synchronizes access to the pteg and pvo
	 * tables.
	 */
	mtx_init(&moea64_table_mutex, "pmap table", NULL, MTX_DEF |
	    MTX_RECURSE);
	mtx_init(&moea64_slb_mutex, "SLB table", NULL, MTX_DEF);

	/*
	 * Initialise the unmanaged pvo pool.
	 */
	moea64_bpvo_pool = (struct pvo_entry *)moea64_bootstrap_alloc(
		BPVO_POOL_SIZE*sizeof(struct pvo_entry), 0);
	moea64_bpvo_pool_index = 0;

	/*
	 * Make sure kernel vsid is allocated as well as VSID 0.
	 */
	#ifndef __powerpc64__
	moea64_vsid_bitmap[(KERNEL_VSIDBITS & (NVSIDS - 1)) / VSID_NBPW]
		|= 1 << (KERNEL_VSIDBITS % VSID_NBPW);
	moea64_vsid_bitmap[0] |= 1;
	#endif

	/*
	 * Initialize the kernel pmap (which is statically allocated).
	 */
	#ifdef __powerpc64__
	for (i = 0; i < 64; i++) {
		pcpup->pc_slb[i].slbv = 0;
		pcpup->pc_slb[i].slbe = 0;
	}
	#else
	for (i = 0; i < 16; i++) 
		kernel_pmap->pm_sr[i] = EMPTY_SEGMENT + i;
	#endif

	kernel_pmap->pmap_phys = kernel_pmap;
	CPU_FILL(&kernel_pmap->pm_active);
	LIST_INIT(&kernel_pmap->pmap_pvo);

	PMAP_LOCK_INIT(kernel_pmap);

	/*
	 * Now map in all the other buffers we allocated earlier
	 */

	moea64_setup_direct_map(mmup, kernelstart, kernelend);
}

void
moea64_late_bootstrap(mmu_t mmup, vm_offset_t kernelstart, vm_offset_t kernelend)
{
	ihandle_t	mmui;
	phandle_t	chosen;
	phandle_t	mmu;
	size_t		sz;
	int		i;
	vm_offset_t	pa, va;
	void		*dpcpu;

	/*
	 * Set up the Open Firmware pmap and add its mappings if not in real
	 * mode.
	 */

	chosen = OF_finddevice("/chosen");
	if (chosen != -1 && OF_getprop(chosen, "mmu", &mmui, 4) != -1) {
	    mmu = OF_instance_to_package(mmui);
	    if (mmu == -1 || (sz = OF_getproplen(mmu, "translations")) == -1)
		sz = 0;
	    if (sz > 6144 /* tmpstksz - 2 KB headroom */)
		panic("moea64_bootstrap: too many ofw translations");

	    if (sz > 0)
		moea64_add_ofw_mappings(mmup, mmu, sz);
	}

	/*
	 * Calculate the last available physical address.
	 */
	for (i = 0; phys_avail[i + 2] != 0; i += 2)
		;
	Maxmem = powerpc_btop(phys_avail[i + 1]);

	/*
	 * Initialize MMU and remap early physical mappings
	 */
	MMU_CPU_BOOTSTRAP(mmup,0);
	mtmsr(mfmsr() | PSL_DR | PSL_IR);
	pmap_bootstrapped++;
	bs_remap_earlyboot();

	/*
	 * Set the start and end of kva.
	 */
	virtual_avail = VM_MIN_KERNEL_ADDRESS;
	virtual_end = VM_MAX_SAFE_KERNEL_ADDRESS; 

	/*
	 * Map the entire KVA range into the SLB. We must not fault there.
	 */
	#ifdef __powerpc64__
	for (va = virtual_avail; va < virtual_end; va += SEGMENT_LENGTH)
		moea64_bootstrap_slb_prefault(va, 0);
	#endif

	/*
	 * Figure out how far we can extend virtual_end into segment 16
	 * without running into existing mappings. Segment 16 is guaranteed
	 * to contain neither RAM nor devices (at least on Apple hardware),
	 * but will generally contain some OFW mappings we should not
	 * step on.
	 */

	#ifndef __powerpc64__	/* KVA is in high memory on PPC64 */
	PMAP_LOCK(kernel_pmap);
	while (virtual_end < VM_MAX_KERNEL_ADDRESS &&
	    moea64_pvo_find_va(kernel_pmap, virtual_end+1) == NULL)
		virtual_end += PAGE_SIZE;
	PMAP_UNLOCK(kernel_pmap);
	#endif

	/*
	 * Allocate a kernel stack with a guard page for thread0 and map it
	 * into the kernel page map.
	 */
	pa = moea64_bootstrap_alloc(KSTACK_PAGES * PAGE_SIZE, PAGE_SIZE);
	va = virtual_avail + KSTACK_GUARD_PAGES * PAGE_SIZE;
	virtual_avail = va + KSTACK_PAGES * PAGE_SIZE;
	CTR2(KTR_PMAP, "moea64_bootstrap: kstack0 at %#x (%#x)", pa, va);
	thread0.td_kstack = va;
	thread0.td_kstack_pages = KSTACK_PAGES;
	for (i = 0; i < KSTACK_PAGES; i++) {
		moea64_kenter(mmup, va, pa);
		pa += PAGE_SIZE;
		va += PAGE_SIZE;
	}

	/*
	 * Allocate virtual address space for the message buffer.
	 */
	pa = msgbuf_phys = moea64_bootstrap_alloc(msgbufsize, PAGE_SIZE);
	msgbufp = (struct msgbuf *)virtual_avail;
	va = virtual_avail;
	virtual_avail += round_page(msgbufsize);
	while (va < virtual_avail) {
		moea64_kenter(mmup, va, pa);
		pa += PAGE_SIZE;
		va += PAGE_SIZE;
	}

	/*
	 * Allocate virtual address space for the dynamic percpu area.
	 */
	pa = moea64_bootstrap_alloc(DPCPU_SIZE, PAGE_SIZE);
	dpcpu = (void *)virtual_avail;
	va = virtual_avail;
	virtual_avail += DPCPU_SIZE;
	while (va < virtual_avail) {
		moea64_kenter(mmup, va, pa);
		pa += PAGE_SIZE;
		va += PAGE_SIZE;
	}
	dpcpu_init(dpcpu, 0);

	/*
	 * Allocate some things for page zeroing. We put this directly
	 * in the page table, marked with LPTE_LOCKED, to avoid any
	 * of the PVO book-keeping or other parts of the VM system
	 * from even knowing that this hack exists.
	 */

	if (!hw_direct_map) {
		mtx_init(&moea64_scratchpage_mtx, "pvo zero page", NULL,
		    MTX_DEF);
		for (i = 0; i < 2; i++) {
			moea64_scratchpage_va[i] = (virtual_end+1) - PAGE_SIZE;
			virtual_end -= PAGE_SIZE;

			moea64_kenter(mmup, moea64_scratchpage_va[i], 0);

			moea64_scratchpage_pvo[i] = moea64_pvo_find_va(
			    kernel_pmap, (vm_offset_t)moea64_scratchpage_va[i]);
			LOCK_TABLE();
			moea64_scratchpage_pte[i] = MOEA64_PVO_TO_PTE(
			    mmup, moea64_scratchpage_pvo[i]);
			moea64_scratchpage_pvo[i]->pvo_pte.lpte.pte_hi
			    |= LPTE_LOCKED;
			MOEA64_PTE_CHANGE(mmup, moea64_scratchpage_pte[i],
			    &moea64_scratchpage_pvo[i]->pvo_pte.lpte,
			    moea64_scratchpage_pvo[i]->pvo_vpn);
			UNLOCK_TABLE();
		}
	}
}

/*
 * Activate a user pmap.  The pmap must be activated before its address
 * space can be accessed in any way.
 */
void
moea64_activate(mmu_t mmu, struct thread *td)
{
	pmap_t	pm;

	pm = &td->td_proc->p_vmspace->vm_pmap;
	CPU_SET(PCPU_GET(cpuid), &pm->pm_active);

	#ifdef __powerpc64__
	PCPU_SET(userslb, pm->pm_slb);
	#else
	PCPU_SET(curpmap, pm->pmap_phys);
	#endif
}

void
moea64_deactivate(mmu_t mmu, struct thread *td)
{
	pmap_t	pm;

	pm = &td->td_proc->p_vmspace->vm_pmap;
	CPU_CLR(PCPU_GET(cpuid), &pm->pm_active);
	#ifdef __powerpc64__
	PCPU_SET(userslb, NULL);
	#else
	PCPU_SET(curpmap, NULL);
	#endif
}

void
moea64_change_wiring(mmu_t mmu, pmap_t pm, vm_offset_t va, boolean_t wired)
{
	struct	pvo_entry *pvo;
	uintptr_t pt;
	uint64_t vsid;
	int	i, ptegidx;

	PMAP_LOCK(pm);
	pvo = moea64_pvo_find_va(pm, va & ~ADDR_POFF);

	if (pvo != NULL) {
		LOCK_TABLE();
		pt = MOEA64_PVO_TO_PTE(mmu, pvo);

		if (wired) {
			if ((pvo->pvo_vaddr & PVO_WIRED) == 0)
				pm->pm_stats.wired_count++;
			pvo->pvo_vaddr |= PVO_WIRED;
			pvo->pvo_pte.lpte.pte_hi |= LPTE_WIRED;
		} else {
			if ((pvo->pvo_vaddr & PVO_WIRED) != 0)
				pm->pm_stats.wired_count--;
			pvo->pvo_vaddr &= ~PVO_WIRED;
			pvo->pvo_pte.lpte.pte_hi &= ~LPTE_WIRED;
		}

		if (pt != -1) {
			/* Update wiring flag in page table. */
			MOEA64_PTE_CHANGE(mmu, pt, &pvo->pvo_pte.lpte,
			    pvo->pvo_vpn);
		} else if (wired) {
			/*
			 * If we are wiring the page, and it wasn't in the
			 * page table before, add it.
			 */
			vsid = PVO_VSID(pvo);
			ptegidx = va_to_pteg(vsid, PVO_VADDR(pvo),
			    pvo->pvo_vaddr & PVO_LARGE);

			i = MOEA64_PTE_INSERT(mmu, ptegidx, &pvo->pvo_pte.lpte);
			
			if (i >= 0) {
				PVO_PTEGIDX_CLR(pvo);
				PVO_PTEGIDX_SET(pvo, i);
			}
		}
			
		UNLOCK_TABLE();
	}
	PMAP_UNLOCK(pm);
}

/*
 * This goes through and sets the physical address of our
 * special scratch PTE to the PA we want to zero or copy. Because
 * of locking issues (this can get called in pvo_enter() by
 * the UMA allocator), we can't use most other utility functions here
 */

static __inline
void moea64_set_scratchpage_pa(mmu_t mmup, int which, vm_offset_t pa) {

	KASSERT(!hw_direct_map, ("Using OEA64 scratchpage with a direct map!"));
	mtx_assert(&moea64_scratchpage_mtx, MA_OWNED);

	moea64_scratchpage_pvo[which]->pvo_pte.lpte.pte_lo &=
	    ~(LPTE_WIMG | LPTE_RPGN);
	moea64_scratchpage_pvo[which]->pvo_pte.lpte.pte_lo |=
	    moea64_calc_wimg(pa, VM_MEMATTR_DEFAULT) | (uint64_t)pa;
	MOEA64_PTE_CHANGE(mmup, moea64_scratchpage_pte[which],
	    &moea64_scratchpage_pvo[which]->pvo_pte.lpte,
	    moea64_scratchpage_pvo[which]->pvo_vpn);
	isync();
}

void
moea64_copy_page(mmu_t mmu, vm_page_t msrc, vm_page_t mdst)
{
	vm_offset_t	dst;
	vm_offset_t	src;

	dst = VM_PAGE_TO_PHYS(mdst);
	src = VM_PAGE_TO_PHYS(msrc);

	if (hw_direct_map) {
		kcopy((void *)src, (void *)dst, PAGE_SIZE);
	} else {
		mtx_lock(&moea64_scratchpage_mtx);

		moea64_set_scratchpage_pa(mmu, 0, src);
		moea64_set_scratchpage_pa(mmu, 1, dst);

		kcopy((void *)moea64_scratchpage_va[0], 
		    (void *)moea64_scratchpage_va[1], PAGE_SIZE);

		mtx_unlock(&moea64_scratchpage_mtx);
	}
}

void
moea64_zero_page_area(mmu_t mmu, vm_page_t m, int off, int size)
{
	vm_offset_t pa = VM_PAGE_TO_PHYS(m);

	if (size + off > PAGE_SIZE)
		panic("moea64_zero_page: size + off > PAGE_SIZE");

	if (hw_direct_map) {
		bzero((caddr_t)pa + off, size);
	} else {
		mtx_lock(&moea64_scratchpage_mtx);
		moea64_set_scratchpage_pa(mmu, 0, pa);
		bzero((caddr_t)moea64_scratchpage_va[0] + off, size);
		mtx_unlock(&moea64_scratchpage_mtx);
	}
}

/*
 * Zero a page of physical memory by temporarily mapping it
 */
void
moea64_zero_page(mmu_t mmu, vm_page_t m)
{
	vm_offset_t pa = VM_PAGE_TO_PHYS(m);
	vm_offset_t va, off;

	if (!hw_direct_map) {
		mtx_lock(&moea64_scratchpage_mtx);

		moea64_set_scratchpage_pa(mmu, 0, pa);
		va = moea64_scratchpage_va[0];
	} else {
		va = pa;
	}

	for (off = 0; off < PAGE_SIZE; off += cacheline_size)
		__asm __volatile("dcbz 0,%0" :: "r"(va + off));

	if (!hw_direct_map)
		mtx_unlock(&moea64_scratchpage_mtx);
}

void
moea64_zero_page_idle(mmu_t mmu, vm_page_t m)
{

	moea64_zero_page(mmu, m);
}

/*
 * Map the given physical page at the specified virtual address in the
 * target pmap with the protection requested.  If specified the page
 * will be wired down.
 */
void
moea64_enter(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_page_t m, 
    vm_prot_t prot, boolean_t wired)
{

	vm_page_lock_queues();
	PMAP_LOCK(pmap);
	moea64_enter_locked(mmu, pmap, va, m, prot, wired);
	vm_page_unlock_queues();
	PMAP_UNLOCK(pmap);
}

/*
 * Map the given physical page at the specified virtual address in the
 * target pmap with the protection requested.  If specified the page
 * will be wired down.
 *
 * The page queues and pmap must be locked.
 */

static void
moea64_enter_locked(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_page_t m,
    vm_prot_t prot, boolean_t wired)
{
	struct		pvo_head *pvo_head;
	uma_zone_t	zone;
	vm_page_t	pg;
	uint64_t	pte_lo;
	u_int		pvo_flags;
	int		error;

	if (!moea64_initialized) {
		pvo_head = &moea64_pvo_kunmanaged;
		pg = NULL;
		zone = moea64_upvo_zone;
		pvo_flags = 0;
	} else {
		pvo_head = vm_page_to_pvoh(m);
		pg = m;
		zone = moea64_mpvo_zone;
		pvo_flags = PVO_MANAGED;
	}

	if (pmap_bootstrapped)
		mtx_assert(&vm_page_queue_mtx, MA_OWNED);
	PMAP_LOCK_ASSERT(pmap, MA_OWNED);
	KASSERT((m->oflags & (VPO_UNMANAGED | VPO_BUSY)) != 0 ||
	    VM_OBJECT_LOCKED(m->object),
	    ("moea64_enter_locked: page %p is not busy", m));

	/* XXX change the pvo head for fake pages */
	if ((m->oflags & VPO_UNMANAGED) != 0) {
		pvo_flags &= ~PVO_MANAGED;
		pvo_head = &moea64_pvo_kunmanaged;
		zone = moea64_upvo_zone;
	}

	pte_lo = moea64_calc_wimg(VM_PAGE_TO_PHYS(m), pmap_page_get_memattr(m));

	if (prot & VM_PROT_WRITE) {
		pte_lo |= LPTE_BW;
		if (pmap_bootstrapped &&
		    (m->oflags & VPO_UNMANAGED) == 0)
			vm_page_aflag_set(m, PGA_WRITEABLE);
	} else
		pte_lo |= LPTE_BR;

	if ((prot & VM_PROT_EXECUTE) == 0)
		pte_lo |= LPTE_NOEXEC;

	if (wired)
		pvo_flags |= PVO_WIRED;

	error = moea64_pvo_enter(mmu, pmap, zone, pvo_head, va,
	    VM_PAGE_TO_PHYS(m), pte_lo, pvo_flags);

	/*
	 * Flush the page from the instruction cache if this page is
	 * mapped executable and cacheable.
	 */
	if ((pte_lo & (LPTE_I | LPTE_G | LPTE_NOEXEC)) == 0)
		moea64_syncicache(mmu, pmap, va, VM_PAGE_TO_PHYS(m), PAGE_SIZE);
}

static void
moea64_syncicache(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_offset_t pa,
    vm_size_t sz)
{

	/*
	 * This is much trickier than on older systems because
	 * we can't sync the icache on physical addresses directly
	 * without a direct map. Instead we check a couple of cases
	 * where the memory is already mapped in and, failing that,
	 * use the same trick we use for page zeroing to create
	 * a temporary mapping for this physical address.
	 */

	if (!pmap_bootstrapped) {
		/*
		 * If PMAP is not bootstrapped, we are likely to be
		 * in real mode.
		 */
		__syncicache((void *)pa, sz);
	} else if (pmap == kernel_pmap) {
		__syncicache((void *)va, sz);
	} else if (hw_direct_map) {
		__syncicache((void *)pa, sz);
	} else {
		/* Use the scratch page to set up a temp mapping */

		mtx_lock(&moea64_scratchpage_mtx);

		moea64_set_scratchpage_pa(mmu, 1, pa & ~ADDR_POFF);
		__syncicache((void *)(moea64_scratchpage_va[1] + 
		    (va & ADDR_POFF)), sz);

		mtx_unlock(&moea64_scratchpage_mtx);
	}
}

/*
 * 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.
 */
void
moea64_enter_object(mmu_t mmu, pmap_t pm, 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;

	psize = atop(end - start);
	m = m_start;
	vm_page_lock_queues();
	PMAP_LOCK(pm);
	while (m != NULL && (diff = m->pindex - m_start->pindex) < psize) {
		moea64_enter_locked(mmu, pm, start + ptoa(diff), m, prot &
		    (VM_PROT_READ | VM_PROT_EXECUTE), FALSE);
		m = TAILQ_NEXT(m, listq);
	}
	vm_page_unlock_queues();
	PMAP_UNLOCK(pm);
}

void
moea64_enter_quick(mmu_t mmu, pmap_t pm, vm_offset_t va, vm_page_t m,
    vm_prot_t prot)
{

	vm_page_lock_queues();
	PMAP_LOCK(pm);
	moea64_enter_locked(mmu, pm, va, m,
	    prot & (VM_PROT_READ | VM_PROT_EXECUTE), FALSE);
	vm_page_unlock_queues();
	PMAP_UNLOCK(pm);
}

vm_paddr_t
moea64_extract(mmu_t mmu, pmap_t pm, vm_offset_t va)
{
	struct	pvo_entry *pvo;
	vm_paddr_t pa;

	PMAP_LOCK(pm);
	pvo = moea64_pvo_find_va(pm, va);
	if (pvo == NULL)
		pa = 0;
	else
		pa = (pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN) |
		    (va - PVO_VADDR(pvo));
	PMAP_UNLOCK(pm);
	return (pa);
}

/*
 * Atomically extract and hold the physical page with the given
 * pmap and virtual address pair if that mapping permits the given
 * protection.
 */
vm_page_t
moea64_extract_and_hold(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_prot_t prot)
{
	struct	pvo_entry *pvo;
	vm_page_t m;
        vm_paddr_t pa;
        
	m = NULL;
	pa = 0;
	PMAP_LOCK(pmap);
retry:
	pvo = moea64_pvo_find_va(pmap, va & ~ADDR_POFF);
	if (pvo != NULL && (pvo->pvo_pte.lpte.pte_hi & LPTE_VALID) &&
	    ((pvo->pvo_pte.lpte.pte_lo & LPTE_PP) == LPTE_RW ||
	     (prot & VM_PROT_WRITE) == 0)) {
		if (vm_page_pa_tryrelock(pmap,
			pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN, &pa))
			goto retry;
		m = PHYS_TO_VM_PAGE(pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN);
		vm_page_hold(m);
	}
	PA_UNLOCK_COND(pa);
	PMAP_UNLOCK(pmap);
	return (m);
}

static mmu_t installed_mmu;

static void *
moea64_uma_page_alloc(uma_zone_t zone, int bytes, u_int8_t *flags, int wait) 
{
	/*
	 * This entire routine is a horrible hack to avoid bothering kmem
	 * for new KVA addresses. Because this can get called from inside
	 * kmem allocation routines, calling kmem for a new address here
	 * can lead to multiply locking non-recursive mutexes.
	 */
        vm_offset_t va;

        vm_page_t m;
        int pflags, needed_lock;

	*flags = UMA_SLAB_PRIV;
	needed_lock = !PMAP_LOCKED(kernel_pmap);

	if (needed_lock)
		PMAP_LOCK(kernel_pmap);

        if ((wait & (M_NOWAIT|M_USE_RESERVE)) == M_NOWAIT)
                pflags = VM_ALLOC_INTERRUPT | VM_ALLOC_WIRED;
        else
                pflags = VM_ALLOC_SYSTEM | VM_ALLOC_WIRED;
        if (wait & M_ZERO)
                pflags |= VM_ALLOC_ZERO;

        for (;;) {
                m = vm_page_alloc(NULL, 0, pflags | VM_ALLOC_NOOBJ);
                if (m == NULL) {
                        if (wait & M_NOWAIT)
                                return (NULL);
                        VM_WAIT;
                } else
                        break;
        }

	va = VM_PAGE_TO_PHYS(m);

	moea64_pvo_enter(installed_mmu, kernel_pmap, moea64_upvo_zone,
	    &moea64_pvo_kunmanaged, va, VM_PAGE_TO_PHYS(m), LPTE_M,
	    PVO_WIRED | PVO_BOOTSTRAP);

	if (needed_lock)
		PMAP_UNLOCK(kernel_pmap);
	
	if ((wait & M_ZERO) && (m->flags & PG_ZERO) == 0)
                bzero((void *)va, PAGE_SIZE);

	return (void *)va;
}

void
moea64_init(mmu_t mmu)
{

	CTR0(KTR_PMAP, "moea64_init");

	moea64_upvo_zone = uma_zcreate("UPVO entry", sizeof (struct pvo_entry),
	    NULL, NULL, NULL, NULL, UMA_ALIGN_PTR,
	    UMA_ZONE_VM | UMA_ZONE_NOFREE);
	moea64_mpvo_zone = uma_zcreate("MPVO entry", sizeof(struct pvo_entry),
	    NULL, NULL, NULL, NULL, UMA_ALIGN_PTR,
	    UMA_ZONE_VM | UMA_ZONE_NOFREE);

	if (!hw_direct_map) {
		installed_mmu = mmu;
		uma_zone_set_allocf(moea64_upvo_zone,moea64_uma_page_alloc);
		uma_zone_set_allocf(moea64_mpvo_zone,moea64_uma_page_alloc);
	}

	moea64_initialized = TRUE;
}

boolean_t
moea64_is_referenced(mmu_t mmu, vm_page_t m)
{

	KASSERT((m->oflags & VPO_UNMANAGED) == 0,
	    ("moea64_is_referenced: page %p is not managed", m));
	return (moea64_query_bit(mmu, m, PTE_REF));
}

boolean_t
moea64_is_modified(mmu_t mmu, vm_page_t m)
{

	KASSERT((m->oflags & VPO_UNMANAGED) == 0,
	    ("moea64_is_modified: page %p is not managed", m));

	/*
	 * If the page is not VPO_BUSY, then PGA_WRITEABLE cannot be
	 * concurrently set while the object is locked.  Thus, if PGA_WRITEABLE
	 * is clear, no PTEs can have LPTE_CHG set.
	 */
	VM_OBJECT_LOCK_ASSERT(m->object, MA_OWNED);
	if ((m->oflags & VPO_BUSY) == 0 &&
	    (m->aflags & PGA_WRITEABLE) == 0)
		return (FALSE);
	return (moea64_query_bit(mmu, m, LPTE_CHG));
}

boolean_t
moea64_is_prefaultable(mmu_t mmu, pmap_t pmap, vm_offset_t va)
{
	struct pvo_entry *pvo;
	boolean_t rv;

	PMAP_LOCK(pmap);
	pvo = moea64_pvo_find_va(pmap, va & ~ADDR_POFF);
	rv = pvo == NULL || (pvo->pvo_pte.lpte.pte_hi & LPTE_VALID) == 0;
	PMAP_UNLOCK(pmap);
	return (rv);
}

void
moea64_clear_reference(mmu_t mmu, vm_page_t m)
{

	KASSERT((m->oflags & VPO_UNMANAGED) == 0,
	    ("moea64_clear_reference: page %p is not managed", m));
	moea64_clear_bit(mmu, m, LPTE_REF);
}

void
moea64_clear_modify(mmu_t mmu, vm_page_t m)
{

	KASSERT((m->oflags & VPO_UNMANAGED) == 0,
	    ("moea64_clear_modify: page %p is not managed", m));
	VM_OBJECT_LOCK_ASSERT(m->object, MA_OWNED);
	KASSERT((m->oflags & VPO_BUSY) == 0,
	    ("moea64_clear_modify: page %p is busy", m));

	/*
	 * If the page is not PGA_WRITEABLE, then no PTEs can have LPTE_CHG
	 * set.  If the object containing the page is locked and the page is
	 * not VPO_BUSY, then PGA_WRITEABLE cannot be concurrently set.
	 */
	if ((m->aflags & PGA_WRITEABLE) == 0)
		return;
	moea64_clear_bit(mmu, m, LPTE_CHG);
}

/*
 * Clear the write and modified bits in each of the given page's mappings.
 */
void
moea64_remove_write(mmu_t mmu, vm_page_t m)
{
	struct	pvo_entry *pvo;
	uintptr_t pt;
	pmap_t	pmap;
	uint64_t lo;

	KASSERT((m->oflags & VPO_UNMANAGED) == 0,
	    ("moea64_remove_write: page %p is not managed", m));

	/*
	 * If the page is not VPO_BUSY, 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_LOCK_ASSERT(m->object, MA_OWNED);
	if ((m->oflags & VPO_BUSY) == 0 &&
	    (m->aflags & PGA_WRITEABLE) == 0)
		return;
	vm_page_lock_queues();
	lo = moea64_attr_fetch(m);
	powerpc_sync();
	LIST_FOREACH(pvo, vm_page_to_pvoh(m), pvo_vlink) {
		pmap = pvo->pvo_pmap;
		PMAP_LOCK(pmap);
		LOCK_TABLE();
		if ((pvo->pvo_pte.lpte.pte_lo & LPTE_PP) != LPTE_BR) {
			pt = MOEA64_PVO_TO_PTE(mmu, pvo);
			pvo->pvo_pte.lpte.pte_lo &= ~LPTE_PP;
			pvo->pvo_pte.lpte.pte_lo |= LPTE_BR;
			if (pt != -1) {
				MOEA64_PTE_SYNCH(mmu, pt, &pvo->pvo_pte.lpte);
				lo |= pvo->pvo_pte.lpte.pte_lo;
				pvo->pvo_pte.lpte.pte_lo &= ~LPTE_CHG;
				MOEA64_PTE_CHANGE(mmu, pt,
				    &pvo->pvo_pte.lpte, pvo->pvo_vpn);
				if (pvo->pvo_pmap == kernel_pmap)
					isync();
			}
		}
		UNLOCK_TABLE();
		PMAP_UNLOCK(pmap);
	}
	if ((lo & LPTE_CHG) != 0) {
		moea64_attr_clear(m, LPTE_CHG);
		vm_page_dirty(m);
	}
	vm_page_aflag_clear(m, PGA_WRITEABLE);
	vm_page_unlock_queues();
}

/*
 *	moea64_ts_referenced:
 *
 *	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.
 */
boolean_t
moea64_ts_referenced(mmu_t mmu, vm_page_t m)
{

	KASSERT((m->oflags & VPO_UNMANAGED) == 0,
	    ("moea64_ts_referenced: page %p is not managed", m));
	return (moea64_clear_bit(mmu, m, LPTE_REF));
}

/*
 * Modify the WIMG settings of all mappings for a page.
 */
void
moea64_page_set_memattr(mmu_t mmu, vm_page_t m, vm_memattr_t ma)
{
	struct	pvo_entry *pvo;
	struct  pvo_head *pvo_head;
	uintptr_t pt;
	pmap_t	pmap;
	uint64_t lo;

	if ((m->oflags & VPO_UNMANAGED) != 0) {
		m->md.mdpg_cache_attrs = ma;
		return;
	}

	vm_page_lock_queues();
	pvo_head = vm_page_to_pvoh(m);
	lo = moea64_calc_wimg(VM_PAGE_TO_PHYS(m), ma);
	LIST_FOREACH(pvo, pvo_head, pvo_vlink) {
		pmap = pvo->pvo_pmap;
		PMAP_LOCK(pmap);
		LOCK_TABLE();
		pt = MOEA64_PVO_TO_PTE(mmu, pvo);
		pvo->pvo_pte.lpte.pte_lo &= ~LPTE_WIMG;
		pvo->pvo_pte.lpte.pte_lo |= lo;
		if (pt != -1) {
			MOEA64_PTE_CHANGE(mmu, pt, &pvo->pvo_pte.lpte,
			    pvo->pvo_vpn);
			if (pvo->pvo_pmap == kernel_pmap)
				isync();
		}
		UNLOCK_TABLE();
		PMAP_UNLOCK(pmap);
	}
	m->md.mdpg_cache_attrs = ma;
	vm_page_unlock_queues();
}

/*
 * Map a wired page into kernel virtual address space.
 */
void
moea64_kenter_attr(mmu_t mmu, vm_offset_t va, vm_offset_t pa, vm_memattr_t ma)
{
	uint64_t	pte_lo;
	int		error;	

	pte_lo = moea64_calc_wimg(pa, ma);

	PMAP_LOCK(kernel_pmap);
	error = moea64_pvo_enter(mmu, kernel_pmap, moea64_upvo_zone,
	    &moea64_pvo_kunmanaged, va, pa, pte_lo, PVO_WIRED);

	if (error != 0 && error != ENOENT)
		panic("moea64_kenter: failed to enter va %#zx pa %#zx: %d", va,
		    pa, error);

	/*
	 * Flush the memory from the instruction cache.
	 */
	if ((pte_lo & (LPTE_I | LPTE_G)) == 0)
		__syncicache((void *)va, PAGE_SIZE);
	PMAP_UNLOCK(kernel_pmap);
}

void
moea64_kenter(mmu_t mmu, vm_offset_t va, vm_offset_t pa)
{

	moea64_kenter_attr(mmu, va, pa, VM_MEMATTR_DEFAULT);
}

/*
 * Extract the physical page address associated with the given kernel virtual
 * address.
 */
vm_offset_t
moea64_kextract(mmu_t mmu, vm_offset_t va)
{
	struct		pvo_entry *pvo;
	vm_paddr_t pa;

	/*
	 * Shortcut the direct-mapped case when applicable.  We never put
	 * anything but 1:1 mappings below VM_MIN_KERNEL_ADDRESS.
	 */
	if (va < VM_MIN_KERNEL_ADDRESS)
		return (va);

	PMAP_LOCK(kernel_pmap);
	pvo = moea64_pvo_find_va(kernel_pmap, va);
	KASSERT(pvo != NULL, ("moea64_kextract: no addr found for %#" PRIxPTR,
	    va));
	pa = (pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN) | (va - PVO_VADDR(pvo));
	PMAP_UNLOCK(kernel_pmap);
	return (pa);
}

/*
 * Remove a wired page from kernel virtual address space.
 */
void
moea64_kremove(mmu_t mmu, vm_offset_t va)
{
	moea64_remove(mmu, kernel_pmap, va, va + PAGE_SIZE);
}

/*
 * Map a range of physical addresses into kernel virtual address space.
 *
 * The value passed in *virt is a suggested virtual address for the mapping.
 * Architectures which can support a direct-mapped physical to virtual region
 * can return the appropriate address within that region, leaving '*virt'
 * unchanged.  We cannot and therefore do not; *virt is updated with the
 * first usable address after the mapped region.
 */
vm_offset_t
moea64_map(mmu_t mmu, vm_offset_t *virt, vm_offset_t pa_start,
    vm_offset_t pa_end, int prot)
{
	vm_offset_t	sva, va;

	sva = *virt;
	va = sva;
	for (; pa_start < pa_end; pa_start += PAGE_SIZE, va += PAGE_SIZE)
		moea64_kenter(mmu, va, pa_start);
	*virt = va;

	return (sva);
}

/*
 * Returns 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.
 */
boolean_t
moea64_page_exists_quick(mmu_t mmu, pmap_t pmap, vm_page_t m)
{
        int loops;
	struct pvo_entry *pvo;
	boolean_t rv;

	KASSERT((m->oflags & VPO_UNMANAGED) == 0,
	    ("moea64_page_exists_quick: page %p is not managed", m));
	loops = 0;
	rv = FALSE;
	vm_page_lock_queues();
	LIST_FOREACH(pvo, vm_page_to_pvoh(m), pvo_vlink) {
		if (pvo->pvo_pmap == pmap) {
			rv = TRUE;
			break;
		}
		if (++loops >= 16)
			break;
	}
	vm_page_unlock_queues();
	return (rv);
}

/*
 * Return the number of managed mappings to the given physical page
 * that are wired.
 */
int
moea64_page_wired_mappings(mmu_t mmu, vm_page_t m)
{
	struct pvo_entry *pvo;
	int count;

	count = 0;
	if ((m->oflags & VPO_UNMANAGED) != 0)
		return (count);
	vm_page_lock_queues();
	LIST_FOREACH(pvo, vm_page_to_pvoh(m), pvo_vlink)
		if ((pvo->pvo_vaddr & PVO_WIRED) != 0)
			count++;
	vm_page_unlock_queues();
	return (count);
}

static uintptr_t	moea64_vsidcontext;

uintptr_t
moea64_get_unique_vsid(void) {
	u_int entropy;
	register_t hash;
	uint32_t mask;
	int i;

	entropy = 0;
	__asm __volatile("mftb %0" : "=r"(entropy));

	mtx_lock(&moea64_slb_mutex);
	for (i = 0; i < NVSIDS; i += VSID_NBPW) {
		u_int	n;

		/*
		 * Create a new value by mutiplying by a prime and adding in
		 * entropy from the timebase register.  This is to make the
		 * VSID more random so that the PT hash function collides
		 * less often.  (Note that the prime casues gcc to do shifts
		 * instead of a multiply.)
		 */
		moea64_vsidcontext = (moea64_vsidcontext * 0x1105) + entropy;
		hash = moea64_vsidcontext & (NVSIDS - 1);
		if (hash == 0)		/* 0 is special, avoid it */
			continue;
		n = hash >> 5;
		mask = 1 << (hash & (VSID_NBPW - 1));
		hash = (moea64_vsidcontext & VSID_HASHMASK);
		if (moea64_vsid_bitmap[n] & mask) {	/* collision? */
			/* anything free in this bucket? */
			if (moea64_vsid_bitmap[n] == 0xffffffff) {
				entropy = (moea64_vsidcontext >> 20);
				continue;
			}
			i = ffs(~moea64_vsid_bitmap[n]) - 1;
			mask = 1 << i;
			hash &= VSID_HASHMASK & ~(VSID_NBPW - 1);
			hash |= i;
		}
		KASSERT(!(moea64_vsid_bitmap[n] & mask),
		    ("Allocating in-use VSID %#zx\n", hash));
		moea64_vsid_bitmap[n] |= mask;
		mtx_unlock(&moea64_slb_mutex);
		return (hash);
	}

	mtx_unlock(&moea64_slb_mutex);
	panic("%s: out of segments",__func__);
}

#ifdef __powerpc64__
void
moea64_pinit(mmu_t mmu, pmap_t pmap)
{
	PMAP_LOCK_INIT(pmap);
	LIST_INIT(&pmap->pmap_pvo);

	pmap->pm_slb_tree_root = slb_alloc_tree();
	pmap->pm_slb = slb_alloc_user_cache();
	pmap->pm_slb_len = 0;
}
#else
void
moea64_pinit(mmu_t mmu, pmap_t pmap)
{
	int	i;
	uint32_t hash;

	PMAP_LOCK_INIT(pmap);
	LIST_INIT(&pmap->pmap_pvo);

	if (pmap_bootstrapped)
		pmap->pmap_phys = (pmap_t)moea64_kextract(mmu,
		    (vm_offset_t)pmap);
	else
		pmap->pmap_phys = pmap;

	/*
	 * Allocate some segment registers for this pmap.
	 */
	hash = moea64_get_unique_vsid();

	for (i = 0; i < 16; i++) 
		pmap->pm_sr[i] = VSID_MAKE(i, hash);

	KASSERT(pmap->pm_sr[0] != 0, ("moea64_pinit: pm_sr[0] = 0"));
}
#endif

/*
 * Initialize the pmap associated with process 0.
 */
void
moea64_pinit0(mmu_t mmu, pmap_t pm)
{
	moea64_pinit(mmu, pm);
	bzero(&pm->pm_stats, sizeof(pm->pm_stats));
}

/*
 * Set the physical protection on the specified range of this map as requested.
 */
void
moea64_protect(mmu_t mmu, pmap_t pm, vm_offset_t sva, vm_offset_t eva,
    vm_prot_t prot)
{
	struct	pvo_entry *pvo;
	uintptr_t pt;

	CTR4(KTR_PMAP, "moea64_protect: pm=%p sva=%#x eva=%#x prot=%#x", pm, sva,
	    eva, prot);


	KASSERT(pm == &curproc->p_vmspace->vm_pmap || pm == kernel_pmap,
	    ("moea64_protect: non current pmap"));

	if ((prot & VM_PROT_READ) == VM_PROT_NONE) {
		moea64_remove(mmu, pm, sva, eva);
		return;
	}

	vm_page_lock_queues();
	PMAP_LOCK(pm);
	for (; sva < eva; sva += PAGE_SIZE) {
		pvo = moea64_pvo_find_va(pm, sva);
		if (pvo == NULL)
			continue;

		/*
		 * Grab the PTE pointer before we diddle with the cached PTE
		 * copy.
		 */
		LOCK_TABLE();
		pt = MOEA64_PVO_TO_PTE(mmu, pvo);

		/*
		 * Change the protection of the page.
		 */
		pvo->pvo_pte.lpte.pte_lo &= ~LPTE_PP;
		pvo->pvo_pte.lpte.pte_lo |= LPTE_BR;
		pvo->pvo_pte.lpte.pte_lo &= ~LPTE_NOEXEC;
		if ((prot & VM_PROT_EXECUTE) == 0) 
			pvo->pvo_pte.lpte.pte_lo |= LPTE_NOEXEC;

		/*
		 * If the PVO is in the page table, update that pte as well.
		 */
		if (pt != -1) {
			MOEA64_PTE_CHANGE(mmu, pt, &pvo->pvo_pte.lpte,
			    pvo->pvo_vpn);
			if ((pvo->pvo_pte.lpte.pte_lo & 
			    (LPTE_I | LPTE_G | LPTE_NOEXEC)) == 0) {
				moea64_syncicache(mmu, pm, sva,
				    pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN,
				    PAGE_SIZE);
			}
		}
		UNLOCK_TABLE();
	}
	vm_page_unlock_queues();
	PMAP_UNLOCK(pm);
}

/*
 * 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.
 */
void
moea64_qenter(mmu_t mmu, vm_offset_t va, vm_page_t *m, int count)
{
	while (count-- > 0) {
		moea64_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 moea64_qenter.
 */
void
moea64_qremove(mmu_t mmu, vm_offset_t va, int count)
{
	while (count-- > 0) {
		moea64_kremove(mmu, va);
		va += PAGE_SIZE;
	}
}

void
moea64_release_vsid(uint64_t vsid)
{
	int idx, mask;

	mtx_lock(&moea64_slb_mutex);
	idx = vsid & (NVSIDS-1);
	mask = 1 << (idx % VSID_NBPW);
	idx /= VSID_NBPW;
	KASSERT(moea64_vsid_bitmap[idx] & mask,
	    ("Freeing unallocated VSID %#jx", vsid));
	moea64_vsid_bitmap[idx] &= ~mask;
	mtx_unlock(&moea64_slb_mutex);
}
	

void
moea64_release(mmu_t mmu, pmap_t pmap)
{
        
	/*
	 * Free segment registers' VSIDs
	 */
    #ifdef __powerpc64__
	slb_free_tree(pmap);
	slb_free_user_cache(pmap->pm_slb);
    #else
	KASSERT(pmap->pm_sr[0] != 0, ("moea64_release: pm_sr[0] = 0"));

	moea64_release_vsid(VSID_TO_HASH(pmap->pm_sr[0]));
    #endif

	PMAP_LOCK_DESTROY(pmap);
}

/*
 * Remove the given range of addresses from the specified map.
 */
void
moea64_remove(mmu_t mmu, pmap_t pm, vm_offset_t sva, vm_offset_t eva)
{
	struct	pvo_entry *pvo;

	vm_page_lock_queues();
	PMAP_LOCK(pm);
	if ((eva - sva)/PAGE_SIZE < 10) {
		for (; sva < eva; sva += PAGE_SIZE) {
			pvo = moea64_pvo_find_va(pm, sva);
			if (pvo != NULL)
				moea64_pvo_remove(mmu, pvo);
		}
	} else {
		LIST_FOREACH(pvo, &pm->pmap_pvo, pvo_plink) {
			if (PVO_VADDR(pvo) < sva || PVO_VADDR(pvo) >= eva)
				continue;
			moea64_pvo_remove(mmu, pvo);
		}
	}
	vm_page_unlock_queues();
	PMAP_UNLOCK(pm);
}

/*
 * Remove physical page from all pmaps in which it resides. moea64_pvo_remove()
 * will reflect changes in pte's back to the vm_page.
 */
void
moea64_remove_all(mmu_t mmu, vm_page_t m)
{
	struct  pvo_head *pvo_head;
	struct	pvo_entry *pvo, *next_pvo;
	pmap_t	pmap;

	vm_page_lock_queues();
	pvo_head = vm_page_to_pvoh(m);
	for (pvo = LIST_FIRST(pvo_head); pvo != NULL; pvo = next_pvo) {
		next_pvo = LIST_NEXT(pvo, pvo_vlink);

		pmap = pvo->pvo_pmap;
		PMAP_LOCK(pmap);
		moea64_pvo_remove(mmu, pvo);
		PMAP_UNLOCK(pmap);
	}
	if ((m->aflags & PGA_WRITEABLE) && moea64_is_modified(mmu, m)) {
		moea64_attr_clear(m, LPTE_CHG);
		vm_page_dirty(m);
	}
	vm_page_aflag_clear(m, PGA_WRITEABLE);
	vm_page_unlock_queues();
}

/*
 * Allocate a physical page of memory directly from the phys_avail map.
 * Can only be called from moea64_bootstrap before avail start and end are
 * calculated.
 */
vm_offset_t
moea64_bootstrap_alloc(vm_size_t size, u_int align)
{
	vm_offset_t	s, e;
	int		i, j;

	size = round_page(size);
	for (i = 0; phys_avail[i + 1] != 0; i += 2) {
		if (align != 0)
			s = (phys_avail[i] + align - 1) & ~(align - 1);
		else
			s = phys_avail[i];
		e = s + size;

		if (s < phys_avail[i] || e > phys_avail[i + 1])
			continue;

		if (s + size > platform_real_maxaddr())
			continue;

		if (s == phys_avail[i]) {
			phys_avail[i] += size;
		} else if (e == phys_avail[i + 1]) {
			phys_avail[i + 1] -= size;
		} else {
			for (j = phys_avail_count * 2; j > i; j -= 2) {
				phys_avail[j] = phys_avail[j - 2];
				phys_avail[j + 1] = phys_avail[j - 1];
			}

			phys_avail[i + 3] = phys_avail[i + 1];
			phys_avail[i + 1] = s;
			phys_avail[i + 2] = e;
			phys_avail_count++;
		}

		return (s);
	}
	panic("moea64_bootstrap_alloc: could not allocate memory");
}

static int
moea64_pvo_enter(mmu_t mmu, pmap_t pm, uma_zone_t zone,
    struct pvo_head *pvo_head, vm_offset_t va, vm_offset_t pa,
    uint64_t pte_lo, int flags)
{
	struct	 pvo_entry *pvo;
	uint64_t vsid;
	int	 first;
	u_int	 ptegidx;
	int	 i;
	int      bootstrap;

	/*
	 * One nasty thing that can happen here is that the UMA calls to
	 * allocate new PVOs need to map more memory, which calls pvo_enter(),
	 * which calls UMA...
	 *
	 * We break the loop by detecting recursion and allocating out of
	 * the bootstrap pool.
	 */

	first = 0;
	bootstrap = (flags & PVO_BOOTSTRAP);

	if (!moea64_initialized)
		bootstrap = 1;

	/*
	 * Compute the PTE Group index.
	 */
	va &= ~ADDR_POFF;
	vsid = va_to_vsid(pm, va);
	ptegidx = va_to_pteg(vsid, va, flags & PVO_LARGE);

	/*
	 * Remove any existing mapping for this page.  Reuse the pvo entry if
	 * there is a mapping.
	 */
	LOCK_TABLE();

	moea64_pvo_enter_calls++;

	LIST_FOREACH(pvo, &moea64_pvo_table[ptegidx], pvo_olink) {
		if (pvo->pvo_pmap == pm && PVO_VADDR(pvo) == va) {
			if ((pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN) == pa &&
			    (pvo->pvo_pte.lpte.pte_lo & (LPTE_NOEXEC | LPTE_PP))
			    == (pte_lo & (LPTE_NOEXEC | LPTE_PP))) {
			    	if (!(pvo->pvo_pte.lpte.pte_hi & LPTE_VALID)) {
					/* Re-insert if spilled */
					i = MOEA64_PTE_INSERT(mmu, ptegidx,
					    &pvo->pvo_pte.lpte);
					if (i >= 0)
						PVO_PTEGIDX_SET(pvo, i);
					moea64_pte_overflow--;
				}
				UNLOCK_TABLE();
				return (0);
			}
			moea64_pvo_remove(mmu, pvo);
			break;
		}
	}

	/*
	 * If we aren't overwriting a mapping, try to allocate.
	 */
	if (bootstrap) {
		if (moea64_bpvo_pool_index >= BPVO_POOL_SIZE) {
			panic("moea64_enter: bpvo pool exhausted, %d, %d, %zd",
			      moea64_bpvo_pool_index, BPVO_POOL_SIZE, 
			      BPVO_POOL_SIZE * sizeof(struct pvo_entry));
		}
		pvo = &moea64_bpvo_pool[moea64_bpvo_pool_index];
		moea64_bpvo_pool_index++;
		bootstrap = 1;
	} else {
		/*
		 * Note: drop the table lock around the UMA allocation in
		 * case the UMA allocator needs to manipulate the page
		 * table. The mapping we are working with is already
		 * protected by the PMAP lock.
		 */
		UNLOCK_TABLE();
		pvo = uma_zalloc(zone, M_NOWAIT);
		LOCK_TABLE();
	}

	if (pvo == NULL) {
		UNLOCK_TABLE();
		return (ENOMEM);
	}

	moea64_pvo_entries++;
	pvo->pvo_vaddr = va;
	pvo->pvo_vpn = (uint64_t)((va & ADDR_PIDX) >> ADDR_PIDX_SHFT)
	    | (vsid << 16);
	pvo->pvo_pmap = pm;
	LIST_INSERT_HEAD(&moea64_pvo_table[ptegidx], pvo, pvo_olink);
	pvo->pvo_vaddr &= ~ADDR_POFF;

	if (flags & PVO_WIRED)
		pvo->pvo_vaddr |= PVO_WIRED;
	if (pvo_head != &moea64_pvo_kunmanaged)
		pvo->pvo_vaddr |= PVO_MANAGED;
	if (bootstrap)
		pvo->pvo_vaddr |= PVO_BOOTSTRAP;
	if (flags & PVO_LARGE)
		pvo->pvo_vaddr |= PVO_LARGE;

	moea64_pte_create(&pvo->pvo_pte.lpte, vsid, va, 
	    (uint64_t)(pa) | pte_lo, flags);

	/*
	 * Add to pmap list
	 */
	LIST_INSERT_HEAD(&pm->pmap_pvo, pvo, pvo_plink);

	/*
	 * Remember if the list was empty and therefore will be the first
	 * item.
	 */
	if (LIST_FIRST(pvo_head) == NULL)
		first = 1;
	LIST_INSERT_HEAD(pvo_head, pvo, pvo_vlink);

	if (pvo->pvo_vaddr & PVO_WIRED) {
		pvo->pvo_pte.lpte.pte_hi |= LPTE_WIRED;
		pm->pm_stats.wired_count++;
	}
	pm->pm_stats.resident_count++;

	/*
	 * We hope this succeeds but it isn't required.
	 */
	i = MOEA64_PTE_INSERT(mmu, ptegidx, &pvo->pvo_pte.lpte);
	if (i >= 0) {
		PVO_PTEGIDX_SET(pvo, i);
	} else {
		panic("moea64_pvo_enter: overflow");
		moea64_pte_overflow++;
	}

	if (pm == kernel_pmap)
		isync();

	UNLOCK_TABLE();

#ifdef __powerpc64__
	/*
	 * Make sure all our bootstrap mappings are in the SLB as soon
	 * as virtual memory is switched on.
	 */
	if (!pmap_bootstrapped)
		moea64_bootstrap_slb_prefault(va, flags & PVO_LARGE);
#endif

	return (first ? ENOENT : 0);
}

static void
moea64_pvo_remove(mmu_t mmu, struct pvo_entry *pvo)
{
	uintptr_t pt;

	/*
	 * If there is an active pte entry, we need to deactivate it (and
	 * save the ref & cfg bits).
	 */
	LOCK_TABLE();
	pt = MOEA64_PVO_TO_PTE(mmu, pvo);
	if (pt != -1) {
		MOEA64_PTE_UNSET(mmu, pt, &pvo->pvo_pte.lpte, pvo->pvo_vpn);
		PVO_PTEGIDX_CLR(pvo);
	} else {
		moea64_pte_overflow--;
	}

	/*
	 * Update our statistics.
	 */
	pvo->pvo_pmap->pm_stats.resident_count--;
	if (pvo->pvo_vaddr & PVO_WIRED)
		pvo->pvo_pmap->pm_stats.wired_count--;

	/*
	 * Save the REF/CHG bits into their cache if the page is managed.
	 */
	if ((pvo->pvo_vaddr & PVO_MANAGED) == PVO_MANAGED) {
		struct	vm_page *pg;

		pg = PHYS_TO_VM_PAGE(pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN);
		if (pg != NULL) {
			moea64_attr_save(pg, pvo->pvo_pte.lpte.pte_lo &
			    (LPTE_REF | LPTE_CHG));
		}
	}

	/*
	 * Remove this PVO from the PV and pmap lists.
	 */
	LIST_REMOVE(pvo, pvo_vlink);
	LIST_REMOVE(pvo, pvo_plink);

	/*
	 * Remove this from the overflow list and return it to the pool
	 * if we aren't going to reuse it.
	 */
	LIST_REMOVE(pvo, pvo_olink);

	moea64_pvo_entries--;
	moea64_pvo_remove_calls++;

	UNLOCK_TABLE();

	if (!(pvo->pvo_vaddr & PVO_BOOTSTRAP))
		uma_zfree((pvo->pvo_vaddr & PVO_MANAGED) ? moea64_mpvo_zone :
		    moea64_upvo_zone, pvo);
}

static struct pvo_entry *
moea64_pvo_find_va(pmap_t pm, vm_offset_t va)
{
	struct		pvo_entry *pvo;
	int		ptegidx;
	uint64_t	vsid;
	#ifdef __powerpc64__
	uint64_t	slbv;

	if (pm == kernel_pmap) {
		slbv = kernel_va_to_slbv(va);
	} else {
		struct slb *slb;
		slb = user_va_to_slb_entry(pm, va);
		/* The page is not mapped if the segment isn't */
		if (slb == NULL)
			return NULL;
		slbv = slb->slbv;
	}

	vsid = (slbv & SLBV_VSID_MASK) >> SLBV_VSID_SHIFT;
	if (slbv & SLBV_L)
		va &= ~moea64_large_page_mask;
	else
		va &= ~ADDR_POFF;
	ptegidx = va_to_pteg(vsid, va, slbv & SLBV_L);
	#else
	va &= ~ADDR_POFF;
	vsid = va_to_vsid(pm, va);
	ptegidx = va_to_pteg(vsid, va, 0);
	#endif

	LOCK_TABLE();
	LIST_FOREACH(pvo, &moea64_pvo_table[ptegidx], pvo_olink) {
		if (pvo->pvo_pmap == pm && PVO_VADDR(pvo) == va)
			break;
	}
	UNLOCK_TABLE();

	return (pvo);
}

static boolean_t
moea64_query_bit(mmu_t mmu, vm_page_t m, u_int64_t ptebit)
{
	struct	pvo_entry *pvo;
	uintptr_t pt;

	if (moea64_attr_fetch(m) & ptebit)
		return (TRUE);

	vm_page_lock_queues();

	LIST_FOREACH(pvo, vm_page_to_pvoh(m), pvo_vlink) {

		/*
		 * See if we saved the bit off.  If so, cache it and return
		 * success.
		 */
		if (pvo->pvo_pte.lpte.pte_lo & ptebit) {
			moea64_attr_save(m, ptebit);
			vm_page_unlock_queues();
			return (TRUE);
		}
	}

	/*
	 * No luck, now go through the hard part of looking at the PTEs
	 * themselves.  Sync so that any pending REF/CHG bits are flushed to
	 * the PTEs.
	 */
	powerpc_sync();
	LIST_FOREACH(pvo, vm_page_to_pvoh(m), pvo_vlink) {

		/*
		 * See if this pvo has a valid PTE.  if so, fetch the
		 * REF/CHG bits from the valid PTE.  If the appropriate
		 * ptebit is set, cache it and return success.
		 */
		LOCK_TABLE();
		pt = MOEA64_PVO_TO_PTE(mmu, pvo);
		if (pt != -1) {
			MOEA64_PTE_SYNCH(mmu, pt, &pvo->pvo_pte.lpte);
			if (pvo->pvo_pte.lpte.pte_lo & ptebit) {
				UNLOCK_TABLE();

				moea64_attr_save(m, ptebit);
				vm_page_unlock_queues();
				return (TRUE);
			}
		}
		UNLOCK_TABLE();
	}

	vm_page_unlock_queues();
	return (FALSE);
}

static u_int
moea64_clear_bit(mmu_t mmu, vm_page_t m, u_int64_t ptebit)
{
	u_int	count;
	struct	pvo_entry *pvo;
	uintptr_t pt;

	vm_page_lock_queues();

	/*
	 * Clear the cached value.
	 */
	moea64_attr_clear(m, ptebit);

	/*
	 * Sync so that any pending REF/CHG bits are flushed to the PTEs (so
	 * we can reset the right ones).  note that since the pvo entries and
	 * list heads are accessed via BAT0 and are never placed in the page
	 * table, we don't have to worry about further accesses setting the
	 * REF/CHG bits.
	 */
	powerpc_sync();

	/*
	 * For each pvo entry, clear the pvo's ptebit.  If this pvo has a
	 * valid pte clear the ptebit from the valid pte.
	 */
	count = 0;
	LIST_FOREACH(pvo, vm_page_to_pvoh(m), pvo_vlink) {

		LOCK_TABLE();
		pt = MOEA64_PVO_TO_PTE(mmu, pvo);
		if (pt != -1) {
			MOEA64_PTE_SYNCH(mmu, pt, &pvo->pvo_pte.lpte);
			if (pvo->pvo_pte.lpte.pte_lo & ptebit) {
				count++;
				MOEA64_PTE_CLEAR(mmu, pt, &pvo->pvo_pte.lpte,
				    pvo->pvo_vpn, ptebit);
			}
		}
		pvo->pvo_pte.lpte.pte_lo &= ~ptebit;
		UNLOCK_TABLE();
	}

	vm_page_unlock_queues();
	return (count);
}

boolean_t
moea64_dev_direct_mapped(mmu_t mmu, vm_offset_t pa, vm_size_t size)
{
	struct pvo_entry *pvo;
	vm_offset_t ppa;
	int error = 0;

	PMAP_LOCK(kernel_pmap);
	for (ppa = pa & ~ADDR_POFF; ppa < pa + size; ppa += PAGE_SIZE) {
		pvo = moea64_pvo_find_va(kernel_pmap, ppa);
		if (pvo == NULL ||
		    (pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN) != ppa) {
			error = EFAULT;
			break;
		}
	}
	PMAP_UNLOCK(kernel_pmap);

	return (error);
}

/*
 * 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.
 */
void *
moea64_mapdev_attr(mmu_t mmu, vm_offset_t pa, vm_size_t size, vm_memattr_t ma)
{
	vm_offset_t va, tmpva, ppa, offset;

	ppa = trunc_page(pa);
	offset = pa & PAGE_MASK;
	size = roundup(offset + size, PAGE_SIZE);

	va = kmem_alloc_nofault(kernel_map, size);

	if (!va)
		panic("moea64_mapdev: Couldn't alloc kernel virtual memory");

	for (tmpva = va; size > 0;) {
		moea64_kenter_attr(mmu, tmpva, ppa, ma);
		size -= PAGE_SIZE;
		tmpva += PAGE_SIZE;
		ppa += PAGE_SIZE;
	}

	return ((void *)(va + offset));
}

void *
moea64_mapdev(mmu_t mmu, vm_offset_t pa, vm_size_t size)
{

	return moea64_mapdev_attr(mmu, pa, size, VM_MEMATTR_DEFAULT);
}

void
moea64_unmapdev(mmu_t mmu, vm_offset_t va, vm_size_t size)
{
	vm_offset_t base, offset;

	base = trunc_page(va);
	offset = va & PAGE_MASK;
	size = roundup(offset + size, PAGE_SIZE);

	kmem_free(kernel_map, base, size);
}

void
moea64_sync_icache(mmu_t mmu, pmap_t pm, vm_offset_t va, vm_size_t sz)
{
	struct pvo_entry *pvo;
	vm_offset_t lim;
	vm_paddr_t pa;
	vm_size_t len;

	PMAP_LOCK(pm);
	while (sz > 0) {
		lim = round_page(va);
		len = MIN(lim - va, sz);
		pvo = moea64_pvo_find_va(pm, va & ~ADDR_POFF);
		if (pvo != NULL && !(pvo->pvo_pte.lpte.pte_lo & LPTE_I)) {
			pa = (pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN) |
			    (va & ADDR_POFF);
			moea64_syncicache(mmu, pm, va, pa, len);
		}
		va += len;
		sz -= len;
	}
	PMAP_UNLOCK(pm);
}
OpenPOWER on IntegriCloud