summaryrefslogtreecommitdiffstats
path: root/arch/sparc/mm/srmmu.c
blob: c664b962987cdde1ad1a5bd17d288dbd2e907c75 (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
/*
 * srmmu.c:  SRMMU specific routines for memory management.
 *
 * Copyright (C) 1995 David S. Miller  (davem@caip.rutgers.edu)
 * Copyright (C) 1995,2002 Pete Zaitcev (zaitcev@yahoo.com)
 * Copyright (C) 1996 Eddie C. Dost    (ecd@skynet.be)
 * Copyright (C) 1997,1998 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
 * Copyright (C) 1999,2000 Anton Blanchard (anton@samba.org)
 */

#include <linux/config.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/pagemap.h>
#include <linux/init.h>
#include <linux/spinlock.h>
#include <linux/bootmem.h>
#include <linux/fs.h>
#include <linux/seq_file.h>

#include <asm/bitext.h>
#include <asm/page.h>
#include <asm/pgalloc.h>
#include <asm/pgtable.h>
#include <asm/io.h>
#include <asm/kdebug.h>
#include <asm/vaddrs.h>
#include <asm/traps.h>
#include <asm/smp.h>
#include <asm/mbus.h>
#include <asm/cache.h>
#include <asm/oplib.h>
#include <asm/sbus.h>
#include <asm/asi.h>
#include <asm/msi.h>
#include <asm/a.out.h>
#include <asm/mmu_context.h>
#include <asm/io-unit.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

/* Now the cpu specific definitions. */
#include <asm/viking.h>
#include <asm/mxcc.h>
#include <asm/ross.h>
#include <asm/tsunami.h>
#include <asm/swift.h>
#include <asm/turbosparc.h>

#include <asm/btfixup.h>

enum mbus_module srmmu_modtype;
unsigned int hwbug_bitmask;
int vac_cache_size;
int vac_line_size;

extern struct resource sparc_iomap;

extern unsigned long last_valid_pfn;

extern unsigned long page_kernel;

pgd_t *srmmu_swapper_pg_dir;

#ifdef CONFIG_SMP
#define FLUSH_BEGIN(mm)
#define FLUSH_END
#else
#define FLUSH_BEGIN(mm) if((mm)->context != NO_CONTEXT) {
#define FLUSH_END	}
#endif

BTFIXUPDEF_CALL(void, flush_page_for_dma, unsigned long)
#define flush_page_for_dma(page) BTFIXUP_CALL(flush_page_for_dma)(page)

int flush_page_for_dma_global = 1;

#ifdef CONFIG_SMP
BTFIXUPDEF_CALL(void, local_flush_page_for_dma, unsigned long)
#define local_flush_page_for_dma(page) BTFIXUP_CALL(local_flush_page_for_dma)(page)
#endif

char *srmmu_name;

ctxd_t *srmmu_ctx_table_phys;
ctxd_t *srmmu_context_table;

int viking_mxcc_present;
static DEFINE_SPINLOCK(srmmu_context_spinlock);

int is_hypersparc;

/*
 * In general all page table modifications should use the V8 atomic
 * swap instruction.  This insures the mmu and the cpu are in sync
 * with respect to ref/mod bits in the page tables.
 */
static inline unsigned long srmmu_swap(unsigned long *addr, unsigned long value)
{
	__asm__ __volatile__("swap [%2], %0" : "=&r" (value) : "0" (value), "r" (addr));
	return value;
}

static inline void srmmu_set_pte(pte_t *ptep, pte_t pteval)
{
	srmmu_swap((unsigned long *)ptep, pte_val(pteval));
}

/* The very generic SRMMU page table operations. */
static inline int srmmu_device_memory(unsigned long x)
{
	return ((x & 0xF0000000) != 0);
}

int srmmu_cache_pagetables;

/* these will be initialized in srmmu_nocache_calcsize() */
unsigned long srmmu_nocache_size;
unsigned long srmmu_nocache_end;

/* 1 bit <=> 256 bytes of nocache <=> 64 PTEs */
#define SRMMU_NOCACHE_BITMAP_SHIFT (PAGE_SHIFT - 4)

/* The context table is a nocache user with the biggest alignment needs. */
#define SRMMU_NOCACHE_ALIGN_MAX (sizeof(ctxd_t)*SRMMU_MAX_CONTEXTS)

void *srmmu_nocache_pool;
void *srmmu_nocache_bitmap;
static struct bit_map srmmu_nocache_map;

static unsigned long srmmu_pte_pfn(pte_t pte)
{
	if (srmmu_device_memory(pte_val(pte))) {
		/* Just return something that will cause
		 * pfn_valid() to return false.  This makes
		 * copy_one_pte() to just directly copy to
		 * PTE over.
		 */
		return ~0UL;
	}
	return (pte_val(pte) & SRMMU_PTE_PMASK) >> (PAGE_SHIFT-4);
}

static struct page *srmmu_pmd_page(pmd_t pmd)
{

	if (srmmu_device_memory(pmd_val(pmd)))
		BUG();
	return pfn_to_page((pmd_val(pmd) & SRMMU_PTD_PMASK) >> (PAGE_SHIFT-4));
}

static inline unsigned long srmmu_pgd_page(pgd_t pgd)
{ return srmmu_device_memory(pgd_val(pgd))?~0:(unsigned long)__nocache_va((pgd_val(pgd) & SRMMU_PTD_PMASK) << 4); }


static inline int srmmu_pte_none(pte_t pte)
{ return !(pte_val(pte) & 0xFFFFFFF); }

static inline int srmmu_pte_present(pte_t pte)
{ return ((pte_val(pte) & SRMMU_ET_MASK) == SRMMU_ET_PTE); }

static inline int srmmu_pte_read(pte_t pte)
{ return !(pte_val(pte) & SRMMU_NOREAD); }

static inline void srmmu_pte_clear(pte_t *ptep)
{ srmmu_set_pte(ptep, __pte(0)); }

static inline int srmmu_pmd_none(pmd_t pmd)
{ return !(pmd_val(pmd) & 0xFFFFFFF); }

static inline int srmmu_pmd_bad(pmd_t pmd)
{ return (pmd_val(pmd) & SRMMU_ET_MASK) != SRMMU_ET_PTD; }

static inline int srmmu_pmd_present(pmd_t pmd)
{ return ((pmd_val(pmd) & SRMMU_ET_MASK) == SRMMU_ET_PTD); }

static inline void srmmu_pmd_clear(pmd_t *pmdp) {
	int i;
	for (i = 0; i < PTRS_PER_PTE/SRMMU_REAL_PTRS_PER_PTE; i++)
		srmmu_set_pte((pte_t *)&pmdp->pmdv[i], __pte(0));
}

static inline int srmmu_pgd_none(pgd_t pgd)          
{ return !(pgd_val(pgd) & 0xFFFFFFF); }

static inline int srmmu_pgd_bad(pgd_t pgd)
{ return (pgd_val(pgd) & SRMMU_ET_MASK) != SRMMU_ET_PTD; }

static inline int srmmu_pgd_present(pgd_t pgd)
{ return ((pgd_val(pgd) & SRMMU_ET_MASK) == SRMMU_ET_PTD); }

static inline void srmmu_pgd_clear(pgd_t * pgdp)
{ srmmu_set_pte((pte_t *)pgdp, __pte(0)); }

static inline pte_t srmmu_pte_wrprotect(pte_t pte)
{ return __pte(pte_val(pte) & ~SRMMU_WRITE);}

static inline pte_t srmmu_pte_mkclean(pte_t pte)
{ return __pte(pte_val(pte) & ~SRMMU_DIRTY);}

static inline pte_t srmmu_pte_mkold(pte_t pte)
{ return __pte(pte_val(pte) & ~SRMMU_REF);}

static inline pte_t srmmu_pte_mkwrite(pte_t pte)
{ return __pte(pte_val(pte) | SRMMU_WRITE);}

static inline pte_t srmmu_pte_mkdirty(pte_t pte)
{ return __pte(pte_val(pte) | SRMMU_DIRTY);}

static inline pte_t srmmu_pte_mkyoung(pte_t pte)
{ return __pte(pte_val(pte) | SRMMU_REF);}

/*
 * Conversion functions: convert a page and protection to a page entry,
 * and a page entry and page directory to the page they refer to.
 */
static pte_t srmmu_mk_pte(struct page *page, pgprot_t pgprot)
{ return __pte((page_to_pfn(page) << (PAGE_SHIFT-4)) | pgprot_val(pgprot)); }

static pte_t srmmu_mk_pte_phys(unsigned long page, pgprot_t pgprot)
{ return __pte(((page) >> 4) | pgprot_val(pgprot)); }

static pte_t srmmu_mk_pte_io(unsigned long page, pgprot_t pgprot, int space)
{ return __pte(((page) >> 4) | (space << 28) | pgprot_val(pgprot)); }

/* XXX should we hyper_flush_whole_icache here - Anton */
static inline void srmmu_ctxd_set(ctxd_t *ctxp, pgd_t *pgdp)
{ srmmu_set_pte((pte_t *)ctxp, (SRMMU_ET_PTD | (__nocache_pa((unsigned long) pgdp) >> 4))); }

static inline void srmmu_pgd_set(pgd_t * pgdp, pmd_t * pmdp)
{ srmmu_set_pte((pte_t *)pgdp, (SRMMU_ET_PTD | (__nocache_pa((unsigned long) pmdp) >> 4))); }

static void srmmu_pmd_set(pmd_t *pmdp, pte_t *ptep)
{
	unsigned long ptp;	/* Physical address, shifted right by 4 */
	int i;

	ptp = __nocache_pa((unsigned long) ptep) >> 4;
	for (i = 0; i < PTRS_PER_PTE/SRMMU_REAL_PTRS_PER_PTE; i++) {
		srmmu_set_pte((pte_t *)&pmdp->pmdv[i], SRMMU_ET_PTD | ptp);
		ptp += (SRMMU_REAL_PTRS_PER_PTE*sizeof(pte_t) >> 4);
	}
}

static void srmmu_pmd_populate(pmd_t *pmdp, struct page *ptep)
{
	unsigned long ptp;	/* Physical address, shifted right by 4 */
	int i;

	ptp = page_to_pfn(ptep) << (PAGE_SHIFT-4);	/* watch for overflow */
	for (i = 0; i < PTRS_PER_PTE/SRMMU_REAL_PTRS_PER_PTE; i++) {
		srmmu_set_pte((pte_t *)&pmdp->pmdv[i], SRMMU_ET_PTD | ptp);
		ptp += (SRMMU_REAL_PTRS_PER_PTE*sizeof(pte_t) >> 4);
	}
}

static inline pte_t srmmu_pte_modify(pte_t pte, pgprot_t newprot)
{ return __pte((pte_val(pte) & SRMMU_CHG_MASK) | pgprot_val(newprot)); }

/* to find an entry in a top-level page table... */
static inline pgd_t *srmmu_pgd_offset(struct mm_struct * mm, unsigned long address)
{ return mm->pgd + (address >> SRMMU_PGDIR_SHIFT); }

/* Find an entry in the second-level page table.. */
static inline pmd_t *srmmu_pmd_offset(pgd_t * dir, unsigned long address)
{
	return (pmd_t *) srmmu_pgd_page(*dir) +
	    ((address >> PMD_SHIFT) & (PTRS_PER_PMD - 1));
}

/* Find an entry in the third-level page table.. */ 
static inline pte_t *srmmu_pte_offset(pmd_t * dir, unsigned long address)
{
	void *pte;

	pte = __nocache_va((dir->pmdv[0] & SRMMU_PTD_PMASK) << 4);
	return (pte_t *) pte +
	    ((address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1));
}

static unsigned long srmmu_swp_type(swp_entry_t entry)
{
	return (entry.val >> SRMMU_SWP_TYPE_SHIFT) & SRMMU_SWP_TYPE_MASK;
}

static unsigned long srmmu_swp_offset(swp_entry_t entry)
{
	return (entry.val >> SRMMU_SWP_OFF_SHIFT) & SRMMU_SWP_OFF_MASK;
}

static swp_entry_t srmmu_swp_entry(unsigned long type, unsigned long offset)
{
	return (swp_entry_t) {
		  (type & SRMMU_SWP_TYPE_MASK) << SRMMU_SWP_TYPE_SHIFT
		| (offset & SRMMU_SWP_OFF_MASK) << SRMMU_SWP_OFF_SHIFT };
}

/*
 * size: bytes to allocate in the nocache area.
 * align: bytes, number to align at.
 * Returns the virtual address of the allocated area.
 */
static unsigned long __srmmu_get_nocache(int size, int align)
{
	int offset;

	if (size < SRMMU_NOCACHE_BITMAP_SHIFT) {
		printk("Size 0x%x too small for nocache request\n", size);
		size = SRMMU_NOCACHE_BITMAP_SHIFT;
	}
	if (size & (SRMMU_NOCACHE_BITMAP_SHIFT-1)) {
		printk("Size 0x%x unaligned int nocache request\n", size);
		size += SRMMU_NOCACHE_BITMAP_SHIFT-1;
	}
	BUG_ON(align > SRMMU_NOCACHE_ALIGN_MAX);

	offset = bit_map_string_get(&srmmu_nocache_map,
		       			size >> SRMMU_NOCACHE_BITMAP_SHIFT,
					align >> SRMMU_NOCACHE_BITMAP_SHIFT);
	if (offset == -1) {
		printk("srmmu: out of nocache %d: %d/%d\n",
		    size, (int) srmmu_nocache_size,
		    srmmu_nocache_map.used << SRMMU_NOCACHE_BITMAP_SHIFT);
		return 0;
	}

	return (SRMMU_NOCACHE_VADDR + (offset << SRMMU_NOCACHE_BITMAP_SHIFT));
}

unsigned inline long srmmu_get_nocache(int size, int align)
{
	unsigned long tmp;

	tmp = __srmmu_get_nocache(size, align);

	if (tmp)
		memset((void *)tmp, 0, size);

	return tmp;
}

void srmmu_free_nocache(unsigned long vaddr, int size)
{
	int offset;

	if (vaddr < SRMMU_NOCACHE_VADDR) {
		printk("Vaddr %lx is smaller than nocache base 0x%lx\n",
		    vaddr, (unsigned long)SRMMU_NOCACHE_VADDR);
		BUG();
	}
	if (vaddr+size > srmmu_nocache_end) {
		printk("Vaddr %lx is bigger than nocache end 0x%lx\n",
		    vaddr, srmmu_nocache_end);
		BUG();
	}
	if (size & (size-1)) {
		printk("Size 0x%x is not a power of 2\n", size);
		BUG();
	}
	if (size < SRMMU_NOCACHE_BITMAP_SHIFT) {
		printk("Size 0x%x is too small\n", size);
		BUG();
	}
	if (vaddr & (size-1)) {
		printk("Vaddr %lx is not aligned to size 0x%x\n", vaddr, size);
		BUG();
	}

	offset = (vaddr - SRMMU_NOCACHE_VADDR) >> SRMMU_NOCACHE_BITMAP_SHIFT;
	size = size >> SRMMU_NOCACHE_BITMAP_SHIFT;

	bit_map_clear(&srmmu_nocache_map, offset, size);
}

void srmmu_early_allocate_ptable_skeleton(unsigned long start, unsigned long end);

extern unsigned long probe_memory(void);	/* in fault.c */

/*
 * Reserve nocache dynamically proportionally to the amount of
 * system RAM. -- Tomas Szepe <szepe@pinerecords.com>, June 2002
 */
void srmmu_nocache_calcsize(void)
{
	unsigned long sysmemavail = probe_memory() / 1024;
	int srmmu_nocache_npages;

	srmmu_nocache_npages =
		sysmemavail / SRMMU_NOCACHE_ALCRATIO / 1024 * 256;

 /* P3 XXX The 4x overuse: corroborated by /proc/meminfo. */
	// if (srmmu_nocache_npages < 256) srmmu_nocache_npages = 256;
	if (srmmu_nocache_npages < SRMMU_MIN_NOCACHE_PAGES)
		srmmu_nocache_npages = SRMMU_MIN_NOCACHE_PAGES;

	/* anything above 1280 blows up */
	if (srmmu_nocache_npages > SRMMU_MAX_NOCACHE_PAGES)
		srmmu_nocache_npages = SRMMU_MAX_NOCACHE_PAGES;

	srmmu_nocache_size = srmmu_nocache_npages * PAGE_SIZE;
	srmmu_nocache_end = SRMMU_NOCACHE_VADDR + srmmu_nocache_size;
}

void srmmu_nocache_init(void)
{
	unsigned int bitmap_bits;
	pgd_t *pgd;
	pmd_t *pmd;
	pte_t *pte;
	unsigned long paddr, vaddr;
	unsigned long pteval;

	bitmap_bits = srmmu_nocache_size >> SRMMU_NOCACHE_BITMAP_SHIFT;

	srmmu_nocache_pool = __alloc_bootmem(srmmu_nocache_size,
		SRMMU_NOCACHE_ALIGN_MAX, 0UL);
	memset(srmmu_nocache_pool, 0, srmmu_nocache_size);

	srmmu_nocache_bitmap = __alloc_bootmem(bitmap_bits >> 3, SMP_CACHE_BYTES, 0UL);
	bit_map_init(&srmmu_nocache_map, srmmu_nocache_bitmap, bitmap_bits);

	srmmu_swapper_pg_dir = (pgd_t *)__srmmu_get_nocache(SRMMU_PGD_TABLE_SIZE, SRMMU_PGD_TABLE_SIZE);
	memset(__nocache_fix(srmmu_swapper_pg_dir), 0, SRMMU_PGD_TABLE_SIZE);
	init_mm.pgd = srmmu_swapper_pg_dir;

	srmmu_early_allocate_ptable_skeleton(SRMMU_NOCACHE_VADDR, srmmu_nocache_end);

	paddr = __pa((unsigned long)srmmu_nocache_pool);
	vaddr = SRMMU_NOCACHE_VADDR;

	while (vaddr < srmmu_nocache_end) {
		pgd = pgd_offset_k(vaddr);
		pmd = srmmu_pmd_offset(__nocache_fix(pgd), vaddr);
		pte = srmmu_pte_offset(__nocache_fix(pmd), vaddr);

		pteval = ((paddr >> 4) | SRMMU_ET_PTE | SRMMU_PRIV);

		if (srmmu_cache_pagetables)
			pteval |= SRMMU_CACHE;

		srmmu_set_pte(__nocache_fix(pte), __pte(pteval));

		vaddr += PAGE_SIZE;
		paddr += PAGE_SIZE;
	}

	flush_cache_all();
	flush_tlb_all();
}

static inline pgd_t *srmmu_get_pgd_fast(void)
{
	pgd_t *pgd = NULL;

	pgd = (pgd_t *)__srmmu_get_nocache(SRMMU_PGD_TABLE_SIZE, SRMMU_PGD_TABLE_SIZE);
	if (pgd) {
		pgd_t *init = pgd_offset_k(0);
		memset(pgd, 0, USER_PTRS_PER_PGD * sizeof(pgd_t));
		memcpy(pgd + USER_PTRS_PER_PGD, init + USER_PTRS_PER_PGD,
						(PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t));
	}

	return pgd;
}

static void srmmu_free_pgd_fast(pgd_t *pgd)
{
	srmmu_free_nocache((unsigned long)pgd, SRMMU_PGD_TABLE_SIZE);
}

static pmd_t *srmmu_pmd_alloc_one(struct mm_struct *mm, unsigned long address)
{
	return (pmd_t *)srmmu_get_nocache(SRMMU_PMD_TABLE_SIZE, SRMMU_PMD_TABLE_SIZE);
}

static void srmmu_pmd_free(pmd_t * pmd)
{
	srmmu_free_nocache((unsigned long)pmd, SRMMU_PMD_TABLE_SIZE);
}

/*
 * Hardware needs alignment to 256 only, but we align to whole page size
 * to reduce fragmentation problems due to the buddy principle.
 * XXX Provide actual fragmentation statistics in /proc.
 *
 * Alignments up to the page size are the same for physical and virtual
 * addresses of the nocache area.
 */
static pte_t *
srmmu_pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
{
	return (pte_t *)srmmu_get_nocache(PTE_SIZE, PTE_SIZE);
}

static struct page *
srmmu_pte_alloc_one(struct mm_struct *mm, unsigned long address)
{
	unsigned long pte;

	if ((pte = (unsigned long)srmmu_pte_alloc_one_kernel(mm, address)) == 0)
		return NULL;
	return pfn_to_page( __nocache_pa(pte) >> PAGE_SHIFT );
}

static void srmmu_free_pte_fast(pte_t *pte)
{
	srmmu_free_nocache((unsigned long)pte, PTE_SIZE);
}

static void srmmu_pte_free(struct page *pte)
{
	unsigned long p;

	p = (unsigned long)page_address(pte);	/* Cached address (for test) */
	if (p == 0)
		BUG();
	p = page_to_pfn(pte) << PAGE_SHIFT;	/* Physical address */
	p = (unsigned long) __nocache_va(p);	/* Nocached virtual */
	srmmu_free_nocache(p, PTE_SIZE);
}

/*
 */
static inline void alloc_context(struct mm_struct *old_mm, struct mm_struct *mm)
{
	struct ctx_list *ctxp;

	ctxp = ctx_free.next;
	if(ctxp != &ctx_free) {
		remove_from_ctx_list(ctxp);
		add_to_used_ctxlist(ctxp);
		mm->context = ctxp->ctx_number;
		ctxp->ctx_mm = mm;
		return;
	}
	ctxp = ctx_used.next;
	if(ctxp->ctx_mm == old_mm)
		ctxp = ctxp->next;
	if(ctxp == &ctx_used)
		panic("out of mmu contexts");
	flush_cache_mm(ctxp->ctx_mm);
	flush_tlb_mm(ctxp->ctx_mm);
	remove_from_ctx_list(ctxp);
	add_to_used_ctxlist(ctxp);
	ctxp->ctx_mm->context = NO_CONTEXT;
	ctxp->ctx_mm = mm;
	mm->context = ctxp->ctx_number;
}

static inline void free_context(int context)
{
	struct ctx_list *ctx_old;

	ctx_old = ctx_list_pool + context;
	remove_from_ctx_list(ctx_old);
	add_to_free_ctxlist(ctx_old);
}


static void srmmu_switch_mm(struct mm_struct *old_mm, struct mm_struct *mm,
    struct task_struct *tsk, int cpu)
{
	if(mm->context == NO_CONTEXT) {
		spin_lock(&srmmu_context_spinlock);
		alloc_context(old_mm, mm);
		spin_unlock(&srmmu_context_spinlock);
		srmmu_ctxd_set(&srmmu_context_table[mm->context], mm->pgd);
	}

	if (is_hypersparc)
		hyper_flush_whole_icache();

	srmmu_set_context(mm->context);
}

/* Low level IO area allocation on the SRMMU. */
static inline void srmmu_mapioaddr(unsigned long physaddr,
    unsigned long virt_addr, int bus_type)
{
	pgd_t *pgdp;
	pmd_t *pmdp;
	pte_t *ptep;
	unsigned long tmp;

	physaddr &= PAGE_MASK;
	pgdp = pgd_offset_k(virt_addr);
	pmdp = srmmu_pmd_offset(pgdp, virt_addr);
	ptep = srmmu_pte_offset(pmdp, virt_addr);
	tmp = (physaddr >> 4) | SRMMU_ET_PTE;

	/*
	 * I need to test whether this is consistent over all
	 * sun4m's.  The bus_type represents the upper 4 bits of
	 * 36-bit physical address on the I/O space lines...
	 */
	tmp |= (bus_type << 28);
	tmp |= SRMMU_PRIV;
	__flush_page_to_ram(virt_addr);
	srmmu_set_pte(ptep, __pte(tmp));
}

static void srmmu_mapiorange(unsigned int bus, unsigned long xpa,
    unsigned long xva, unsigned int len)
{
	while (len != 0) {
		len -= PAGE_SIZE;
		srmmu_mapioaddr(xpa, xva, bus);
		xva += PAGE_SIZE;
		xpa += PAGE_SIZE;
	}
	flush_tlb_all();
}

static inline void srmmu_unmapioaddr(unsigned long virt_addr)
{
	pgd_t *pgdp;
	pmd_t *pmdp;
	pte_t *ptep;

	pgdp = pgd_offset_k(virt_addr);
	pmdp = srmmu_pmd_offset(pgdp, virt_addr);
	ptep = srmmu_pte_offset(pmdp, virt_addr);

	/* No need to flush uncacheable page. */
	srmmu_pte_clear(ptep);
}

static void srmmu_unmapiorange(unsigned long virt_addr, unsigned int len)
{
	while (len != 0) {
		len -= PAGE_SIZE;
		srmmu_unmapioaddr(virt_addr);
		virt_addr += PAGE_SIZE;
	}
	flush_tlb_all();
}

/*
 * On the SRMMU we do not have the problems with limited tlb entries
 * for mapping kernel pages, so we just take things from the free page
 * pool.  As a side effect we are putting a little too much pressure
 * on the gfp() subsystem.  This setup also makes the logic of the
 * iommu mapping code a lot easier as we can transparently handle
 * mappings on the kernel stack without any special code as we did
 * need on the sun4c.
 */
struct thread_info *srmmu_alloc_thread_info(void)
{
	struct thread_info *ret;

	ret = (struct thread_info *)__get_free_pages(GFP_KERNEL,
						     THREAD_INFO_ORDER);
#ifdef CONFIG_DEBUG_STACK_USAGE
	if (ret)
		memset(ret, 0, PAGE_SIZE << THREAD_INFO_ORDER);
#endif /* DEBUG_STACK_USAGE */

	return ret;
}

static void srmmu_free_thread_info(struct thread_info *ti)
{
	free_pages((unsigned long)ti, THREAD_INFO_ORDER);
}

/* tsunami.S */
extern void tsunami_flush_cache_all(void);
extern void tsunami_flush_cache_mm(struct mm_struct *mm);
extern void tsunami_flush_cache_range(struct vm_area_struct *vma, unsigned long start, unsigned long end);
extern void tsunami_flush_cache_page(struct vm_area_struct *vma, unsigned long page);
extern void tsunami_flush_page_to_ram(unsigned long page);
extern void tsunami_flush_page_for_dma(unsigned long page);
extern void tsunami_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr);
extern void tsunami_flush_tlb_all(void);
extern void tsunami_flush_tlb_mm(struct mm_struct *mm);
extern void tsunami_flush_tlb_range(struct vm_area_struct *vma, unsigned long start, unsigned long end);
extern void tsunami_flush_tlb_page(struct vm_area_struct *vma, unsigned long page);
extern void tsunami_setup_blockops(void);

/*
 * Workaround, until we find what's going on with Swift. When low on memory,
 * it sometimes loops in fault/handle_mm_fault incl. flush_tlb_page to find
 * out it is already in page tables/ fault again on the same instruction.
 * I really don't understand it, have checked it and contexts
 * are right, flush_tlb_all is done as well, and it faults again...
 * Strange. -jj
 *
 * The following code is a deadwood that may be necessary when
 * we start to make precise page flushes again. --zaitcev
 */
static void swift_update_mmu_cache(struct vm_area_struct * vma, unsigned long address, pte_t pte)
{
#if 0
	static unsigned long last;
	unsigned int val;
	/* unsigned int n; */

	if (address == last) {
		val = srmmu_hwprobe(address);
		if (val != 0 && pte_val(pte) != val) {
			printk("swift_update_mmu_cache: "
			    "addr %lx put %08x probed %08x from %p\n",
			    address, pte_val(pte), val,
			    __builtin_return_address(0));
			srmmu_flush_whole_tlb();
		}
	}
	last = address;
#endif
}

/* swift.S */
extern void swift_flush_cache_all(void);
extern void swift_flush_cache_mm(struct mm_struct *mm);
extern void swift_flush_cache_range(struct vm_area_struct *vma,
				    unsigned long start, unsigned long end);
extern void swift_flush_cache_page(struct vm_area_struct *vma, unsigned long page);
extern void swift_flush_page_to_ram(unsigned long page);
extern void swift_flush_page_for_dma(unsigned long page);
extern void swift_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr);
extern void swift_flush_tlb_all(void);
extern void swift_flush_tlb_mm(struct mm_struct *mm);
extern void swift_flush_tlb_range(struct vm_area_struct *vma,
				  unsigned long start, unsigned long end);
extern void swift_flush_tlb_page(struct vm_area_struct *vma, unsigned long page);

#if 0  /* P3: deadwood to debug precise flushes on Swift. */
void swift_flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
{
	int cctx, ctx1;

	page &= PAGE_MASK;
	if ((ctx1 = vma->vm_mm->context) != -1) {
		cctx = srmmu_get_context();
/* Is context # ever different from current context? P3 */
		if (cctx != ctx1) {
			printk("flush ctx %02x curr %02x\n", ctx1, cctx);
			srmmu_set_context(ctx1);
			swift_flush_page(page);
			__asm__ __volatile__("sta %%g0, [%0] %1\n\t" : :
					"r" (page), "i" (ASI_M_FLUSH_PROBE));
			srmmu_set_context(cctx);
		} else {
			 /* Rm. prot. bits from virt. c. */
			/* swift_flush_cache_all(); */
			/* swift_flush_cache_page(vma, page); */
			swift_flush_page(page);

			__asm__ __volatile__("sta %%g0, [%0] %1\n\t" : :
				"r" (page), "i" (ASI_M_FLUSH_PROBE));
			/* same as above: srmmu_flush_tlb_page() */
		}
	}
}
#endif

/*
 * The following are all MBUS based SRMMU modules, and therefore could
 * be found in a multiprocessor configuration.  On the whole, these
 * chips seems to be much more touchy about DVMA and page tables
 * with respect to cache coherency.
 */

/* Cypress flushes. */
static void cypress_flush_cache_all(void)
{
	volatile unsigned long cypress_sucks;
	unsigned long faddr, tagval;

	flush_user_windows();
	for(faddr = 0; faddr < 0x10000; faddr += 0x20) {
		__asm__ __volatile__("lda [%1 + %2] %3, %0\n\t" :
				     "=r" (tagval) :
				     "r" (faddr), "r" (0x40000),
				     "i" (ASI_M_DATAC_TAG));

		/* If modified and valid, kick it. */
		if((tagval & 0x60) == 0x60)
			cypress_sucks = *(unsigned long *)(0xf0020000 + faddr);
	}
}

static void cypress_flush_cache_mm(struct mm_struct *mm)
{
	register unsigned long a, b, c, d, e, f, g;
	unsigned long flags, faddr;
	int octx;

	FLUSH_BEGIN(mm)
	flush_user_windows();
	local_irq_save(flags);
	octx = srmmu_get_context();
	srmmu_set_context(mm->context);
	a = 0x20; b = 0x40; c = 0x60;
	d = 0x80; e = 0xa0; f = 0xc0; g = 0xe0;

	faddr = (0x10000 - 0x100);
	goto inside;
	do {
		faddr -= 0x100;
	inside:
		__asm__ __volatile__("sta %%g0, [%0] %1\n\t"
				     "sta %%g0, [%0 + %2] %1\n\t"
				     "sta %%g0, [%0 + %3] %1\n\t"
				     "sta %%g0, [%0 + %4] %1\n\t"
				     "sta %%g0, [%0 + %5] %1\n\t"
				     "sta %%g0, [%0 + %6] %1\n\t"
				     "sta %%g0, [%0 + %7] %1\n\t"
				     "sta %%g0, [%0 + %8] %1\n\t" : :
				     "r" (faddr), "i" (ASI_M_FLUSH_CTX),
				     "r" (a), "r" (b), "r" (c), "r" (d),
				     "r" (e), "r" (f), "r" (g));
	} while(faddr);
	srmmu_set_context(octx);
	local_irq_restore(flags);
	FLUSH_END
}

static void cypress_flush_cache_range(struct vm_area_struct *vma, unsigned long start, unsigned long end)
{
	struct mm_struct *mm = vma->vm_mm;
	register unsigned long a, b, c, d, e, f, g;
	unsigned long flags, faddr;
	int octx;

	FLUSH_BEGIN(mm)
	flush_user_windows();
	local_irq_save(flags);
	octx = srmmu_get_context();
	srmmu_set_context(mm->context);
	a = 0x20; b = 0x40; c = 0x60;
	d = 0x80; e = 0xa0; f = 0xc0; g = 0xe0;

	start &= SRMMU_REAL_PMD_MASK;
	while(start < end) {
		faddr = (start + (0x10000 - 0x100));
		goto inside;
		do {
			faddr -= 0x100;
		inside:
			__asm__ __volatile__("sta %%g0, [%0] %1\n\t"
					     "sta %%g0, [%0 + %2] %1\n\t"
					     "sta %%g0, [%0 + %3] %1\n\t"
					     "sta %%g0, [%0 + %4] %1\n\t"
					     "sta %%g0, [%0 + %5] %1\n\t"
					     "sta %%g0, [%0 + %6] %1\n\t"
					     "sta %%g0, [%0 + %7] %1\n\t"
					     "sta %%g0, [%0 + %8] %1\n\t" : :
					     "r" (faddr),
					     "i" (ASI_M_FLUSH_SEG),
					     "r" (a), "r" (b), "r" (c), "r" (d),
					     "r" (e), "r" (f), "r" (g));
		} while (faddr != start);
		start += SRMMU_REAL_PMD_SIZE;
	}
	srmmu_set_context(octx);
	local_irq_restore(flags);
	FLUSH_END
}

static void cypress_flush_cache_page(struct vm_area_struct *vma, unsigned long page)
{
	register unsigned long a, b, c, d, e, f, g;
	struct mm_struct *mm = vma->vm_mm;
	unsigned long flags, line;
	int octx;

	FLUSH_BEGIN(mm)
	flush_user_windows();
	local_irq_save(flags);
	octx = srmmu_get_context();
	srmmu_set_context(mm->context);
	a = 0x20; b = 0x40; c = 0x60;
	d = 0x80; e = 0xa0; f = 0xc0; g = 0xe0;

	page &= PAGE_MASK;
	line = (page + PAGE_SIZE) - 0x100;
	goto inside;
	do {
		line -= 0x100;
	inside:
			__asm__ __volatile__("sta %%g0, [%0] %1\n\t"
					     "sta %%g0, [%0 + %2] %1\n\t"
					     "sta %%g0, [%0 + %3] %1\n\t"
					     "sta %%g0, [%0 + %4] %1\n\t"
					     "sta %%g0, [%0 + %5] %1\n\t"
					     "sta %%g0, [%0 + %6] %1\n\t"
					     "sta %%g0, [%0 + %7] %1\n\t"
					     "sta %%g0, [%0 + %8] %1\n\t" : :
					     "r" (line),
					     "i" (ASI_M_FLUSH_PAGE),
					     "r" (a), "r" (b), "r" (c), "r" (d),
					     "r" (e), "r" (f), "r" (g));
	} while(line != page);
	srmmu_set_context(octx);
	local_irq_restore(flags);
	FLUSH_END
}

/* Cypress is copy-back, at least that is how we configure it. */
static void cypress_flush_page_to_ram(unsigned long page)
{
	register unsigned long a, b, c, d, e, f, g;
	unsigned long line;

	a = 0x20; b = 0x40; c = 0x60; d = 0x80; e = 0xa0; f = 0xc0; g = 0xe0;
	page &= PAGE_MASK;
	line = (page + PAGE_SIZE) - 0x100;
	goto inside;
	do {
		line -= 0x100;
	inside:
		__asm__ __volatile__("sta %%g0, [%0] %1\n\t"
				     "sta %%g0, [%0 + %2] %1\n\t"
				     "sta %%g0, [%0 + %3] %1\n\t"
				     "sta %%g0, [%0 + %4] %1\n\t"
				     "sta %%g0, [%0 + %5] %1\n\t"
				     "sta %%g0, [%0 + %6] %1\n\t"
				     "sta %%g0, [%0 + %7] %1\n\t"
				     "sta %%g0, [%0 + %8] %1\n\t" : :
				     "r" (line),
				     "i" (ASI_M_FLUSH_PAGE),
				     "r" (a), "r" (b), "r" (c), "r" (d),
				     "r" (e), "r" (f), "r" (g));
	} while(line != page);
}

/* Cypress is also IO cache coherent. */
static void cypress_flush_page_for_dma(unsigned long page)
{
}

/* Cypress has unified L2 VIPT, from which both instructions and data
 * are stored.  It does not have an onboard icache of any sort, therefore
 * no flush is necessary.
 */
static void cypress_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr)
{
}

static void cypress_flush_tlb_all(void)
{
	srmmu_flush_whole_tlb();
}

static void cypress_flush_tlb_mm(struct mm_struct *mm)
{
	FLUSH_BEGIN(mm)
	__asm__ __volatile__(
	"lda	[%0] %3, %%g5\n\t"
	"sta	%2, [%0] %3\n\t"
	"sta	%%g0, [%1] %4\n\t"
	"sta	%%g5, [%0] %3\n"
	: /* no outputs */
	: "r" (SRMMU_CTX_REG), "r" (0x300), "r" (mm->context),
	  "i" (ASI_M_MMUREGS), "i" (ASI_M_FLUSH_PROBE)
	: "g5");
	FLUSH_END
}

static void cypress_flush_tlb_range(struct vm_area_struct *vma, unsigned long start, unsigned long end)
{
	struct mm_struct *mm = vma->vm_mm;
	unsigned long size;

	FLUSH_BEGIN(mm)
	start &= SRMMU_PGDIR_MASK;
	size = SRMMU_PGDIR_ALIGN(end) - start;
	__asm__ __volatile__(
		"lda	[%0] %5, %%g5\n\t"
		"sta	%1, [%0] %5\n"
		"1:\n\t"
		"subcc	%3, %4, %3\n\t"
		"bne	1b\n\t"
		" sta	%%g0, [%2 + %3] %6\n\t"
		"sta	%%g5, [%0] %5\n"
	: /* no outputs */
	: "r" (SRMMU_CTX_REG), "r" (mm->context), "r" (start | 0x200),
	  "r" (size), "r" (SRMMU_PGDIR_SIZE), "i" (ASI_M_MMUREGS),
	  "i" (ASI_M_FLUSH_PROBE)
	: "g5", "cc");
	FLUSH_END
}

static void cypress_flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
{
	struct mm_struct *mm = vma->vm_mm;

	FLUSH_BEGIN(mm)
	__asm__ __volatile__(
	"lda	[%0] %3, %%g5\n\t"
	"sta	%1, [%0] %3\n\t"
	"sta	%%g0, [%2] %4\n\t"
	"sta	%%g5, [%0] %3\n"
	: /* no outputs */
	: "r" (SRMMU_CTX_REG), "r" (mm->context), "r" (page & PAGE_MASK),
	  "i" (ASI_M_MMUREGS), "i" (ASI_M_FLUSH_PROBE)
	: "g5");
	FLUSH_END
}

/* viking.S */
extern void viking_flush_cache_all(void);
extern void viking_flush_cache_mm(struct mm_struct *mm);
extern void viking_flush_cache_range(struct vm_area_struct *vma, unsigned long start,
				     unsigned long end);
extern void viking_flush_cache_page(struct vm_area_struct *vma, unsigned long page);
extern void viking_flush_page_to_ram(unsigned long page);
extern void viking_flush_page_for_dma(unsigned long page);
extern void viking_flush_sig_insns(struct mm_struct *mm, unsigned long addr);
extern void viking_flush_page(unsigned long page);
extern void viking_mxcc_flush_page(unsigned long page);
extern void viking_flush_tlb_all(void);
extern void viking_flush_tlb_mm(struct mm_struct *mm);
extern void viking_flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
				   unsigned long end);
extern void viking_flush_tlb_page(struct vm_area_struct *vma,
				  unsigned long page);
extern void sun4dsmp_flush_tlb_all(void);
extern void sun4dsmp_flush_tlb_mm(struct mm_struct *mm);
extern void sun4dsmp_flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
				   unsigned long end);
extern void sun4dsmp_flush_tlb_page(struct vm_area_struct *vma,
				  unsigned long page);

/* hypersparc.S */
extern void hypersparc_flush_cache_all(void);
extern void hypersparc_flush_cache_mm(struct mm_struct *mm);
extern void hypersparc_flush_cache_range(struct vm_area_struct *vma, unsigned long start, unsigned long end);
extern void hypersparc_flush_cache_page(struct vm_area_struct *vma, unsigned long page);
extern void hypersparc_flush_page_to_ram(unsigned long page);
extern void hypersparc_flush_page_for_dma(unsigned long page);
extern void hypersparc_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr);
extern void hypersparc_flush_tlb_all(void);
extern void hypersparc_flush_tlb_mm(struct mm_struct *mm);
extern void hypersparc_flush_tlb_range(struct vm_area_struct *vma, unsigned long start, unsigned long end);
extern void hypersparc_flush_tlb_page(struct vm_area_struct *vma, unsigned long page);
extern void hypersparc_setup_blockops(void);

/*
 * NOTE: All of this startup code assumes the low 16mb (approx.) of
 *       kernel mappings are done with one single contiguous chunk of
 *       ram.  On small ram machines (classics mainly) we only get
 *       around 8mb mapped for us.
 */

void __init early_pgtable_allocfail(char *type)
{
	prom_printf("inherit_prom_mappings: Cannot alloc kernel %s.\n", type);
	prom_halt();
}

void __init srmmu_early_allocate_ptable_skeleton(unsigned long start, unsigned long end)
{
	pgd_t *pgdp;
	pmd_t *pmdp;
	pte_t *ptep;

	while(start < end) {
		pgdp = pgd_offset_k(start);
		if(srmmu_pgd_none(*(pgd_t *)__nocache_fix(pgdp))) {
			pmdp = (pmd_t *) __srmmu_get_nocache(
			    SRMMU_PMD_TABLE_SIZE, SRMMU_PMD_TABLE_SIZE);
			if (pmdp == NULL)
				early_pgtable_allocfail("pmd");
			memset(__nocache_fix(pmdp), 0, SRMMU_PMD_TABLE_SIZE);
			srmmu_pgd_set(__nocache_fix(pgdp), pmdp);
		}
		pmdp = srmmu_pmd_offset(__nocache_fix(pgdp), start);
		if(srmmu_pmd_none(*(pmd_t *)__nocache_fix(pmdp))) {
			ptep = (pte_t *)__srmmu_get_nocache(PTE_SIZE, PTE_SIZE);
			if (ptep == NULL)
				early_pgtable_allocfail("pte");
			memset(__nocache_fix(ptep), 0, PTE_SIZE);
			srmmu_pmd_set(__nocache_fix(pmdp), ptep);
		}
		if (start > (0xffffffffUL - PMD_SIZE))
			break;
		start = (start + PMD_SIZE) & PMD_MASK;
	}
}

void __init srmmu_allocate_ptable_skeleton(unsigned long start, unsigned long end)
{
	pgd_t *pgdp;
	pmd_t *pmdp;
	pte_t *ptep;

	while(start < end) {
		pgdp = pgd_offset_k(start);
		if(srmmu_pgd_none(*pgdp)) {
			pmdp = (pmd_t *)__srmmu_get_nocache(SRMMU_PMD_TABLE_SIZE, SRMMU_PMD_TABLE_SIZE);
			if (pmdp == NULL)
				early_pgtable_allocfail("pmd");
			memset(pmdp, 0, SRMMU_PMD_TABLE_SIZE);
			srmmu_pgd_set(pgdp, pmdp);
		}
		pmdp = srmmu_pmd_offset(pgdp, start);
		if(srmmu_pmd_none(*pmdp)) {
			ptep = (pte_t *) __srmmu_get_nocache(PTE_SIZE,
							     PTE_SIZE);
			if (ptep == NULL)
				early_pgtable_allocfail("pte");
			memset(ptep, 0, PTE_SIZE);
			srmmu_pmd_set(pmdp, ptep);
		}
		if (start > (0xffffffffUL - PMD_SIZE))
			break;
		start = (start + PMD_SIZE) & PMD_MASK;
	}
}

/*
 * This is much cleaner than poking around physical address space
 * looking at the prom's page table directly which is what most
 * other OS's do.  Yuck... this is much better.
 */
void __init srmmu_inherit_prom_mappings(unsigned long start,unsigned long end)
{
	pgd_t *pgdp;
	pmd_t *pmdp;
	pte_t *ptep;
	int what = 0; /* 0 = normal-pte, 1 = pmd-level pte, 2 = pgd-level pte */
	unsigned long prompte;

	while(start <= end) {
		if (start == 0)
			break; /* probably wrap around */
		if(start == 0xfef00000)
			start = KADB_DEBUGGER_BEGVM;
		if(!(prompte = srmmu_hwprobe(start))) {
			start += PAGE_SIZE;
			continue;
		}
    
		/* A red snapper, see what it really is. */
		what = 0;
    
		if(!(start & ~(SRMMU_REAL_PMD_MASK))) {
			if(srmmu_hwprobe((start-PAGE_SIZE) + SRMMU_REAL_PMD_SIZE) == prompte)
				what = 1;
		}
    
		if(!(start & ~(SRMMU_PGDIR_MASK))) {
			if(srmmu_hwprobe((start-PAGE_SIZE) + SRMMU_PGDIR_SIZE) ==
			   prompte)
				what = 2;
		}
    
		pgdp = pgd_offset_k(start);
		if(what == 2) {
			*(pgd_t *)__nocache_fix(pgdp) = __pgd(prompte);
			start += SRMMU_PGDIR_SIZE;
			continue;
		}
		if(srmmu_pgd_none(*(pgd_t *)__nocache_fix(pgdp))) {
			pmdp = (pmd_t *)__srmmu_get_nocache(SRMMU_PMD_TABLE_SIZE, SRMMU_PMD_TABLE_SIZE);
			if (pmdp == NULL)
				early_pgtable_allocfail("pmd");
			memset(__nocache_fix(pmdp), 0, SRMMU_PMD_TABLE_SIZE);
			srmmu_pgd_set(__nocache_fix(pgdp), pmdp);
		}
		pmdp = srmmu_pmd_offset(__nocache_fix(pgdp), start);
		if(srmmu_pmd_none(*(pmd_t *)__nocache_fix(pmdp))) {
			ptep = (pte_t *) __srmmu_get_nocache(PTE_SIZE,
							     PTE_SIZE);
			if (ptep == NULL)
				early_pgtable_allocfail("pte");
			memset(__nocache_fix(ptep), 0, PTE_SIZE);
			srmmu_pmd_set(__nocache_fix(pmdp), ptep);
		}
		if(what == 1) {
			/*
			 * We bend the rule where all 16 PTPs in a pmd_t point
			 * inside the same PTE page, and we leak a perfectly
			 * good hardware PTE piece. Alternatives seem worse.
			 */
			unsigned int x;	/* Index of HW PMD in soft cluster */
			x = (start >> PMD_SHIFT) & 15;
			*(unsigned long *)__nocache_fix(&pmdp->pmdv[x]) = prompte;
			start += SRMMU_REAL_PMD_SIZE;
			continue;
		}
		ptep = srmmu_pte_offset(__nocache_fix(pmdp), start);
		*(pte_t *)__nocache_fix(ptep) = __pte(prompte);
		start += PAGE_SIZE;
	}
}

#define KERNEL_PTE(page_shifted) ((page_shifted)|SRMMU_CACHE|SRMMU_PRIV|SRMMU_VALID)

/* Create a third-level SRMMU 16MB page mapping. */
static void __init do_large_mapping(unsigned long vaddr, unsigned long phys_base)
{
	pgd_t *pgdp = pgd_offset_k(vaddr);
	unsigned long big_pte;

	big_pte = KERNEL_PTE(phys_base >> 4);
	*(pgd_t *)__nocache_fix(pgdp) = __pgd(big_pte);
}

/* Map sp_bank entry SP_ENTRY, starting at virtual address VBASE. */
static unsigned long __init map_spbank(unsigned long vbase, int sp_entry)
{
	unsigned long pstart = (sp_banks[sp_entry].base_addr & SRMMU_PGDIR_MASK);
	unsigned long vstart = (vbase & SRMMU_PGDIR_MASK);
	unsigned long vend = SRMMU_PGDIR_ALIGN(vbase + sp_banks[sp_entry].num_bytes);
	/* Map "low" memory only */
	const unsigned long min_vaddr = PAGE_OFFSET;
	const unsigned long max_vaddr = PAGE_OFFSET + SRMMU_MAXMEM;

	if (vstart < min_vaddr || vstart >= max_vaddr)
		return vstart;
	
	if (vend > max_vaddr || vend < min_vaddr)
		vend = max_vaddr;

	while(vstart < vend) {
		do_large_mapping(vstart, pstart);
		vstart += SRMMU_PGDIR_SIZE; pstart += SRMMU_PGDIR_SIZE;
	}
	return vstart;
}

static inline void memprobe_error(char *msg)
{
	prom_printf(msg);
	prom_printf("Halting now...\n");
	prom_halt();
}

static inline void map_kernel(void)
{
	int i;

	if (phys_base > 0) {
		do_large_mapping(PAGE_OFFSET, phys_base);
	}

	for (i = 0; sp_banks[i].num_bytes != 0; i++) {
		map_spbank((unsigned long)__va(sp_banks[i].base_addr), i);
	}

	BTFIXUPSET_SIMM13(user_ptrs_per_pgd, PAGE_OFFSET / SRMMU_PGDIR_SIZE);
}

/* Paging initialization on the Sparc Reference MMU. */
extern void sparc_context_init(int);

void (*poke_srmmu)(void) __initdata = NULL;

extern unsigned long bootmem_init(unsigned long *pages_avail);

void __init srmmu_paging_init(void)
{
	int i, cpunode;
	char node_str[128];
	pgd_t *pgd;
	pmd_t *pmd;
	pte_t *pte;
	unsigned long pages_avail;

	sparc_iomap.start = SUN4M_IOBASE_VADDR;	/* 16MB of IOSPACE on all sun4m's. */

	if (sparc_cpu_model == sun4d)
		num_contexts = 65536; /* We know it is Viking */
	else {
		/* Find the number of contexts on the srmmu. */
		cpunode = prom_getchild(prom_root_node);
		num_contexts = 0;
		while(cpunode != 0) {
			prom_getstring(cpunode, "device_type", node_str, sizeof(node_str));
			if(!strcmp(node_str, "cpu")) {
				num_contexts = prom_getintdefault(cpunode, "mmu-nctx", 0x8);
				break;
			}
			cpunode = prom_getsibling(cpunode);
		}
	}

	if(!num_contexts) {
		prom_printf("Something wrong, can't find cpu node in paging_init.\n");
		prom_halt();
	}

	pages_avail = 0;
	last_valid_pfn = bootmem_init(&pages_avail);

	srmmu_nocache_calcsize();
	srmmu_nocache_init();
        srmmu_inherit_prom_mappings(0xfe400000,(LINUX_OPPROM_ENDVM-PAGE_SIZE));
	map_kernel();

	/* ctx table has to be physically aligned to its size */
	srmmu_context_table = (ctxd_t *)__srmmu_get_nocache(num_contexts*sizeof(ctxd_t), num_contexts*sizeof(ctxd_t));
	srmmu_ctx_table_phys = (ctxd_t *)__nocache_pa((unsigned long)srmmu_context_table);

	for(i = 0; i < num_contexts; i++)
		srmmu_ctxd_set((ctxd_t *)__nocache_fix(&srmmu_context_table[i]), srmmu_swapper_pg_dir);

	flush_cache_all();
	srmmu_set_ctable_ptr((unsigned long)srmmu_ctx_table_phys);
	flush_tlb_all();
	poke_srmmu();

#ifdef CONFIG_SUN_IO
	srmmu_allocate_ptable_skeleton(sparc_iomap.start, IOBASE_END);
	srmmu_allocate_ptable_skeleton(DVMA_VADDR, DVMA_END);
#endif

	srmmu_allocate_ptable_skeleton(
		__fix_to_virt(__end_of_fixed_addresses - 1), FIXADDR_TOP);
	srmmu_allocate_ptable_skeleton(PKMAP_BASE, PKMAP_END);

	pgd = pgd_offset_k(PKMAP_BASE);
	pmd = srmmu_pmd_offset(pgd, PKMAP_BASE);
	pte = srmmu_pte_offset(pmd, PKMAP_BASE);
	pkmap_page_table = pte;

	flush_cache_all();
	flush_tlb_all();

	sparc_context_init(num_contexts);

	kmap_init();

	{
		unsigned long zones_size[MAX_NR_ZONES];
		unsigned long zholes_size[MAX_NR_ZONES];
		unsigned long npages;
		int znum;

		for (znum = 0; znum < MAX_NR_ZONES; znum++)
			zones_size[znum] = zholes_size[znum] = 0;

		npages = max_low_pfn - pfn_base;

		zones_size[ZONE_DMA] = npages;
		zholes_size[ZONE_DMA] = npages - pages_avail;

		npages = highend_pfn - max_low_pfn;
		zones_size[ZONE_HIGHMEM] = npages;
		zholes_size[ZONE_HIGHMEM] = npages - calc_highpages();

		free_area_init_node(0, &contig_page_data, zones_size,
				    pfn_base, zholes_size);
	}
}

static void srmmu_mmu_info(struct seq_file *m)
{
	seq_printf(m, 
		   "MMU type\t: %s\n"
		   "contexts\t: %d\n"
		   "nocache total\t: %ld\n"
		   "nocache used\t: %d\n",
		   srmmu_name,
		   num_contexts,
		   srmmu_nocache_size,
		   srmmu_nocache_map.used << SRMMU_NOCACHE_BITMAP_SHIFT);
}

static void srmmu_update_mmu_cache(struct vm_area_struct * vma, unsigned long address, pte_t pte)
{
}

static void srmmu_destroy_context(struct mm_struct *mm)
{

	if(mm->context != NO_CONTEXT) {
		flush_cache_mm(mm);
		srmmu_ctxd_set(&srmmu_context_table[mm->context], srmmu_swapper_pg_dir);
		flush_tlb_mm(mm);
		spin_lock(&srmmu_context_spinlock);
		free_context(mm->context);
		spin_unlock(&srmmu_context_spinlock);
		mm->context = NO_CONTEXT;
	}
}

/* Init various srmmu chip types. */
static void __init srmmu_is_bad(void)
{
	prom_printf("Could not determine SRMMU chip type.\n");
	prom_halt();
}

static void __init init_vac_layout(void)
{
	int nd, cache_lines;
	char node_str[128];
#ifdef CONFIG_SMP
	int cpu = 0;
	unsigned long max_size = 0;
	unsigned long min_line_size = 0x10000000;
#endif

	nd = prom_getchild(prom_root_node);
	while((nd = prom_getsibling(nd)) != 0) {
		prom_getstring(nd, "device_type", node_str, sizeof(node_str));
		if(!strcmp(node_str, "cpu")) {
			vac_line_size = prom_getint(nd, "cache-line-size");
			if (vac_line_size == -1) {
				prom_printf("can't determine cache-line-size, "
					    "halting.\n");
				prom_halt();
			}
			cache_lines = prom_getint(nd, "cache-nlines");
			if (cache_lines == -1) {
				prom_printf("can't determine cache-nlines, halting.\n");
				prom_halt();
			}

			vac_cache_size = cache_lines * vac_line_size;
#ifdef CONFIG_SMP
			if(vac_cache_size > max_size)
				max_size = vac_cache_size;
			if(vac_line_size < min_line_size)
				min_line_size = vac_line_size;
			cpu++;
			if (cpu >= NR_CPUS || !cpu_online(cpu))
				break;
#else
			break;
#endif
		}
	}
	if(nd == 0) {
		prom_printf("No CPU nodes found, halting.\n");
		prom_halt();
	}
#ifdef CONFIG_SMP
	vac_cache_size = max_size;
	vac_line_size = min_line_size;
#endif
	printk("SRMMU: Using VAC size of %d bytes, line size %d bytes.\n",
	       (int)vac_cache_size, (int)vac_line_size);
}

static void __init poke_hypersparc(void)
{
	volatile unsigned long clear;
	unsigned long mreg = srmmu_get_mmureg();

	hyper_flush_unconditional_combined();

	mreg &= ~(HYPERSPARC_CWENABLE);
	mreg |= (HYPERSPARC_CENABLE | HYPERSPARC_WBENABLE);
	mreg |= (HYPERSPARC_CMODE);

	srmmu_set_mmureg(mreg);

#if 0 /* XXX I think this is bad news... -DaveM */
	hyper_clear_all_tags();
#endif

	put_ross_icr(HYPERSPARC_ICCR_FTD | HYPERSPARC_ICCR_ICE);
	hyper_flush_whole_icache();
	clear = srmmu_get_faddr();
	clear = srmmu_get_fstatus();
}

static void __init init_hypersparc(void)
{
	srmmu_name = "ROSS HyperSparc";
	srmmu_modtype = HyperSparc;

	init_vac_layout();

	is_hypersparc = 1;

	BTFIXUPSET_CALL(pte_clear, srmmu_pte_clear, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pmd_clear, srmmu_pmd_clear, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pgd_clear, srmmu_pgd_clear, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_all, hypersparc_flush_cache_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_mm, hypersparc_flush_cache_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_range, hypersparc_flush_cache_range, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_page, hypersparc_flush_cache_page, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(flush_tlb_all, hypersparc_flush_tlb_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_mm, hypersparc_flush_tlb_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_range, hypersparc_flush_tlb_range, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_page, hypersparc_flush_tlb_page, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(__flush_page_to_ram, hypersparc_flush_page_to_ram, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_sig_insns, hypersparc_flush_sig_insns, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_page_for_dma, hypersparc_flush_page_for_dma, BTFIXUPCALL_NOP);


	poke_srmmu = poke_hypersparc;

	hypersparc_setup_blockops();
}

static void __init poke_cypress(void)
{
	unsigned long mreg = srmmu_get_mmureg();
	unsigned long faddr, tagval;
	volatile unsigned long cypress_sucks;
	volatile unsigned long clear;

	clear = srmmu_get_faddr();
	clear = srmmu_get_fstatus();

	if (!(mreg & CYPRESS_CENABLE)) {
		for(faddr = 0x0; faddr < 0x10000; faddr += 20) {
			__asm__ __volatile__("sta %%g0, [%0 + %1] %2\n\t"
					     "sta %%g0, [%0] %2\n\t" : :
					     "r" (faddr), "r" (0x40000),
					     "i" (ASI_M_DATAC_TAG));
		}
	} else {
		for(faddr = 0; faddr < 0x10000; faddr += 0x20) {
			__asm__ __volatile__("lda [%1 + %2] %3, %0\n\t" :
					     "=r" (tagval) :
					     "r" (faddr), "r" (0x40000),
					     "i" (ASI_M_DATAC_TAG));

			/* If modified and valid, kick it. */
			if((tagval & 0x60) == 0x60)
				cypress_sucks = *(unsigned long *)
							(0xf0020000 + faddr);
		}
	}

	/* And one more, for our good neighbor, Mr. Broken Cypress. */
	clear = srmmu_get_faddr();
	clear = srmmu_get_fstatus();

	mreg |= (CYPRESS_CENABLE | CYPRESS_CMODE);
	srmmu_set_mmureg(mreg);
}

static void __init init_cypress_common(void)
{
	init_vac_layout();

	BTFIXUPSET_CALL(pte_clear, srmmu_pte_clear, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pmd_clear, srmmu_pmd_clear, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pgd_clear, srmmu_pgd_clear, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_all, cypress_flush_cache_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_mm, cypress_flush_cache_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_range, cypress_flush_cache_range, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_page, cypress_flush_cache_page, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(flush_tlb_all, cypress_flush_tlb_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_mm, cypress_flush_tlb_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_page, cypress_flush_tlb_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_range, cypress_flush_tlb_range, BTFIXUPCALL_NORM);


	BTFIXUPSET_CALL(__flush_page_to_ram, cypress_flush_page_to_ram, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_sig_insns, cypress_flush_sig_insns, BTFIXUPCALL_NOP);
	BTFIXUPSET_CALL(flush_page_for_dma, cypress_flush_page_for_dma, BTFIXUPCALL_NOP);

	poke_srmmu = poke_cypress;
}

static void __init init_cypress_604(void)
{
	srmmu_name = "ROSS Cypress-604(UP)";
	srmmu_modtype = Cypress;
	init_cypress_common();
}

static void __init init_cypress_605(unsigned long mrev)
{
	srmmu_name = "ROSS Cypress-605(MP)";
	if(mrev == 0xe) {
		srmmu_modtype = Cypress_vE;
		hwbug_bitmask |= HWBUG_COPYBACK_BROKEN;
	} else {
		if(mrev == 0xd) {
			srmmu_modtype = Cypress_vD;
			hwbug_bitmask |= HWBUG_ASIFLUSH_BROKEN;
		} else {
			srmmu_modtype = Cypress;
		}
	}
	init_cypress_common();
}

static void __init poke_swift(void)
{
	unsigned long mreg;

	/* Clear any crap from the cache or else... */
	swift_flush_cache_all();

	/* Enable I & D caches */
	mreg = srmmu_get_mmureg();
	mreg |= (SWIFT_IE | SWIFT_DE);
	/*
	 * The Swift branch folding logic is completely broken.  At
	 * trap time, if things are just right, if can mistakenly
	 * think that a trap is coming from kernel mode when in fact
	 * it is coming from user mode (it mis-executes the branch in
	 * the trap code).  So you see things like crashme completely
	 * hosing your machine which is completely unacceptable.  Turn
	 * this shit off... nice job Fujitsu.
	 */
	mreg &= ~(SWIFT_BF);
	srmmu_set_mmureg(mreg);
}

#define SWIFT_MASKID_ADDR  0x10003018
static void __init init_swift(void)
{
	unsigned long swift_rev;

	__asm__ __volatile__("lda [%1] %2, %0\n\t"
			     "srl %0, 0x18, %0\n\t" :
			     "=r" (swift_rev) :
			     "r" (SWIFT_MASKID_ADDR), "i" (ASI_M_BYPASS));
	srmmu_name = "Fujitsu Swift";
	switch(swift_rev) {
	case 0x11:
	case 0x20:
	case 0x23:
	case 0x30:
		srmmu_modtype = Swift_lots_o_bugs;
		hwbug_bitmask |= (HWBUG_KERN_ACCBROKEN | HWBUG_KERN_CBITBROKEN);
		/*
		 * Gee george, I wonder why Sun is so hush hush about
		 * this hardware bug... really braindamage stuff going
		 * on here.  However I think we can find a way to avoid
		 * all of the workaround overhead under Linux.  Basically,
		 * any page fault can cause kernel pages to become user
		 * accessible (the mmu gets confused and clears some of
		 * the ACC bits in kernel ptes).  Aha, sounds pretty
		 * horrible eh?  But wait, after extensive testing it appears
		 * that if you use pgd_t level large kernel pte's (like the
		 * 4MB pages on the Pentium) the bug does not get tripped
		 * at all.  This avoids almost all of the major overhead.
		 * Welcome to a world where your vendor tells you to,
		 * "apply this kernel patch" instead of "sorry for the
		 * broken hardware, send it back and we'll give you
		 * properly functioning parts"
		 */
		break;
	case 0x25:
	case 0x31:
		srmmu_modtype = Swift_bad_c;
		hwbug_bitmask |= HWBUG_KERN_CBITBROKEN;
		/*
		 * You see Sun allude to this hardware bug but never
		 * admit things directly, they'll say things like,
		 * "the Swift chip cache problems" or similar.
		 */
		break;
	default:
		srmmu_modtype = Swift_ok;
		break;
	};

	BTFIXUPSET_CALL(flush_cache_all, swift_flush_cache_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_mm, swift_flush_cache_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_page, swift_flush_cache_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_range, swift_flush_cache_range, BTFIXUPCALL_NORM);


	BTFIXUPSET_CALL(flush_tlb_all, swift_flush_tlb_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_mm, swift_flush_tlb_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_page, swift_flush_tlb_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_range, swift_flush_tlb_range, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(__flush_page_to_ram, swift_flush_page_to_ram, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_sig_insns, swift_flush_sig_insns, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_page_for_dma, swift_flush_page_for_dma, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(update_mmu_cache, swift_update_mmu_cache, BTFIXUPCALL_NORM);

	flush_page_for_dma_global = 0;

	/*
	 * Are you now convinced that the Swift is one of the
	 * biggest VLSI abortions of all time?  Bravo Fujitsu!
	 * Fujitsu, the !#?!%$'d up processor people.  I bet if
	 * you examined the microcode of the Swift you'd find
	 * XXX's all over the place.
	 */
	poke_srmmu = poke_swift;
}

static void turbosparc_flush_cache_all(void)
{
	flush_user_windows();
	turbosparc_idflash_clear();
}

static void turbosparc_flush_cache_mm(struct mm_struct *mm)
{
	FLUSH_BEGIN(mm)
	flush_user_windows();
	turbosparc_idflash_clear();
	FLUSH_END
}

static void turbosparc_flush_cache_range(struct vm_area_struct *vma, unsigned long start, unsigned long end)
{
	FLUSH_BEGIN(vma->vm_mm)
	flush_user_windows();
	turbosparc_idflash_clear();
	FLUSH_END
}

static void turbosparc_flush_cache_page(struct vm_area_struct *vma, unsigned long page)
{
	FLUSH_BEGIN(vma->vm_mm)
	flush_user_windows();
	if (vma->vm_flags & VM_EXEC)
		turbosparc_flush_icache();
	turbosparc_flush_dcache();
	FLUSH_END
}

/* TurboSparc is copy-back, if we turn it on, but this does not work. */
static void turbosparc_flush_page_to_ram(unsigned long page)
{
#ifdef TURBOSPARC_WRITEBACK
	volatile unsigned long clear;

	if (srmmu_hwprobe(page))
		turbosparc_flush_page_cache(page);
	clear = srmmu_get_fstatus();
#endif
}

static void turbosparc_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr)
{
}

static void turbosparc_flush_page_for_dma(unsigned long page)
{
	turbosparc_flush_dcache();
}

static void turbosparc_flush_tlb_all(void)
{
	srmmu_flush_whole_tlb();
}

static void turbosparc_flush_tlb_mm(struct mm_struct *mm)
{
	FLUSH_BEGIN(mm)
	srmmu_flush_whole_tlb();
	FLUSH_END
}

static void turbosparc_flush_tlb_range(struct vm_area_struct *vma, unsigned long start, unsigned long end)
{
	FLUSH_BEGIN(vma->vm_mm)
	srmmu_flush_whole_tlb();
	FLUSH_END
}

static void turbosparc_flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
{
	FLUSH_BEGIN(vma->vm_mm)
	srmmu_flush_whole_tlb();
	FLUSH_END
}


static void __init poke_turbosparc(void)
{
	unsigned long mreg = srmmu_get_mmureg();
	unsigned long ccreg;

	/* Clear any crap from the cache or else... */
	turbosparc_flush_cache_all();
	mreg &= ~(TURBOSPARC_ICENABLE | TURBOSPARC_DCENABLE); /* Temporarily disable I & D caches */
	mreg &= ~(TURBOSPARC_PCENABLE);		/* Don't check parity */
	srmmu_set_mmureg(mreg);
	
	ccreg = turbosparc_get_ccreg();

#ifdef TURBOSPARC_WRITEBACK
	ccreg |= (TURBOSPARC_SNENABLE);		/* Do DVMA snooping in Dcache */
	ccreg &= ~(TURBOSPARC_uS2 | TURBOSPARC_WTENABLE);
			/* Write-back D-cache, emulate VLSI
			 * abortion number three, not number one */
#else
	/* For now let's play safe, optimize later */
	ccreg |= (TURBOSPARC_SNENABLE | TURBOSPARC_WTENABLE);
			/* Do DVMA snooping in Dcache, Write-thru D-cache */
	ccreg &= ~(TURBOSPARC_uS2);
			/* Emulate VLSI abortion number three, not number one */
#endif

	switch (ccreg & 7) {
	case 0: /* No SE cache */
	case 7: /* Test mode */
		break;
	default:
		ccreg |= (TURBOSPARC_SCENABLE);
	}
	turbosparc_set_ccreg (ccreg);

	mreg |= (TURBOSPARC_ICENABLE | TURBOSPARC_DCENABLE); /* I & D caches on */
	mreg |= (TURBOSPARC_ICSNOOP);		/* Icache snooping on */
	srmmu_set_mmureg(mreg);
}

static void __init init_turbosparc(void)
{
	srmmu_name = "Fujitsu TurboSparc";
	srmmu_modtype = TurboSparc;

	BTFIXUPSET_CALL(flush_cache_all, turbosparc_flush_cache_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_mm, turbosparc_flush_cache_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_page, turbosparc_flush_cache_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_range, turbosparc_flush_cache_range, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(flush_tlb_all, turbosparc_flush_tlb_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_mm, turbosparc_flush_tlb_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_page, turbosparc_flush_tlb_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_range, turbosparc_flush_tlb_range, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(__flush_page_to_ram, turbosparc_flush_page_to_ram, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(flush_sig_insns, turbosparc_flush_sig_insns, BTFIXUPCALL_NOP);
	BTFIXUPSET_CALL(flush_page_for_dma, turbosparc_flush_page_for_dma, BTFIXUPCALL_NORM);

	poke_srmmu = poke_turbosparc;
}

static void __init poke_tsunami(void)
{
	unsigned long mreg = srmmu_get_mmureg();

	tsunami_flush_icache();
	tsunami_flush_dcache();
	mreg &= ~TSUNAMI_ITD;
	mreg |= (TSUNAMI_IENAB | TSUNAMI_DENAB);
	srmmu_set_mmureg(mreg);
}

static void __init init_tsunami(void)
{
	/*
	 * Tsunami's pretty sane, Sun and TI actually got it
	 * somewhat right this time.  Fujitsu should have
	 * taken some lessons from them.
	 */

	srmmu_name = "TI Tsunami";
	srmmu_modtype = Tsunami;

	BTFIXUPSET_CALL(flush_cache_all, tsunami_flush_cache_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_mm, tsunami_flush_cache_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_page, tsunami_flush_cache_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_range, tsunami_flush_cache_range, BTFIXUPCALL_NORM);


	BTFIXUPSET_CALL(flush_tlb_all, tsunami_flush_tlb_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_mm, tsunami_flush_tlb_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_page, tsunami_flush_tlb_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_tlb_range, tsunami_flush_tlb_range, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(__flush_page_to_ram, tsunami_flush_page_to_ram, BTFIXUPCALL_NOP);
	BTFIXUPSET_CALL(flush_sig_insns, tsunami_flush_sig_insns, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_page_for_dma, tsunami_flush_page_for_dma, BTFIXUPCALL_NORM);

	poke_srmmu = poke_tsunami;

	tsunami_setup_blockops();
}

static void __init poke_viking(void)
{
	unsigned long mreg = srmmu_get_mmureg();
	static int smp_catch;

	if(viking_mxcc_present) {
		unsigned long mxcc_control = mxcc_get_creg();

		mxcc_control |= (MXCC_CTL_ECE | MXCC_CTL_PRE | MXCC_CTL_MCE);
		mxcc_control &= ~(MXCC_CTL_RRC);
		mxcc_set_creg(mxcc_control);

		/*
		 * We don't need memory parity checks.
		 * XXX This is a mess, have to dig out later. ecd.
		viking_mxcc_turn_off_parity(&mreg, &mxcc_control);
		 */

		/* We do cache ptables on MXCC. */
		mreg |= VIKING_TCENABLE;
	} else {
		unsigned long bpreg;

		mreg &= ~(VIKING_TCENABLE);
		if(smp_catch++) {
			/* Must disable mixed-cmd mode here for other cpu's. */
			bpreg = viking_get_bpreg();
			bpreg &= ~(VIKING_ACTION_MIX);
			viking_set_bpreg(bpreg);

			/* Just in case PROM does something funny. */
			msi_set_sync();
		}
	}

	mreg |= VIKING_SPENABLE;
	mreg |= (VIKING_ICENABLE | VIKING_DCENABLE);
	mreg |= VIKING_SBENABLE;
	mreg &= ~(VIKING_ACENABLE);
	srmmu_set_mmureg(mreg);

#ifdef CONFIG_SMP
	/* Avoid unnecessary cross calls. */
	BTFIXUPCOPY_CALL(flush_cache_all, local_flush_cache_all);
	BTFIXUPCOPY_CALL(flush_cache_mm, local_flush_cache_mm);
	BTFIXUPCOPY_CALL(flush_cache_range, local_flush_cache_range);
	BTFIXUPCOPY_CALL(flush_cache_page, local_flush_cache_page);
	BTFIXUPCOPY_CALL(__flush_page_to_ram, local_flush_page_to_ram);
	BTFIXUPCOPY_CALL(flush_sig_insns, local_flush_sig_insns);
	BTFIXUPCOPY_CALL(flush_page_for_dma, local_flush_page_for_dma);
	btfixup();
#endif
}

static void __init init_viking(void)
{
	unsigned long mreg = srmmu_get_mmureg();

	/* Ahhh, the viking.  SRMMU VLSI abortion number two... */
	if(mreg & VIKING_MMODE) {
		srmmu_name = "TI Viking";
		viking_mxcc_present = 0;
		msi_set_sync();

		BTFIXUPSET_CALL(pte_clear, srmmu_pte_clear, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(pmd_clear, srmmu_pmd_clear, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(pgd_clear, srmmu_pgd_clear, BTFIXUPCALL_NORM);

		/*
		 * We need this to make sure old viking takes no hits
		 * on it's cache for dma snoops to workaround the
		 * "load from non-cacheable memory" interrupt bug.
		 * This is only necessary because of the new way in
		 * which we use the IOMMU.
		 */
		BTFIXUPSET_CALL(flush_page_for_dma, viking_flush_page, BTFIXUPCALL_NORM);

		flush_page_for_dma_global = 0;
	} else {
		srmmu_name = "TI Viking/MXCC";
		viking_mxcc_present = 1;

		srmmu_cache_pagetables = 1;

		/* MXCC vikings lack the DMA snooping bug. */
		BTFIXUPSET_CALL(flush_page_for_dma, viking_flush_page_for_dma, BTFIXUPCALL_NOP);
	}

	BTFIXUPSET_CALL(flush_cache_all, viking_flush_cache_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_mm, viking_flush_cache_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_page, viking_flush_cache_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_range, viking_flush_cache_range, BTFIXUPCALL_NORM);

#ifdef CONFIG_SMP
	if (sparc_cpu_model == sun4d) {
		BTFIXUPSET_CALL(flush_tlb_all, sun4dsmp_flush_tlb_all, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_mm, sun4dsmp_flush_tlb_mm, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_page, sun4dsmp_flush_tlb_page, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_range, sun4dsmp_flush_tlb_range, BTFIXUPCALL_NORM);
	} else
#endif
	{
		BTFIXUPSET_CALL(flush_tlb_all, viking_flush_tlb_all, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_mm, viking_flush_tlb_mm, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_page, viking_flush_tlb_page, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_range, viking_flush_tlb_range, BTFIXUPCALL_NORM);
	}

	BTFIXUPSET_CALL(__flush_page_to_ram, viking_flush_page_to_ram, BTFIXUPCALL_NOP);
	BTFIXUPSET_CALL(flush_sig_insns, viking_flush_sig_insns, BTFIXUPCALL_NOP);

	poke_srmmu = poke_viking;
}

/* Probe for the srmmu chip version. */
static void __init get_srmmu_type(void)
{
	unsigned long mreg, psr;
	unsigned long mod_typ, mod_rev, psr_typ, psr_vers;

	srmmu_modtype = SRMMU_INVAL_MOD;
	hwbug_bitmask = 0;

	mreg = srmmu_get_mmureg(); psr = get_psr();
	mod_typ = (mreg & 0xf0000000) >> 28;
	mod_rev = (mreg & 0x0f000000) >> 24;
	psr_typ = (psr >> 28) & 0xf;
	psr_vers = (psr >> 24) & 0xf;

	/* First, check for HyperSparc or Cypress. */
	if(mod_typ == 1) {
		switch(mod_rev) {
		case 7:
			/* UP or MP Hypersparc */
			init_hypersparc();
			break;
		case 0:
		case 2:
			/* Uniprocessor Cypress */
			init_cypress_604();
			break;
		case 10:
		case 11:
		case 12:
			/* _REALLY OLD_ Cypress MP chips... */
		case 13:
		case 14:
		case 15:
			/* MP Cypress mmu/cache-controller */
			init_cypress_605(mod_rev);
			break;
		default:
			/* Some other Cypress revision, assume a 605. */
			init_cypress_605(mod_rev);
			break;
		};
		return;
	}
	
	/*
	 * Now Fujitsu TurboSparc. It might happen that it is
	 * in Swift emulation mode, so we will check later...
	 */
	if (psr_typ == 0 && psr_vers == 5) {
		init_turbosparc();
		return;
	}

	/* Next check for Fujitsu Swift. */
	if(psr_typ == 0 && psr_vers == 4) {
		int cpunode;
		char node_str[128];

		/* Look if it is not a TurboSparc emulating Swift... */
		cpunode = prom_getchild(prom_root_node);
		while((cpunode = prom_getsibling(cpunode)) != 0) {
			prom_getstring(cpunode, "device_type", node_str, sizeof(node_str));
			if(!strcmp(node_str, "cpu")) {
				if (!prom_getintdefault(cpunode, "psr-implementation", 1) &&
				    prom_getintdefault(cpunode, "psr-version", 1) == 5) {
					init_turbosparc();
					return;
				}
				break;
			}
		}
		
		init_swift();
		return;
	}

	/* Now the Viking family of srmmu. */
	if(psr_typ == 4 &&
	   ((psr_vers == 0) ||
	    ((psr_vers == 1) && (mod_typ == 0) && (mod_rev == 0)))) {
		init_viking();
		return;
	}

	/* Finally the Tsunami. */
	if(psr_typ == 4 && psr_vers == 1 && (mod_typ || mod_rev)) {
		init_tsunami();
		return;
	}

	/* Oh well */
	srmmu_is_bad();
}

/* don't laugh, static pagetables */
static void srmmu_check_pgt_cache(int low, int high)
{
}

extern unsigned long spwin_mmu_patchme, fwin_mmu_patchme,
	tsetup_mmu_patchme, rtrap_mmu_patchme;

extern unsigned long spwin_srmmu_stackchk, srmmu_fwin_stackchk,
	tsetup_srmmu_stackchk, srmmu_rett_stackchk;

extern unsigned long srmmu_fault;

#define PATCH_BRANCH(insn, dest) do { \
		iaddr = &(insn); \
		daddr = &(dest); \
		*iaddr = SPARC_BRANCH((unsigned long) daddr, (unsigned long) iaddr); \
	} while(0)

static void __init patch_window_trap_handlers(void)
{
	unsigned long *iaddr, *daddr;
	
	PATCH_BRANCH(spwin_mmu_patchme, spwin_srmmu_stackchk);
	PATCH_BRANCH(fwin_mmu_patchme, srmmu_fwin_stackchk);
	PATCH_BRANCH(tsetup_mmu_patchme, tsetup_srmmu_stackchk);
	PATCH_BRANCH(rtrap_mmu_patchme, srmmu_rett_stackchk);
	PATCH_BRANCH(sparc_ttable[SP_TRAP_TFLT].inst_three, srmmu_fault);
	PATCH_BRANCH(sparc_ttable[SP_TRAP_DFLT].inst_three, srmmu_fault);
	PATCH_BRANCH(sparc_ttable[SP_TRAP_DACC].inst_three, srmmu_fault);
}

#ifdef CONFIG_SMP
/* Local cross-calls. */
static void smp_flush_page_for_dma(unsigned long page)
{
	xc1((smpfunc_t) BTFIXUP_CALL(local_flush_page_for_dma), page);
	local_flush_page_for_dma(page);
}

#endif

static pte_t srmmu_pgoff_to_pte(unsigned long pgoff)
{
	return __pte((pgoff << SRMMU_PTE_FILE_SHIFT) | SRMMU_FILE);
}

static unsigned long srmmu_pte_to_pgoff(pte_t pte)
{
	return pte_val(pte) >> SRMMU_PTE_FILE_SHIFT;
}

/* Load up routines and constants for sun4m and sun4d mmu */
void __init ld_mmu_srmmu(void)
{
	extern void ld_mmu_iommu(void);
	extern void ld_mmu_iounit(void);
	extern void ___xchg32_sun4md(void);

	BTFIXUPSET_SIMM13(pgdir_shift, SRMMU_PGDIR_SHIFT);
	BTFIXUPSET_SETHI(pgdir_size, SRMMU_PGDIR_SIZE);
	BTFIXUPSET_SETHI(pgdir_mask, SRMMU_PGDIR_MASK);

	BTFIXUPSET_SIMM13(ptrs_per_pmd, SRMMU_PTRS_PER_PMD);
	BTFIXUPSET_SIMM13(ptrs_per_pgd, SRMMU_PTRS_PER_PGD);

	BTFIXUPSET_INT(page_none, pgprot_val(SRMMU_PAGE_NONE));
	BTFIXUPSET_INT(page_shared, pgprot_val(SRMMU_PAGE_SHARED));
	BTFIXUPSET_INT(page_copy, pgprot_val(SRMMU_PAGE_COPY));
	BTFIXUPSET_INT(page_readonly, pgprot_val(SRMMU_PAGE_RDONLY));
	BTFIXUPSET_INT(page_kernel, pgprot_val(SRMMU_PAGE_KERNEL));
	page_kernel = pgprot_val(SRMMU_PAGE_KERNEL);
	pg_iobits = SRMMU_VALID | SRMMU_WRITE | SRMMU_REF;

	/* Functions */
#ifndef CONFIG_SMP	
	BTFIXUPSET_CALL(___xchg32, ___xchg32_sun4md, BTFIXUPCALL_SWAPG1G2);
#endif
	BTFIXUPSET_CALL(do_check_pgt_cache, srmmu_check_pgt_cache, BTFIXUPCALL_NOP);

	BTFIXUPSET_CALL(set_pte, srmmu_set_pte, BTFIXUPCALL_SWAPO0O1);
	BTFIXUPSET_CALL(switch_mm, srmmu_switch_mm, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(pte_pfn, srmmu_pte_pfn, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pmd_page, srmmu_pmd_page, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pgd_page, srmmu_pgd_page, BTFIXUPCALL_NORM);

	BTFIXUPSET_SETHI(none_mask, 0xF0000000);

	BTFIXUPSET_CALL(pte_present, srmmu_pte_present, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pte_clear, srmmu_pte_clear, BTFIXUPCALL_SWAPO0G0);
	BTFIXUPSET_CALL(pte_read, srmmu_pte_read, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(pmd_bad, srmmu_pmd_bad, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pmd_present, srmmu_pmd_present, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pmd_clear, srmmu_pmd_clear, BTFIXUPCALL_SWAPO0G0);

	BTFIXUPSET_CALL(pgd_none, srmmu_pgd_none, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pgd_bad, srmmu_pgd_bad, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pgd_present, srmmu_pgd_present, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pgd_clear, srmmu_pgd_clear, BTFIXUPCALL_SWAPO0G0);

	BTFIXUPSET_CALL(mk_pte, srmmu_mk_pte, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(mk_pte_phys, srmmu_mk_pte_phys, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(mk_pte_io, srmmu_mk_pte_io, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pgd_set, srmmu_pgd_set, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pmd_set, srmmu_pmd_set, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pmd_populate, srmmu_pmd_populate, BTFIXUPCALL_NORM);
	
	BTFIXUPSET_INT(pte_modify_mask, SRMMU_CHG_MASK);
	BTFIXUPSET_CALL(pmd_offset, srmmu_pmd_offset, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pte_offset_kernel, srmmu_pte_offset, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(free_pte_fast, srmmu_free_pte_fast, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pte_free, srmmu_pte_free, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pte_alloc_one_kernel, srmmu_pte_alloc_one_kernel, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pte_alloc_one, srmmu_pte_alloc_one, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(free_pmd_fast, srmmu_pmd_free, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pmd_alloc_one, srmmu_pmd_alloc_one, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(free_pgd_fast, srmmu_free_pgd_fast, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(get_pgd_fast, srmmu_get_pgd_fast, BTFIXUPCALL_NORM);

	BTFIXUPSET_HALF(pte_writei, SRMMU_WRITE);
	BTFIXUPSET_HALF(pte_dirtyi, SRMMU_DIRTY);
	BTFIXUPSET_HALF(pte_youngi, SRMMU_REF);
	BTFIXUPSET_HALF(pte_filei, SRMMU_FILE);
	BTFIXUPSET_HALF(pte_wrprotecti, SRMMU_WRITE);
	BTFIXUPSET_HALF(pte_mkcleani, SRMMU_DIRTY);
	BTFIXUPSET_HALF(pte_mkoldi, SRMMU_REF);
	BTFIXUPSET_CALL(pte_mkwrite, srmmu_pte_mkwrite, BTFIXUPCALL_ORINT(SRMMU_WRITE));
	BTFIXUPSET_CALL(pte_mkdirty, srmmu_pte_mkdirty, BTFIXUPCALL_ORINT(SRMMU_DIRTY));
	BTFIXUPSET_CALL(pte_mkyoung, srmmu_pte_mkyoung, BTFIXUPCALL_ORINT(SRMMU_REF));
	BTFIXUPSET_CALL(update_mmu_cache, srmmu_update_mmu_cache, BTFIXUPCALL_NOP);
	BTFIXUPSET_CALL(destroy_context, srmmu_destroy_context, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(sparc_mapiorange, srmmu_mapiorange, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(sparc_unmapiorange, srmmu_unmapiorange, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(__swp_type, srmmu_swp_type, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(__swp_offset, srmmu_swp_offset, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(__swp_entry, srmmu_swp_entry, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(mmu_info, srmmu_mmu_info, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(alloc_thread_info, srmmu_alloc_thread_info, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(free_thread_info, srmmu_free_thread_info, BTFIXUPCALL_NORM);

	BTFIXUPSET_CALL(pte_to_pgoff, srmmu_pte_to_pgoff, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(pgoff_to_pte, srmmu_pgoff_to_pte, BTFIXUPCALL_NORM);

	get_srmmu_type();
	patch_window_trap_handlers();

#ifdef CONFIG_SMP
	/* El switcheroo... */

	BTFIXUPCOPY_CALL(local_flush_cache_all, flush_cache_all);
	BTFIXUPCOPY_CALL(local_flush_cache_mm, flush_cache_mm);
	BTFIXUPCOPY_CALL(local_flush_cache_range, flush_cache_range);
	BTFIXUPCOPY_CALL(local_flush_cache_page, flush_cache_page);
	BTFIXUPCOPY_CALL(local_flush_tlb_all, flush_tlb_all);
	BTFIXUPCOPY_CALL(local_flush_tlb_mm, flush_tlb_mm);
	BTFIXUPCOPY_CALL(local_flush_tlb_range, flush_tlb_range);
	BTFIXUPCOPY_CALL(local_flush_tlb_page, flush_tlb_page);
	BTFIXUPCOPY_CALL(local_flush_page_to_ram, __flush_page_to_ram);
	BTFIXUPCOPY_CALL(local_flush_sig_insns, flush_sig_insns);
	BTFIXUPCOPY_CALL(local_flush_page_for_dma, flush_page_for_dma);

	BTFIXUPSET_CALL(flush_cache_all, smp_flush_cache_all, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_mm, smp_flush_cache_mm, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_range, smp_flush_cache_range, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_cache_page, smp_flush_cache_page, BTFIXUPCALL_NORM);
	if (sparc_cpu_model != sun4d) {
		BTFIXUPSET_CALL(flush_tlb_all, smp_flush_tlb_all, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_mm, smp_flush_tlb_mm, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_range, smp_flush_tlb_range, BTFIXUPCALL_NORM);
		BTFIXUPSET_CALL(flush_tlb_page, smp_flush_tlb_page, BTFIXUPCALL_NORM);
	}
	BTFIXUPSET_CALL(__flush_page_to_ram, smp_flush_page_to_ram, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_sig_insns, smp_flush_sig_insns, BTFIXUPCALL_NORM);
	BTFIXUPSET_CALL(flush_page_for_dma, smp_flush_page_for_dma, BTFIXUPCALL_NORM);
#endif

	if (sparc_cpu_model == sun4d)
		ld_mmu_iounit();
	else
		ld_mmu_iommu();
#ifdef CONFIG_SMP
	if (sparc_cpu_model == sun4d)
		sun4d_init_smp();
	else
		sun4m_init_smp();
#endif
}