summaryrefslogtreecommitdiff
path: root/i386/intel/pmap.c
blob: 5302b0922e99b5b3a8c292556e4f23ca9f6a3889 (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
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
/*
 * Mach Operating System
 * Copyright (c) 1991,1990,1989,1988 Carnegie Mellon University
 * All Rights Reserved.
 *
 * Permission to use, copy, modify and distribute this software and its
 * documentation is hereby granted, provided that both the copyright
 * notice and this permission notice appear in all copies of the
 * software, derivative works or modified versions, and any portions
 * thereof, and that both notices appear in supporting documentation.
 *
 * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
 * CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND FOR
 * ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
 *
 * Carnegie Mellon requests users of this software to return to
 *
 *  Software Distribution Coordinator  or  Software.Distribution@CS.CMU.EDU
 *  School of Computer Science
 *  Carnegie Mellon University
 *  Pittsburgh PA 15213-3890
 *
 * any improvements or extensions that they make and grant Carnegie Mellon
 * the rights to redistribute these changes.
 */
/*
 *	File:	pmap.c
 *	Author:	Avadis Tevanian, Jr., Michael Wayne Young
 *	(These guys wrote the Vax version)
 *
 *	Physical Map management code for Intel i386, and i486.
 *
 *	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 virtual-to-physical 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 or 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 <string.h>

#include <mach/machine/vm_types.h>

#include <mach/boolean.h>
#include <kern/debug.h>
#include <kern/printf.h>
#include <kern/thread.h>
#include <kern/slab.h>

#include <kern/lock.h>

#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_kern.h>
#include <i386/vm_param.h>
#include <mach/vm_prot.h>
#include <vm/vm_object.h>
#include <vm/vm_page.h>
#include <vm/vm_user.h>

#include <mach/machine/vm_param.h>
#include <mach/xen.h>
#include <machine/thread.h>
#include <i386/cpu_number.h>
#include <i386/proc_reg.h>
#include <i386/locore.h>
#include <i386/model_dep.h>
#include <i386/spl.h>
#include <i386at/biosmem.h>
#include <i386at/model_dep.h>

#if	NCPUS > 1
#include <i386/mp_desc.h>
#endif

#ifdef	MACH_PSEUDO_PHYS
#define	WRITE_PTE(pte_p, pte_entry)		*(pte_p) = pte_entry?pa_to_ma(pte_entry):0;
#else	/* MACH_PSEUDO_PHYS */
#define	WRITE_PTE(pte_p, pte_entry)		*(pte_p) = (pte_entry);
#endif	/* MACH_PSEUDO_PHYS */

/*
 *	Private data structures.
 */

/*
 *	For each vm_page_t, there is a list of all currently
 *	valid virtual mappings of that page.  An entry is
 *	a pv_entry_t; the list is the pv_table.
 */

typedef struct pv_entry {
	struct pv_entry	*next;		/* next pv_entry */
	pmap_t		pmap;		/* pmap where mapping lies */
	vm_offset_t	va;		/* virtual address for mapping */
} *pv_entry_t;

#define PV_ENTRY_NULL	((pv_entry_t) 0)

pv_entry_t	pv_head_table;		/* array of entries, one per page */

/*
 *	pv_list entries are kept on a list that can only be accessed
 *	with the pmap system locked (at SPLVM, not in the cpus_active set).
 *	The list is refilled from the pv_list_cache if it becomes empty.
 */
pv_entry_t	pv_free_list;		/* free list at SPLVM */
decl_simple_lock_data(, pv_free_list_lock)

#define	PV_ALLOC(pv_e) { \
	simple_lock(&pv_free_list_lock); \
	if ((pv_e = pv_free_list) != 0) { \
	    pv_free_list = pv_e->next; \
	} \
	simple_unlock(&pv_free_list_lock); \
}

#define	PV_FREE(pv_e) { \
	simple_lock(&pv_free_list_lock); \
	pv_e->next = pv_free_list; \
	pv_free_list = pv_e; \
	simple_unlock(&pv_free_list_lock); \
}

struct kmem_cache	pv_list_cache;		/* cache of pv_entry structures */

/*
 *	Each entry in the pv_head_table is locked by a bit in the
 *	pv_lock_table.  The lock bits are accessed by the physical
 *	address of the page they lock.
 */

char	*pv_lock_table;		/* pointer to array of bits */
#define pv_lock_table_size(n)	(((n)+BYTE_SIZE-1)/BYTE_SIZE)

/* Has pmap_init completed? */
boolean_t	pmap_initialized = FALSE;

/*
 *	Range of kernel virtual addresses available for kernel memory mapping.
 *	Does not include the virtual addresses used to map physical memory 1-1.
 *	Initialized by pmap_bootstrap.
 */
vm_offset_t kernel_virtual_start;
vm_offset_t kernel_virtual_end;

/*
 *	Index into pv_head table, its lock bits, and the modify/reference
 *	bits.
 */
#define pa_index(pa)	vm_page_table_index(pa)

#define pai_to_pvh(pai)		(&pv_head_table[pai])
#define lock_pvh_pai(pai)	(bit_lock(pai, pv_lock_table))
#define unlock_pvh_pai(pai)	(bit_unlock(pai, pv_lock_table))

/*
 *	Array of physical page attributes for managed pages.
 *	One byte per physical page.
 */
char	*pmap_phys_attributes;

/*
 *	Physical page attributes.  Copy bits from PTE definition.
 */
#define	PHYS_MODIFIED	INTEL_PTE_MOD	/* page modified */
#define	PHYS_REFERENCED	INTEL_PTE_REF	/* page referenced */

/*
 *	Amount of virtual memory mapped by one
 *	page-directory entry.
 */
#define	PDE_MAPPED_SIZE		(pdenum2lin(1))

/*
 *	We allocate page table pages directly from the VM system
 *	through this object.  It maps physical memory.
 */
vm_object_t	pmap_object = VM_OBJECT_NULL;

/*
 *	Locking and TLB invalidation
 */

/*
 *	Locking Protocols:
 *
 *	There are two structures in the pmap module that need locking:
 *	the pmaps themselves, and the per-page pv_lists (which are locked
 *	by locking the pv_lock_table entry that corresponds to the pv_head
 *	for the list in question.)  Most routines want to lock a pmap and
 *	then do operations in it that require pv_list locking -- however
 *	pmap_remove_all and pmap_copy_on_write operate on a physical page
 *	basis and want to do the locking in the reverse order, i.e. lock
 *	a pv_list and then go through all the pmaps referenced by that list.
 *	To protect against deadlock between these two cases, the pmap_lock
 *	is used.  There are three different locking protocols as a result:
 *
 *  1.  pmap operations only (pmap_extract, pmap_access, ...)  Lock only
 *		the pmap.
 *
 *  2.  pmap-based operations (pmap_enter, pmap_remove, ...)  Get a read
 *		lock on the pmap_lock (shared read), then lock the pmap
 *		and finally the pv_lists as needed [i.e. pmap lock before
 *		pv_list lock.]
 *
 *  3.  pv_list-based operations (pmap_remove_all, pmap_copy_on_write, ...)
 *		Get a write lock on the pmap_lock (exclusive write); this
 *		also guaranteees exclusive access to the pv_lists.  Lock the
 *		pmaps as needed.
 *
 *	At no time may any routine hold more than one pmap lock or more than
 *	one pv_list lock.  Because interrupt level routines can allocate
 *	mbufs and cause pmap_enter's, the pmap_lock and the lock on the
 *	kernel_pmap can only be held at splvm.
 */

#if	NCPUS > 1
/*
 *	We raise the interrupt level to splvm, to block interprocessor
 *	interrupts during pmap operations.  We must take the CPU out of
 *	the cpus_active set while interrupts are blocked.
 */
#define SPLVM(spl)	{ \
	spl = splvm(); \
	i_bit_clear(cpu_number(), &cpus_active); \
}

#define SPLX(spl)	{ \
	i_bit_set(cpu_number(), &cpus_active); \
	splx(spl); \
}

/*
 *	Lock on pmap system
 */
lock_data_t	pmap_system_lock;

#define PMAP_READ_LOCK(pmap, spl) { \
	SPLVM(spl); \
	lock_read(&pmap_system_lock); \
	simple_lock(&(pmap)->lock); \
}

#define PMAP_WRITE_LOCK(spl) { \
	SPLVM(spl); \
	lock_write(&pmap_system_lock); \
}

#define PMAP_READ_UNLOCK(pmap, spl) { \
	simple_unlock(&(pmap)->lock); \
	lock_read_done(&pmap_system_lock); \
	SPLX(spl); \
}

#define PMAP_WRITE_UNLOCK(spl) { \
	lock_write_done(&pmap_system_lock); \
	SPLX(spl); \
}

#define PMAP_WRITE_TO_READ_LOCK(pmap) { \
	simple_lock(&(pmap)->lock); \
	lock_write_to_read(&pmap_system_lock); \
}

#define LOCK_PVH(index)		(lock_pvh_pai(index))

#define UNLOCK_PVH(index)	(unlock_pvh_pai(index))

#define PMAP_UPDATE_TLBS(pmap, s, e) \
{ \
	cpu_set	cpu_mask = 1 << cpu_number(); \
	cpu_set	users; \
 \
	/* Since the pmap is locked, other updates are locked */ \
	/* out, and any pmap_activate has finished. */ \
 \
	/* find other cpus using the pmap */ \
	users = (pmap)->cpus_using & ~cpu_mask; \
	if (users) { \
	    /* signal them, and wait for them to finish */ \
	    /* using the pmap */ \
	    signal_cpus(users, (pmap), (s), (e)); \
	    while ((pmap)->cpus_using & cpus_active & ~cpu_mask) \
		continue; \
	} \
 \
	/* invalidate our own TLB if pmap is in use */ \
	if ((pmap)->cpus_using & cpu_mask) { \
	    INVALIDATE_TLB((pmap), (s), (e)); \
	} \
}

#else	/* NCPUS > 1 */

#define SPLVM(spl) ((void)(spl))
#define SPLX(spl) ((void)(spl))

#define PMAP_READ_LOCK(pmap, spl)	SPLVM(spl)
#define PMAP_WRITE_LOCK(spl)		SPLVM(spl)
#define PMAP_READ_UNLOCK(pmap, spl)	SPLX(spl)
#define PMAP_WRITE_UNLOCK(spl)		SPLX(spl)
#define PMAP_WRITE_TO_READ_LOCK(pmap)

#define LOCK_PVH(index)
#define UNLOCK_PVH(index)

#define PMAP_UPDATE_TLBS(pmap, s, e) { \
	/* invalidate our own TLB if pmap is in use */ \
	if ((pmap)->cpus_using) { \
	    INVALIDATE_TLB((pmap), (s), (e)); \
	} \
}

#endif	/* NCPUS > 1 */

#ifdef	MACH_PV_PAGETABLES
#define INVALIDATE_TLB(pmap, s, e) do { \
	if (__builtin_constant_p((e) - (s)) \
		&& (e) - (s) == PAGE_SIZE) \
		hyp_invlpg((pmap) == kernel_pmap ? kvtolin(s) : (s)); \
	else \
		hyp_mmuext_op_void(MMUEXT_TLB_FLUSH_LOCAL); \
} while(0)
#else	/* MACH_PV_PAGETABLES */
/* It is hard to know when a TLB flush becomes less expensive than a bunch of
 * invlpgs.  But it surely is more expensive than just one invlpg.  */
#define INVALIDATE_TLB(pmap, s, e) do { \
	if (__builtin_constant_p((e) - (s)) \
		&& (e) - (s) == PAGE_SIZE) \
		invlpg_linear((pmap) == kernel_pmap ? kvtolin(s) : (s)); \
	else \
		flush_tlb(); \
} while (0)
#endif	/* MACH_PV_PAGETABLES */


#if	NCPUS > 1
/*
 *	Structures to keep track of pending TLB invalidations
 */

#define UPDATE_LIST_SIZE	4

struct pmap_update_item {
	pmap_t		pmap;		/* pmap to invalidate */
	vm_offset_t	start;		/* start address to invalidate */
	vm_offset_t	end;		/* end address to invalidate */
} ;

typedef	struct pmap_update_item	*pmap_update_item_t;

/*
 *	List of pmap updates.  If the list overflows,
 *	the last entry is changed to invalidate all.
 */
struct pmap_update_list {
	decl_simple_lock_data(,	lock)
	int			count;
	struct pmap_update_item	item[UPDATE_LIST_SIZE];
} ;
typedef	struct pmap_update_list	*pmap_update_list_t;

struct pmap_update_list	cpu_update_list[NCPUS];

cpu_set		cpus_active;
cpu_set		cpus_idle;
volatile
boolean_t	cpu_update_needed[NCPUS];

#endif	/* NCPUS > 1 */

/*
 *	Other useful macros.
 */
#define current_pmap()		(vm_map_pmap(current_thread()->task->map))
#define pmap_in_use(pmap, cpu)	(((pmap)->cpus_using & (1 << (cpu))) != 0)

struct pmap	kernel_pmap_store;
pmap_t		kernel_pmap;

struct kmem_cache	pmap_cache;		/* cache of pmap structures */
struct kmem_cache	pd_cache;		/* cache of page directories */
#if PAE
struct kmem_cache	pdpt_cache;		/* cache of page
						   directory pointer
						   tables */
#endif

boolean_t		pmap_debug = FALSE;	/* flag for debugging prints */

#if 0
int		ptes_per_vm_page;	/* number of hardware ptes needed
					   to map one VM page. */
#else
#define		ptes_per_vm_page	1
#endif

unsigned int	inuse_ptepages_count = 0;	/* debugging */

/*
 * Pointer to the basic page directory for the kernel.
 * Initialized by pmap_bootstrap().
 */
pt_entry_t *kernel_page_dir;

/*
 * Two slots for temporary physical page mapping, to allow for
 * physical-to-physical transfers.
 */
static pmap_mapwindow_t mapwindows[PMAP_NMAPWINDOWS];

static inline pt_entry_t *
pmap_pde(const pmap_t pmap, vm_offset_t addr)
{
	pt_entry_t *page_dir;
	if (pmap == kernel_pmap)
		addr = kvtolin(addr);
#if PAE
	page_dir = (pt_entry_t *) ptetokv(pmap->pdpbase[lin2pdpnum(addr)]);
#else
	page_dir = pmap->dirbase;
#endif
	return &page_dir[lin2pdenum(addr)];
}

/*
 *	Given an offset and a map, compute the address of the
 *	pte.  If the address is invalid with respect to the map
 *	then PT_ENTRY_NULL is returned (and the map may need to grow).
 *
 *	This is only used internally.
 */
pt_entry_t *
pmap_pte(const pmap_t pmap, vm_offset_t addr)
{
	pt_entry_t	*ptp;
	pt_entry_t	pte;

#if PAE
	if (pmap->pdpbase == 0)
		return(PT_ENTRY_NULL);
#else
	if (pmap->dirbase == 0)
		return(PT_ENTRY_NULL);
#endif
	pte = *pmap_pde(pmap, addr);
	if ((pte & INTEL_PTE_VALID) == 0)
		return(PT_ENTRY_NULL);
	ptp = (pt_entry_t *)ptetokv(pte);
	return(&ptp[ptenum(addr)]);
}

#define DEBUG_PTE_PAGE	0

#if	DEBUG_PTE_PAGE
void ptep_check(ptep_t ptep)
{
	pt_entry_t		*pte, *epte;
	int			ctu, ctw;

	/* check the use and wired counts */
	if (ptep == PTE_PAGE_NULL)
		return;
	pte = pmap_pte(ptep->pmap, ptep->va);
	epte = pte + INTEL_PGBYTES/sizeof(pt_entry_t);
	ctu = 0;
	ctw = 0;
	while (pte < epte) {
		if (pte->pfn != 0) {
			ctu++;
			if (pte->wired)
				ctw++;
		}
		pte += ptes_per_vm_page;
	}

	if (ctu != ptep->use_count || ctw != ptep->wired_count) {
		printf("use %d wired %d - actual use %d wired %d\n",
		    	ptep->use_count, ptep->wired_count, ctu, ctw);
		panic("pte count");
	}
}
#endif	/* DEBUG_PTE_PAGE */

/*
 *	Map memory at initialization.  The physical addresses being
 *	mapped are not managed and are never unmapped.
 *
 *	For now, VM is already on, we only need to map the
 *	specified memory.
 */
vm_offset_t pmap_map(
	vm_offset_t	virt,
	phys_addr_t	start,
	phys_addr_t	end,
	int		prot)
{
	int		ps;

	ps = PAGE_SIZE;
	while (start < end) {
		pmap_enter(kernel_pmap, virt, start, prot, FALSE);
		virt += ps;
		start += ps;
	}
	return(virt);
}

/*
 *	Back-door routine for mapping kernel VM at initialization.
 * 	Useful for mapping memory outside the range of direct mapped
 *	physical memory (i.e., devices).
 *	Otherwise like pmap_map.
 */
vm_offset_t pmap_map_bd(
	vm_offset_t	virt,
	phys_addr_t	start,
	phys_addr_t	end,
	vm_prot_t	prot)
{
	pt_entry_t	template;
	pt_entry_t	*pte;
	int		spl;
#ifdef	MACH_PV_PAGETABLES
	int n, i = 0;
	struct mmu_update update[HYP_BATCH_MMU_UPDATES];
#endif	/* MACH_PV_PAGETABLES */

	template = pa_to_pte(start)
		| INTEL_PTE_NCACHE|INTEL_PTE_WTHRU
		| INTEL_PTE_VALID;
	if (CPU_HAS_FEATURE(CPU_FEATURE_PGE))
		template |= INTEL_PTE_GLOBAL;
	if (prot & VM_PROT_WRITE)
	    template |= INTEL_PTE_WRITE;

	PMAP_READ_LOCK(kernel_pmap, spl);
	while (start < end) {
		pte = pmap_pte(kernel_pmap, virt);
		if (pte == PT_ENTRY_NULL)
			panic("pmap_map_bd: Invalid kernel address\n");
#ifdef	MACH_PV_PAGETABLES
		update[i].ptr = kv_to_ma(pte);
		update[i].val = pa_to_ma(template);
		i++;
		if (i == HYP_BATCH_MMU_UPDATES) {
			hyp_mmu_update(kvtolin(&update), i, kvtolin(&n), DOMID_SELF);
			if (n != i)
				panic("couldn't pmap_map_bd\n");
			i = 0;
		}
#else	/* MACH_PV_PAGETABLES */
		WRITE_PTE(pte, template)
#endif	/* MACH_PV_PAGETABLES */
		pte_increment_pa(template);
		virt += PAGE_SIZE;
		start += PAGE_SIZE;
	}
#ifdef	MACH_PV_PAGETABLES
	if (i > HYP_BATCH_MMU_UPDATES)
		panic("overflowed array in pmap_map_bd");
	hyp_mmu_update(kvtolin(&update), i, kvtolin(&n), DOMID_SELF);
	if (n != i)
		panic("couldn't pmap_map_bd\n");
#endif	/* MACH_PV_PAGETABLES */
	PMAP_READ_UNLOCK(kernel_pmap, spl);
	return(virt);
}

/*
 *	Bootstrap the system enough to run with virtual memory.
 *	Allocate the kernel page directory and page tables,
 *	and direct-map all physical memory.
 *	Called with mapping off.
 */
void pmap_bootstrap(void)
{
	/*
	 * Mapping is turned off; we must reference only physical addresses.
	 * The load image of the system is to be mapped 1-1 physical = virtual.
	 */

	/*
	 *	Set ptes_per_vm_page for general use.
	 */
#if 0
	ptes_per_vm_page = PAGE_SIZE / INTEL_PGBYTES;
#endif

	/*
	 *	The kernel's pmap is statically allocated so we don't
	 *	have to use pmap_create, which is unlikely to work
	 *	correctly at this part of the boot sequence.
	 */

	kernel_pmap = &kernel_pmap_store;

#if	NCPUS > 1
	lock_init(&pmap_system_lock, FALSE);	/* NOT a sleep lock */
#endif	/* NCPUS > 1 */

	simple_lock_init(&kernel_pmap->lock);

	kernel_pmap->ref_count = 1;

	/*
	 * Determine the kernel virtual address range.
	 * It starts at the end of the physical memory
	 * mapped into the kernel address space,
	 * and extends to a stupid arbitrary limit beyond that.
	 */
	kernel_virtual_start = phystokv(biosmem_directmap_end());
	kernel_virtual_end = kernel_virtual_start + VM_KERNEL_MAP_SIZE;

	if (kernel_virtual_end < kernel_virtual_start
			|| kernel_virtual_end > VM_MAX_KERNEL_ADDRESS)
		kernel_virtual_end = VM_MAX_KERNEL_ADDRESS;

	/*
	 * Allocate and clear a kernel page directory.
	 */
	/* Note: initial Xen mapping holds at least 512kB free mapped page.
	 * We use that for directly building our linear mapping. */
#if PAE
	{
		vm_offset_t addr;
		init_alloc_aligned(PDPNUM * INTEL_PGBYTES, &addr);
		kernel_page_dir = (pt_entry_t*)phystokv(addr);
	}
	kernel_pmap->pdpbase = (pt_entry_t*)phystokv(pmap_grab_page());
	memset(kernel_pmap->pdpbase, 0, INTEL_PGBYTES);
	{
		int i;
		for (i = 0; i < PDPNUM; i++)
			WRITE_PTE(&kernel_pmap->pdpbase[i],
				  pa_to_pte(_kvtophys((void *) kernel_page_dir
						      + i * INTEL_PGBYTES))
				  | INTEL_PTE_VALID
#ifdef	MACH_PV_PAGETABLES
				  | INTEL_PTE_WRITE
#endif
				  );
	}
#ifdef __x86_64__
#ifdef MACH_HYP
	kernel_pmap->user_l4base = NULL;
	kernel_pmap->user_pdpbase = NULL;
#endif
	kernel_pmap->l4base = (pt_entry_t*)phystokv(pmap_grab_page());
	memset(kernel_pmap->l4base, 0, INTEL_PGBYTES);
	WRITE_PTE(&kernel_pmap->l4base[0], pa_to_pte(_kvtophys(kernel_pmap->pdpbase)) | INTEL_PTE_VALID | INTEL_PTE_WRITE);
#ifdef	MACH_PV_PAGETABLES
	pmap_set_page_readonly_init(kernel_pmap->l4base);
#endif
#endif	/* x86_64 */
#else	/* PAE */
	kernel_pmap->dirbase = kernel_page_dir = (pt_entry_t*)phystokv(pmap_grab_page());
#endif	/* PAE */
	{
		unsigned i;
		for (i = 0; i < NPDES; i++)
			kernel_page_dir[i] = 0;
	}

#ifdef	MACH_PV_PAGETABLES
	/* We don't actually deal with the CR3 register content at all */
	hyp_vm_assist(VMASST_CMD_enable, VMASST_TYPE_pae_extended_cr3);
	/*
	 * Xen may only provide as few as 512KB extra bootstrap linear memory,
	 * which is far from enough to map all available memory, so we need to
	 * map more bootstrap linear memory. We here map 1 (resp. 4 for PAE)
	 * other L1 table(s), thus 4MiB extra memory (resp. 8MiB), which is
	 * enough for a pagetable mapping 4GiB.
	 */
#ifdef PAE
#define NSUP_L1 4
#else
#define NSUP_L1 1
#endif
	pt_entry_t *l1_map[NSUP_L1];
	{
		vm_offset_t la;
		int n_l1map;
		for (n_l1map = 0, la = VM_MIN_KERNEL_ADDRESS; la >= VM_MIN_KERNEL_ADDRESS; la += NPTES * PAGE_SIZE) {
			pt_entry_t *base = (pt_entry_t*) boot_info.pt_base;
#ifdef	PAE
#ifdef __x86_64__
			base = (pt_entry_t*) ptetokv(base[0]);
#endif /* x86_64 */
			pt_entry_t *l2_map = (pt_entry_t*) ptetokv(base[lin2pdpnum(la)]);
#else	/* PAE */
			pt_entry_t *l2_map = base;
#endif	/* PAE */
			/* Like lin2pdenum, but works with non-contiguous boot L3 */
			l2_map += (la >> PDESHIFT) & PDEMASK;
			if (!(*l2_map & INTEL_PTE_VALID)) {
				struct mmu_update update;
				unsigned j, n;

				l1_map[n_l1map] = (pt_entry_t*) phystokv(pmap_grab_page());
				for (j = 0; j < NPTES; j++)
					l1_map[n_l1map][j] = (((pt_entry_t)pfn_to_mfn(lin2pdenum(la - VM_MIN_KERNEL_ADDRESS) * NPTES + j)) << PAGE_SHIFT) | INTEL_PTE_VALID | INTEL_PTE_WRITE;
				pmap_set_page_readonly_init(l1_map[n_l1map]);
				if (!hyp_mmuext_op_mfn (MMUEXT_PIN_L1_TABLE, kv_to_mfn (l1_map[n_l1map])))
					panic("couldn't pin page %p(%lx)", l1_map[n_l1map], (vm_offset_t) kv_to_ma (l1_map[n_l1map]));
				update.ptr = kv_to_ma(l2_map);
				update.val = kv_to_ma(l1_map[n_l1map]) | INTEL_PTE_VALID | INTEL_PTE_WRITE;
				hyp_mmu_update(kv_to_la(&update), 1, kv_to_la(&n), DOMID_SELF);
				if (n != 1)
					panic("couldn't complete bootstrap map");
				/* added the last L1 table, can stop */
				if (++n_l1map >= NSUP_L1)
					break;
			}
		}
	}
#endif	/* MACH_PV_PAGETABLES */

	/*
	 * Allocate and set up the kernel page tables.
	 */
	{
		vm_offset_t va;
		pt_entry_t global = CPU_HAS_FEATURE(CPU_FEATURE_PGE) ? INTEL_PTE_GLOBAL : 0;

		/*
		 * Map virtual memory for all directly mappable physical memory, 1-1,
		 * Make any mappings completely in the kernel's text segment read-only.
		 *
		 * Also allocate some additional all-null page tables afterwards
		 * for kernel virtual memory allocation,
		 * because this PMAP module is too stupid
		 * to allocate new kernel page tables later.
		 * XX fix this
		 */
		for (va = phystokv(0); va >= phystokv(0) && va < kernel_virtual_end; )
		{
			pt_entry_t *pde = kernel_page_dir + lin2pdenum_cont(kvtolin(va));
			pt_entry_t *ptable = (pt_entry_t*)phystokv(pmap_grab_page());
			pt_entry_t *pte;

			/* Initialize the page directory entry.  */
			WRITE_PTE(pde, pa_to_pte((vm_offset_t)_kvtophys(ptable))
				| INTEL_PTE_VALID | INTEL_PTE_WRITE);

			/* Initialize the page table.  */
			for (pte = ptable; (va < phystokv(biosmem_directmap_end())) && (pte < ptable+NPTES); pte++)
			{
				if ((pte - ptable) < ptenum(va))
				{
					WRITE_PTE(pte, 0);
				}
				else
#ifdef	MACH_PV_PAGETABLES
				if (va == (vm_offset_t) &hyp_shared_info)
				{
					*pte = boot_info.shared_info | INTEL_PTE_VALID | INTEL_PTE_WRITE;
					va += INTEL_PGBYTES;
				}
				else
#endif	/* MACH_PV_PAGETABLES */
				{
					extern char _start[], etext[];

					if (((va >= (vm_offset_t) _start)
					    && (va + INTEL_PGBYTES <= (vm_offset_t)etext))
#ifdef	MACH_PV_PAGETABLES
					    || (va >= (vm_offset_t) boot_info.pt_base
					    && (va + INTEL_PGBYTES <=
					    (vm_offset_t) ptable + INTEL_PGBYTES))
#endif	/* MACH_PV_PAGETABLES */
					    )
					{
						WRITE_PTE(pte, pa_to_pte(_kvtophys(va))
							| INTEL_PTE_VALID | global);
					}
					else
					{
#ifdef	MACH_PV_PAGETABLES
						/* Keep supplementary L1 pages read-only */
						int i;
						for (i = 0; i < NSUP_L1; i++)
							if (va == (vm_offset_t) l1_map[i]) {
								WRITE_PTE(pte, pa_to_pte(_kvtophys(va))
									| INTEL_PTE_VALID | global);
								break;
							}
						if (i == NSUP_L1)
#endif	/* MACH_PV_PAGETABLES */
							WRITE_PTE(pte, pa_to_pte(_kvtophys(va))
								| INTEL_PTE_VALID | INTEL_PTE_WRITE | global)

					}
					va += INTEL_PGBYTES;
				}
			}
			for (; pte < ptable+NPTES; pte++)
			{
				if (va >= kernel_virtual_end - PMAP_NMAPWINDOWS * PAGE_SIZE && va < kernel_virtual_end)
				{
					pmap_mapwindow_t *win = &mapwindows[atop(va - (kernel_virtual_end - PMAP_NMAPWINDOWS * PAGE_SIZE))];
					win->entry = pte;
					win->vaddr = va;
				}
				WRITE_PTE(pte, 0);
				va += INTEL_PGBYTES;
			}
#ifdef	MACH_PV_PAGETABLES
			pmap_set_page_readonly_init(ptable);
			if (!hyp_mmuext_op_mfn (MMUEXT_PIN_L1_TABLE, kv_to_mfn (ptable)))
				panic("couldn't pin page %p(%lx)\n", ptable, (vm_offset_t) kv_to_ma (ptable));
#endif	/* MACH_PV_PAGETABLES */
		}
	}

	/* Architecture-specific code will turn on paging
	   soon after we return from here.  */
}

#ifdef	MACH_PV_PAGETABLES
/* These are only required because of Xen security policies */

/* Set back a page read write */
void pmap_set_page_readwrite(void *_vaddr) {
	vm_offset_t vaddr = (vm_offset_t) _vaddr;
	phys_addr_t paddr = kvtophys(vaddr);
	vm_offset_t canon_vaddr = phystokv(paddr);
	if (hyp_do_update_va_mapping (kvtolin(vaddr), pa_to_pte (pa_to_ma(paddr)) | INTEL_PTE_VALID | INTEL_PTE_WRITE, UVMF_NONE))
		panic("couldn't set hiMMU readwrite for addr %lx(%lx)\n", vaddr, (vm_offset_t) pa_to_ma (paddr));
	if (canon_vaddr != vaddr)
		if (hyp_do_update_va_mapping (kvtolin(canon_vaddr), pa_to_pte (pa_to_ma(paddr)) | INTEL_PTE_VALID | INTEL_PTE_WRITE, UVMF_NONE))
			panic("couldn't set hiMMU readwrite for paddr %lx(%lx)\n", canon_vaddr, (vm_offset_t) pa_to_ma (paddr));
}

/* Set a page read only (so as to pin it for instance) */
void pmap_set_page_readonly(void *_vaddr) {
	vm_offset_t vaddr = (vm_offset_t) _vaddr;
	phys_addr_t paddr = kvtophys(vaddr);
	vm_offset_t canon_vaddr = phystokv(paddr);
	if (*pmap_pde(kernel_pmap, vaddr) & INTEL_PTE_VALID) {
		if (hyp_do_update_va_mapping (kvtolin(vaddr), pa_to_pte (pa_to_ma(paddr)) | INTEL_PTE_VALID, UVMF_NONE))
			panic("couldn't set hiMMU readonly for vaddr %lx(%lx)\n", vaddr, (vm_offset_t) pa_to_ma (paddr));
	}
	if (canon_vaddr != vaddr &&
		*pmap_pde(kernel_pmap, canon_vaddr) & INTEL_PTE_VALID) {
		if (hyp_do_update_va_mapping (kvtolin(canon_vaddr), pa_to_pte (pa_to_ma(paddr)) | INTEL_PTE_VALID, UVMF_NONE))
			panic("couldn't set hiMMU readonly for vaddr %lx canon_vaddr %lx paddr %lx (%lx)\n", vaddr, canon_vaddr, paddr, (vm_offset_t) pa_to_ma (paddr));
	}
}

/* This needs to be called instead of pmap_set_page_readonly as long as RC3
 * still points to the bootstrap dirbase, to also fix the bootstrap table.  */
void pmap_set_page_readonly_init(void *_vaddr) {
	vm_offset_t vaddr = (vm_offset_t) _vaddr;
#if PAE
	pt_entry_t *pdpbase = (void*) boot_info.pt_base;
#ifdef __x86_64__
	pdpbase = (pt_entry_t *) ptetokv(pdpbase[lin2l4num(vaddr)]);
#endif
	/* The bootstrap table does not necessarily use contiguous pages for the pde tables */
	pt_entry_t *dirbase = (void*) ptetokv(pdpbase[lin2pdpnum(vaddr)]);
#else
	pt_entry_t *dirbase = (void*) boot_info.pt_base;
#endif
	pt_entry_t *pte = &dirbase[lin2pdenum(vaddr) & PTEMASK];
	/* Modify our future kernel map (can't use update_va_mapping for this)... */
	if (*pmap_pde(kernel_pmap, vaddr) & INTEL_PTE_VALID) {
		if (!hyp_mmu_update_la (kvtolin(vaddr), pa_to_pte (kv_to_ma(vaddr)) | INTEL_PTE_VALID))
			panic("couldn't set hiMMU readonly for vaddr %lx(%lx)\n", vaddr, (vm_offset_t) kv_to_ma (vaddr));
	}
	/* ... and the bootstrap map.  */
	if (*pte & INTEL_PTE_VALID) {
		if (hyp_do_update_va_mapping (vaddr, pa_to_pte (kv_to_ma(vaddr)) | INTEL_PTE_VALID, UVMF_NONE))
			panic("couldn't set MMU readonly for vaddr %lx(%lx)\n", vaddr, (vm_offset_t) kv_to_ma (vaddr));
	}
}

void pmap_clear_bootstrap_pagetable(pt_entry_t *base) {
	unsigned i;
	pt_entry_t *dir;
	vm_offset_t va = 0;
#ifdef __x86_64__
	int l4i, l3i;
#else
#if PAE
	unsigned j;
#endif	/* PAE */
#endif
	if (!hyp_mmuext_op_mfn (MMUEXT_UNPIN_TABLE, kv_to_mfn(base)))
		panic("pmap_clear_bootstrap_pagetable: couldn't unpin page %p(%lx)\n", base, (vm_offset_t) kv_to_ma(base));
#ifdef __x86_64__
	/* 4-level page table */
	for (l4i = 0; l4i < NPTES && va < HYP_VIRT_START && va < 0x0000800000000000UL; l4i++) {
		pt_entry_t l4e = base[l4i];
		pt_entry_t *l3;
		if (!(l4e & INTEL_PTE_VALID)) {
			va += NPTES * NPTES * NPTES * INTEL_PGBYTES;
			continue;
		}
		l3 = (pt_entry_t *) ptetokv(l4e);

		for (l3i = 0; l3i < NPTES && va < HYP_VIRT_START; l3i++) {
			pt_entry_t l3e = l3[l3i];
			if (!(l3e & INTEL_PTE_VALID)) {
				va += NPTES * NPTES * INTEL_PGBYTES;
				continue;
			}
			dir = (pt_entry_t *) ptetokv(l3e);
#else
#if PAE
	/* 3-level page table */
	for (j = 0; j < PDPNUM && va < HYP_VIRT_START; j++)
	{
			pt_entry_t pdpe = base[j];
			if (!(pdpe & INTEL_PTE_VALID)) {
				va += NPTES * NPTES * INTEL_PGBYTES;
				continue;
			}
			dir = (pt_entry_t *) ptetokv(pdpe);
#else	/* PAE */
			/* 2-level page table */
			dir = base;
#endif	/* PAE */
#endif
			for (i = 0; i < NPTES && va < HYP_VIRT_START; i++) {
				pt_entry_t pde = dir[i];
				unsigned long pfn = atop(pte_to_pa(pde));
				void *pgt = (void*) phystokv(ptoa(pfn));
				if (pde & INTEL_PTE_VALID)
					hyp_free_page(pfn, pgt);
				va += NPTES * INTEL_PGBYTES;
			}
#ifndef __x86_64__
#if PAE
			hyp_free_page(atop(_kvtophys(dir)), dir);
	}
#endif	/* PAE */
#else
			hyp_free_page(atop(_kvtophys(dir)), dir);
		}
		hyp_free_page(atop(_kvtophys(l3)), l3);
	}
#endif
	hyp_free_page(atop(_kvtophys(base)), base);
}
#endif	/* MACH_PV_PAGETABLES */

/*
 * Create a temporary mapping for a given physical entry
 *
 * This can be used to access physical pages which are not mapped 1:1 by
 * phystokv().
 */
pmap_mapwindow_t *pmap_get_mapwindow(pt_entry_t entry)
{
	pmap_mapwindow_t *map;

	assert(entry != 0);

	/* Find an empty one.  */
	for (map = &mapwindows[0]; map < &mapwindows[sizeof (mapwindows) / sizeof (*mapwindows)]; map++)
		if (!(*map->entry))
			break;
	assert(map < &mapwindows[sizeof (mapwindows) / sizeof (*mapwindows)]);

#ifdef MACH_PV_PAGETABLES
	if (!hyp_mmu_update_pte(kv_to_ma(map->entry), pa_to_ma(entry)))
		panic("pmap_get_mapwindow");
#else /* MACH_PV_PAGETABLES */
	WRITE_PTE(map->entry, entry);
#endif /* MACH_PV_PAGETABLES */
	return map;
}

/*
 * Destroy a temporary mapping for a physical entry
 */
void pmap_put_mapwindow(pmap_mapwindow_t *map)
{
#ifdef MACH_PV_PAGETABLES
	if (!hyp_mmu_update_pte(kv_to_ma(map->entry), 0))
		panic("pmap_put_mapwindow");
#else /* MACH_PV_PAGETABLES */
	WRITE_PTE(map->entry, 0);
#endif /* MACH_PV_PAGETABLES */
	PMAP_UPDATE_TLBS(kernel_pmap, map->vaddr, map->vaddr + PAGE_SIZE);
}

void pmap_virtual_space(
	vm_offset_t *startp,
	vm_offset_t *endp)
{
	*startp = kernel_virtual_start;
	*endp = kernel_virtual_end - PMAP_NMAPWINDOWS * PAGE_SIZE;
}

/*
 *	Initialize the pmap module.
 *	Called by vm_init, to initialize any structures that the pmap
 *	system needs to map virtual memory.
 */
void pmap_init(void)
{
	unsigned long		npages;
	vm_offset_t		addr;
	vm_size_t		s;
#if	NCPUS > 1
	int			i;
#endif	/* NCPUS > 1 */

	/*
	 *	Allocate memory for the pv_head_table and its lock bits,
	 *	the modify bit array, and the pte_page table.
	 */

	npages = vm_page_table_size();
	s = (vm_size_t) (sizeof(struct pv_entry) * npages
				+ pv_lock_table_size(npages)
				+ npages);

	s = round_page(s);
	if (kmem_alloc_wired(kernel_map, &addr, s) != KERN_SUCCESS)
		panic("pmap_init");
	memset((void *) addr, 0, s);

	/*
	 *	Allocate the structures first to preserve word-alignment.
	 */
	pv_head_table = (pv_entry_t) addr;
	addr = (vm_offset_t) (pv_head_table + npages);

	pv_lock_table = (char *) addr;
	addr = (vm_offset_t) (pv_lock_table + pv_lock_table_size(npages));

	pmap_phys_attributes = (char *) addr;

	/*
	 *	Create the cache of physical maps,
	 *	and of the physical-to-virtual entries.
	 */
	s = (vm_size_t) sizeof(struct pmap);
	kmem_cache_init(&pmap_cache, "pmap", s, 0, NULL, 0);
	kmem_cache_init(&pd_cache, "pd",
			INTEL_PGBYTES, INTEL_PGBYTES, NULL,
			KMEM_CACHE_PHYSMEM);
#if PAE
	kmem_cache_init(&pdpt_cache, "pdpt",
			INTEL_PGBYTES, INTEL_PGBYTES, NULL,
			KMEM_CACHE_PHYSMEM);
#endif
	s = (vm_size_t) sizeof(struct pv_entry);
	kmem_cache_init(&pv_list_cache, "pv_entry", s, 0, NULL, 0);

#if	NCPUS > 1
	/*
	 *	Set up the pmap request lists
	 */
	for (i = 0; i < NCPUS; i++) {
	    pmap_update_list_t	up = &cpu_update_list[i];

	    simple_lock_init(&up->lock);
	    up->count = 0;
	}
#endif	/* NCPUS > 1 */

	/*
	 * Indicate that the PMAP module is now fully initialized.
	 */
	pmap_initialized = TRUE;
}

static inline boolean_t
valid_page(phys_addr_t addr)
{
	struct vm_page *p;

	if (!pmap_initialized)
		return FALSE;

	p = vm_page_lookup_pa(addr);
	return (p != NULL);
}

/*
 *	Routine:	pmap_page_table_page_alloc
 *
 *	Allocates a new physical page to be used as a page-table page.
 *
 *	Must be called with the pmap system and the pmap unlocked,
 *	since these must be unlocked to use vm_page_grab.
 */
vm_offset_t
pmap_page_table_page_alloc(void)
{
	vm_page_t	m;
	phys_addr_t	pa;

	check_simple_locks();

	/*
	 *	We cannot allocate the pmap_object in pmap_init,
	 *	because it is called before the cache package is up.
	 *	Allocate it now if it is missing.
	 */
	if (pmap_object == VM_OBJECT_NULL)
	    pmap_object = vm_object_allocate(vm_page_table_size() * PAGE_SIZE);

	/*
	 *	Allocate a VM page for the level 2 page table entries.
	 */
	while ((m = vm_page_grab(VM_PAGE_DIRECTMAP)) == VM_PAGE_NULL)
		VM_PAGE_WAIT((void (*)()) 0);

	/*
	 *	Map the page to its physical address so that it
	 *	can be found later.
	 */
	pa = m->phys_addr;
	assert(pa == (vm_offset_t) pa);
	vm_object_lock(pmap_object);
	vm_page_insert(m, pmap_object, pa);
	vm_page_lock_queues();
	vm_page_wire(m);
	inuse_ptepages_count++;
	vm_page_unlock_queues();
	vm_object_unlock(pmap_object);

	/*
	 *	Zero the page.
	 */
	memset((void *)phystokv(pa), 0, PAGE_SIZE);

	return pa;
}

#ifdef	MACH_XEN
void pmap_map_mfn(void *_addr, unsigned long mfn) {
	vm_offset_t addr = (vm_offset_t) _addr;
	pt_entry_t	*pte, *pdp;
	vm_offset_t	ptp;
	pt_entry_t ma = ((pt_entry_t) mfn) << PAGE_SHIFT;

	/* Add a ptp if none exist yet for this pte */
	if ((pte = pmap_pte(kernel_pmap, addr)) == PT_ENTRY_NULL) {
		ptp = phystokv(pmap_page_table_page_alloc());
#ifdef	MACH_PV_PAGETABLES
		pmap_set_page_readonly((void*) ptp);
		if (!hyp_mmuext_op_mfn (MMUEXT_PIN_L1_TABLE, pa_to_mfn(ptp)))
			panic("couldn't pin page %lx(%lx)\n",ptp,(vm_offset_t) kv_to_ma(ptp));
#endif	/* MACH_PV_PAGETABLES */
		pdp = pmap_pde(kernel_pmap, addr);

#ifdef	MACH_PV_PAGETABLES
		if (!hyp_mmu_update_pte(kv_to_ma(pdp),
			pa_to_pte(kv_to_ma(ptp)) | INTEL_PTE_VALID
					      | INTEL_PTE_USER
					      | INTEL_PTE_WRITE))
			panic("%s:%d could not set pde %llx(%lx) to %lx(%lx)\n",__FILE__,__LINE__,kvtophys((vm_offset_t)pdp),(vm_offset_t) kv_to_ma(pdp), ptp, (vm_offset_t) pa_to_ma(ptp));
#else	/* MACH_PV_PAGETABLES */
		*pdp = pa_to_pte(kvtophys(ptp)) | INTEL_PTE_VALID
						| INTEL_PTE_USER
						| INTEL_PTE_WRITE;
#endif	/* MACH_PV_PAGETABLES */
		pte = pmap_pte(kernel_pmap, addr);
	}

#ifdef	MACH_PV_PAGETABLES
	if (!hyp_mmu_update_pte(kv_to_ma(pte), ma | INTEL_PTE_VALID | INTEL_PTE_WRITE))
		panic("%s:%d could not set pte %p(%lx) to %llx(%llx)\n",__FILE__,__LINE__,pte,(vm_offset_t) kv_to_ma(pte), ma, ma_to_pa(ma));
#else	/* MACH_PV_PAGETABLES */
	/* Note: in this case, mfn is actually a pfn.  */
	WRITE_PTE(pte, ma | INTEL_PTE_VALID | INTEL_PTE_WRITE);
#endif	/* MACH_PV_PAGETABLES */
}
#endif	/* MACH_XEN */

/*
 *	Deallocate a page-table page.
 *	The page-table page must have all mappings removed,
 *	and be removed from its page directory.
 */
void
pmap_page_table_page_dealloc(vm_offset_t pa)
{
	vm_page_t	m;

	vm_object_lock(pmap_object);
	m = vm_page_lookup(pmap_object, pa);
	vm_page_lock_queues();
	vm_page_free(m);
	inuse_ptepages_count--;
	vm_page_unlock_queues();
	vm_object_unlock(pmap_object);
}

/*
 *	Create and return a physical map.
 *
 *	If the size specified for the map
 *	is zero, the map is an actual physical
 *	map, and may be referenced by the
 *	hardware.
 *
 *	If the size specified is non-zero,
 *	the map will be used in software only, and
 *	is bounded by that size.
 */
pmap_t pmap_create(vm_size_t size)
{
	pt_entry_t		*page_dir[PDPNUM];
	int			i;
	pmap_t			p;
	pmap_statistics_t	stats;

	/*
	 *	A software use-only map doesn't even need a map.
	 */

	if (size != 0) {
		return(PMAP_NULL);
	}

/*
 *	Allocate a pmap struct from the pmap_cache.  Then allocate
 *	the page descriptor table.
 */

	p = (pmap_t) kmem_cache_alloc(&pmap_cache);
	if (p == PMAP_NULL)
		return PMAP_NULL;

	for (i = 0; i < PDPNUM; i++) {
		page_dir[i] = (pt_entry_t *) kmem_cache_alloc(&pd_cache);
		if (page_dir[i] == NULL) {
			i -= 1;
			while (i >= 0) {
				kmem_cache_free(&pd_cache,
						(vm_address_t) page_dir[i]);
				i -= 1;
			}
			kmem_cache_free(&pmap_cache, (vm_address_t) p);
			return PMAP_NULL;
		}
		memcpy(page_dir[i],
		       (void *) kernel_page_dir + i * INTEL_PGBYTES,
		       INTEL_PGBYTES);
	}

#if VM_MIN_KERNEL_ADDRESS != 0
	/* Do not map BIOS in user tasks */
	page_dir
#if PAE
		[lin2pdpnum(LINEAR_MIN_KERNEL_ADDRESS - VM_MIN_KERNEL_ADDRESS)]
#else
		[0]
#endif
		[lin2pdenum(LINEAR_MIN_KERNEL_ADDRESS - VM_MIN_KERNEL_ADDRESS)]
		= 0;
#endif
#ifdef	MACH_PV_PAGETABLES
	{
		for (i = 0; i < PDPNUM; i++)
			pmap_set_page_readonly((void *) page_dir[i]);
	}
#endif	/* MACH_PV_PAGETABLES */

#if PAE
	p->pdpbase = (pt_entry_t *) kmem_cache_alloc(&pdpt_cache);
	if (p->pdpbase == NULL) {
		for (i = 0; i < PDPNUM; i++)
			kmem_cache_free(&pd_cache, (vm_address_t) page_dir[i]);
		kmem_cache_free(&pmap_cache, (vm_address_t) p);
		return PMAP_NULL;
	}

	memset(p->pdpbase, 0, INTEL_PGBYTES);
	{
		for (i = 0; i < PDPNUM; i++)
			WRITE_PTE(&p->pdpbase[i],
				  pa_to_pte(kvtophys((vm_offset_t) page_dir[i]))
				  | INTEL_PTE_VALID
#ifdef	MACH_PV_PAGETABLES
				  | INTEL_PTE_WRITE
#endif
				  );
	}
#ifdef __x86_64__
	// FIXME: use kmem_cache_alloc instead
	if (kmem_alloc_wired(kernel_map,
			     (vm_offset_t *)&p->l4base, INTEL_PGBYTES)
							!= KERN_SUCCESS)
		panic("pmap_create");
	memset(p->l4base, 0, INTEL_PGBYTES);
	WRITE_PTE(&p->l4base[0], pa_to_pte(kvtophys((vm_offset_t) p->pdpbase)) | INTEL_PTE_VALID | INTEL_PTE_WRITE);
#ifdef	MACH_PV_PAGETABLES
	// FIXME: use kmem_cache_alloc instead
	if (kmem_alloc_wired(kernel_map,
			     (vm_offset_t *)&p->user_pdpbase, INTEL_PGBYTES)
							!= KERN_SUCCESS)
		panic("pmap_create");
	memset(p->user_pdpbase, 0, INTEL_PGBYTES);
	{
		int i;
		for (i = 0; i < lin2pdpnum(VM_MAX_ADDRESS); i++)
			WRITE_PTE(&p->user_pdpbase[i], pa_to_pte(kvtophys((vm_offset_t) page_dir[i])) | INTEL_PTE_VALID | INTEL_PTE_WRITE);
	}
	// FIXME: use kmem_cache_alloc instead
	if (kmem_alloc_wired(kernel_map,
			     (vm_offset_t *)&p->user_l4base, INTEL_PGBYTES)
							!= KERN_SUCCESS)
		panic("pmap_create");
	memset(p->user_l4base, 0, INTEL_PGBYTES);
	WRITE_PTE(&p->user_l4base[0], pa_to_pte(kvtophys((vm_offset_t) p->user_pdpbase)) | INTEL_PTE_VALID | INTEL_PTE_WRITE);
#endif	/* MACH_PV_PAGETABLES */
#endif	/* _x86_64 */
#ifdef	MACH_PV_PAGETABLES 
#ifdef __x86_64__
	pmap_set_page_readonly(p->l4base);
	pmap_set_page_readonly(p->user_l4base);
	pmap_set_page_readonly(p->user_pdpbase);
#endif
	pmap_set_page_readonly(p->pdpbase);
#endif	/* MACH_PV_PAGETABLES */
#else	/* PAE */
	p->dirbase = page_dir[0];
#endif	/* PAE */

	p->ref_count = 1;

	simple_lock_init(&p->lock);
	p->cpus_using = 0;

	/*
	 *	Initialize statistics.
	 */

	stats = &p->stats;
	stats->resident_count = 0;
	stats->wired_count = 0;

	return(p);
}

/*
 *	Retire the given physical map from service.
 *	Should only be called if the map contains
 *	no valid mappings.
 */

void pmap_destroy(pmap_t p)
{
#if PAE
	int		i;
#endif
	boolean_t	free_all;
	pt_entry_t     	*page_dir;
	pt_entry_t	*pdep;
	phys_addr_t	pa;
	int		c, s;
	vm_page_t	m;

	if (p == PMAP_NULL)
		return;

	SPLVM(s);
	simple_lock(&p->lock);
	c = --p->ref_count;
	simple_unlock(&p->lock);
	SPLX(s);

	if (c != 0) {
	    return;	/* still in use */
	}

#if PAE
	for (i = 0; i <= lin2pdpnum(LINEAR_MIN_KERNEL_ADDRESS); i++) {
	    free_all = i < lin2pdpnum(LINEAR_MIN_KERNEL_ADDRESS);
	    page_dir = (pt_entry_t *) ptetokv(p->pdpbase[i]);
#else
	    free_all = FALSE;
	    page_dir = p->dirbase;
#endif

#ifdef __x86_64__
#warning FIXME 64bit need to free l3
#endif
	    /*
	     *	Free the memory maps, then the
	     *	pmap structure.
	     */
	    for (pdep = page_dir;
		 (free_all
		  || pdep < &page_dir[lin2pdenum(LINEAR_MIN_KERNEL_ADDRESS)])
		     && pdep < &page_dir[NPTES];
		 pdep += ptes_per_vm_page) {
		if (*pdep & INTEL_PTE_VALID) {
		    pa = pte_to_pa(*pdep);
		    assert(pa == (vm_offset_t) pa);
		    vm_object_lock(pmap_object);
		    m = vm_page_lookup(pmap_object, pa);
		    if (m == VM_PAGE_NULL)
			panic("pmap_destroy: pte page not in object");
		    vm_page_lock_queues();
#ifdef	MACH_PV_PAGETABLES
		    if (!hyp_mmuext_op_mfn (MMUEXT_UNPIN_TABLE, pa_to_mfn(pa)))
			panic("pmap_destroy: couldn't unpin page %llx(%lx)\n", pa, (vm_offset_t) kv_to_ma(pa));
		    pmap_set_page_readwrite((void*) phystokv(pa));
#endif	/* MACH_PV_PAGETABLES */
		    vm_page_free(m);
		    inuse_ptepages_count--;
		    vm_page_unlock_queues();
		    vm_object_unlock(pmap_object);
		}
	    }
#ifdef	MACH_PV_PAGETABLES
	    pmap_set_page_readwrite((void*) page_dir);
#endif	/* MACH_PV_PAGETABLES */
	    kmem_cache_free(&pd_cache, (vm_offset_t) page_dir);
#if PAE
	}

#ifdef	MACH_PV_PAGETABLES
#ifdef __x86_64__
	pmap_set_page_readwrite(p->l4base);
	pmap_set_page_readwrite(p->user_l4base);
	pmap_set_page_readwrite(p->user_pdpbase);
#endif
	pmap_set_page_readwrite(p->pdpbase);
#endif	/* MACH_PV_PAGETABLES */
#ifdef __x86_64__
	kmem_free(kernel_map, (vm_offset_t)p->l4base, INTEL_PGBYTES);
#ifdef MACH_PV_PAGETABLES
	kmem_free(kernel_map, (vm_offset_t)p->user_l4base, INTEL_PGBYTES);
	kmem_free(kernel_map, (vm_offset_t)p->user_pdpbase, INTEL_PGBYTES);
#endif
#endif
	kmem_cache_free(&pdpt_cache, (vm_offset_t) p->pdpbase);
#endif	/* PAE */
	kmem_cache_free(&pmap_cache, (vm_offset_t) p);
}

/*
 *	Add a reference to the specified pmap.
 */

void pmap_reference(pmap_t p)
{
	int	s;
	if (p != PMAP_NULL) {
		SPLVM(s);
		simple_lock(&p->lock);
		p->ref_count++;
		simple_unlock(&p->lock);
		SPLX(s);
	}
}

/*
 *	Remove a range of hardware page-table entries.
 *	The entries given are the first (inclusive)
 *	and last (exclusive) entries for the VM pages.
 *	The virtual address is the va for the first pte.
 *
 *	The pmap must be locked.
 *	If the pmap is not the kernel pmap, the range must lie
 *	entirely within one pte-page.  This is NOT checked.
 *	Assumes that the pte-page exists.
 */

static
void pmap_remove_range(
	pmap_t			pmap,
	vm_offset_t		va,
	pt_entry_t		*spte,
	pt_entry_t		*epte)
{
	pt_entry_t		*cpte;
	unsigned long		num_removed, num_unwired;
	unsigned long		pai;
	phys_addr_t		pa;
#ifdef	MACH_PV_PAGETABLES
	int n, ii = 0;
	struct mmu_update update[HYP_BATCH_MMU_UPDATES];
#endif	/* MACH_PV_PAGETABLES */

#if	DEBUG_PTE_PAGE
	if (pmap != kernel_pmap)
		ptep_check(get_pte_page(spte));
#endif	/* DEBUG_PTE_PAGE */
	num_removed = 0;
	num_unwired = 0;

	for (cpte = spte; cpte < epte;
	     cpte += ptes_per_vm_page, va += PAGE_SIZE) {

	    if (*cpte == 0)
		continue;
	    pa = pte_to_pa(*cpte);

	    num_removed++;
	    if (*cpte & INTEL_PTE_WIRED)
		num_unwired++;

	    if (!valid_page(pa)) {

		/*
		 *	Outside range of managed physical memory.
		 *	Just remove the mappings.
		 */
		int		i = ptes_per_vm_page;
		pt_entry_t	*lpte = cpte;
		do {
#ifdef	MACH_PV_PAGETABLES
		    update[ii].ptr = kv_to_ma(lpte);
		    update[ii].val = 0;
		    ii++;
		    if (ii == HYP_BATCH_MMU_UPDATES) {
			hyp_mmu_update(kvtolin(&update), ii, kvtolin(&n), DOMID_SELF);
			if (n != ii)
				panic("couldn't pmap_remove_range\n");
			ii = 0;
		    }
#else	/* MACH_PV_PAGETABLES */
		    *lpte = 0;
#endif	/* MACH_PV_PAGETABLES */
		    lpte++;
		} while (--i > 0);
		continue;
	    }

	    pai = pa_index(pa);
	    LOCK_PVH(pai);

	    /*
	     *	Get the modify and reference bits.
	     */
	    {
		int		i;
		pt_entry_t	*lpte;

		i = ptes_per_vm_page;
		lpte = cpte;
		do {
		    pmap_phys_attributes[pai] |=
			*lpte & (PHYS_MODIFIED|PHYS_REFERENCED);
#ifdef	MACH_PV_PAGETABLES
		    update[ii].ptr = kv_to_ma(lpte);
		    update[ii].val = 0;
		    ii++;
		    if (ii == HYP_BATCH_MMU_UPDATES) {
			hyp_mmu_update(kvtolin(&update), ii, kvtolin(&n), DOMID_SELF);
			if (n != ii)
				panic("couldn't pmap_remove_range\n");
			ii = 0;
		    }
#else	/* MACH_PV_PAGETABLES */
		    *lpte = 0;
#endif	/* MACH_PV_PAGETABLES */
		    lpte++;
		} while (--i > 0);
	    }

	    /*
	     *	Remove the mapping from the pvlist for
	     *	this physical page.
	     */
	    {
		pv_entry_t	pv_h, prev, cur;

		pv_h = pai_to_pvh(pai);
		if (pv_h->pmap == PMAP_NULL) {
		    panic("pmap_remove: null pv_list!");
		}
		if (pv_h->va == va && pv_h->pmap == pmap) {
		    /*
		     * Header is the pv_entry.  Copy the next one
		     * to header and free the next one (we cannot
		     * free the header)
		     */
		    cur = pv_h->next;
		    if (cur != PV_ENTRY_NULL) {
			*pv_h = *cur;
			PV_FREE(cur);
		    }
		    else {
			pv_h->pmap = PMAP_NULL;
		    }
		}
		else {
		    cur = pv_h;
		    do {
			prev = cur;
			if ((cur = prev->next) == PV_ENTRY_NULL) {
			    panic("pmap-remove: mapping not in pv_list!");
			}
		    } while (cur->va != va || cur->pmap != pmap);
		    prev->next = cur->next;
		    PV_FREE(cur);
		}
		UNLOCK_PVH(pai);
	    }
	}

#ifdef	MACH_PV_PAGETABLES
	if (ii > HYP_BATCH_MMU_UPDATES)
		panic("overflowed array in pmap_remove_range");
	hyp_mmu_update(kvtolin(&update), ii, kvtolin(&n), DOMID_SELF);
	if (n != ii)
		panic("couldn't pmap_remove_range\n");
#endif	/* MACH_PV_PAGETABLES */

	/*
	 *	Update the counts
	 */
	pmap->stats.resident_count -= num_removed;
	pmap->stats.wired_count -= num_unwired;
}

/*
 *	Remove the given range of addresses
 *	from the specified map.
 *
 *	It is assumed that the start and end are properly
 *	rounded to the hardware page size.
 */

void pmap_remove(
	pmap_t		map,
	vm_offset_t	s, 
	vm_offset_t	e)
{
	int			spl;
	pt_entry_t		*spte, *epte;
	vm_offset_t		l;
	vm_offset_t		_s = s;

	if (map == PMAP_NULL)
		return;

	PMAP_READ_LOCK(map, spl);

	while (s < e) {
	    pt_entry_t *pde = pmap_pde(map, s);

	    l = (s + PDE_MAPPED_SIZE) & ~(PDE_MAPPED_SIZE-1);
	    if (l > e)
		l = e;
	    if (*pde & INTEL_PTE_VALID) {
		spte = (pt_entry_t *)ptetokv(*pde);
		spte = &spte[ptenum(s)];
		epte = &spte[intel_btop(l-s)];
		pmap_remove_range(map, s, spte, epte);
	    }
	    s = l;
	}
	PMAP_UPDATE_TLBS(map, _s, e);

	PMAP_READ_UNLOCK(map, spl);
}

/*
 *	Routine:	pmap_page_protect
 *
 *	Function:
 *		Lower the permission for all mappings to a given
 *		page.
 */
void pmap_page_protect(
	phys_addr_t	phys,
	vm_prot_t	prot)
{
	pv_entry_t		pv_h, prev;
	pv_entry_t		pv_e;
	pt_entry_t		*pte;
	unsigned long		pai;
	pmap_t			pmap;
	int			spl;
	boolean_t		remove;

	assert(phys != vm_page_fictitious_addr);
	if (!valid_page(phys)) {
	    /*
	     *	Not a managed page.
	     */
	    return;
	}

	/*
	 * Determine the new protection.
	 */
	switch (prot) {
	    case VM_PROT_READ:
	    case VM_PROT_READ|VM_PROT_EXECUTE:
		remove = FALSE;
		break;
	    case VM_PROT_ALL:
		return;	/* nothing to do */
	    default:
		remove = TRUE;
		break;
	}

	/*
	 *	Lock the pmap system first, since we will be changing
	 *	several pmaps.
	 */

	PMAP_WRITE_LOCK(spl);

	pai = pa_index(phys);
	pv_h = pai_to_pvh(pai);

	/*
	 * Walk down PV list, changing or removing all mappings.
	 * We do not have to lock the pv_list because we have
	 * the entire pmap system locked.
	 */
	if (pv_h->pmap != PMAP_NULL) {

	    prev = pv_e = pv_h;
	    do {
		vm_offset_t va;

		pmap = pv_e->pmap;
		/*
		 * Lock the pmap to block pmap_extract and similar routines.
		 */
		simple_lock(&pmap->lock);

		va = pv_e->va;
		pte = pmap_pte(pmap, va);

		/*
		 * Consistency checks.
		 */
		assert(*pte & INTEL_PTE_VALID);
		assert(pte_to_pa(*pte) == phys);

		/*
		 * Remove the mapping if new protection is NONE
		 * or if write-protecting a kernel mapping.
		 */
		if (remove || pmap == kernel_pmap) {
		    /*
		     * Remove the mapping, collecting any modify bits.
		     */

		    if (*pte & INTEL_PTE_WIRED) {
			pmap->stats.wired_count--;
		    }

		    {
			int	i = ptes_per_vm_page;

			do {
			    pmap_phys_attributes[pai] |=
				*pte & (PHYS_MODIFIED|PHYS_REFERENCED);
#ifdef	MACH_PV_PAGETABLES
			    if (!hyp_mmu_update_pte(kv_to_ma(pte++), 0))
			    	panic("%s:%d could not clear pte %p\n",__FILE__,__LINE__,pte-1);
#else	/* MACH_PV_PAGETABLES */
			    *pte++ = 0;
#endif	/* MACH_PV_PAGETABLES */
			} while (--i > 0);
		    }

		    pmap->stats.resident_count--;

		    /*
		     * Remove the pv_entry.
		     */
		    if (pv_e == pv_h) {
			/*
			 * Fix up head later.
			 */
			pv_h->pmap = PMAP_NULL;
		    }
		    else {
			/*
			 * Delete this entry.
			 */
			prev->next = pv_e->next;
			PV_FREE(pv_e);
		    }
		}
		else {
		    /*
		     * Write-protect.
		     */
		    int i = ptes_per_vm_page;

		    do {
#ifdef	MACH_PV_PAGETABLES
			if (!hyp_mmu_update_pte(kv_to_ma(pte), *pte & ~INTEL_PTE_WRITE))
			    	panic("%s:%d could not disable write on pte %p\n",__FILE__,__LINE__,pte);
#else	/* MACH_PV_PAGETABLES */
			*pte &= ~INTEL_PTE_WRITE;
#endif	/* MACH_PV_PAGETABLES */
			pte++;
		    } while (--i > 0);

		    /*
		     * Advance prev.
		     */
		    prev = pv_e;
		}
		PMAP_UPDATE_TLBS(pmap, va, va + PAGE_SIZE);

		simple_unlock(&pmap->lock);

	    } while ((pv_e = prev->next) != PV_ENTRY_NULL);

	    /*
	     * If pv_head mapping was removed, fix it up.
	     */
	    if (pv_h->pmap == PMAP_NULL) {
		pv_e = pv_h->next;
		if (pv_e != PV_ENTRY_NULL) {
		    *pv_h = *pv_e;
		    PV_FREE(pv_e);
		}
	    }
	}

	PMAP_WRITE_UNLOCK(spl);
}

/*
 *	Set the physical protection on the
 *	specified range of this map as requested.
 *	Will not increase permissions.
 */
void pmap_protect(
	pmap_t		map,
	vm_offset_t	s, 
	vm_offset_t	e,
	vm_prot_t	prot)
{
	pt_entry_t	*pde;
	pt_entry_t	*spte, *epte;
	vm_offset_t	l;
	int		spl;
	vm_offset_t	_s = s;

	if (map == PMAP_NULL)
		return;

	/*
	 * Determine the new protection.
	 */
	switch (prot) {
	    case VM_PROT_READ:
	    case VM_PROT_READ|VM_PROT_EXECUTE:
		break;
	    case VM_PROT_READ|VM_PROT_WRITE:
	    case VM_PROT_ALL:
		return;	/* nothing to do */
	    default:
		pmap_remove(map, s, e);
		return;
	}

#if !(__i486__ || __i586__ || __i686__)
	/*
	 * If write-protecting in the kernel pmap,
	 * remove the mappings; the i386 ignores
	 * the write-permission bit in kernel mode.
	 */
	if (map == kernel_pmap) {
	    pmap_remove(map, s, e);
	    return;
	}
#endif

	SPLVM(spl);
	simple_lock(&map->lock);

	pde = pmap_pde(map, s);
	while (s < e) {
	    l = (s + PDE_MAPPED_SIZE) & ~(PDE_MAPPED_SIZE-1);
	    if (l > e)
		l = e;
	    if (*pde & INTEL_PTE_VALID) {
		spte = (pt_entry_t *)ptetokv(*pde);
		spte = &spte[ptenum(s)];
		epte = &spte[intel_btop(l-s)];

#ifdef	MACH_PV_PAGETABLES
		int n, i = 0;
		struct mmu_update update[HYP_BATCH_MMU_UPDATES];
#endif	/* MACH_PV_PAGETABLES */

		while (spte < epte) {
		    if (*spte & INTEL_PTE_VALID) {
#ifdef	MACH_PV_PAGETABLES
			update[i].ptr = kv_to_ma(spte);
			update[i].val = *spte & ~INTEL_PTE_WRITE;
			i++;
			if (i == HYP_BATCH_MMU_UPDATES) {
			    hyp_mmu_update(kvtolin(&update), i, kvtolin(&n), DOMID_SELF);
			    if (n != i)
				    panic("couldn't pmap_protect\n");
			    i = 0;
			}
#else	/* MACH_PV_PAGETABLES */
			*spte &= ~INTEL_PTE_WRITE;
#endif	/* MACH_PV_PAGETABLES */
		    }
		    spte++;
		}
#ifdef	MACH_PV_PAGETABLES
		if (i > HYP_BATCH_MMU_UPDATES)
			panic("overflowed array in pmap_protect");
		hyp_mmu_update(kvtolin(&update), i, kvtolin(&n), DOMID_SELF);
		if (n != i)
			panic("couldn't pmap_protect\n");
#endif	/* MACH_PV_PAGETABLES */
	    }
	    s = l;
	    pde++;
	}
	PMAP_UPDATE_TLBS(map, _s, e);

	simple_unlock(&map->lock);
	SPLX(spl);
}

/*
 *	Insert the given physical page (p) at
 *	the specified virtual address (v) in the
 *	target physical map with the protection requested.
 *
 *	If specified, the page will be wired down, meaning
 *	that the related pte can not be reclaimed.
 *
 *	NB:  This is the only routine which MAY NOT lazy-evaluate
 *	or lose information.  That is, this routine must actually
 *	insert this page into the given map NOW.
 */
void pmap_enter(
	pmap_t			pmap,
	vm_offset_t		v,
	phys_addr_t		pa,
	vm_prot_t		prot,
	boolean_t		wired)
{
	boolean_t		is_physmem;
	pt_entry_t		*pte;
	pv_entry_t		pv_h;
	unsigned long		i, pai;
	pv_entry_t		pv_e;
	pt_entry_t		template;
	int			spl;
	phys_addr_t		old_pa;

	assert(pa != vm_page_fictitious_addr);
	if (pmap_debug) printf("pmap(%lx, %llx)\n", v, (unsigned long long) pa);
	if (pmap == PMAP_NULL)
		return;

#if !MACH_KDB
	if (pmap == kernel_pmap && (v < kernel_virtual_start || v >= kernel_virtual_end))
		panic("pmap_enter(%lx, %llx) falls in physical memory area!\n", v, (unsigned long long) pa);
#endif
#if !(__i486__ || __i586__ || __i686__)
	if (pmap == kernel_pmap && (prot & VM_PROT_WRITE) == 0
	    && !wired /* hack for io_wire */ ) {
	    /*
	     *	Because the 386 ignores write protection in kernel mode,
	     *	we cannot enter a read-only kernel mapping, and must
	     *	remove an existing mapping if changing it.
	     */
	    PMAP_READ_LOCK(pmap, spl);

	    pte = pmap_pte(pmap, v);
	    if (pte != PT_ENTRY_NULL && *pte != 0) {
		/*
		 *	Invalidate the translation buffer,
		 *	then remove the mapping.
		 */
		pmap_remove_range(pmap, v, pte,
				  pte + ptes_per_vm_page);
		PMAP_UPDATE_TLBS(pmap, v, v + PAGE_SIZE);
	    }
	    PMAP_READ_UNLOCK(pmap, spl);
	    return;
	}
#endif

	/*
	 *	Must allocate a new pvlist entry while we're unlocked;
	 *	Allocating may cause pageout (which will lock the pmap system).
	 *	If we determine we need a pvlist entry, we will unlock
	 *	and allocate one.  Then we will retry, throughing away
	 *	the allocated entry later (if we no longer need it).
	 */
	pv_e = PV_ENTRY_NULL;
Retry:
	PMAP_READ_LOCK(pmap, spl);

	/*
	 *	Expand pmap to include this pte.  Assume that
	 *	pmap is always expanded to include enough hardware
	 *	pages to map one VM page.
	 */

	while ((pte = pmap_pte(pmap, v)) == PT_ENTRY_NULL) {
	    /*
	     * Need to allocate a new page-table page.
	     */
	    vm_offset_t	ptp;
	    pt_entry_t	*pdp;
	    int		i;

	    if (pmap == kernel_pmap) {
		/*
		 * Would have to enter the new page-table page in
		 * EVERY pmap.
		 */
		panic("pmap_expand kernel pmap to %#lx", v);
	    }

	    /*
	     * Unlock the pmap and allocate a new page-table page.
	     */
	    PMAP_READ_UNLOCK(pmap, spl);

	    ptp = phystokv(pmap_page_table_page_alloc());

	    /*
	     * Re-lock the pmap and check that another thread has
	     * not already allocated the page-table page.  If it
	     * has, discard the new page-table page (and try
	     * again to make sure).
	     */
	    PMAP_READ_LOCK(pmap, spl);

	    if (pmap_pte(pmap, v) != PT_ENTRY_NULL) {
		/*
		 * Oops...
		 */
		PMAP_READ_UNLOCK(pmap, spl);
		pmap_page_table_page_dealloc(kvtophys(ptp));
		PMAP_READ_LOCK(pmap, spl);
		continue;
	    }

	    /*
	     * Enter the new page table page in the page directory.
	     */
	    i = ptes_per_vm_page;
	    pdp = pmap_pde(pmap, v);
	    do {
#ifdef	MACH_PV_PAGETABLES
		pmap_set_page_readonly((void *) ptp);
		if (!hyp_mmuext_op_mfn (MMUEXT_PIN_L1_TABLE, kv_to_mfn(ptp)))
			panic("couldn't pin page %lx(%lx)\n",ptp,(vm_offset_t) kv_to_ma(ptp));
		if (!hyp_mmu_update_pte(pa_to_ma(kvtophys((vm_offset_t)pdp)),
			pa_to_pte(pa_to_ma(kvtophys(ptp))) | INTEL_PTE_VALID
					      | INTEL_PTE_USER
					      | INTEL_PTE_WRITE))
			panic("%s:%d could not set pde %p(%llx,%lx) to %lx(%llx,%lx) %lx\n",__FILE__,__LINE__, pdp, kvtophys((vm_offset_t)pdp), (vm_offset_t) pa_to_ma(kvtophys((vm_offset_t)pdp)), ptp, kvtophys(ptp), (vm_offset_t) pa_to_ma(kvtophys(ptp)), (vm_offset_t) pa_to_pte(kv_to_ma(ptp)));
#else	/* MACH_PV_PAGETABLES */
		*pdp = pa_to_pte(kvtophys(ptp)) | INTEL_PTE_VALID
					        | INTEL_PTE_USER
					        | INTEL_PTE_WRITE;
#endif	/* MACH_PV_PAGETABLES */
		pdp++;	/* Note: This is safe b/c we stay in one page.  */
		ptp += INTEL_PGBYTES;
	    } while (--i > 0);

	    /*
	     * Now, get the address of the page-table entry.
	     */
	    continue;
	}

	if (vm_page_ready())
		is_physmem = (vm_page_lookup_pa(pa) != NULL);
	else
		is_physmem = (pa < biosmem_directmap_end());

	/*
	 *	Special case if the physical page is already mapped
	 *	at this address.
	 */
	old_pa = pte_to_pa(*pte);
	if (*pte && old_pa == pa) {
	    /*
	     *	May be changing its wired attribute or protection
	     */

	    if (wired && !(*pte & INTEL_PTE_WIRED))
		pmap->stats.wired_count++;
	    else if (!wired && (*pte & INTEL_PTE_WIRED))
		pmap->stats.wired_count--;

	    template = pa_to_pte(pa) | INTEL_PTE_VALID;
	    if (pmap != kernel_pmap)
		template |= INTEL_PTE_USER;
	    if (prot & VM_PROT_WRITE)
		template |= INTEL_PTE_WRITE;
	    if (machine_slot[cpu_number()].cpu_type >= CPU_TYPE_I486
		&& !is_physmem)
		template |= INTEL_PTE_NCACHE|INTEL_PTE_WTHRU;
	    if (wired)
		template |= INTEL_PTE_WIRED;
	    i = ptes_per_vm_page;
	    do {
		if (*pte & INTEL_PTE_MOD)
		    template |= INTEL_PTE_MOD;
#ifdef	MACH_PV_PAGETABLES
		if (!hyp_mmu_update_pte(kv_to_ma(pte), pa_to_ma(template)))
			panic("%s:%d could not set pte %p to %llx\n",__FILE__,__LINE__,pte,template);
#else	/* MACH_PV_PAGETABLES */
		WRITE_PTE(pte, template)
#endif	/* MACH_PV_PAGETABLES */
		pte++;
		pte_increment_pa(template);
	    } while (--i > 0);
	    PMAP_UPDATE_TLBS(pmap, v, v + PAGE_SIZE);
	}
	else {

	    /*
	     *	Remove old mapping from the PV list if necessary.
	     */
	    if (*pte) {
		/*
		 *	Don't free the pte page if removing last
		 *	mapping - we will immediately replace it.
		 */
		pmap_remove_range(pmap, v, pte,
				  pte + ptes_per_vm_page);
		PMAP_UPDATE_TLBS(pmap, v, v + PAGE_SIZE);
	    }

	    if (valid_page(pa)) {

		/*
		 *	Enter the mapping in the PV list for this
		 *	physical page.
		 */

		pai = pa_index(pa);
		LOCK_PVH(pai);
		pv_h = pai_to_pvh(pai);

		if (pv_h->pmap == PMAP_NULL) {
		    /*
		     *	No mappings yet
		     */
		    pv_h->va = v;
		    pv_h->pmap = pmap;
		    pv_h->next = PV_ENTRY_NULL;
		}
		else {
#if	DEBUG
		    {
			/* check that this mapping is not already there */
			pv_entry_t	e = pv_h;
			while (e != PV_ENTRY_NULL) {
			    if (e->pmap == pmap && e->va == v)
				panic("pmap_enter: already in pv_list");
			    e = e->next;
			}
		    }
#endif	/* DEBUG */

		    /*
		     *	Add new pv_entry after header.
		     */
		    if (pv_e == PV_ENTRY_NULL) {
			PV_ALLOC(pv_e);
			if (pv_e == PV_ENTRY_NULL) {
			    UNLOCK_PVH(pai);
			    PMAP_READ_UNLOCK(pmap, spl);

			    /*
			     * Refill from cache.
			     */
			    pv_e = (pv_entry_t) kmem_cache_alloc(&pv_list_cache);
			    goto Retry;
			}
		    }
		    pv_e->va = v;
		    pv_e->pmap = pmap;
		    pv_e->next = pv_h->next;
		    pv_h->next = pv_e;
		    /*
		     *	Remember that we used the pvlist entry.
		     */
		    pv_e = PV_ENTRY_NULL;
		}
		UNLOCK_PVH(pai);
	    }

	    /*
	     *	And count the mapping.
	     */

	    pmap->stats.resident_count++;
	    if (wired)
		pmap->stats.wired_count++;

	    /*
	     *	Build a template to speed up entering -
	     *	only the pfn changes.
	     */
	    template = pa_to_pte(pa) | INTEL_PTE_VALID;
	    if (pmap != kernel_pmap)
		template |= INTEL_PTE_USER;
	    if (prot & VM_PROT_WRITE)
		template |= INTEL_PTE_WRITE;
	    if (machine_slot[cpu_number()].cpu_type >= CPU_TYPE_I486
		&& !is_physmem)
		template |= INTEL_PTE_NCACHE|INTEL_PTE_WTHRU;
	    if (wired)
		template |= INTEL_PTE_WIRED;
	    i = ptes_per_vm_page;
	    do {
#ifdef	MACH_PV_PAGETABLES
		if (!(hyp_mmu_update_pte(kv_to_ma(pte), pa_to_ma(template))))
			panic("%s:%d could not set pte %p to %llx\n",__FILE__,__LINE__,pte,template);
#else	/* MACH_PV_PAGETABLES */
		WRITE_PTE(pte, template)
#endif	/* MACH_PV_PAGETABLES */
		pte++;
		pte_increment_pa(template);
	    } while (--i > 0);
	}

	if (pv_e != PV_ENTRY_NULL) {
	    PV_FREE(pv_e);
	}

	PMAP_READ_UNLOCK(pmap, spl);
}

/*
 *	Routine:	pmap_change_wiring
 *	Function:	Change the wiring attribute for a map/virtual-address
 *			pair.
 *	In/out conditions:
 *			The mapping must already exist in the pmap.
 */
void pmap_change_wiring(
	pmap_t		map,
	vm_offset_t	v,
	boolean_t	wired)
{
	pt_entry_t	*pte;
	int		i;
	int		spl;

	/*
	 *	We must grab the pmap system lock because we may
	 *	change a pte_page queue.
	 */
	PMAP_READ_LOCK(map, spl);

	if ((pte = pmap_pte(map, v)) == PT_ENTRY_NULL)
		panic("pmap_change_wiring: pte missing");

	if (wired && !(*pte & INTEL_PTE_WIRED)) {
	    /*
	     *	wiring down mapping
	     */
	    map->stats.wired_count++;
	    i = ptes_per_vm_page;
	    do {
		*pte++ |= INTEL_PTE_WIRED;
	    } while (--i > 0);
	}
	else if (!wired && (*pte & INTEL_PTE_WIRED)) {
	    /*
	     *	unwiring mapping
	     */
	    map->stats.wired_count--;
	    i = ptes_per_vm_page;
	    do {
#ifdef	MACH_PV_PAGETABLES
		if (!(hyp_mmu_update_pte(kv_to_ma(pte), *pte & ~INTEL_PTE_WIRED)))
			panic("%s:%d could not wire down pte %p\n",__FILE__,__LINE__,pte);
#else	/* MACH_PV_PAGETABLES */
		*pte &= ~INTEL_PTE_WIRED;
#endif	/* MACH_PV_PAGETABLES */
		pte++;
	    } while (--i > 0);
	}

	PMAP_READ_UNLOCK(map, spl);
}

/*
 *	Routine:	pmap_extract
 *	Function:
 *		Extract the physical page address associated
 *		with the given map/virtual_address pair.
 */

phys_addr_t pmap_extract(
	pmap_t		pmap,
	vm_offset_t	va)
{
	pt_entry_t	*pte;
	phys_addr_t	pa;
	int		spl;

	SPLVM(spl);
	simple_lock(&pmap->lock);
	if ((pte = pmap_pte(pmap, va)) == PT_ENTRY_NULL)
	    pa = 0;
	else if (!(*pte & INTEL_PTE_VALID))
	    pa = 0;
	else
	    pa = pte_to_pa(*pte) + (va & INTEL_OFFMASK);
	simple_unlock(&pmap->lock);
	SPLX(spl);
	return(pa);
}

/*
 *	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.
 */
#if	0
void pmap_copy(
	pmap_t		dst_pmap,
	pmap_t		src_pmap,
	vm_offset_t	dst_addr,
	vm_size_t	len,
	vm_offset_t	src_addr)
{
}
#endif	/* 0 */

/*
 *	Routine:	pmap_collect
 *	Function:
 *		Garbage collects the physical map system for
 *		pages which are no longer used.
 *		Success need not be guaranteed -- that is, there
 *		may well be pages which are not referenced, but
 *		others may be collected.
 *	Usage:
 *		Called by the pageout daemon when pages are scarce.
 */
void pmap_collect(pmap_t p)
{
	int			i;
	boolean_t		free_all;
	pt_entry_t		*page_dir;
	pt_entry_t		*pdp, *ptp;
	pt_entry_t		*eptp;
	phys_addr_t		pa;
	int			spl, wired;

	if (p == PMAP_NULL)
		return;

	if (p == kernel_pmap)
		return;

#if PAE
	for (i = 0; i <= lin2pdpnum(LINEAR_MIN_KERNEL_ADDRESS); i++) {
	    free_all = i < lin2pdpnum(LINEAR_MIN_KERNEL_ADDRESS);
	    page_dir = (pt_entry_t *) ptetokv(p->pdpbase[i]);
#else
	    i = 0;
	    free_all = FALSE;
	    page_dir = p->dirbase;
#endif

	    /*
	     *	Garbage collect map.
	     */
	    PMAP_READ_LOCK(p, spl);
	    for (pdp = page_dir;
		 (free_all
		  || pdp < &page_dir[lin2pdenum(LINEAR_MIN_KERNEL_ADDRESS)])
		     && pdp < &page_dir[NPTES];
		 pdp += ptes_per_vm_page) {
		if (*pdp & INTEL_PTE_VALID) {

		    pa = pte_to_pa(*pdp);
		    ptp = (pt_entry_t *)phystokv(pa);
		    eptp = ptp + NPTES*ptes_per_vm_page;

		    /*
		     * If the pte page has any wired mappings, we cannot
		     * free it.
		     */
		    wired = 0;
		    {
			pt_entry_t *ptep;
			for (ptep = ptp; ptep < eptp; ptep++) {
			    if (*ptep & INTEL_PTE_WIRED) {
				wired = 1;
				break;
			    }
			}
		    }
		    if (!wired) {
			/*
			 * Remove the virtual addresses mapped by this pte page.
			 */
			{ /*XXX big hack*/
			    vm_offset_t va = pdenum2lin(pdp - page_dir
							+ i * NPTES);
			    if (p == kernel_pmap)
				va = lintokv(va);
			    pmap_remove_range(p,
					      va,
					      ptp,
					      eptp);
			}

			/*
			 * Invalidate the page directory pointer.
			 */
			{
			    int i = ptes_per_vm_page;
			    pt_entry_t *pdep = pdp;
			    do {
#ifdef	MACH_PV_PAGETABLES
				unsigned long pte = *pdep;
				void *ptable = (void*) ptetokv(pte);
				if (!(hyp_mmu_update_pte(pa_to_ma(kvtophys((vm_offset_t)pdep++)), 0)))
				    panic("%s:%d could not clear pde %p\n",__FILE__,__LINE__,pdep-1);
				if (!hyp_mmuext_op_mfn (MMUEXT_UNPIN_TABLE, kv_to_mfn(ptable)))
				    panic("couldn't unpin page %p(%lx)\n", ptable, (vm_offset_t) pa_to_ma(kvtophys((vm_offset_t)ptable)));
				pmap_set_page_readwrite(ptable);
#else	/* MACH_PV_PAGETABLES */
				*pdep++ = 0;
#endif	/* MACH_PV_PAGETABLES */
			    } while (--i > 0);
			}

			PMAP_READ_UNLOCK(p, spl);

			/*
			 * And free the pte page itself.
			 */
			{
			    vm_page_t m;

			    vm_object_lock(pmap_object);
			    assert(pa == (vm_offset_t) pa);
			    m = vm_page_lookup(pmap_object, pa);
			    if (m == VM_PAGE_NULL)
				panic("pmap_collect: pte page not in object");
			    vm_page_lock_queues();
			    vm_page_free(m);
			    inuse_ptepages_count--;
			    vm_page_unlock_queues();
			    vm_object_unlock(pmap_object);
			}

			PMAP_READ_LOCK(p, spl);
		    }
		}
	    }
#if PAE
	}
#endif
	PMAP_UPDATE_TLBS(p, VM_MIN_ADDRESS, VM_MAX_ADDRESS);

	PMAP_READ_UNLOCK(p, spl);
	return;

}

/*
 *	Routine:	pmap_activate
 *	Function:
 *		Binds the given physical map to the given
 *		processor, and returns a hardware map description.
 */
#if	0
void pmap_activate(pmap_t my_pmap, thread_t th, int my_cpu)
{
	PMAP_ACTIVATE(my_pmap, th, my_cpu);
}
#endif	/* 0 */

/*
 *	Routine:	pmap_deactivate
 *	Function:
 *		Indicates that the given physical map is no longer
 *		in use on the specified processor.  (This is a macro
 *		in pmap.h)
 */
#if	0
void pmap_deactivate(pmap_t pmap, thread_t th, int which_cpu)
{
	PMAP_DEACTIVATE(pmap, th, which_cpu);
}
#endif	/* 0 */

/*
 *	Routine:	pmap_kernel
 *	Function:
 *		Returns the physical map handle for the kernel.
 */
#if	0
pmap_t pmap_kernel()
{
    	return (kernel_pmap);
}
#endif	/* 0 */

/*
 *	pmap_zero_page zeros the specified (machine independent) page.
 *	See machine/phys.c or machine/phys.s for implementation.
 */
#if	0
pmap_zero_page(vm_offset_t phys)
{
	int	i;

	assert(phys != vm_page_fictitious_addr);
	i = PAGE_SIZE / INTEL_PGBYTES;
	phys = intel_pfn(phys);

	while (i--)
		zero_phys(phys++);
}
#endif	/* 0 */

/*
 *	pmap_copy_page copies the specified (machine independent) page.
 *	See machine/phys.c or machine/phys.s for implementation.
 */
#if	0
pmap_copy_page(vm_offset_t src, vm_offset_t dst)
{
	int	i;

	assert(src != vm_page_fictitious_addr);
	assert(dst != vm_page_fictitious_addr);
	i = PAGE_SIZE / INTEL_PGBYTES;

	while (i--) {
		copy_phys(intel_pfn(src), intel_pfn(dst));
		src += INTEL_PGBYTES;
		dst += INTEL_PGBYTES;
	}
}
#endif	/* 0 */

/*
 *	Routine:	pmap_pageable
 *	Function:
 *		Make the specified pages (by pmap, offset)
 *		pageable (or not) as requested.
 *
 *		A page which is not pageable may not take
 *		a fault; therefore, its page table entry
 *		must remain valid for the duration.
 *
 *		This routine is merely advisory; pmap_enter
 *		will specify that these pages are to be wired
 *		down (or not) as appropriate.
 */
void
pmap_pageable(
	pmap_t		pmap,
	vm_offset_t	start,
	vm_offset_t	end,
	boolean_t	pageable)
{
}

/*
 *	Clear specified attribute bits.
 */
void
phys_attribute_clear(
	phys_addr_t	phys,
	int		bits)
{
	pv_entry_t		pv_h;
	pv_entry_t		pv_e;
	pt_entry_t		*pte;
	unsigned long		pai;
	pmap_t			pmap;
	int			spl;

	assert(phys != vm_page_fictitious_addr);
	if (!valid_page(phys)) {
	    /*
	     *	Not a managed page.
	     */
	    return;
	}

	/*
	 *	Lock the pmap system first, since we will be changing
	 *	several pmaps.
	 */

	PMAP_WRITE_LOCK(spl);

	pai = pa_index(phys);
	pv_h = pai_to_pvh(pai);

	/*
	 * Walk down PV list, clearing all modify or reference bits.
	 * We do not have to lock the pv_list because we have
	 * the entire pmap system locked.
	 */
	if (pv_h->pmap != PMAP_NULL) {
	    /*
	     * There are some mappings.
	     */
	    for (pv_e = pv_h; pv_e != PV_ENTRY_NULL; pv_e = pv_e->next) {
		vm_offset_t va;

		pmap = pv_e->pmap;
		/*
		 * Lock the pmap to block pmap_extract and similar routines.
		 */
		simple_lock(&pmap->lock);

		va = pv_e->va;
		pte = pmap_pte(pmap, va);

		/*
		 * Consistency checks.
		 */
		assert(*pte & INTEL_PTE_VALID);
		assert(pte_to_pa(*pte) == phys);

		/*
		 * Clear modify or reference bits.
		 */
		{
		    int	i = ptes_per_vm_page;
		    do {
#ifdef	MACH_PV_PAGETABLES
			if (!(hyp_mmu_update_pte(kv_to_ma(pte), *pte & ~bits)))
			    panic("%s:%d could not clear bits %x from pte %p\n",__FILE__,__LINE__,bits,pte);
#else	/* MACH_PV_PAGETABLES */
			*pte &= ~bits;
#endif	/* MACH_PV_PAGETABLES */
		    } while (--i > 0);
		}
		PMAP_UPDATE_TLBS(pmap, va, va + PAGE_SIZE);
		simple_unlock(&pmap->lock);
	    }
	}

	pmap_phys_attributes[pai] &= ~bits;

	PMAP_WRITE_UNLOCK(spl);
}

/*
 *	Check specified attribute bits.
 */
boolean_t
phys_attribute_test(
	phys_addr_t	phys,
	int		bits)
{
	pv_entry_t		pv_h;
	pv_entry_t		pv_e;
	pt_entry_t		*pte;
	unsigned long		pai;
	pmap_t			pmap;
	int			spl;

	assert(phys != vm_page_fictitious_addr);
	if (!valid_page(phys)) {
	    /*
	     *	Not a managed page.
	     */
	    return (FALSE);
	}

	/*
	 *	Lock the pmap system first, since we will be checking
	 *	several pmaps.
	 */

	PMAP_WRITE_LOCK(spl);

	pai = pa_index(phys);
	pv_h = pai_to_pvh(pai);

	if (pmap_phys_attributes[pai] & bits) {
	    PMAP_WRITE_UNLOCK(spl);
	    return (TRUE);
	}

	/*
	 * Walk down PV list, checking all mappings.
	 * We do not have to lock the pv_list because we have
	 * the entire pmap system locked.
	 */
	if (pv_h->pmap != PMAP_NULL) {
	    /*
	     * There are some mappings.
	     */
	    for (pv_e = pv_h; pv_e != PV_ENTRY_NULL; pv_e = pv_e->next) {

		pmap = pv_e->pmap;
		/*
		 * Lock the pmap to block pmap_extract and similar routines.
		 */
		simple_lock(&pmap->lock);

		{
		    vm_offset_t va;

		    va = pv_e->va;
		    pte = pmap_pte(pmap, va);

		    /*
		     * Consistency checks.
		     */
		    assert(*pte & INTEL_PTE_VALID);
		    assert(pte_to_pa(*pte) == phys);
		}

		/*
		 * Check modify or reference bits.
		 */
		{
		    int	i = ptes_per_vm_page;

		    do {
			if (*pte & bits) {
			    simple_unlock(&pmap->lock);
			    PMAP_WRITE_UNLOCK(spl);
			    return (TRUE);
			}
		    } while (--i > 0);
		}
		simple_unlock(&pmap->lock);
	    }
	}
	PMAP_WRITE_UNLOCK(spl);
	return (FALSE);
}

/*
 *	Clear the modify bits on the specified physical page.
 */

void pmap_clear_modify(phys_addr_t phys)
{
	phys_attribute_clear(phys, PHYS_MODIFIED);
}

/*
 *	pmap_is_modified:
 *
 *	Return whether or not the specified physical page is modified
 *	by any physical maps.
 */

boolean_t pmap_is_modified(phys_addr_t phys)
{
	return (phys_attribute_test(phys, PHYS_MODIFIED));
}

/*
 *	pmap_clear_reference:
 *
 *	Clear the reference bit on the specified physical page.
 */

void pmap_clear_reference(phys_addr_t phys)
{
	phys_attribute_clear(phys, PHYS_REFERENCED);
}

/*
 *	pmap_is_referenced:
 *
 *	Return whether or not the specified physical page is referenced
 *	by any physical maps.
 */

boolean_t pmap_is_referenced(phys_addr_t phys)
{
	return (phys_attribute_test(phys, PHYS_REFERENCED));
}

#if	NCPUS > 1
/*
*	    TLB Coherence Code (TLB "shootdown" code)
*
* Threads that belong to the same task share the same address space and
* hence share a pmap.  However, they  may run on distinct cpus and thus
* have distinct TLBs that cache page table entries. In order to guarantee
* the TLBs are consistent, whenever a pmap is changed, all threads that
* are active in that pmap must have their TLB updated. To keep track of
* this information, the set of cpus that are currently using a pmap is
* maintained within each pmap structure (cpus_using). Pmap_activate() and
* pmap_deactivate add and remove, respectively, a cpu from this set.
* Since the TLBs are not addressable over the bus, each processor must
* flush its own TLB; a processor that needs to invalidate another TLB
* needs to interrupt the processor that owns that TLB to signal the
* update.
*
* Whenever a pmap is updated, the lock on that pmap is locked, and all
* cpus using the pmap are signaled to invalidate. All threads that need
* to activate a pmap must wait for the lock to clear to await any updates
* in progress before using the pmap. They must ACQUIRE the lock to add
* their cpu to the cpus_using set. An implicit assumption made
* throughout the TLB code is that all kernel code that runs at or higher
* than splvm blocks out update interrupts, and that such code does not
* touch pageable pages.
*
* A shootdown interrupt serves another function besides signaling a
* processor to invalidate. The interrupt routine (pmap_update_interrupt)
* waits for the both the pmap lock (and the kernel pmap lock) to clear,
* preventing user code from making implicit pmap updates while the
* sending processor is performing its update. (This could happen via a
* user data write reference that turns on the modify bit in the page
* table). It must wait for any kernel updates that may have started
* concurrently with a user pmap update because the IPC code
* changes mappings.
* Spinning on the VALUES of the locks is sufficient (rather than
* having to acquire the locks) because any updates that occur subsequent
* to finding the lock unlocked will be signaled via another interrupt.
* (This assumes the interrupt is cleared before the low level interrupt code
* calls pmap_update_interrupt()).
*
* The signaling processor must wait for any implicit updates in progress
* to terminate before continuing with its update. Thus it must wait for an
* acknowledgement of the interrupt from each processor for which such
* references could be made. For maintaining this information, a set
* cpus_active is used. A cpu is in this set if and only if it can
* use a pmap. When pmap_update_interrupt() is entered, a cpu is removed from
* this set; when all such cpus are removed, it is safe to update.
*
* Before attempting to acquire the update lock on a pmap, a cpu (A) must
* be at least at the priority of the interprocessor interrupt
* (splip<=splvm). Otherwise, A could grab a lock and be interrupted by a
* kernel update; it would spin forever in pmap_update_interrupt() trying
* to acquire the user pmap lock it had already acquired. Furthermore A
* must remove itself from cpus_active.  Otherwise, another cpu holding
* the lock (B) could be in the process of sending an update signal to A,
* and thus be waiting for A to remove itself from cpus_active. If A is
* spinning on the lock at priority this will never happen and a deadlock
* will result.
*/

/*
 *	Signal another CPU that it must flush its TLB
 */
void    signal_cpus(
	cpu_set		use_list,
	pmap_t		pmap,
	vm_offset_t	start, 
	vm_offset_t	end)
{
	int			which_cpu, j;
	pmap_update_list_t	update_list_p;

	while ((which_cpu = ffs(use_list)) != 0) {
	    which_cpu -= 1;	/* convert to 0 origin */

	    update_list_p = &cpu_update_list[which_cpu];
	    simple_lock(&update_list_p->lock);

	    j = update_list_p->count;
	    if (j >= UPDATE_LIST_SIZE) {
		/*
		 *	list overflowed.  Change last item to
		 *	indicate overflow.
		 */
		update_list_p->item[UPDATE_LIST_SIZE-1].pmap  = kernel_pmap;
		update_list_p->item[UPDATE_LIST_SIZE-1].start = VM_MIN_ADDRESS;
		update_list_p->item[UPDATE_LIST_SIZE-1].end   = VM_MAX_KERNEL_ADDRESS;
	    }
	    else {
		update_list_p->item[j].pmap  = pmap;
		update_list_p->item[j].start = start;
		update_list_p->item[j].end   = end;
		update_list_p->count = j+1;
	    }
	    cpu_update_needed[which_cpu] = TRUE;
	    simple_unlock(&update_list_p->lock);

	    if ((cpus_idle & (1 << which_cpu)) == 0)
		interrupt_processor(which_cpu);
	    use_list &= ~(1 << which_cpu);
	}
}

void process_pmap_updates(pmap_t my_pmap)
{
	int			my_cpu = cpu_number();
	pmap_update_list_t	update_list_p;
	int			j;
	pmap_t			pmap;

	update_list_p = &cpu_update_list[my_cpu];
	simple_lock(&update_list_p->lock);

	for (j = 0; j < update_list_p->count; j++) {
	    pmap = update_list_p->item[j].pmap;
	    if (pmap == my_pmap ||
		pmap == kernel_pmap) {

		INVALIDATE_TLB(pmap,
				update_list_p->item[j].start,
				update_list_p->item[j].end);
	    }
	}
	update_list_p->count = 0;
	cpu_update_needed[my_cpu] = FALSE;
	simple_unlock(&update_list_p->lock);
}

/*
 *	Interrupt routine for TBIA requested from other processor.
 */
void pmap_update_interrupt(void)
{
	int		my_cpu;
	pmap_t		my_pmap;
	int		s;

	my_cpu = cpu_number();

	/*
	 *	Exit now if we're idle.  We'll pick up the update request
	 *	when we go active, and we must not put ourselves back in
	 *	the active set because we'll never process the interrupt
	 *	while we're idle (thus hanging the system).
	 */
	if (cpus_idle & (1 << my_cpu))
	    return;

	if (current_thread() == THREAD_NULL)
	    my_pmap = kernel_pmap;
	else {
	    my_pmap = current_pmap();
	    if (!pmap_in_use(my_pmap, my_cpu))
		my_pmap = kernel_pmap;
	}

	/*
	 *	Raise spl to splvm (above splip) to block out pmap_extract
	 *	from IO code (which would put this cpu back in the active
	 *	set).
	 */
	s = splvm();

	do {

	    /*
	     *	Indicate that we're not using either user or kernel
	     *	pmap.
	     */
	    i_bit_clear(my_cpu, &cpus_active);

	    /*
	     *	Wait for any pmap updates in progress, on either user
	     *	or kernel pmap.
	     */
	    while (*(volatile int *)&my_pmap->lock.lock_data ||
		   *(volatile int *)&kernel_pmap->lock.lock_data)
		continue;

	    process_pmap_updates(my_pmap);

	    i_bit_set(my_cpu, &cpus_active);

	} while (cpu_update_needed[my_cpu]);

	splx(s);
}
#else	/* NCPUS > 1 */
/*
 *	Dummy routine to satisfy external reference.
 */
void pmap_update_interrupt(void)
{
	/* should never be called. */
}
#endif	/* NCPUS > 1 */

#if defined(__i386__) || defined (__x86_64__)
/* Unmap page 0 to trap NULL references.  */
void
pmap_unmap_page_zero (void)
{
  int *pte;

  printf("Unmapping the zero page.  Some BIOS functions may not be working any more.\n");
  pte = (int *) pmap_pte (kernel_pmap, 0);
  if (!pte)
    return;
  assert (pte);
#ifdef	MACH_PV_PAGETABLES
  if (!hyp_mmu_update_pte(kv_to_ma(pte), 0))
    printf("couldn't unmap page 0\n");
#else	/* MACH_PV_PAGETABLES */
  *pte = 0;
  INVALIDATE_TLB(kernel_pmap, 0, PAGE_SIZE);
#endif	/* MACH_PV_PAGETABLES */
}
#endif /* __i386__ */