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
|
/*
* Copyright (C) 2005 Intel Co
* Kun Tian (Kevin Tian) <kevin.tian@intel.com>
*
* 05/04/29 Kun Tian (Kevin Tian) <kevin.tian@intel.com> Add VTI domain support
*
* Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
* VA Linux Systems Japan K.K.
* dom0 vp model support
*/
/*
* NOTES on SMP
*
* * shared structures
* There are some structures which are accessed by CPUs concurrently.
* Here is the list of shared structures and operations on them which
* read/write the structures.
*
* - struct page_info
* This is a xen global resource. This structure is accessed by
* any CPUs.
*
* operations on this structure:
* - get_page() and its variant
* - put_page() and its variant
*
* - vTLB
* vcpu->arch.{d, i}tlb: Software tlb cache. These are per VCPU data.
* DEFINE_PER_CPU (unsigned long, vhpt_paddr): VHPT table per physical CPU.
*
* domain_flush_vtlb_range() and domain_flush_vtlb_all()
* write vcpu->arch.{d, i}tlb and VHPT table of vcpu which isn't current.
* So there are potential races to read/write VHPT and vcpu->arch.{d, i}tlb.
* Please note that reading VHPT is done by hardware page table walker.
*
* operations on this structure:
* - global tlb purge
* vcpu_ptc_g(), vcpu_ptc_ga() and domain_page_flush_and_put()
* I.e. callers of domain_flush_vtlb_range() and domain_flush_vtlb_all()
* These functions invalidate VHPT entry and vcpu->arch.{i, d}tlb
*
* - tlb insert and fc
* vcpu_itc_i()
* vcpu_itc_d()
* ia64_do_page_fault()
* vcpu_fc()
* These functions set VHPT entry and vcpu->arch.{i, d}tlb.
* Actually vcpu_itc_no_srlz() does.
*
* - the P2M table
* domain->mm and pgd, pud, pmd, pte table page.
* This structure is used to convert domain pseudo physical address
* to machine address. This is per domain resource.
*
* operations on this structure:
* - populate the P2M table tree
* lookup_alloc_domain_pte() and its variants.
* - set p2m entry
* assign_new_domain_page() and its variants.
* assign_domain_page() and its variants.
* - xchg p2m entry
* assign_domain_page_replace()
* - cmpxchg p2m entry
* assign_domain_page_cmpxchg_rel()
* replace_grant_host_mapping()
* steal_page()
* zap_domain_page_one()
* - read p2m entry
* lookup_alloc_domain_pte() and its variants.
*
* - the M2P table
* mpt_table (or machine_to_phys_mapping)
* This is a table which converts from machine address to pseudo physical
* address. This is a global structure.
*
* operations on this structure:
* - set m2p entry
* set_gpfn_from_mfn()
* - zap m2p entry
* set_gpfn_from_mfn(INVALID_P2M_ENTRY)
* - get m2p entry
* get_gpfn_from_mfn()
*
*
* * avoiding races
* The resources which are shared by CPUs must be accessed carefully
* to avoid race.
* IA64 has weak memory ordering so that attention must be paid
* to access shared structures. [SDM vol2 PartII chap. 2]
*
* - struct page_info memory ordering
* get_page() has acquire semantics.
* put_page() has release semantics.
*
* - populating the p2m table
* pgd, pud, pmd are append only.
*
* - races when updating the P2M tables and the M2P table
* The P2M entry are shared by more than one vcpu.
* So they are accessed atomic operations.
* I.e. xchg or cmpxchg must be used to update the p2m entry.
* NOTE: When creating/destructing a domain, we don't need to take care of
* this race.
*
* The M2P table is inverse of the P2M table.
* I.e. P2M(M2P(p)) = p and M2P(P2M(m)) = m
* The M2P table and P2M table must be updated consistently.
* Here is the update sequence
*
* xchg or cmpxchg case
* - set_gpfn_from_mfn(new_mfn, gpfn)
* - memory barrier
* - atomic update of the p2m entry (xchg or cmpxchg the p2m entry)
* get old_mfn entry as a result.
* - memory barrier
* - set_gpfn_from_mfn(old_mfn, INVALID_P2M_ENTRY)
*
* Here memory barrier can be achieved by release semantics.
*
* - races between global tlb purge and tlb insert
* This is a race between reading/writing vcpu->arch.{d, i}tlb or VHPT entry.
* When a vcpu is about to insert tlb, another vcpu may purge tlb
* cache globally. Inserting tlb (vcpu_itc_no_srlz()) or global tlb purge
* (domain_flush_vtlb_range() and domain_flush_vtlb_all()) can't update
* cpu->arch.{d, i}tlb, VHPT and mTLB. So there is a race here.
*
* Here check vcpu->arch.{d, i}tlb.p bit
* After inserting tlb entry, check the p bit and retry to insert.
* This means that when global tlb purge and tlb insert are issued
* simultaneously, always global tlb purge happens after tlb insert.
*
* - races between p2m entry update and tlb insert
* This is a race between reading/writing the p2m entry.
* reader: vcpu_itc_i(), vcpu_itc_d(), ia64_do_page_fault(), vcpu_fc()
* writer: assign_domain_page_cmpxchg_rel(), replace_grant_host_mapping(),
* steal_page(), zap_domain_page_one()
*
* For example, vcpu_itc_i() is about to insert tlb by calling
* vcpu_itc_no_srlz() after reading the p2m entry.
* At the same time, the p2m entry is replaced by xchg or cmpxchg and
* tlb cache of the page is flushed.
* There is a possibility that the p2m entry doesn't already point to the
* old page, but tlb cache still points to the old page.
* This can be detected similar to sequence lock using the p2m entry itself.
* reader remember the read value of the p2m entry, and insert tlb.
* Then read the p2m entry again. If the new p2m entry value is different
* from the used p2m entry value, the retry.
*
* - races between referencing page and p2m entry update
* This is a race between reading/writing the p2m entry.
* reader: vcpu_get_domain_bundle(), vmx_get_domain_bundle(),
* efi_emulate_get_time()
* writer: assign_domain_page_cmpxchg_rel(), replace_grant_host_mapping(),
* steal_page(), zap_domain_page_one()
*
* A page which assigned to a domain can be de-assigned by another vcpu.
* So before read/write to a domain page, the page's reference count
* must be incremented.
* vcpu_get_domain_bundle(), vmx_get_domain_bundle() and
* efi_emulate_get_time()
*
*/
#include <xen/config.h>
#include <xen/sched.h>
#include <xen/domain.h>
#include <asm/xentypes.h>
#include <xen/mm.h>
#include <xen/errno.h>
#include <asm/pgalloc.h>
#include <asm/vhpt.h>
#include <asm/vcpu.h>
#include <asm/shadow.h>
#include <asm/p2m_entry.h>
#include <asm/tlb_track.h>
#include <linux/efi.h>
#include <linux/sort.h>
#include <xen/guest_access.h>
#include <asm/page.h>
#include <asm/dom_fw_common.h>
#include <public/memory.h>
#include <asm/event.h>
#include <asm/debugger.h>
static void domain_page_flush_and_put(struct domain* d, unsigned long mpaddr,
volatile pte_t* ptep, pte_t old_pte,
struct page_info* page);
extern unsigned long ia64_iobase;
static struct domain *dom_xen, *dom_io;
/*
* This number is bigger than DOMID_SELF, DOMID_XEN and DOMID_IO.
* If more reserved domain ids are introduced, this might be increased.
*/
#define DOMID_P2M (0x7FF8U)
static struct domain *dom_p2m;
// followings are stolen from arch_init_memory() @ xen/arch/x86/mm.c
void
alloc_dom_xen_and_dom_io(void)
{
/*
* Initialise our DOMID_XEN domain.
* Any Xen-heap pages that we will allow to be mapped will have
* their domain field set to dom_xen.
*/
dom_xen = alloc_domain(DOMID_XEN);
BUG_ON(dom_xen == NULL);
/*
* Initialise our DOMID_IO domain.
* This domain owns I/O pages that are within the range of the page_info
* array. Mappings occur at the priv of the caller.
*/
dom_io = alloc_domain(DOMID_IO);
BUG_ON(dom_io == NULL);
}
static int
mm_teardown_can_skip(struct domain* d, unsigned long offset)
{
return d->arch.mm_teardown_offset > offset;
}
static void
mm_teardown_update_offset(struct domain* d, unsigned long offset)
{
d->arch.mm_teardown_offset = offset;
}
static void
mm_teardown_pte(struct domain* d, volatile pte_t* pte, unsigned long offset)
{
pte_t old_pte;
unsigned long mfn;
struct page_info* page;
old_pte = ptep_get_and_clear(&d->arch.mm, offset, pte);// acquire semantics
// vmx domain use bit[58:56] to distinguish io region from memory.
// see vmx_build_physmap_table() in vmx_init.c
if (!pte_mem(old_pte))
return;
// domain might map IO space or acpi table pages. check it.
mfn = pte_pfn(old_pte);
if (!mfn_valid(mfn))
return;
page = mfn_to_page(mfn);
BUG_ON(page_get_owner(page) == NULL);
// struct page_info corresponding to mfn may exist or not depending
// on CONFIG_VIRTUAL_FRAME_TABLE.
// The above check is too easy.
// The right way is to check whether this page is of io area or acpi pages
if (pte_pgc_allocated(old_pte)) {
BUG_ON(page_get_owner(page) != d);
BUG_ON(get_gpfn_from_mfn(mfn) == INVALID_M2P_ENTRY);
set_gpfn_from_mfn(mfn, INVALID_M2P_ENTRY);
if (test_and_clear_bit(_PGC_allocated, &page->count_info))
put_page(page);
} else {
put_page(page);
}
}
static int
mm_teardown_pmd(struct domain* d, volatile pmd_t* pmd, unsigned long offset)
{
unsigned long i;
volatile pte_t* pte = pte_offset_map(pmd, offset);
for (i = 0; i < PTRS_PER_PTE; i++, pte++) {
unsigned long cur_offset = offset + (i << PAGE_SHIFT);
if (mm_teardown_can_skip(d, cur_offset + PAGE_SIZE))
continue;
if (!pte_present(*pte)) { // acquire semantics
mm_teardown_update_offset(d, cur_offset);
continue;
}
mm_teardown_update_offset(d, cur_offset);
mm_teardown_pte(d, pte, cur_offset);
if (hypercall_preempt_check())
return -EAGAIN;
}
return 0;
}
static int
mm_teardown_pud(struct domain* d, volatile pud_t *pud, unsigned long offset)
{
unsigned long i;
volatile pmd_t *pmd = pmd_offset(pud, offset);
for (i = 0; i < PTRS_PER_PMD; i++, pmd++) {
unsigned long cur_offset = offset + (i << PMD_SHIFT);
if (mm_teardown_can_skip(d, cur_offset + PMD_SIZE))
continue;
if (!pmd_present(*pmd)) { // acquire semantics
mm_teardown_update_offset(d, cur_offset);
continue;
}
if (mm_teardown_pmd(d, pmd, cur_offset))
return -EAGAIN;
}
return 0;
}
static int
mm_teardown_pgd(struct domain* d, volatile pgd_t *pgd, unsigned long offset)
{
unsigned long i;
volatile pud_t *pud = pud_offset(pgd, offset);
for (i = 0; i < PTRS_PER_PUD; i++, pud++) {
unsigned long cur_offset = offset + (i << PUD_SHIFT);
#ifndef __PAGETABLE_PUD_FOLDED
if (mm_teardown_can_skip(d, cur_offset + PUD_SIZE))
continue;
#endif
if (!pud_present(*pud)) { // acquire semantics
#ifndef __PAGETABLE_PUD_FOLDED
mm_teardown_update_offset(d, cur_offset);
#endif
continue;
}
if (mm_teardown_pud(d, pud, cur_offset))
return -EAGAIN;
}
return 0;
}
int
mm_teardown(struct domain* d)
{
struct mm_struct* mm = &d->arch.mm;
unsigned long i;
volatile pgd_t* pgd;
if (mm->pgd == NULL)
return 0;
pgd = pgd_offset(mm, 0);
for (i = 0; i < PTRS_PER_PGD; i++, pgd++) {
unsigned long cur_offset = i << PGDIR_SHIFT;
if (mm_teardown_can_skip(d, cur_offset + PGDIR_SIZE))
continue;
if (!pgd_present(*pgd)) { // acquire semantics
mm_teardown_update_offset(d, cur_offset);
continue;
}
if (mm_teardown_pgd(d, pgd, cur_offset))
return -EAGAIN;
}
foreign_p2m_destroy(d);
return 0;
}
static void
mm_p2m_teardown_pmd(struct domain* d, volatile pmd_t* pmd,
unsigned long offset)
{
pte_free_kernel(pte_offset_map(pmd, offset));
}
static void
mm_p2m_teardown_pud(struct domain* d, volatile pud_t *pud,
unsigned long offset)
{
unsigned long i;
volatile pmd_t *pmd = pmd_offset(pud, offset);
for (i = 0; i < PTRS_PER_PMD; i++, pmd++) {
if (!pmd_present(*pmd))
continue;
mm_p2m_teardown_pmd(d, pmd, offset + (i << PMD_SHIFT));
}
pmd_free(pmd_offset(pud, offset));
}
static void
mm_p2m_teardown_pgd(struct domain* d, volatile pgd_t *pgd,
unsigned long offset)
{
unsigned long i;
volatile pud_t *pud = pud_offset(pgd, offset);
for (i = 0; i < PTRS_PER_PUD; i++, pud++) {
if (!pud_present(*pud))
continue;
mm_p2m_teardown_pud(d, pud, offset + (i << PUD_SHIFT));
}
pud_free(pud_offset(pgd, offset));
}
static void
mm_p2m_teardown(struct domain* d)
{
struct mm_struct* mm = &d->arch.mm;
unsigned long i;
volatile pgd_t* pgd;
BUG_ON(mm->pgd == NULL);
pgd = pgd_offset(mm, 0);
for (i = 0; i < PTRS_PER_PGD; i++, pgd++) {
if (!pgd_present(*pgd))
continue;
mm_p2m_teardown_pgd(d, pgd, i << PGDIR_SHIFT);
}
pgd_free(mm->pgd);
mm->pgd = NULL;
}
void
mm_final_teardown(struct domain* d)
{
if (d->arch.shadow_bitmap != NULL) {
xfree(d->arch.shadow_bitmap);
d->arch.shadow_bitmap = NULL;
}
mm_p2m_teardown(d);
}
unsigned long
domain_get_maximum_gpfn(struct domain *d)
{
return (d->arch.convmem_end + PAGE_SIZE - 1) >> PAGE_SHIFT;
}
// stolen from share_xen_page_with_guest() in xen/arch/x86/mm.c
void
share_xen_page_with_guest(struct page_info *page,
struct domain *d, int readonly)
{
if ( page_get_owner(page) == d )
return;
#if 1
if (readonly) {
printk("%s:%d readonly is not supported yet\n", __func__, __LINE__);
}
#endif
// alloc_xenheap_pages() doesn't initialize page owner.
//BUG_ON(page_get_owner(page) != NULL);
spin_lock(&d->page_alloc_lock);
#ifndef __ia64__
/* The incremented type count pins as writable or read-only. */
page->u.inuse.type_info = (readonly ? PGT_none : PGT_writable_page);
page->u.inuse.type_info |= PGT_validated | 1;
#endif
page_set_owner(page, d);
wmb(); /* install valid domain ptr before updating refcnt. */
ASSERT(page->count_info == 0);
/* Only add to the allocation list if the domain isn't dying. */
if ( !d->is_dying )
{
page->count_info |= PGC_allocated | 1;
if ( unlikely(d->xenheap_pages++ == 0) )
get_knownalive_domain(d);
list_add_tail(&page->list, &d->xenpage_list);
}
// grant_table_destroy() releases these pages.
// but it doesn't clear their m2p entry. So there might remain stale
// entries. such a stale entry is cleared here.
set_gpfn_from_mfn(page_to_mfn(page), INVALID_M2P_ENTRY);
spin_unlock(&d->page_alloc_lock);
}
void
share_xen_page_with_privileged_guests(struct page_info *page, int readonly)
{
share_xen_page_with_guest(page, dom_xen, readonly);
}
unsigned long
gmfn_to_mfn_foreign(struct domain *d, unsigned long gpfn)
{
unsigned long pte;
pte = lookup_domain_mpa(d,gpfn << PAGE_SHIFT, NULL);
if (!pte) {
panic("gmfn_to_mfn_foreign: bad gpfn. spinning...\n");
}
return ((pte & _PFN_MASK) >> PAGE_SHIFT);
}
// given a domain virtual address, pte and pagesize, extract the metaphysical
// address, convert the pte for a physical address for (possibly different)
// Xen PAGE_SIZE and return modified pte. (NOTE: TLB insert should use
// current->arch.vhpt_pg_shift!)
u64 translate_domain_pte(u64 pteval, u64 address, u64 itir__, u64* itir,
struct p2m_entry* entry)
{
struct domain *d = current->domain;
ia64_itir_t _itir = {.itir = itir__};
u64 mask, mpaddr, pteval2;
u64 arflags;
u64 arflags2;
u64 maflags2;
pteval &= ((1UL << 53) - 1);// ignore [63:53] bits
// FIXME address had better be pre-validated on insert
mask = ~itir_mask(_itir.itir);
mpaddr = ((pteval & _PAGE_PPN_MASK) & ~mask) | (address & mask);
if (_itir.ps > PAGE_SHIFT)
_itir.ps = PAGE_SHIFT;
((ia64_itir_t*)itir)->itir = _itir.itir;/* Copy the whole register. */
((ia64_itir_t*)itir)->ps = _itir.ps; /* Overwrite ps part! */
pteval2 = lookup_domain_mpa(d, mpaddr, entry);
if (_itir.ps < PAGE_SHIFT)
pteval2 |= mpaddr & ~PAGE_MASK & ~((1L << _itir.ps) - 1);
/* Check access rights. */
arflags = pteval & _PAGE_AR_MASK;
arflags2 = pteval2 & _PAGE_AR_MASK;
if (arflags != _PAGE_AR_R && arflags2 == _PAGE_AR_R) {
#if 0
dprintk(XENLOG_WARNING,
"%s:%d "
"pteval 0x%lx arflag 0x%lx address 0x%lx itir 0x%lx "
"pteval2 0x%lx arflags2 0x%lx mpaddr 0x%lx\n",
__func__, __LINE__,
pteval, arflags, address, itir__,
pteval2, arflags2, mpaddr);
#endif
pteval = (pteval & ~_PAGE_AR_MASK) | _PAGE_AR_R;
}
/* Check memory attribute. The switch is on the *requested* memory
attribute. */
maflags2 = pteval2 & _PAGE_MA_MASK;
switch (pteval & _PAGE_MA_MASK) {
case _PAGE_MA_NAT:
/* NaT pages are always accepted! */
break;
case _PAGE_MA_UC:
case _PAGE_MA_UCE:
case _PAGE_MA_WC:
if (maflags2 == _PAGE_MA_WB) {
/* Don't let domains WB-map uncached addresses.
This can happen when domU tries to touch i/o
port space. Also prevents possible address
aliasing issues. */
if (!(mpaddr - IO_PORTS_PADDR < IO_PORTS_SIZE)) {
u64 ucwb;
/*
* If dom0 page has both UC & WB attributes
* don't warn about attempted UC access.
*/
ucwb = efi_mem_attribute(mpaddr, PAGE_SIZE);
ucwb &= EFI_MEMORY_UC | EFI_MEMORY_WB;
ucwb ^= EFI_MEMORY_UC | EFI_MEMORY_WB;
if (d != dom0 || ucwb != 0)
gdprintk(XENLOG_WARNING, "Warning: UC"
" to WB for mpaddr=%lx\n",
mpaddr);
}
pteval = (pteval & ~_PAGE_MA_MASK) | _PAGE_MA_WB;
}
break;
case _PAGE_MA_WB:
if (maflags2 != _PAGE_MA_WB) {
/* Forbid non-coherent access to coherent memory. */
panic_domain(NULL, "try to use WB mem attr on "
"UC page, mpaddr=%lx\n", mpaddr);
}
break;
default:
panic_domain(NULL, "try to use unknown mem attribute\n");
}
/* If shadow mode is enabled, virtualize dirty bit. */
if (shadow_mode_enabled(d) && (pteval & _PAGE_D)) {
u64 mp_page = mpaddr >> PAGE_SHIFT;
pteval |= _PAGE_VIRT_D;
/* If the page is not already dirty, don't set the dirty bit! */
if (mp_page < d->arch.shadow_bitmap_size * 8
&& !test_bit(mp_page, d->arch.shadow_bitmap))
pteval &= ~_PAGE_D;
}
/* Ignore non-addr bits of pteval2 and force PL0->1
(PL3 is unaffected) */
return (pteval & ~(_PAGE_PPN_MASK | _PAGE_PL_MASK)) |
(pteval2 & _PAGE_PPN_MASK) |
(vcpu_pl_adjust(pteval, 7) & _PAGE_PL_MASK);
}
// given a current domain metaphysical address, return the physical address
unsigned long translate_domain_mpaddr(unsigned long mpaddr,
struct p2m_entry* entry)
{
unsigned long pteval;
pteval = lookup_domain_mpa(current->domain, mpaddr, entry);
return ((pteval & _PAGE_PPN_MASK) | (mpaddr & ~PAGE_MASK));
}
//XXX !xxx_present() should be used instread of !xxx_none()?
// pud, pmd, pte page is zero cleared when they are allocated.
// Their area must be visible before population so that
// cmpxchg must have release semantics.
static volatile pte_t*
lookup_alloc_domain_pte(struct domain* d, unsigned long mpaddr)
{
struct mm_struct *mm = &d->arch.mm;
volatile pgd_t *pgd;
volatile pud_t *pud;
volatile pmd_t *pmd;
BUG_ON(mm->pgd == NULL);
pgd = pgd_offset(mm, mpaddr);
again_pgd:
if (unlikely(pgd_none(*pgd))) { // acquire semantics
pud_t *old_pud = NULL;
pud = pud_alloc_one(mm, mpaddr);
if (unlikely(!pgd_cmpxchg_rel(mm, pgd, old_pud, pud))) {
pud_free(pud);
goto again_pgd;
}
}
pud = pud_offset(pgd, mpaddr);
again_pud:
if (unlikely(pud_none(*pud))) { // acquire semantics
pmd_t* old_pmd = NULL;
pmd = pmd_alloc_one(mm, mpaddr);
if (unlikely(!pud_cmpxchg_rel(mm, pud, old_pmd, pmd))) {
pmd_free(pmd);
goto again_pud;
}
}
pmd = pmd_offset(pud, mpaddr);
again_pmd:
if (unlikely(pmd_none(*pmd))) { // acquire semantics
pte_t* old_pte = NULL;
pte_t* pte = pte_alloc_one_kernel(mm, mpaddr);
if (unlikely(!pmd_cmpxchg_kernel_rel(mm, pmd, old_pte, pte))) {
pte_free_kernel(pte);
goto again_pmd;
}
}
return pte_offset_map(pmd, mpaddr);
}
//XXX xxx_none() should be used instread of !xxx_present()?
volatile pte_t*
lookup_noalloc_domain_pte(struct domain* d, unsigned long mpaddr)
{
struct mm_struct *mm = &d->arch.mm;
volatile pgd_t *pgd;
volatile pud_t *pud;
volatile pmd_t *pmd;
BUG_ON(mm->pgd == NULL);
pgd = pgd_offset(mm, mpaddr);
if (unlikely(!pgd_present(*pgd))) // acquire semantics
return NULL;
pud = pud_offset(pgd, mpaddr);
if (unlikely(!pud_present(*pud))) // acquire semantics
return NULL;
pmd = pmd_offset(pud, mpaddr);
if (unlikely(!pmd_present(*pmd))) // acquire semantics
return NULL;
return pte_offset_map(pmd, mpaddr);
}
static volatile pte_t*
lookup_noalloc_domain_pte_none(struct domain* d, unsigned long mpaddr)
{
struct mm_struct *mm = &d->arch.mm;
volatile pgd_t *pgd;
volatile pud_t *pud;
volatile pmd_t *pmd;
BUG_ON(mm->pgd == NULL);
pgd = pgd_offset(mm, mpaddr);
if (unlikely(pgd_none(*pgd))) // acquire semantics
return NULL;
pud = pud_offset(pgd, mpaddr);
if (unlikely(pud_none(*pud))) // acquire semantics
return NULL;
pmd = pmd_offset(pud, mpaddr);
if (unlikely(pmd_none(*pmd))) // acquire semantics
return NULL;
return pte_offset_map(pmd, mpaddr);
}
unsigned long
____lookup_domain_mpa(struct domain *d, unsigned long mpaddr)
{
volatile pte_t *pte;
pte = lookup_noalloc_domain_pte(d, mpaddr);
if (pte == NULL)
return INVALID_MFN;
if (pte_present(*pte))
return (pte->pte & _PFN_MASK);
else if (VMX_DOMAIN(d->vcpu[0]))
return GPFN_INV_MASK;
return INVALID_MFN;
}
unsigned long lookup_domain_mpa(struct domain *d, unsigned long mpaddr,
struct p2m_entry* entry)
{
volatile pte_t *pte = lookup_noalloc_domain_pte(d, mpaddr);
if (pte != NULL) {
pte_t tmp_pte = *pte;// pte is volatile. copy the value.
if (pte_present(tmp_pte)) {
if (entry != NULL)
p2m_entry_set(entry, pte, tmp_pte);
return pte_val(tmp_pte);
} else if (VMX_DOMAIN(d->vcpu[0]))
return GPFN_INV_MASK;
}
if (mpaddr < d->arch.convmem_end && !d->is_dying) {
gdprintk(XENLOG_WARNING, "vcpu %d iip 0x%016lx: non-allocated mpa "
"d %"PRId16" 0x%lx (< 0x%lx)\n",
current->vcpu_id, PSCB(current, iip),
d->domain_id, mpaddr, d->arch.convmem_end);
} else if (mpaddr - IO_PORTS_PADDR < IO_PORTS_SIZE) {
/* Log I/O port probing, but complain less loudly about it */
gdprintk(XENLOG_INFO, "vcpu %d iip 0x%016lx: bad I/O port access "
"d %"PRId16" 0x%lx\n",
current->vcpu_id, PSCB(current, iip), d->domain_id,
IO_SPACE_SPARSE_DECODING(mpaddr - IO_PORTS_PADDR));
} else {
gdprintk(XENLOG_WARNING, "vcpu %d iip 0x%016lx: bad mpa "
"d %"PRId16" 0x%lx (=> 0x%lx)\n",
current->vcpu_id, PSCB(current, iip),
d->domain_id, mpaddr, d->arch.convmem_end);
}
debugger_event (XEN_IA64_DEBUG_ON_BAD_MPA);
if (entry != NULL)
p2m_entry_set(entry, NULL, __pte(0));
//XXX This is a work around until the emulation memory access to a region
// where memory or device are attached is implemented.
return pte_val(pfn_pte(0, __pgprot(__DIRTY_BITS | _PAGE_PL_PRIV |
_PAGE_AR_RWX)));
}
// FIXME: ONLY USE FOR DOMAIN PAGE_SIZE == PAGE_SIZE
#if 1
void *domain_mpa_to_imva(struct domain *d, unsigned long mpaddr)
{
unsigned long pte = lookup_domain_mpa(d, mpaddr, NULL);
unsigned long imva;
pte &= _PAGE_PPN_MASK;
imva = (unsigned long) __va(pte);
imva |= mpaddr & ~PAGE_MASK;
return (void*)imva;
}
#else
void *domain_mpa_to_imva(struct domain *d, unsigned long mpaddr)
{
unsigned long imva = __gpa_to_mpa(d, mpaddr);
return (void *)__va(imva);
}
#endif
unsigned long
paddr_to_maddr(unsigned long paddr)
{
struct vcpu *v = current;
struct domain *d = v->domain;
u64 pa;
pa = ____lookup_domain_mpa(d, paddr);
if (pa == INVALID_MFN) {
printk("%s: called with bad memory address: 0x%lx - iip=%lx\n",
__func__, paddr, vcpu_regs(v)->cr_iip);
return 0;
}
return (pa & _PFN_MASK) | (paddr & ~PAGE_MASK);
}
/* Allocate a new page for domain and map it to the specified metaphysical
address. */
static struct page_info *
__assign_new_domain_page(struct domain *d, unsigned long mpaddr,
volatile pte_t* pte)
{
struct page_info *p;
unsigned long maddr;
BUG_ON(!pte_none(*pte));
p = alloc_domheap_page(d);
if (unlikely(!p)) {
printk("assign_new_domain_page: Can't alloc!!!! Aaaargh!\n");
return(p);
}
// zero out pages for security reasons
clear_page(page_to_virt(p));
maddr = page_to_maddr (p);
if (unlikely(maddr > __get_cpu_var(vhpt_paddr)
&& maddr < __get_cpu_var(vhpt_pend))) {
/* FIXME: how can this happen ?
vhpt is allocated by alloc_domheap_page. */
printk("assign_new_domain_page: reassigned vhpt page %lx!!\n",
maddr);
}
set_gpfn_from_mfn(page_to_mfn(p), mpaddr >> PAGE_SHIFT);
// clear_page() and set_gpfn_from_mfn() become visible before set_pte_rel()
// because set_pte_rel() has release semantics
set_pte_rel(pte,
pfn_pte(maddr >> PAGE_SHIFT,
__pgprot(_PAGE_PGC_ALLOCATED | __DIRTY_BITS |
_PAGE_PL_PRIV | _PAGE_AR_RWX)));
smp_mb();
return p;
}
struct page_info *
assign_new_domain_page(struct domain *d, unsigned long mpaddr)
{
volatile pte_t *pte = lookup_alloc_domain_pte(d, mpaddr);
if (!pte_none(*pte))
return NULL;
return __assign_new_domain_page(d, mpaddr, pte);
}
void __init
assign_new_domain0_page(struct domain *d, unsigned long mpaddr)
{
volatile pte_t *pte;
BUG_ON(d != dom0);
pte = lookup_alloc_domain_pte(d, mpaddr);
if (pte_none(*pte)) {
struct page_info *p = __assign_new_domain_page(d, mpaddr, pte);
if (p == NULL) {
panic("%s: can't allocate page for dom0\n", __func__);
}
}
}
static unsigned long
flags_to_prot (unsigned long flags)
{
unsigned long res = _PAGE_PL_PRIV | __DIRTY_BITS;
res |= flags & ASSIGN_readonly ? _PAGE_AR_R: _PAGE_AR_RWX;
res |= flags & ASSIGN_nocache ? _PAGE_MA_UC: _PAGE_MA_WB;
#ifdef CONFIG_XEN_IA64_TLB_TRACK
res |= flags & ASSIGN_tlb_track ? _PAGE_TLB_TRACKING: 0;
#endif
res |= flags & ASSIGN_pgc_allocated ? _PAGE_PGC_ALLOCATED: 0;
return res;
}
/* map a physical address to the specified metaphysical addr */
// flags: currently only ASSIGN_readonly, ASSIGN_nocache, ASSIGN_tlb_tack
// This is called by assign_domain_mmio_page().
// So accessing to pte is racy.
int
__assign_domain_page(struct domain *d,
unsigned long mpaddr, unsigned long physaddr,
unsigned long flags)
{
volatile pte_t *pte;
pte_t old_pte;
pte_t new_pte;
pte_t ret_pte;
unsigned long prot = flags_to_prot(flags);
pte = lookup_alloc_domain_pte(d, mpaddr);
old_pte = __pte(0);
new_pte = pfn_pte(physaddr >> PAGE_SHIFT, __pgprot(prot));
ret_pte = ptep_cmpxchg_rel(&d->arch.mm, mpaddr, pte, old_pte, new_pte);
if (pte_val(ret_pte) == pte_val(old_pte)) {
smp_mb();
return 0;
}
// dom0 tries to map real machine's I/O region, but failed.
// It is very likely that dom0 doesn't boot correctly because
// it can't access I/O. So complain here.
if (flags & ASSIGN_nocache) {
int warn = 0;
if (pte_pfn(ret_pte) != (physaddr >> PAGE_SHIFT))
warn = 1;
else if (!(pte_val(ret_pte) & _PAGE_MA_UC)) {
u32 type;
u64 attr;
warn = 1;
/*
* See
* complete_dom0_memmap()
* case EFI_RUNTIME_SERVICES_CODE:
* case EFI_RUNTIME_SERVICES_DATA:
* case EFI_ACPI_RECLAIM_MEMORY:
* case EFI_ACPI_MEMORY_NVS:
* case EFI_RESERVED_TYPE:
*
* Currently only EFI_RUNTIME_SERVICES_CODE is found
* so that we suppress only EFI_RUNTIME_SERVICES_CODE case.
*/
type = efi_mem_type(physaddr);
attr = efi_mem_attributes(physaddr);
if (type == EFI_RUNTIME_SERVICES_CODE &&
(attr & EFI_MEMORY_UC) && (attr & EFI_MEMORY_WB))
warn = 0;
}
if (warn)
printk("%s:%d WARNING can't assign page domain 0x%p id %d\n"
"\talready assigned pte_val 0x%016lx\n"
"\tmpaddr 0x%016lx physaddr 0x%016lx flags 0x%lx\n",
__func__, __LINE__,
d, d->domain_id, pte_val(ret_pte),
mpaddr, physaddr, flags);
}
return -EAGAIN;
}
/* get_page() and map a physical address to the specified metaphysical addr */
void
assign_domain_page(struct domain *d,
unsigned long mpaddr, unsigned long physaddr)
{
struct page_info* page = mfn_to_page(physaddr >> PAGE_SHIFT);
BUG_ON((physaddr & GPFN_IO_MASK) != GPFN_MEM);
BUG_ON(page->count_info != (PGC_allocated | 1));
set_gpfn_from_mfn(physaddr >> PAGE_SHIFT, mpaddr >> PAGE_SHIFT);
// because __assign_domain_page() uses set_pte_rel() which has
// release semantics, smp_mb() isn't needed.
(void)__assign_domain_page(d, mpaddr, physaddr,
ASSIGN_writable | ASSIGN_pgc_allocated);
}
int
ioports_permit_access(struct domain *d, unsigned int fp, unsigned int lp)
{
struct io_space *space;
unsigned long mmio_start, mmio_end, mach_start;
int ret;
if (IO_SPACE_NR(fp) >= num_io_spaces) {
dprintk(XENLOG_WARNING, "Unknown I/O Port range 0x%x - 0x%x\n", fp, lp);
return -EFAULT;
}
/*
* The ioport_cap rangeset tracks the I/O port address including
* the port space ID. This means port space IDs need to match
* between Xen and dom0. This is also a requirement because
* the hypercall to pass these port ranges only uses a u32.
*
* NB - non-dom0 driver domains may only have a subset of the
* I/O port spaces and thus will number port spaces differently.
* This is ok, they don't make use of this interface.
*/
ret = rangeset_add_range(d->arch.ioport_caps, fp, lp);
if (ret != 0)
return ret;
space = &io_space[IO_SPACE_NR(fp)];
/* Legacy I/O on dom0 is already setup */
if (d == dom0 && space == &io_space[0])
return 0;
fp = IO_SPACE_PORT(fp);
lp = IO_SPACE_PORT(lp);
if (space->sparse) {
mmio_start = IO_SPACE_SPARSE_ENCODING(fp) & ~PAGE_MASK;
mmio_end = PAGE_ALIGN(IO_SPACE_SPARSE_ENCODING(lp));
} else {
mmio_start = fp & ~PAGE_MASK;
mmio_end = PAGE_ALIGN(lp);
}
/*
* The "machine first port" is not necessarily identity mapped
* to the guest first port. At least for the legacy range.
*/
mach_start = mmio_start | __pa(space->mmio_base);
if (space == &io_space[0]) {
mmio_start |= IO_PORTS_PADDR;
mmio_end |= IO_PORTS_PADDR;
} else {
mmio_start |= __pa(space->mmio_base);
mmio_end |= __pa(space->mmio_base);
}
while (mmio_start <= mmio_end) {
(void)__assign_domain_page(d, mmio_start, mach_start, ASSIGN_nocache);
mmio_start += PAGE_SIZE;
mach_start += PAGE_SIZE;
}
return 0;
}
static int
ioports_has_allowed(struct domain *d, unsigned int fp, unsigned int lp)
{
for (; fp < lp; fp++)
if (rangeset_contains_singleton(d->arch.ioport_caps, fp))
return 1;
return 0;
}
int
ioports_deny_access(struct domain *d, unsigned int fp, unsigned int lp)
{
int ret;
struct mm_struct *mm = &d->arch.mm;
unsigned long mmio_start, mmio_end, mmio_base;
unsigned int fp_base, lp_base;
struct io_space *space;
if (IO_SPACE_NR(fp) >= num_io_spaces) {
dprintk(XENLOG_WARNING, "Unknown I/O Port range 0x%x - 0x%x\n", fp, lp);
return -EFAULT;
}
ret = rangeset_remove_range(d->arch.ioport_caps, fp, lp);
if (ret != 0)
return ret;
space = &io_space[IO_SPACE_NR(fp)];
fp_base = IO_SPACE_PORT(fp);
lp_base = IO_SPACE_PORT(lp);
if (space->sparse) {
mmio_start = IO_SPACE_SPARSE_ENCODING(fp_base) & ~PAGE_MASK;
mmio_end = PAGE_ALIGN(IO_SPACE_SPARSE_ENCODING(lp_base));
} else {
mmio_start = fp_base & ~PAGE_MASK;
mmio_end = PAGE_ALIGN(lp_base);
}
if (space == &io_space[0] && d != dom0)
mmio_base = IO_PORTS_PADDR;
else
mmio_base = __pa(space->mmio_base);
for (; mmio_start < mmio_end; mmio_start += PAGE_SIZE) {
unsigned int port, range;
unsigned long mpaddr;
volatile pte_t *pte;
pte_t old_pte;
if (space->sparse) {
port = IO_SPACE_SPARSE_DECODING(mmio_start);
range = IO_SPACE_SPARSE_PORTS_PER_PAGE - 1;
} else {
port = mmio_start;
range = PAGE_SIZE - 1;
}
port |= IO_SPACE_BASE(IO_SPACE_NR(fp));
if (port < fp || port + range > lp) {
/* Maybe this covers an allowed port. */
if (ioports_has_allowed(d, port, port + range))
continue;
}
mpaddr = mmio_start | mmio_base;
pte = lookup_noalloc_domain_pte_none(d, mpaddr);
BUG_ON(pte == NULL);
BUG_ON(pte_none(*pte));
/* clear pte */
old_pte = ptep_get_and_clear(mm, mpaddr, pte);
}
domain_flush_vtlb_all(d);
return 0;
}
static void
assign_domain_same_page(struct domain *d,
unsigned long mpaddr, unsigned long size,
unsigned long flags)
{
//XXX optimization
unsigned long end = PAGE_ALIGN(mpaddr + size);
for (mpaddr &= PAGE_MASK; mpaddr < end; mpaddr += PAGE_SIZE) {
(void)__assign_domain_page(d, mpaddr, mpaddr, flags);
}
}
int
efi_mmio(unsigned long physaddr, unsigned long size)
{
void *efi_map_start, *efi_map_end;
u64 efi_desc_size;
void* p;
efi_map_start = __va(ia64_boot_param->efi_memmap);
efi_map_end = efi_map_start + ia64_boot_param->efi_memmap_size;
efi_desc_size = ia64_boot_param->efi_memdesc_size;
for (p = efi_map_start; p < efi_map_end; p += efi_desc_size) {
efi_memory_desc_t* md = (efi_memory_desc_t *)p;
unsigned long start = md->phys_addr;
unsigned long end = md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT);
if (start <= physaddr && physaddr < end) {
if ((physaddr + size) > end) {
gdprintk(XENLOG_INFO, "%s: physaddr 0x%lx size = 0x%lx\n",
__func__, physaddr, size);
return 0;
}
// for io space
if (md->type == EFI_MEMORY_MAPPED_IO ||
md->type == EFI_MEMORY_MAPPED_IO_PORT_SPACE) {
return 1;
}
// for runtime
// see efi_enter_virtual_mode(void)
// in linux/arch/ia64/kernel/efi.c
if ((md->attribute & EFI_MEMORY_RUNTIME) &&
!(md->attribute & EFI_MEMORY_WB)) {
return 1;
}
return 0;
}
if (physaddr < start) {
break;
}
}
return 1;
}
unsigned long
assign_domain_mmio_page(struct domain *d, unsigned long mpaddr,
unsigned long phys_addr, unsigned long size,
unsigned long flags)
{
unsigned long addr = mpaddr & PAGE_MASK;
unsigned long end = PAGE_ALIGN(mpaddr + size);
if (size == 0) {
gdprintk(XENLOG_INFO, "%s: domain %p mpaddr 0x%lx size = 0x%lx\n",
__func__, d, mpaddr, size);
}
if (!efi_mmio(phys_addr, size)) {
#ifndef NDEBUG
gdprintk(XENLOG_INFO, "%s: domain %p mpaddr 0x%lx size = 0x%lx\n",
__func__, d, mpaddr, size);
#endif
return -EINVAL;
}
for (phys_addr &= PAGE_MASK; addr < end;
addr += PAGE_SIZE, phys_addr += PAGE_SIZE) {
__assign_domain_page(d, addr, phys_addr, flags);
}
return mpaddr;
}
unsigned long
assign_domain_mach_page(struct domain *d,
unsigned long mpaddr, unsigned long size,
unsigned long flags)
{
BUG_ON(flags & ASSIGN_pgc_allocated);
assign_domain_same_page(d, mpaddr, size, flags);
return mpaddr;
}
static void
adjust_page_count_info(struct page_info* page)
{
struct domain* d = page_get_owner(page);
BUG_ON((page->count_info & PGC_count_mask) != 1);
if (d != NULL) {
int ret = get_page(page, d);
BUG_ON(ret == 0);
} else {
u64 x, nx, y;
y = *((u64*)&page->count_info);
do {
x = y;
nx = x + 1;
BUG_ON((x >> 32) != 0);
BUG_ON((nx & PGC_count_mask) != 2);
y = cmpxchg((u64*)&page->count_info, x, nx);
} while (unlikely(y != x));
}
}
static void
domain_put_page(struct domain* d, unsigned long mpaddr,
volatile pte_t* ptep, pte_t old_pte, int clear_PGC_allocate)
{
unsigned long mfn = pte_pfn(old_pte);
struct page_info* page = mfn_to_page(mfn);
if (pte_pgc_allocated(old_pte)) {
if (page_get_owner(page) == d || page_get_owner(page) == NULL) {
BUG_ON(get_gpfn_from_mfn(mfn) != (mpaddr >> PAGE_SHIFT));
set_gpfn_from_mfn(mfn, INVALID_M2P_ENTRY);
} else {
BUG();
}
if (likely(clear_PGC_allocate)) {
if (!test_and_clear_bit(_PGC_allocated, &page->count_info))
BUG();
/* put_page() is done by domain_page_flush_and_put() */
} else {
// In this case, page reference count mustn't touched.
// domain_page_flush_and_put() decrements it, we increment
// it in advence. This patch is slow path.
//
// guest_remove_page(): owner = d, count_info = 1
// memory_exchange(): owner = NULL, count_info = 1
adjust_page_count_info(page);
}
}
domain_page_flush_and_put(d, mpaddr, ptep, old_pte, page);
}
// caller must get_page(mfn_to_page(mfn)) before call.
// caller must call set_gpfn_from_mfn() before call if necessary.
// because set_gpfn_from_mfn() result must be visible before pte xchg
// caller must use memory barrier. NOTE: xchg has acquire semantics.
// flags: ASSIGN_xxx
static void
assign_domain_page_replace(struct domain *d, unsigned long mpaddr,
unsigned long mfn, unsigned long flags)
{
struct mm_struct *mm = &d->arch.mm;
volatile pte_t* pte;
pte_t old_pte;
pte_t npte;
unsigned long prot = flags_to_prot(flags);
pte = lookup_alloc_domain_pte(d, mpaddr);
// update pte
npte = pfn_pte(mfn, __pgprot(prot));
old_pte = ptep_xchg(mm, mpaddr, pte, npte);
if (pte_mem(old_pte)) {
unsigned long old_mfn = pte_pfn(old_pte);
// mfn = old_mfn case can happen when domain maps a granted page
// twice with the same pseudo physial address.
// It's non sense, but allowed.
// __gnttab_map_grant_ref()
// => create_host_mapping()
// => assign_domain_page_replace()
if (mfn != old_mfn) {
domain_put_page(d, mpaddr, pte, old_pte, 1);
}
}
perfc_incr(assign_domain_page_replace);
}
// caller must get_page(new_page) before
// Only steal_page() calls this function.
static int
assign_domain_page_cmpxchg_rel(struct domain* d, unsigned long mpaddr,
struct page_info* old_page,
struct page_info* new_page,
unsigned long flags, int clear_PGC_allocate)
{
struct mm_struct *mm = &d->arch.mm;
volatile pte_t* pte;
unsigned long old_mfn;
unsigned long old_prot;
pte_t old_pte;
unsigned long new_mfn;
unsigned long new_prot;
pte_t new_pte;
pte_t ret_pte;
BUG_ON((flags & ASSIGN_pgc_allocated) == 0);
pte = lookup_alloc_domain_pte(d, mpaddr);
again:
old_prot = pte_val(*pte) & ~_PAGE_PPN_MASK;
old_mfn = page_to_mfn(old_page);
old_pte = pfn_pte(old_mfn, __pgprot(old_prot));
if (!pte_present(old_pte)) {
gdprintk(XENLOG_INFO,
"%s: old_pte 0x%lx old_prot 0x%lx old_mfn 0x%lx\n",
__func__, pte_val(old_pte), old_prot, old_mfn);
return -EINVAL;
}
new_prot = flags_to_prot(flags);
new_mfn = page_to_mfn(new_page);
new_pte = pfn_pte(new_mfn, __pgprot(new_prot));
// update pte
ret_pte = ptep_cmpxchg_rel(mm, mpaddr, pte, old_pte, new_pte);
if (unlikely(pte_val(old_pte) != pte_val(ret_pte))) {
if (pte_pfn(old_pte) == pte_pfn(ret_pte)) {
goto again;
}
gdprintk(XENLOG_INFO,
"%s: old_pte 0x%lx old_prot 0x%lx old_mfn 0x%lx "
"ret_pte 0x%lx ret_mfn 0x%lx\n",
__func__,
pte_val(old_pte), old_prot, old_mfn,
pte_val(ret_pte), pte_pfn(ret_pte));
return -EINVAL;
}
BUG_ON(!pte_mem(old_pte));
BUG_ON(!pte_pgc_allocated(old_pte));
BUG_ON(page_get_owner(old_page) != d);
BUG_ON(get_gpfn_from_mfn(old_mfn) != (mpaddr >> PAGE_SHIFT));
BUG_ON(old_mfn == new_mfn);
set_gpfn_from_mfn(old_mfn, INVALID_M2P_ENTRY);
if (likely(clear_PGC_allocate)) {
if (!test_and_clear_bit(_PGC_allocated, &old_page->count_info))
BUG();
} else {
int ret;
// adjust for count_info for domain_page_flush_and_put()
// This is slow path.
BUG_ON(!test_bit(_PGC_allocated, &old_page->count_info));
BUG_ON(d == NULL);
ret = get_page(old_page, d);
BUG_ON(ret == 0);
}
domain_page_flush_and_put(d, mpaddr, pte, old_pte, old_page);
perfc_incr(assign_domain_pge_cmpxchg_rel);
return 0;
}
static void
zap_domain_page_one(struct domain *d, unsigned long mpaddr,
int clear_PGC_allocate, unsigned long mfn)
{
struct mm_struct *mm = &d->arch.mm;
volatile pte_t *pte;
pte_t old_pte;
struct page_info *page;
pte = lookup_noalloc_domain_pte_none(d, mpaddr);
if (pte == NULL)
return;
if (pte_none(*pte))
return;
if (mfn == INVALID_MFN) {
// clear pte
old_pte = ptep_get_and_clear(mm, mpaddr, pte);
mfn = pte_pfn(old_pte);
} else {
unsigned long old_arflags;
pte_t new_pte;
pte_t ret_pte;
again:
// memory_exchange() calls guest_physmap_remove_page() with
// a stealed page. i.e. page owner = NULL.
BUG_ON(page_get_owner(mfn_to_page(mfn)) != d &&
page_get_owner(mfn_to_page(mfn)) != NULL);
old_arflags = pte_val(*pte) & ~_PAGE_PPN_MASK;
old_pte = pfn_pte(mfn, __pgprot(old_arflags));
new_pte = __pte(0);
// update pte
ret_pte = ptep_cmpxchg_rel(mm, mpaddr, pte, old_pte, new_pte);
if (unlikely(pte_val(old_pte) != pte_val(ret_pte))) {
if (pte_pfn(old_pte) == pte_pfn(ret_pte)) {
goto again;
}
gdprintk(XENLOG_INFO, "%s: old_pte 0x%lx old_arflags 0x%lx mfn 0x%lx "
"ret_pte 0x%lx ret_mfn 0x%lx\n",
__func__,
pte_val(old_pte), old_arflags, mfn,
pte_val(ret_pte), pte_pfn(ret_pte));
return;
}
BUG_ON(mfn != pte_pfn(ret_pte));
}
page = mfn_to_page(mfn);
BUG_ON((page->count_info & PGC_count_mask) == 0);
BUG_ON(clear_PGC_allocate && (page_get_owner(page) == NULL));
domain_put_page(d, mpaddr, pte, old_pte, clear_PGC_allocate);
perfc_incr(zap_domain_page_one);
}
unsigned long
dom0vp_zap_physmap(struct domain *d, unsigned long gpfn,
unsigned int extent_order)
{
if (extent_order != 0) {
//XXX
return -ENOSYS;
}
zap_domain_page_one(d, gpfn << PAGE_SHIFT, 1, INVALID_MFN);
perfc_incr(dom0vp_zap_physmap);
return 0;
}
static unsigned long
__dom0vp_add_physmap(struct domain* d, unsigned long gpfn,
unsigned long mfn_or_gmfn,
unsigned long flags, domid_t domid, int is_gmfn)
{
int error = -EINVAL;
struct domain* rd;
unsigned long mfn;
/* Not allowed by a domain. */
if (flags & (ASSIGN_nocache | ASSIGN_pgc_allocated))
return -EINVAL;
rd = rcu_lock_domain_by_id(domid);
if (unlikely(rd == NULL)) {
switch (domid) {
case DOMID_XEN:
rd = dom_xen;
break;
case DOMID_IO:
rd = dom_io;
break;
default:
gdprintk(XENLOG_INFO, "d 0x%p domid %d "
"gpfn 0x%lx mfn_or_gmfn 0x%lx flags 0x%lx domid %d\n",
d, d->domain_id, gpfn, mfn_or_gmfn, flags, domid);
return -ESRCH;
}
BUG_ON(rd == NULL);
rcu_lock_domain(rd);
}
if (unlikely(rd == d))
goto out1;
/*
* DOMID_XEN and DOMID_IO don't have their own p2m table.
* It can be considered that their p2m conversion is p==m.
*/
if (likely(is_gmfn && domid != DOMID_XEN && domid != DOMID_IO))
mfn = gmfn_to_mfn(rd, mfn_or_gmfn);
else
mfn = mfn_or_gmfn;
if (unlikely(!mfn_valid(mfn) || get_page(mfn_to_page(mfn), rd) == 0))
goto out1;
error = 0;
BUG_ON(page_get_owner(mfn_to_page(mfn)) == d &&
get_gpfn_from_mfn(mfn) != INVALID_M2P_ENTRY);
assign_domain_page_replace(d, gpfn << PAGE_SHIFT, mfn, flags);
//don't update p2m table because this page belongs to rd, not d.
perfc_incr(dom0vp_add_physmap);
out1:
rcu_unlock_domain(rd);
return error;
}
unsigned long
dom0vp_add_physmap(struct domain* d, unsigned long gpfn, unsigned long mfn,
unsigned long flags, domid_t domid)
{
return __dom0vp_add_physmap(d, gpfn, mfn, flags, domid, 0);
}
unsigned long
dom0vp_add_physmap_with_gmfn(struct domain* d, unsigned long gpfn,
unsigned long gmfn, unsigned long flags,
domid_t domid)
{
return __dom0vp_add_physmap(d, gpfn, gmfn, flags, domid, 1);
}
#ifdef CONFIG_XEN_IA64_EXPOSE_P2M
#define P2M_PFN_ROUNDUP(x) (((x) + PTRS_PER_PTE - 1) & \
~(PTRS_PER_PTE - 1))
#define P2M_PFN_ROUNDDOWN(x) ((x) & ~(PTRS_PER_PTE - 1))
#define P2M_NUM_PFN(x) (((x) + PTRS_PER_PTE - 1) / PTRS_PER_PTE)
#define MD_END(md) ((md)->phys_addr + \
((md)->num_pages << EFI_PAGE_SHIFT))
static struct page_info* p2m_pte_zero_page = NULL;
/* This must called before dom0 p2m table allocation */
void __init
expose_p2m_init(void)
{
pte_t* pte;
/*
* Initialise our DOMID_P2M domain.
* This domain owns m2p table pages.
*/
dom_p2m = alloc_domain(DOMID_P2M);
BUG_ON(dom_p2m == NULL);
dom_p2m->max_pages = ~0U;
pte = pte_alloc_one_kernel(NULL, 0);
BUG_ON(pte == NULL);
smp_mb();// make contents of the page visible.
p2m_pte_zero_page = virt_to_page(pte);
}
// allocate pgd, pmd of dest_dom if necessary
static int
allocate_pgd_pmd(struct domain* dest_dom, unsigned long dest_gpfn,
struct domain* src_dom,
unsigned long src_gpfn, unsigned long num_src_gpfn)
{
unsigned long i = 0;
BUG_ON((src_gpfn % PTRS_PER_PTE) != 0);
BUG_ON((num_src_gpfn % PTRS_PER_PTE) != 0);
while (i < num_src_gpfn) {
volatile pte_t* src_pte;
volatile pte_t* dest_pte;
src_pte = lookup_noalloc_domain_pte(src_dom,
(src_gpfn + i) << PAGE_SHIFT);
if (src_pte == NULL) {
i++;
continue;
}
dest_pte = lookup_alloc_domain_pte(dest_dom,
(dest_gpfn << PAGE_SHIFT) +
i * sizeof(pte_t));
if (dest_pte == NULL) {
gdprintk(XENLOG_INFO, "%s failed to allocate pte page\n",
__func__);
return -ENOMEM;
}
// skip to next pte page
i = P2M_PFN_ROUNDDOWN(i + PTRS_PER_PTE);
}
return 0;
}
static int
expose_p2m_page(struct domain* d, unsigned long mpaddr, struct page_info* page)
{
int ret = get_page(page, dom_p2m);
BUG_ON(ret != 1);
return __assign_domain_page(d, mpaddr, page_to_maddr(page),
ASSIGN_readonly);
}
// expose pte page
static int
expose_p2m_range(struct domain* dest_dom, unsigned long dest_gpfn,
struct domain* src_dom,
unsigned long src_gpfn, unsigned long num_src_gpfn)
{
unsigned long i = 0;
BUG_ON((src_gpfn % PTRS_PER_PTE) != 0);
BUG_ON((num_src_gpfn % PTRS_PER_PTE) != 0);
while (i < num_src_gpfn) {
volatile pte_t* pte;
pte = lookup_noalloc_domain_pte(src_dom, (src_gpfn + i) << PAGE_SHIFT);
if (pte == NULL) {
i++;
continue;
}
if (expose_p2m_page(dest_dom,
(dest_gpfn << PAGE_SHIFT) + i * sizeof(pte_t),
virt_to_page(pte)) < 0) {
gdprintk(XENLOG_INFO, "%s failed to assign page\n", __func__);
return -EAGAIN;
}
// skip to next pte page
i = P2M_PFN_ROUNDDOWN(i + PTRS_PER_PTE);
}
return 0;
}
// expose p2m_pte_zero_page
static int
expose_zero_page(struct domain* dest_dom, unsigned long dest_gpfn,
unsigned long num_src_gpfn)
{
unsigned long i;
for (i = 0; i < P2M_NUM_PFN(num_src_gpfn); i++) {
volatile pte_t* pte;
pte = lookup_noalloc_domain_pte(dest_dom,
(dest_gpfn + i) << PAGE_SHIFT);
if (pte == NULL || pte_present(*pte))
continue;
if (expose_p2m_page(dest_dom, (dest_gpfn + i) << PAGE_SHIFT,
p2m_pte_zero_page) < 0) {
gdprintk(XENLOG_INFO, "%s failed to assign zero-pte page\n",
__func__);
return -EAGAIN;
}
}
return 0;
}
static int
expose_p2m(struct domain* dest_dom, unsigned long dest_gpfn,
struct domain* src_dom,
unsigned long src_gpfn, unsigned long num_src_gpfn)
{
if (allocate_pgd_pmd(dest_dom, dest_gpfn,
src_dom, src_gpfn, num_src_gpfn))
return -ENOMEM;
if (expose_p2m_range(dest_dom, dest_gpfn,
src_dom, src_gpfn, num_src_gpfn))
return -EAGAIN;
if (expose_zero_page(dest_dom, dest_gpfn, num_src_gpfn))
return -EAGAIN;
return 0;
}
static void
unexpose_p2m(struct domain* dest_dom,
unsigned long dest_gpfn, unsigned long num_dest_gpfn)
{
unsigned long i;
for (i = 0; i < num_dest_gpfn; i++) {
zap_domain_page_one(dest_dom, (dest_gpfn + i) << PAGE_SHIFT,
0, INVALID_MFN);
}
}
// It is possible to optimize loop, But this isn't performance critical.
unsigned long
dom0vp_expose_p2m(struct domain* d,
unsigned long conv_start_gpfn,
unsigned long assign_start_gpfn,
unsigned long expose_size, unsigned long granule_pfn)
{
unsigned long ret;
unsigned long expose_num_pfn = expose_size >> PAGE_SHIFT;
if ((expose_size % PAGE_SIZE) != 0 ||
(granule_pfn % PTRS_PER_PTE) != 0 ||
(expose_num_pfn % PTRS_PER_PTE) != 0 ||
(conv_start_gpfn % granule_pfn) != 0 ||
(assign_start_gpfn % granule_pfn) != 0 ||
(expose_num_pfn % granule_pfn) != 0) {
gdprintk(XENLOG_INFO,
"%s conv_start_gpfn 0x%016lx assign_start_gpfn 0x%016lx "
"expose_size 0x%016lx granulte_pfn 0x%016lx\n", __func__,
conv_start_gpfn, assign_start_gpfn, expose_size, granule_pfn);
return -EINVAL;
}
if (granule_pfn != PTRS_PER_PTE) {
gdprintk(XENLOG_INFO,
"%s granule_pfn 0x%016lx PTRS_PER_PTE 0x%016lx\n",
__func__, granule_pfn, PTRS_PER_PTE);
return -ENOSYS;
}
ret = expose_p2m(d, assign_start_gpfn,
d, conv_start_gpfn, expose_num_pfn);
return ret;
}
static int
memmap_info_copy_from_guest(struct xen_ia64_memmap_info* memmap_info,
char** memmap_p,
XEN_GUEST_HANDLE(char) buffer)
{
char *memmap;
char *p;
char *memmap_end;
efi_memory_desc_t *md;
unsigned long start;
unsigned long end;
efi_memory_desc_t *prev_md;
if (copy_from_guest((char*)memmap_info, buffer, sizeof(*memmap_info)))
return -EFAULT;
if (memmap_info->efi_memdesc_size < sizeof(efi_memory_desc_t) ||
memmap_info->efi_memmap_size < memmap_info->efi_memdesc_size ||
(memmap_info->efi_memmap_size % memmap_info->efi_memdesc_size) != 0)
return -EINVAL;
memmap = _xmalloc(memmap_info->efi_memmap_size,
__alignof__(efi_memory_desc_t));
if (memmap == NULL)
return -ENOMEM;
if (copy_from_guest_offset(memmap, buffer, sizeof(*memmap_info),
memmap_info->efi_memmap_size)) {
xfree(memmap);
return -EFAULT;
}
/* intergirty check & simplify */
sort(memmap, memmap_info->efi_memmap_size / memmap_info->efi_memdesc_size,
memmap_info->efi_memdesc_size, efi_mdt_cmp, NULL);
/* alignement & overlap check */
prev_md = NULL;
p = memmap;
memmap_end = memmap + memmap_info->efi_memmap_size;
for (p = memmap; p < memmap_end; p += memmap_info->efi_memmap_size) {
md = (efi_memory_desc_t*)p;
start = md->phys_addr;
if (start & ((1UL << EFI_PAGE_SHIFT) - 1) || md->num_pages == 0) {
xfree(memmap);
return -EINVAL;
}
if (prev_md != NULL) {
unsigned long prev_end = MD_END(prev_md);
if (prev_end > start) {
xfree(memmap);
return -EINVAL;
}
}
prev_md = (efi_memory_desc_t *)p;
}
/* coalease */
prev_md = NULL;
p = memmap;
while (p < memmap_end) {
md = (efi_memory_desc_t*)p;
start = md->phys_addr;
end = MD_END(md);
start = P2M_PFN_ROUNDDOWN(start >> PAGE_SHIFT) << PAGE_SHIFT;
end = P2M_PFN_ROUNDUP(end >> PAGE_SHIFT) << PAGE_SHIFT;
md->phys_addr = start;
md->num_pages = (end - start) >> EFI_PAGE_SHIFT;
if (prev_md != NULL) {
unsigned long prev_end = MD_END(prev_md);
if (prev_end >= start) {
size_t left;
end = max(prev_end, end);
prev_md->num_pages = (end - prev_md->phys_addr) >> EFI_PAGE_SHIFT;
left = memmap_end - p;
if (left > memmap_info->efi_memdesc_size) {
left -= memmap_info->efi_memdesc_size;
memmove(p, p + memmap_info->efi_memdesc_size, left);
}
memmap_info->efi_memmap_size -= memmap_info->efi_memdesc_size;
memmap_end -= memmap_info->efi_memdesc_size;
continue;
}
}
prev_md = md;
p += memmap_info->efi_memdesc_size;
}
if (copy_to_guest(buffer, (char*)memmap_info, sizeof(*memmap_info)) ||
copy_to_guest_offset(buffer, sizeof(*memmap_info),
(char*)memmap, memmap_info->efi_memmap_size)) {
xfree(memmap);
return -EFAULT;
}
*memmap_p = memmap;
return 0;
}
static int
foreign_p2m_allocate_pte(struct domain* d,
const struct xen_ia64_memmap_info* memmap_info,
const void* memmap)
{
const void* memmap_end = memmap + memmap_info->efi_memmap_size;
const void* p;
for (p = memmap; p < memmap_end; p += memmap_info->efi_memdesc_size) {
const efi_memory_desc_t* md = p;
unsigned long start = md->phys_addr;
unsigned long end = MD_END(md);
unsigned long gpaddr;
for (gpaddr = start; gpaddr < end; gpaddr += PAGE_SIZE) {
if (lookup_alloc_domain_pte(d, gpaddr) == NULL) {
return -ENOMEM;
}
}
}
return 0;
}
struct foreign_p2m_region {
unsigned long gpfn;
unsigned long num_gpfn;
};
struct foreign_p2m_entry {
struct list_head list;
int busy;
/* src domain */
struct domain* src_dom;
/* region into which foreign p2m table is mapped */
unsigned long gpfn;
unsigned long num_gpfn;
unsigned int num_region;
struct foreign_p2m_region region[0];
};
/* caller must increment the reference count of src_dom */
static int
foreign_p2m_alloc(struct foreign_p2m* foreign_p2m,
unsigned long dest_gpfn, struct domain* src_dom,
struct xen_ia64_memmap_info* memmap_info, void* memmap,
struct foreign_p2m_entry** entryp)
{
void* memmap_end = memmap + memmap_info->efi_memmap_size;
efi_memory_desc_t* md;
unsigned long dest_gpfn_end;
unsigned long src_gpfn;
unsigned long src_gpfn_end;
unsigned int num_region;
struct foreign_p2m_entry* entry;
struct foreign_p2m_entry* prev;
struct foreign_p2m_entry* pos;
num_region = (memmap_end - memmap) / memmap_info->efi_memdesc_size;
md = memmap;
src_gpfn = P2M_PFN_ROUNDDOWN(md->phys_addr >> PAGE_SHIFT);
md = memmap + (num_region - 1) * memmap_info->efi_memdesc_size;
src_gpfn_end = MD_END(md) >> PAGE_SHIFT;
if (src_gpfn_end >
P2M_PFN_ROUNDUP(src_dom->arch.convmem_end >> PAGE_SHIFT))
return -EINVAL;
src_gpfn_end = P2M_PFN_ROUNDUP(src_gpfn_end);
dest_gpfn_end = dest_gpfn + P2M_NUM_PFN(src_gpfn_end - src_gpfn);
entry = _xmalloc(sizeof(*entry) + num_region * sizeof(entry->region[0]),
__alignof__(*entry));
if (entry == NULL)
return -ENOMEM;
entry->busy = 1;
entry->gpfn = dest_gpfn;
entry->num_gpfn = dest_gpfn_end - dest_gpfn;
entry->src_dom = src_dom;
entry->num_region = 0;
memset(entry->region, 0, sizeof(entry->region[0]) * num_region);
prev = NULL;
spin_lock(&foreign_p2m->lock);
if (list_empty(&foreign_p2m->head))
prev = (struct foreign_p2m_entry*)&foreign_p2m->head;
list_for_each_entry(pos, &foreign_p2m->head, list) {
if (pos->gpfn + pos->num_gpfn < dest_gpfn) {
prev = pos;
continue;
}
if (dest_gpfn_end < pos->gpfn) {
if (prev != NULL && prev->gpfn + prev->num_gpfn > dest_gpfn)
prev = NULL;/* overlap */
break;
}
/* overlap */
prev = NULL;
break;
}
if (prev != NULL) {
list_add(&entry->list, &prev->list);
spin_unlock(&foreign_p2m->lock);
*entryp = entry;
return 0;
}
spin_unlock(&foreign_p2m->lock);
xfree(entry);
return -EBUSY;
}
static void
foreign_p2m_unexpose(struct domain* dest_dom, struct foreign_p2m_entry* entry)
{
unsigned int i;
BUG_ON(!entry->busy);
for (i = 0; i < entry->num_region; i++)
unexpose_p2m(dest_dom,
entry->region[i].gpfn, entry->region[i].num_gpfn);
}
static void
foreign_p2m_unbusy(struct foreign_p2m* foreign_p2m,
struct foreign_p2m_entry* entry)
{
spin_lock(&foreign_p2m->lock);
BUG_ON(!entry->busy);
entry->busy = 0;
spin_unlock(&foreign_p2m->lock);
}
static void
foreign_p2m_free(struct foreign_p2m* foreign_p2m,
struct foreign_p2m_entry* entry)
{
spin_lock(&foreign_p2m->lock);
BUG_ON(!entry->busy);
list_del(&entry->list);
spin_unlock(&foreign_p2m->lock);
put_domain(entry->src_dom);
xfree(entry);
}
void
foreign_p2m_init(struct domain* d)
{
struct foreign_p2m* foreign_p2m = &d->arch.foreign_p2m;
INIT_LIST_HEAD(&foreign_p2m->head);
spin_lock_init(&foreign_p2m->lock);
}
void
foreign_p2m_destroy(struct domain* d)
{
struct foreign_p2m* foreign_p2m = &d->arch.foreign_p2m;
struct foreign_p2m_entry* entry;
struct foreign_p2m_entry* n;
spin_lock(&foreign_p2m->lock);
list_for_each_entry_safe(entry, n, &foreign_p2m->head, list) {
/* mm_teardown() cleared p2m table already */
/* foreign_p2m_unexpose(d, entry);*/
list_del(&entry->list);
put_domain(entry->src_dom);
xfree(entry);
}
spin_unlock(&foreign_p2m->lock);
}
unsigned long
dom0vp_expose_foreign_p2m(struct domain* dest_dom,
unsigned long dest_gpfn,
domid_t domid,
XEN_GUEST_HANDLE(char) buffer,
unsigned long flags)
{
unsigned long ret = 0;
struct domain* src_dom;
struct xen_ia64_memmap_info memmap_info;
char* memmap;
void* memmap_end;
void* p;
struct foreign_p2m_entry* entry;
ret = memmap_info_copy_from_guest(&memmap_info, &memmap, buffer);
if (ret != 0)
return ret;
dest_dom = rcu_lock_domain(dest_dom);
if (dest_dom == NULL) {
ret = -EINVAL;
goto out;
}
#if 1
// Self foreign domain p2m exposure isn't allowed.
// Otherwise the domain can't be destroyed because
// no one decrements the domain reference count.
if (domid == dest_dom->domain_id) {
ret = -EINVAL;
goto out;
}
#endif
src_dom = get_domain_by_id(domid);
if (src_dom == NULL) {
ret = -EINVAL;
goto out_unlock;
}
if (flags & IA64_DOM0VP_EFP_ALLOC_PTE) {
ret = foreign_p2m_allocate_pte(src_dom, &memmap_info, memmap);
if (ret != 0)
goto out_unlock;
}
ret = foreign_p2m_alloc(&dest_dom->arch.foreign_p2m, dest_gpfn,
src_dom, &memmap_info, memmap, &entry);
if (ret != 0)
goto out_unlock;
memmap_end = memmap + memmap_info.efi_memmap_size;
for (p = memmap; p < memmap_end; p += memmap_info.efi_memdesc_size) {
efi_memory_desc_t* md = p;
unsigned long src_gpfn =
P2M_PFN_ROUNDDOWN(md->phys_addr >> PAGE_SHIFT);
unsigned long src_gpfn_end =
P2M_PFN_ROUNDUP(MD_END(md) >> PAGE_SHIFT);
unsigned long num_src_gpfn = src_gpfn_end - src_gpfn;
ret = expose_p2m(dest_dom, dest_gpfn + src_gpfn / PTRS_PER_PTE,
src_dom, src_gpfn, num_src_gpfn);
if (ret != 0)
break;
entry->region[entry->num_region].gpfn =
dest_gpfn + src_gpfn / PTRS_PER_PTE;
entry->region[entry->num_region].num_gpfn = P2M_NUM_PFN(num_src_gpfn);
entry->num_region++;
}
if (ret == 0) {
foreign_p2m_unbusy(&dest_dom->arch.foreign_p2m, entry);
} else {
foreign_p2m_unexpose(dest_dom, entry);
foreign_p2m_free(&dest_dom->arch.foreign_p2m, entry);
}
out_unlock:
rcu_unlock_domain(dest_dom);
out:
xfree(memmap);
return ret;
}
unsigned long
dom0vp_unexpose_foreign_p2m(struct domain* dest_dom,
unsigned long dest_gpfn,
domid_t domid)
{
int ret = -ENOENT;
struct foreign_p2m* foreign_p2m = &dest_dom->arch.foreign_p2m;
struct foreign_p2m_entry* entry;
dest_dom = rcu_lock_domain(dest_dom);
if (dest_dom == NULL)
return ret;
spin_lock(&foreign_p2m->lock);
list_for_each_entry(entry, &foreign_p2m->head, list) {
if (entry->gpfn < dest_gpfn)
continue;
if (dest_gpfn < entry->gpfn)
break;
if (domid == entry->src_dom->domain_id)
ret = 0;
else
ret = -EINVAL;
break;
}
if (ret == 0) {
if (entry->busy == 0)
entry->busy = 1;
else
ret = -EBUSY;
}
spin_unlock(&foreign_p2m->lock);
if (ret == 0) {
foreign_p2m_unexpose(dest_dom, entry);
foreign_p2m_free(&dest_dom->arch.foreign_p2m, entry);
}
rcu_unlock_domain(dest_dom);
return ret;
}
#endif
// grant table host mapping
// mpaddr: host_addr: pseudo physical address
// mfn: frame: machine page frame
// flags: GNTMAP_readonly | GNTMAP_application_map | GNTMAP_contains_pte
int
create_grant_host_mapping(unsigned long gpaddr, unsigned long mfn,
unsigned int flags, unsigned int cache_flags)
{
struct domain* d = current->domain;
struct page_info* page;
int ret;
if ((flags & (GNTMAP_device_map |
GNTMAP_application_map | GNTMAP_contains_pte)) ||
(cache_flags)) {
gdprintk(XENLOG_INFO, "%s: flags 0x%x cache_flags 0x%x\n",
__func__, flags, cache_flags);
return GNTST_general_error;
}
BUG_ON(!mfn_valid(mfn));
page = mfn_to_page(mfn);
ret = get_page(page, page_get_owner(page));
BUG_ON(ret == 0);
assign_domain_page_replace(d, gpaddr, mfn,
#ifdef CONFIG_XEN_IA64_TLB_TRACK
ASSIGN_tlb_track |
#endif
((flags & GNTMAP_readonly) ?
ASSIGN_readonly : ASSIGN_writable));
perfc_incr(create_grant_host_mapping);
return GNTST_okay;
}
// grant table host unmapping
int
replace_grant_host_mapping(unsigned long gpaddr,
unsigned long mfn, unsigned long new_gpaddr, unsigned int flags)
{
struct domain* d = current->domain;
unsigned long gpfn = gpaddr >> PAGE_SHIFT;
volatile pte_t* pte;
unsigned long cur_arflags;
pte_t cur_pte;
pte_t new_pte = __pte(0);
pte_t old_pte;
struct page_info* page = mfn_to_page(mfn);
struct page_info* new_page = NULL;
volatile pte_t* new_page_pte = NULL;
if (new_gpaddr) {
new_page_pte = lookup_noalloc_domain_pte_none(d, new_gpaddr);
if (likely(new_page_pte != NULL)) {
new_pte = ptep_get_and_clear(&d->arch.mm,
new_gpaddr, new_page_pte);
if (likely(pte_present(new_pte))) {
unsigned long new_page_mfn;
struct domain* page_owner;
new_page_mfn = pte_pfn(new_pte);
new_page = mfn_to_page(new_page_mfn);
page_owner = page_get_owner(new_page);
if (unlikely(page_owner == NULL)) {
gdprintk(XENLOG_INFO,
"%s: page_owner == NULL "
"gpaddr 0x%lx mfn 0x%lx "
"new_gpaddr 0x%lx mfn 0x%lx\n",
__func__, gpaddr, mfn, new_gpaddr, new_page_mfn);
new_page = NULL; /* prevent domain_put_page() */
goto out;
}
/*
* domain_put_page(clear_PGC_allcoated = 0)
* doesn't decrement refcount of page with
* pte_ptc_allocated() = 1. Be carefull.
*/
if (unlikely(!pte_pgc_allocated(new_pte))) {
/* domain_put_page() decrements page refcount. adjust it. */
if (get_page(new_page, page_owner)) {
gdprintk(XENLOG_INFO,
"%s: get_page() failed. "
"gpaddr 0x%lx mfn 0x%lx "
"new_gpaddr 0x%lx mfn 0x%lx\n",
__func__, gpaddr, mfn,
new_gpaddr, new_page_mfn);
goto out;
}
}
domain_put_page(d, new_gpaddr, new_page_pte, new_pte, 0);
} else
new_pte = __pte(0);
}
}
if (flags & (GNTMAP_application_map | GNTMAP_contains_pte)) {
gdprintk(XENLOG_INFO, "%s: flags 0x%x\n", __func__, flags);
return GNTST_general_error;
}
pte = lookup_noalloc_domain_pte(d, gpaddr);
if (pte == NULL) {
gdprintk(XENLOG_INFO, "%s: gpaddr 0x%lx mfn 0x%lx\n",
__func__, gpaddr, mfn);
goto out;
}
again:
cur_arflags = pte_val(*pte) & ~_PAGE_PPN_MASK;
cur_pte = pfn_pte(mfn, __pgprot(cur_arflags));
if (!pte_present(cur_pte) ||
(page_get_owner(page) == d && get_gpfn_from_mfn(mfn) == gpfn)) {
gdprintk(XENLOG_INFO, "%s: gpaddr 0x%lx mfn 0x%lx cur_pte 0x%lx\n",
__func__, gpaddr, mfn, pte_val(cur_pte));
goto out;
}
old_pte = ptep_cmpxchg_rel(&d->arch.mm, gpaddr, pte, cur_pte, new_pte);
if (unlikely(!pte_present(old_pte))) {
gdprintk(XENLOG_INFO, "%s: gpaddr 0x%lx mfn 0x%lx"
" cur_pte 0x%lx old_pte 0x%lx\n",
__func__, gpaddr, mfn, pte_val(cur_pte), pte_val(old_pte));
goto out;
}
if (unlikely(pte_val(cur_pte) != pte_val(old_pte))) {
if (pte_pfn(old_pte) == mfn) {
goto again;
}
gdprintk(XENLOG_INFO, "%s gpaddr 0x%lx mfn 0x%lx cur_pte "
"0x%lx old_pte 0x%lx\n",
__func__, gpaddr, mfn, pte_val(cur_pte), pte_val(old_pte));
goto out;
}
BUG_ON(pte_pfn(old_pte) != mfn);
/* try_to_clear_PGC_allocate(d, page) is not needed. */
BUG_ON(page_get_owner(page) == d &&
get_gpfn_from_mfn(mfn) == gpfn);
BUG_ON(pte_pgc_allocated(old_pte));
domain_page_flush_and_put(d, gpaddr, pte, old_pte, page);
perfc_incr(replace_grant_host_mapping);
return GNTST_okay;
out:
if (new_page)
domain_put_page(d, new_gpaddr, new_page_pte, new_pte, 1);
return GNTST_general_error;
}
// heavily depends on the struct page layout.
// gnttab_transfer() calls steal_page() with memflags = 0
// For grant table transfer, we must fill the page.
// memory_exchange() calls steal_page() with memflags = MEMF_no_refcount
// For memory exchange, we don't have to fill the page because
// memory_exchange() does it.
int
steal_page(struct domain *d, struct page_info *page, unsigned int memflags)
{
#if 0 /* if big endian */
# error "implement big endian version of steal_page()"
#endif
u32 _d, _nd;
u64 x, nx, y;
if (page_get_owner(page) != d) {
gdprintk(XENLOG_INFO, "%s d 0x%p owner 0x%p\n",
__func__, d, page_get_owner(page));
return -1;
}
if (!(memflags & MEMF_no_refcount)) {
unsigned long gpfn;
struct page_info *new;
unsigned long new_mfn;
int ret;
new = alloc_domheap_page(d);
if (new == NULL) {
gdprintk(XENLOG_INFO, "alloc_domheap_page() failed\n");
return -1;
}
// zero out pages for security reasons
clear_page(page_to_virt(new));
// assign_domain_page_cmpxchg_rel() has release semantics
// so smp_mb() isn't needed.
gpfn = get_gpfn_from_mfn(page_to_mfn(page));
if (gpfn == INVALID_M2P_ENTRY) {
free_domheap_page(new);
return -1;
}
new_mfn = page_to_mfn(new);
set_gpfn_from_mfn(new_mfn, gpfn);
// smp_mb() isn't needed because assign_domain_pge_cmpxchg_rel()
// has release semantics.
ret = assign_domain_page_cmpxchg_rel(d, gpfn << PAGE_SHIFT, page, new,
ASSIGN_writable |
ASSIGN_pgc_allocated, 0);
if (ret < 0) {
gdprintk(XENLOG_INFO, "assign_domain_page_cmpxchg_rel failed %d\n",
ret);
set_gpfn_from_mfn(new_mfn, INVALID_M2P_ENTRY);
free_domheap_page(new);
return -1;
}
perfc_incr(steal_page_refcount);
}
spin_lock(&d->page_alloc_lock);
/*
* The tricky bit: atomically release ownership while there is just one
* benign reference to the page (PGC_allocated). If that reference
* disappears then the deallocation routine will safely spin.
*/
_d = pickle_domptr(d);
y = *((u64*)&page->count_info);
do {
x = y;
nx = x & 0xffffffff;
// page->count_info: untouched
// page->u.inused._domain = 0;
_nd = x >> 32;
if (unlikely(((x & (PGC_count_mask | PGC_allocated)) !=
(1 | PGC_allocated))) ||
unlikely(_nd != _d)) {
struct domain* nd = unpickle_domptr(_nd);
if (nd == NULL) {
gdprintk(XENLOG_INFO, "gnttab_transfer: "
"Bad page %p: ed=%p(%u) 0x%x, "
"sd=%p 0x%x,"
" caf=%016lx, taf=%" PRtype_info
" memflags 0x%x\n",
(void *) page_to_mfn(page),
d, d->domain_id, _d,
nd, _nd,
x,
page->u.inuse.type_info,
memflags);
} else {
gdprintk(XENLOG_WARNING, "gnttab_transfer: "
"Bad page %p: ed=%p(%u) 0x%x, "
"sd=%p(%u) 0x%x,"
" caf=%016lx, taf=%" PRtype_info
" memflags 0x%x\n",
(void *) page_to_mfn(page),
d, d->domain_id, _d,
nd, nd->domain_id, _nd,
x,
page->u.inuse.type_info,
memflags);
}
spin_unlock(&d->page_alloc_lock);
return -1;
}
y = cmpxchg((u64*)&page->count_info, x, nx);
} while (unlikely(y != x));
/*
* Unlink from 'd'. At least one reference remains (now anonymous), so
* noone else is spinning to try to delete this page from 'd'.
*/
if ( !(memflags & MEMF_no_refcount) )
d->tot_pages--;
list_del(&page->list);
spin_unlock(&d->page_alloc_lock);
perfc_incr(steal_page);
return 0;
}
int
guest_physmap_add_page(struct domain *d, unsigned long gpfn,
unsigned long mfn)
{
BUG_ON(!mfn_valid(mfn));
BUG_ON(mfn_to_page(mfn)->count_info != (PGC_allocated | 1));
set_gpfn_from_mfn(mfn, gpfn);
smp_mb();
assign_domain_page_replace(d, gpfn << PAGE_SHIFT, mfn,
ASSIGN_writable | ASSIGN_pgc_allocated);
//BUG_ON(mfn != ((lookup_domain_mpa(d, gpfn << PAGE_SHIFT) & _PFN_MASK) >> PAGE_SHIFT));
perfc_incr(guest_physmap_add_page);
return 0;
}
void
guest_physmap_remove_page(struct domain *d, unsigned long gpfn,
unsigned long mfn)
{
BUG_ON(mfn == 0);//XXX
zap_domain_page_one(d, gpfn << PAGE_SHIFT, 0, mfn);
perfc_incr(guest_physmap_remove_page);
}
static void
domain_page_flush_and_put(struct domain* d, unsigned long mpaddr,
volatile pte_t* ptep, pte_t old_pte,
struct page_info* page)
{
#ifdef CONFIG_XEN_IA64_TLB_TRACK
struct tlb_track_entry* entry;
#endif
if (shadow_mode_enabled(d))
shadow_mark_page_dirty(d, mpaddr >> PAGE_SHIFT);
#ifndef CONFIG_XEN_IA64_TLB_TRACK
//XXX sledgehammer.
// flush finer range.
domain_flush_vtlb_all(d);
put_page(page);
#else
switch (tlb_track_search_and_remove(d->arch.tlb_track,
ptep, old_pte, &entry)) {
case TLB_TRACK_NOT_TRACKED:
// dprintk(XENLOG_WARNING, "%s TLB_TRACK_NOT_TRACKED\n", __func__);
/* This page is zapped from this domain
* by memory decrease or exchange or dom0vp_zap_physmap.
* I.e. the page is zapped for returning this page to xen
* (balloon driver or DMA page allocation) or
* foreign domain mapped page is unmapped from the domain.
* In the former case the page is to be freed so that
* we can defer freeing page to batch.
* In the latter case the page is unmapped so that
* we need to flush it. But to optimize it, we
* queue the page and flush vTLB only once.
* I.e. The caller must call dfree_flush() explicitly.
*/
domain_flush_vtlb_all(d);
put_page(page);
break;
case TLB_TRACK_NOT_FOUND:
// dprintk(XENLOG_WARNING, "%s TLB_TRACK_NOT_FOUND\n", __func__);
/* This page is zapped from this domain
* by grant table page unmap.
* Luckily the domain that mapped this page didn't
* access this page so that we don't have to flush vTLB.
* Probably the domain did only DMA.
*/
/* do nothing */
put_page(page);
break;
case TLB_TRACK_FOUND:
// dprintk(XENLOG_WARNING, "%s TLB_TRACK_FOUND\n", __func__);
/* This page is zapped from this domain
* by grant table page unmap.
* Fortunately this page is accessced via only one virtual
* memory address. So it is easy to flush it.
*/
domain_flush_vtlb_track_entry(d, entry);
tlb_track_free_entry(d->arch.tlb_track, entry);
put_page(page);
break;
case TLB_TRACK_MANY:
gdprintk(XENLOG_INFO, "%s TLB_TRACK_MANY\n", __func__);
/* This page is zapped from this domain
* by grant table page unmap.
* Unfortunately this page is accessced via many virtual
* memory address (or too many times with single virtual address).
* So we abondaned to track virtual addresses.
* full vTLB flush is necessary.
*/
domain_flush_vtlb_all(d);
put_page(page);
break;
case TLB_TRACK_AGAIN:
gdprintk(XENLOG_ERR, "%s TLB_TRACK_AGAIN\n", __func__);
BUG();
break;
}
#endif
perfc_incr(domain_page_flush_and_put);
}
int
domain_page_mapped(struct domain* d, unsigned long mpaddr)
{
volatile pte_t * pte;
pte = lookup_noalloc_domain_pte(d, mpaddr);
if(pte != NULL && !pte_none(*pte))
return 1;
return 0;
}
/* Flush cache of domain d. */
void domain_cache_flush (struct domain *d, int sync_only)
{
struct mm_struct *mm = &d->arch.mm;
volatile pgd_t *pgd = mm->pgd;
unsigned long maddr;
int i,j,k, l;
int nbr_page = 0;
void (*flush_func)(unsigned long start, unsigned long end);
extern void flush_dcache_range (unsigned long, unsigned long);
if (sync_only)
flush_func = &flush_icache_range;
else
flush_func = &flush_dcache_range;
for (i = 0; i < PTRS_PER_PGD; pgd++, i++) {
volatile pud_t *pud;
if (!pgd_present(*pgd)) // acquire semantics
continue;
pud = pud_offset(pgd, 0);
for (j = 0; j < PTRS_PER_PUD; pud++, j++) {
volatile pmd_t *pmd;
if (!pud_present(*pud)) // acquire semantics
continue;
pmd = pmd_offset(pud, 0);
for (k = 0; k < PTRS_PER_PMD; pmd++, k++) {
volatile pte_t *pte;
if (!pmd_present(*pmd)) // acquire semantics
continue;
pte = pte_offset_map(pmd, 0);
for (l = 0; l < PTRS_PER_PTE; pte++, l++) {
if (!pte_present(*pte)) // acquire semantics
continue;
/* Convert PTE to maddr. */
maddr = __va_ul (pte_val(*pte)
& _PAGE_PPN_MASK);
(*flush_func)(maddr, maddr+ PAGE_SIZE);
nbr_page++;
}
}
}
}
//printk ("domain_cache_flush: %d %d pages\n", d->domain_id, nbr_page);
}
#ifdef VERBOSE
#define MEM_LOG(_f, _a...) \
printk("DOM%u: (file=mm.c, line=%d) " _f "\n", \
current->domain->domain_id , __LINE__ , ## _a )
#else
#define MEM_LOG(_f, _a...) ((void)0)
#endif
static void free_page_type(struct page_info *page, u32 type)
{
}
static int alloc_page_type(struct page_info *page, u32 type)
{
return 1;
}
static int opt_p2m_xenheap;
boolean_param("p2m_xenheap", opt_p2m_xenheap);
void *pgtable_quicklist_alloc(void)
{
void *p;
BUG_ON(dom_p2m == NULL);
if (!opt_p2m_xenheap) {
struct page_info *page = alloc_domheap_page(dom_p2m);
if (page == NULL)
return NULL;
p = page_to_virt(page);
clear_page(p);
return p;
}
p = alloc_xenheap_pages(0);
if (p) {
clear_page(p);
/*
* This page should be read only. At this moment, the third
* argument doesn't make sense. It should be 1 when supported.
*/
share_xen_page_with_guest(virt_to_page(p), dom_p2m, 0);
}
return p;
}
void pgtable_quicklist_free(void *pgtable_entry)
{
struct page_info* page = virt_to_page(pgtable_entry);
BUG_ON(page_get_owner(page) != dom_p2m);
BUG_ON(page->count_info != (1 | PGC_allocated));
put_page(page);
if (opt_p2m_xenheap)
free_xenheap_page(pgtable_entry);
}
void put_page_type(struct page_info *page)
{
u64 nx, x, y = page->u.inuse.type_info;
again:
do {
x = y;
nx = x - 1;
ASSERT((x & PGT_count_mask) != 0);
/*
* The page should always be validated while a reference is held. The
* exception is during domain destruction, when we forcibly invalidate
* page-table pages if we detect a referential loop.
* See domain.c:relinquish_list().
*/
ASSERT((x & PGT_validated) || page_get_owner(page)->is_dying);
if ( unlikely((nx & PGT_count_mask) == 0) )
{
/* Record TLB information for flush later. Races are harmless. */
page->tlbflush_timestamp = tlbflush_current_time();
if ( unlikely((nx & PGT_type_mask) <= PGT_l4_page_table) &&
likely(nx & PGT_validated) )
{
/*
* Page-table pages must be unvalidated when count is zero. The
* 'free' is safe because the refcnt is non-zero and validated
* bit is clear => other ops will spin or fail.
*/
if ( unlikely((y = cmpxchg(&page->u.inuse.type_info, x,
x & ~PGT_validated)) != x) )
goto again;
/* We cleared the 'valid bit' so we do the clean up. */
free_page_type(page, x);
/* Carry on, but with the 'valid bit' now clear. */
x &= ~PGT_validated;
nx &= ~PGT_validated;
}
}
}
while ( unlikely((y = cmpxchg_rel(&page->u.inuse.type_info, x, nx)) != x) );
}
int get_page_type(struct page_info *page, u32 type)
{
u64 nx, x, y = page->u.inuse.type_info;
ASSERT(!(type & ~PGT_type_mask));
again:
do {
x = y;
nx = x + 1;
if ( unlikely((nx & PGT_count_mask) == 0) )
{
MEM_LOG("Type count overflow on pfn %lx", page_to_mfn(page));
return 0;
}
else if ( unlikely((x & PGT_count_mask) == 0) )
{
if ( (x & PGT_type_mask) != type )
{
/*
* On type change we check to flush stale TLB entries. This
* may be unnecessary (e.g., page was GDT/LDT) but those
* circumstances should be very rare.
*/
cpumask_t mask =
page_get_owner(page)->domain_dirty_cpumask;
tlbflush_filter(mask, page->tlbflush_timestamp);
if ( unlikely(!cpus_empty(mask)) )
{
perfc_incr(need_flush_tlb_flush);
flush_tlb_mask(mask);
}
/* We lose existing type, back pointer, and validity. */
nx &= ~(PGT_type_mask | PGT_validated);
nx |= type;
/* No special validation needed for writable pages. */
/* Page tables and GDT/LDT need to be scanned for validity. */
if ( type == PGT_writable_page )
nx |= PGT_validated;
}
}
else if ( unlikely((x & PGT_type_mask) != type) )
{
if ( ((x & PGT_type_mask) != PGT_l2_page_table) ||
(type != PGT_l1_page_table) )
MEM_LOG("Bad type (saw %08lx != exp %08x) "
"for mfn %016lx (pfn %016lx)",
x, type, page_to_mfn(page),
get_gpfn_from_mfn(page_to_mfn(page)));
return 0;
}
else if ( unlikely(!(x & PGT_validated)) )
{
/* Someone else is updating validation of this page. Wait... */
while ( (y = page->u.inuse.type_info) == x )
cpu_relax();
goto again;
}
}
while ( unlikely((y = cmpxchg_acq(&page->u.inuse.type_info, x, nx)) != x) );
if ( unlikely(!(nx & PGT_validated)) )
{
/* Try to validate page type; drop the new reference on failure. */
if ( unlikely(!alloc_page_type(page, type)) )
{
MEM_LOG("Error while validating mfn %lx (pfn %lx) for type %08x"
": caf=%08x taf=%" PRtype_info,
page_to_mfn(page), get_gpfn_from_mfn(page_to_mfn(page)),
type, page->count_info, page->u.inuse.type_info);
/* Noone else can get a reference. We hold the only ref. */
page->u.inuse.type_info = 0;
return 0;
}
/* Noone else is updating simultaneously. */
__set_bit(_PGT_validated, &page->u.inuse.type_info);
}
return 1;
}
int memory_is_conventional_ram(paddr_t p)
{
return (efi_mem_type(p) == EFI_CONVENTIONAL_MEMORY);
}
long
arch_memory_op(int op, XEN_GUEST_HANDLE(void) arg)
{
switch (op) {
case XENMEM_add_to_physmap:
{
struct xen_add_to_physmap xatp;
unsigned long prev_mfn, mfn = 0, gpfn;
struct domain *d;
if (copy_from_guest(&xatp, arg, 1))
return -EFAULT;
if (xatp.domid == DOMID_SELF)
d = rcu_lock_current_domain();
else {
if ((d = rcu_lock_domain_by_id(xatp.domid)) == NULL)
return -ESRCH;
if (!IS_PRIV_FOR(current->domain,d)) {
rcu_lock_domain(d);
return -EPERM;
}
}
/* This hypercall is used for VT-i domain only */
if (!VMX_DOMAIN(d->vcpu[0])) {
rcu_unlock_domain(d);
return -ENOSYS;
}
switch (xatp.space) {
case XENMAPSPACE_shared_info:
if (xatp.idx == 0)
mfn = virt_to_mfn(d->shared_info);
break;
case XENMAPSPACE_grant_table:
spin_lock(&d->grant_table->lock);
if ((xatp.idx >= nr_grant_frames(d->grant_table)) &&
(xatp.idx < max_nr_grant_frames))
gnttab_grow_table(d, xatp.idx + 1);
if (xatp.idx < nr_grant_frames(d->grant_table))
mfn = virt_to_mfn(d->grant_table->shared[xatp.idx]);
spin_unlock(&d->grant_table->lock);
break;
default:
break;
}
if (mfn == 0) {
rcu_unlock_domain(d);
return -EINVAL;
}
LOCK_BIGLOCK(d);
/* Check remapping necessity */
prev_mfn = gmfn_to_mfn(d, xatp.gpfn);
if (mfn == prev_mfn)
goto out;
/* Remove previously mapped page if it was present. */
if (prev_mfn && mfn_valid(prev_mfn)) {
if (is_xen_heap_mfn(prev_mfn))
/* Xen heap frames are simply unhooked from this phys slot. */
guest_physmap_remove_page(d, xatp.gpfn, prev_mfn);
else
/* Normal domain memory is freed, to avoid leaking memory. */
guest_remove_page(d, xatp.gpfn);
}
/* Unmap from old location, if any. */
gpfn = get_gpfn_from_mfn(mfn);
if (gpfn != INVALID_M2P_ENTRY)
guest_physmap_remove_page(d, gpfn, mfn);
/* Map at new location. */
guest_physmap_add_page(d, xatp.gpfn, mfn);
out:
UNLOCK_BIGLOCK(d);
rcu_unlock_domain(d);
break;
}
case XENMEM_machine_memory_map:
{
struct xen_memory_map memmap;
struct xen_ia64_memmap_info memmap_info;
XEN_GUEST_HANDLE(char) buffer;
if (!IS_PRIV(current->domain))
return -EINVAL;
if (copy_from_guest(&memmap, arg, 1))
return -EFAULT;
if (memmap.nr_entries <
sizeof(memmap_info) + ia64_boot_param->efi_memmap_size)
return -EINVAL;
memmap.nr_entries =
sizeof(memmap_info) + ia64_boot_param->efi_memmap_size;
memset(&memmap_info, 0, sizeof(memmap_info));
memmap_info.efi_memmap_size = ia64_boot_param->efi_memmap_size;
memmap_info.efi_memdesc_size = ia64_boot_param->efi_memdesc_size;
memmap_info.efi_memdesc_version = ia64_boot_param->efi_memdesc_version;
buffer = guest_handle_cast(memmap.buffer, char);
if (copy_to_guest(buffer, (char*)&memmap_info, sizeof(memmap_info)) ||
copy_to_guest_offset(buffer, sizeof(memmap_info),
(char*)__va(ia64_boot_param->efi_memmap),
ia64_boot_param->efi_memmap_size) ||
copy_to_guest(arg, &memmap, 1))
return -EFAULT;
return 0;
}
default:
return -ENOSYS;
}
return 0;
}
int is_iomem_page(unsigned long mfn)
{
return (!mfn_valid(mfn) || (page_get_owner(mfn_to_page(mfn)) == dom_io));
}
/*
* Local variables:
* mode: C
* c-set-style: "BSD"
* c-basic-offset: 4
* tab-width: 4
* indent-tabs-mode: nil
* End:
*/
|