summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/i915/intel_ringbuffer.c
blob: 8da1bde442dd94a677335ee8f3ffc44983ee33b8 (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
/*
 * Copyright © 2008-2010 Intel Corporation
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice (including the next
 * paragraph) shall be included in all copies or substantial portions of the
 * Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 *
 * Authors:
 *    Eric Anholt <eric@anholt.net>
 *    Zou Nan hai <nanhai.zou@intel.com>
 *    Xiang Hai hao<haihao.xiang@intel.com>
 *
 */

#include <linux/log2.h>
#include <drm/drmP.h>
#include "i915_drv.h"
#include <drm/i915_drm.h>
#include "i915_trace.h"
#include "intel_drv.h"

/* Rough estimate of the typical request size, performing a flush,
 * set-context and then emitting the batch.
 */
#define LEGACY_REQUEST_SIZE 200

static unsigned int __intel_ring_space(unsigned int head,
				       unsigned int tail,
				       unsigned int size)
{
	/*
	 * "If the Ring Buffer Head Pointer and the Tail Pointer are on the
	 * same cacheline, the Head Pointer must not be greater than the Tail
	 * Pointer."
	 */
	GEM_BUG_ON(!is_power_of_2(size));
	return (head - tail - CACHELINE_BYTES) & (size - 1);
}

unsigned int intel_ring_update_space(struct intel_ring *ring)
{
	unsigned int space;

	space = __intel_ring_space(ring->head, ring->emit, ring->size);

	ring->space = space;
	return space;
}

static int
gen2_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
{
	u32 cmd, *cs;

	cmd = MI_FLUSH;

	if (mode & EMIT_INVALIDATE)
		cmd |= MI_READ_FLUSH;

	cs = intel_ring_begin(req, 2);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = cmd;
	*cs++ = MI_NOOP;
	intel_ring_advance(req, cs);

	return 0;
}

static int
gen4_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
{
	u32 cmd, *cs;

	/*
	 * read/write caches:
	 *
	 * I915_GEM_DOMAIN_RENDER is always invalidated, but is
	 * only flushed if MI_NO_WRITE_FLUSH is unset.  On 965, it is
	 * also flushed at 2d versus 3d pipeline switches.
	 *
	 * read-only caches:
	 *
	 * I915_GEM_DOMAIN_SAMPLER is flushed on pre-965 if
	 * MI_READ_FLUSH is set, and is always flushed on 965.
	 *
	 * I915_GEM_DOMAIN_COMMAND may not exist?
	 *
	 * I915_GEM_DOMAIN_INSTRUCTION, which exists on 965, is
	 * invalidated when MI_EXE_FLUSH is set.
	 *
	 * I915_GEM_DOMAIN_VERTEX, which exists on 965, is
	 * invalidated with every MI_FLUSH.
	 *
	 * TLBs:
	 *
	 * On 965, TLBs associated with I915_GEM_DOMAIN_COMMAND
	 * and I915_GEM_DOMAIN_CPU in are invalidated at PTE write and
	 * I915_GEM_DOMAIN_RENDER and I915_GEM_DOMAIN_SAMPLER
	 * are flushed at any MI_FLUSH.
	 */

	cmd = MI_FLUSH;
	if (mode & EMIT_INVALIDATE) {
		cmd |= MI_EXE_FLUSH;
		if (IS_G4X(req->i915) || IS_GEN5(req->i915))
			cmd |= MI_INVALIDATE_ISP;
	}

	cs = intel_ring_begin(req, 2);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = cmd;
	*cs++ = MI_NOOP;
	intel_ring_advance(req, cs);

	return 0;
}

/**
 * Emits a PIPE_CONTROL with a non-zero post-sync operation, for
 * implementing two workarounds on gen6.  From section 1.4.7.1
 * "PIPE_CONTROL" of the Sandy Bridge PRM volume 2 part 1:
 *
 * [DevSNB-C+{W/A}] Before any depth stall flush (including those
 * produced by non-pipelined state commands), software needs to first
 * send a PIPE_CONTROL with no bits set except Post-Sync Operation !=
 * 0.
 *
 * [Dev-SNB{W/A}]: Before a PIPE_CONTROL with Write Cache Flush Enable
 * =1, a PIPE_CONTROL with any non-zero post-sync-op is required.
 *
 * And the workaround for these two requires this workaround first:
 *
 * [Dev-SNB{W/A}]: Pipe-control with CS-stall bit set must be sent
 * BEFORE the pipe-control with a post-sync op and no write-cache
 * flushes.
 *
 * And this last workaround is tricky because of the requirements on
 * that bit.  From section 1.4.7.2.3 "Stall" of the Sandy Bridge PRM
 * volume 2 part 1:
 *
 *     "1 of the following must also be set:
 *      - Render Target Cache Flush Enable ([12] of DW1)
 *      - Depth Cache Flush Enable ([0] of DW1)
 *      - Stall at Pixel Scoreboard ([1] of DW1)
 *      - Depth Stall ([13] of DW1)
 *      - Post-Sync Operation ([13] of DW1)
 *      - Notify Enable ([8] of DW1)"
 *
 * The cache flushes require the workaround flush that triggered this
 * one, so we can't use it.  Depth stall would trigger the same.
 * Post-sync nonzero is what triggered this second workaround, so we
 * can't use that one either.  Notify enable is IRQs, which aren't
 * really our business.  That leaves only stall at scoreboard.
 */
static int
intel_emit_post_sync_nonzero_flush(struct drm_i915_gem_request *req)
{
	u32 scratch_addr =
		i915_ggtt_offset(req->engine->scratch) + 2 * CACHELINE_BYTES;
	u32 *cs;

	cs = intel_ring_begin(req, 6);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = GFX_OP_PIPE_CONTROL(5);
	*cs++ = PIPE_CONTROL_CS_STALL | PIPE_CONTROL_STALL_AT_SCOREBOARD;
	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
	*cs++ = 0; /* low dword */
	*cs++ = 0; /* high dword */
	*cs++ = MI_NOOP;
	intel_ring_advance(req, cs);

	cs = intel_ring_begin(req, 6);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = GFX_OP_PIPE_CONTROL(5);
	*cs++ = PIPE_CONTROL_QW_WRITE;
	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
	*cs++ = 0;
	*cs++ = 0;
	*cs++ = MI_NOOP;
	intel_ring_advance(req, cs);

	return 0;
}

static int
gen6_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
{
	u32 scratch_addr =
		i915_ggtt_offset(req->engine->scratch) + 2 * CACHELINE_BYTES;
	u32 *cs, flags = 0;
	int ret;

	/* Force SNB workarounds for PIPE_CONTROL flushes */
	ret = intel_emit_post_sync_nonzero_flush(req);
	if (ret)
		return ret;

	/* Just flush everything.  Experiments have shown that reducing the
	 * number of bits based on the write domains has little performance
	 * impact.
	 */
	if (mode & EMIT_FLUSH) {
		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
		/*
		 * Ensure that any following seqno writes only happen
		 * when the render cache is indeed flushed.
		 */
		flags |= PIPE_CONTROL_CS_STALL;
	}
	if (mode & EMIT_INVALIDATE) {
		flags |= PIPE_CONTROL_TLB_INVALIDATE;
		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
		/*
		 * TLB invalidate requires a post-sync write.
		 */
		flags |= PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_CS_STALL;
	}

	cs = intel_ring_begin(req, 4);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = GFX_OP_PIPE_CONTROL(4);
	*cs++ = flags;
	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
	*cs++ = 0;
	intel_ring_advance(req, cs);

	return 0;
}

static int
gen7_render_ring_cs_stall_wa(struct drm_i915_gem_request *req)
{
	u32 *cs;

	cs = intel_ring_begin(req, 4);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = GFX_OP_PIPE_CONTROL(4);
	*cs++ = PIPE_CONTROL_CS_STALL | PIPE_CONTROL_STALL_AT_SCOREBOARD;
	*cs++ = 0;
	*cs++ = 0;
	intel_ring_advance(req, cs);

	return 0;
}

static int
gen7_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
{
	u32 scratch_addr =
		i915_ggtt_offset(req->engine->scratch) + 2 * CACHELINE_BYTES;
	u32 *cs, flags = 0;

	/*
	 * Ensure that any following seqno writes only happen when the render
	 * cache is indeed flushed.
	 *
	 * Workaround: 4th PIPE_CONTROL command (except the ones with only
	 * read-cache invalidate bits set) must have the CS_STALL bit set. We
	 * don't try to be clever and just set it unconditionally.
	 */
	flags |= PIPE_CONTROL_CS_STALL;

	/* Just flush everything.  Experiments have shown that reducing the
	 * number of bits based on the write domains has little performance
	 * impact.
	 */
	if (mode & EMIT_FLUSH) {
		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
		flags |= PIPE_CONTROL_FLUSH_ENABLE;
	}
	if (mode & EMIT_INVALIDATE) {
		flags |= PIPE_CONTROL_TLB_INVALIDATE;
		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_MEDIA_STATE_CLEAR;
		/*
		 * TLB invalidate requires a post-sync write.
		 */
		flags |= PIPE_CONTROL_QW_WRITE;
		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;

		flags |= PIPE_CONTROL_STALL_AT_SCOREBOARD;

		/* Workaround: we must issue a pipe_control with CS-stall bit
		 * set before a pipe_control command that has the state cache
		 * invalidate bit set. */
		gen7_render_ring_cs_stall_wa(req);
	}

	cs = intel_ring_begin(req, 4);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = GFX_OP_PIPE_CONTROL(4);
	*cs++ = flags;
	*cs++ = scratch_addr;
	*cs++ = 0;
	intel_ring_advance(req, cs);

	return 0;
}

static int
gen8_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
{
	u32 flags;
	u32 *cs;

	cs = intel_ring_begin(req, mode & EMIT_INVALIDATE ? 12 : 6);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	flags = PIPE_CONTROL_CS_STALL;

	if (mode & EMIT_FLUSH) {
		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
		flags |= PIPE_CONTROL_FLUSH_ENABLE;
	}
	if (mode & EMIT_INVALIDATE) {
		flags |= PIPE_CONTROL_TLB_INVALIDATE;
		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
		flags |= PIPE_CONTROL_QW_WRITE;
		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;

		/* WaCsStallBeforeStateCacheInvalidate:bdw,chv */
		cs = gen8_emit_pipe_control(cs,
					    PIPE_CONTROL_CS_STALL |
					    PIPE_CONTROL_STALL_AT_SCOREBOARD,
					    0);
	}

	cs = gen8_emit_pipe_control(cs, flags,
				    i915_ggtt_offset(req->engine->scratch) +
				    2 * CACHELINE_BYTES);

	intel_ring_advance(req, cs);

	return 0;
}

static void ring_setup_phys_status_page(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;
	u32 addr;

	addr = dev_priv->status_page_dmah->busaddr;
	if (INTEL_GEN(dev_priv) >= 4)
		addr |= (dev_priv->status_page_dmah->busaddr >> 28) & 0xf0;
	I915_WRITE(HWS_PGA, addr);
}

static void intel_ring_setup_status_page(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;
	i915_reg_t mmio;

	/* The ring status page addresses are no longer next to the rest of
	 * the ring registers as of gen7.
	 */
	if (IS_GEN7(dev_priv)) {
		switch (engine->id) {
		/*
		 * No more rings exist on Gen7. Default case is only to shut up
		 * gcc switch check warning.
		 */
		default:
			GEM_BUG_ON(engine->id);
		case RCS:
			mmio = RENDER_HWS_PGA_GEN7;
			break;
		case BCS:
			mmio = BLT_HWS_PGA_GEN7;
			break;
		case VCS:
			mmio = BSD_HWS_PGA_GEN7;
			break;
		case VECS:
			mmio = VEBOX_HWS_PGA_GEN7;
			break;
		}
	} else if (IS_GEN6(dev_priv)) {
		mmio = RING_HWS_PGA_GEN6(engine->mmio_base);
	} else {
		/* XXX: gen8 returns to sanity */
		mmio = RING_HWS_PGA(engine->mmio_base);
	}

	if (INTEL_GEN(dev_priv) >= 6)
		I915_WRITE(RING_HWSTAM(engine->mmio_base), 0xffffffff);

	I915_WRITE(mmio, engine->status_page.ggtt_offset);
	POSTING_READ(mmio);

	/*
	 * Flush the TLB for this page
	 *
	 * FIXME: These two bits have disappeared on gen8, so a question
	 * arises: do we still need this and if so how should we go about
	 * invalidating the TLB?
	 */
	if (IS_GEN(dev_priv, 6, 7)) {
		i915_reg_t reg = RING_INSTPM(engine->mmio_base);

		/* ring should be idle before issuing a sync flush*/
		WARN_ON((I915_READ_MODE(engine) & MODE_IDLE) == 0);

		I915_WRITE(reg,
			   _MASKED_BIT_ENABLE(INSTPM_TLB_INVALIDATE |
					      INSTPM_SYNC_FLUSH));
		if (intel_wait_for_register(dev_priv,
					    reg, INSTPM_SYNC_FLUSH, 0,
					    1000))
			DRM_ERROR("%s: wait for SyncFlush to complete for TLB invalidation timed out\n",
				  engine->name);
	}
}

static bool stop_ring(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	if (INTEL_GEN(dev_priv) > 2) {
		I915_WRITE_MODE(engine, _MASKED_BIT_ENABLE(STOP_RING));
		if (intel_wait_for_register(dev_priv,
					    RING_MI_MODE(engine->mmio_base),
					    MODE_IDLE,
					    MODE_IDLE,
					    1000)) {
			DRM_ERROR("%s : timed out trying to stop ring\n",
				  engine->name);
			/* Sometimes we observe that the idle flag is not
			 * set even though the ring is empty. So double
			 * check before giving up.
			 */
			if (I915_READ_HEAD(engine) != I915_READ_TAIL(engine))
				return false;
		}
	}

	I915_WRITE_CTL(engine, 0);
	I915_WRITE_HEAD(engine, 0);
	I915_WRITE_TAIL(engine, 0);

	return (I915_READ_HEAD(engine) & HEAD_ADDR) == 0;
}

static int init_ring_common(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;
	struct intel_ring *ring = engine->buffer;
	int ret = 0;

	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);

	if (!stop_ring(engine)) {
		/* G45 ring initialization often fails to reset head to zero */
		DRM_DEBUG_KMS("%s head not reset to zero "
			      "ctl %08x head %08x tail %08x start %08x\n",
			      engine->name,
			      I915_READ_CTL(engine),
			      I915_READ_HEAD(engine),
			      I915_READ_TAIL(engine),
			      I915_READ_START(engine));

		if (!stop_ring(engine)) {
			DRM_ERROR("failed to set %s head to zero "
				  "ctl %08x head %08x tail %08x start %08x\n",
				  engine->name,
				  I915_READ_CTL(engine),
				  I915_READ_HEAD(engine),
				  I915_READ_TAIL(engine),
				  I915_READ_START(engine));
			ret = -EIO;
			goto out;
		}
	}

	if (HWS_NEEDS_PHYSICAL(dev_priv))
		ring_setup_phys_status_page(engine);
	else
		intel_ring_setup_status_page(engine);

	intel_engine_reset_breadcrumbs(engine);

	/* Enforce ordering by reading HEAD register back */
	I915_READ_HEAD(engine);

	/* Initialize the ring. This must happen _after_ we've cleared the ring
	 * registers with the above sequence (the readback of the HEAD registers
	 * also enforces ordering), otherwise the hw might lose the new ring
	 * register values. */
	I915_WRITE_START(engine, i915_ggtt_offset(ring->vma));

	/* WaClearRingBufHeadRegAtInit:ctg,elk */
	if (I915_READ_HEAD(engine))
		DRM_DEBUG("%s initialization failed [head=%08x], fudging\n",
			  engine->name, I915_READ_HEAD(engine));

	intel_ring_update_space(ring);
	I915_WRITE_HEAD(engine, ring->head);
	I915_WRITE_TAIL(engine, ring->tail);
	(void)I915_READ_TAIL(engine);

	I915_WRITE_CTL(engine, RING_CTL_SIZE(ring->size) | RING_VALID);

	/* If the head is still not zero, the ring is dead */
	if (intel_wait_for_register(dev_priv, RING_CTL(engine->mmio_base),
				    RING_VALID, RING_VALID,
				    50)) {
		DRM_ERROR("%s initialization failed "
			  "ctl %08x (valid? %d) head %08x [%08x] tail %08x [%08x] start %08x [expected %08x]\n",
			  engine->name,
			  I915_READ_CTL(engine),
			  I915_READ_CTL(engine) & RING_VALID,
			  I915_READ_HEAD(engine), ring->head,
			  I915_READ_TAIL(engine), ring->tail,
			  I915_READ_START(engine),
			  i915_ggtt_offset(ring->vma));
		ret = -EIO;
		goto out;
	}

	intel_engine_init_hangcheck(engine);

	if (INTEL_GEN(dev_priv) > 2)
		I915_WRITE_MODE(engine, _MASKED_BIT_DISABLE(STOP_RING));

out:
	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);

	return ret;
}

static void reset_ring_common(struct intel_engine_cs *engine,
			      struct drm_i915_gem_request *request)
{
	/*
	 * RC6 must be prevented until the reset is complete and the engine
	 * reinitialised. If it occurs in the middle of this sequence, the
	 * state written to/loaded from the power context is ill-defined (e.g.
	 * the PP_BASE_DIR may be lost).
	 */
	assert_forcewakes_active(engine->i915, FORCEWAKE_ALL);

	/*
	 * Try to restore the logical GPU state to match the continuation
	 * of the request queue. If we skip the context/PD restore, then
	 * the next request may try to execute assuming that its context
	 * is valid and loaded on the GPU and so may try to access invalid
	 * memory, prompting repeated GPU hangs.
	 *
	 * If the request was guilty, we still restore the logical state
	 * in case the next request requires it (e.g. the aliasing ppgtt),
	 * but skip over the hung batch.
	 *
	 * If the request was innocent, we try to replay the request with
	 * the restored context.
	 */
	if (request) {
		struct drm_i915_private *dev_priv = request->i915;
		struct intel_context *ce = &request->ctx->engine[engine->id];
		struct i915_hw_ppgtt *ppgtt;

		/* FIXME consider gen8 reset */

		if (ce->state) {
			I915_WRITE(CCID,
				   i915_ggtt_offset(ce->state) |
				   BIT(8) /* must be set! */ |
				   CCID_EXTENDED_STATE_SAVE |
				   CCID_EXTENDED_STATE_RESTORE |
				   CCID_EN);
		}

		ppgtt = request->ctx->ppgtt ?: engine->i915->mm.aliasing_ppgtt;
		if (ppgtt) {
			u32 pd_offset = ppgtt->pd.base.ggtt_offset << 10;

			I915_WRITE(RING_PP_DIR_DCLV(engine), PP_DIR_DCLV_2G);
			I915_WRITE(RING_PP_DIR_BASE(engine), pd_offset);

			/* Wait for the PD reload to complete */
			if (intel_wait_for_register(dev_priv,
						    RING_PP_DIR_BASE(engine),
						    BIT(0), 0,
						    10))
				DRM_ERROR("Wait for reload of ppgtt page-directory timed out\n");

			ppgtt->pd_dirty_rings &= ~intel_engine_flag(engine);
		}

		/* If the rq hung, jump to its breadcrumb and skip the batch */
		if (request->fence.error == -EIO)
			request->ring->head = request->postfix;
	} else {
		engine->legacy_active_context = NULL;
	}
}

static int intel_rcs_ctx_init(struct drm_i915_gem_request *req)
{
	int ret;

	ret = intel_ring_workarounds_emit(req);
	if (ret != 0)
		return ret;

	ret = i915_gem_render_state_emit(req);
	if (ret)
		return ret;

	return 0;
}

static int init_render_ring(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;
	int ret = init_ring_common(engine);
	if (ret)
		return ret;

	/* WaTimedSingleVertexDispatch:cl,bw,ctg,elk,ilk,snb */
	if (IS_GEN(dev_priv, 4, 6))
		I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(VS_TIMER_DISPATCH));

	/* We need to disable the AsyncFlip performance optimisations in order
	 * to use MI_WAIT_FOR_EVENT within the CS. It should already be
	 * programmed to '1' on all products.
	 *
	 * WaDisableAsyncFlipPerfMode:snb,ivb,hsw,vlv
	 */
	if (IS_GEN(dev_priv, 6, 7))
		I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(ASYNC_FLIP_PERF_DISABLE));

	/* Required for the hardware to program scanline values for waiting */
	/* WaEnableFlushTlbInvalidationMode:snb */
	if (IS_GEN6(dev_priv))
		I915_WRITE(GFX_MODE,
			   _MASKED_BIT_ENABLE(GFX_TLB_INVALIDATE_EXPLICIT));

	/* WaBCSVCSTlbInvalidationMode:ivb,vlv,hsw */
	if (IS_GEN7(dev_priv))
		I915_WRITE(GFX_MODE_GEN7,
			   _MASKED_BIT_ENABLE(GFX_TLB_INVALIDATE_EXPLICIT) |
			   _MASKED_BIT_ENABLE(GFX_REPLAY_MODE));

	if (IS_GEN6(dev_priv)) {
		/* From the Sandybridge PRM, volume 1 part 3, page 24:
		 * "If this bit is set, STCunit will have LRA as replacement
		 *  policy. [...] This bit must be reset.  LRA replacement
		 *  policy is not supported."
		 */
		I915_WRITE(CACHE_MODE_0,
			   _MASKED_BIT_DISABLE(CM0_STC_EVICT_DISABLE_LRA_SNB));
	}

	if (IS_GEN(dev_priv, 6, 7))
		I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING));

	if (INTEL_INFO(dev_priv)->gen >= 6)
		I915_WRITE_IMR(engine, ~engine->irq_keep_mask);

	return init_workarounds_ring(engine);
}

static void render_ring_cleanup(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	i915_vma_unpin_and_release(&dev_priv->semaphore);
}

static u32 *gen8_rcs_signal(struct drm_i915_gem_request *req, u32 *cs)
{
	struct drm_i915_private *dev_priv = req->i915;
	struct intel_engine_cs *waiter;
	enum intel_engine_id id;

	for_each_engine(waiter, dev_priv, id) {
		u64 gtt_offset = req->engine->semaphore.signal_ggtt[id];
		if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID)
			continue;

		*cs++ = GFX_OP_PIPE_CONTROL(6);
		*cs++ = PIPE_CONTROL_GLOBAL_GTT_IVB | PIPE_CONTROL_QW_WRITE |
			PIPE_CONTROL_CS_STALL;
		*cs++ = lower_32_bits(gtt_offset);
		*cs++ = upper_32_bits(gtt_offset);
		*cs++ = req->global_seqno;
		*cs++ = 0;
		*cs++ = MI_SEMAPHORE_SIGNAL |
			MI_SEMAPHORE_TARGET(waiter->hw_id);
		*cs++ = 0;
	}

	return cs;
}

static u32 *gen8_xcs_signal(struct drm_i915_gem_request *req, u32 *cs)
{
	struct drm_i915_private *dev_priv = req->i915;
	struct intel_engine_cs *waiter;
	enum intel_engine_id id;

	for_each_engine(waiter, dev_priv, id) {
		u64 gtt_offset = req->engine->semaphore.signal_ggtt[id];
		if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID)
			continue;

		*cs++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW;
		*cs++ = lower_32_bits(gtt_offset) | MI_FLUSH_DW_USE_GTT;
		*cs++ = upper_32_bits(gtt_offset);
		*cs++ = req->global_seqno;
		*cs++ = MI_SEMAPHORE_SIGNAL |
			MI_SEMAPHORE_TARGET(waiter->hw_id);
		*cs++ = 0;
	}

	return cs;
}

static u32 *gen6_signal(struct drm_i915_gem_request *req, u32 *cs)
{
	struct drm_i915_private *dev_priv = req->i915;
	struct intel_engine_cs *engine;
	enum intel_engine_id id;
	int num_rings = 0;

	for_each_engine(engine, dev_priv, id) {
		i915_reg_t mbox_reg;

		if (!(BIT(engine->hw_id) & GEN6_SEMAPHORES_MASK))
			continue;

		mbox_reg = req->engine->semaphore.mbox.signal[engine->hw_id];
		if (i915_mmio_reg_valid(mbox_reg)) {
			*cs++ = MI_LOAD_REGISTER_IMM(1);
			*cs++ = i915_mmio_reg_offset(mbox_reg);
			*cs++ = req->global_seqno;
			num_rings++;
		}
	}
	if (num_rings & 1)
		*cs++ = MI_NOOP;

	return cs;
}

static void cancel_requests(struct intel_engine_cs *engine)
{
	struct drm_i915_gem_request *request;
	unsigned long flags;

	spin_lock_irqsave(&engine->timeline->lock, flags);

	/* Mark all submitted requests as skipped. */
	list_for_each_entry(request, &engine->timeline->requests, link) {
		GEM_BUG_ON(!request->global_seqno);
		if (!i915_gem_request_completed(request))
			dma_fence_set_error(&request->fence, -EIO);
	}
	/* Remaining _unready_ requests will be nop'ed when submitted */

	spin_unlock_irqrestore(&engine->timeline->lock, flags);
}

static void i9xx_submit_request(struct drm_i915_gem_request *request)
{
	struct drm_i915_private *dev_priv = request->i915;

	i915_gem_request_submit(request);

	I915_WRITE_TAIL(request->engine,
			intel_ring_set_tail(request->ring, request->tail));
}

static void i9xx_emit_breadcrumb(struct drm_i915_gem_request *req, u32 *cs)
{
	*cs++ = MI_STORE_DWORD_INDEX;
	*cs++ = I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT;
	*cs++ = req->global_seqno;
	*cs++ = MI_USER_INTERRUPT;

	req->tail = intel_ring_offset(req, cs);
	assert_ring_tail_valid(req->ring, req->tail);
}

static const int i9xx_emit_breadcrumb_sz = 4;

/**
 * gen6_sema_emit_breadcrumb - Update the semaphore mailbox registers
 *
 * @request - request to write to the ring
 *
 * Update the mailbox registers in the *other* rings with the current seqno.
 * This acts like a signal in the canonical semaphore.
 */
static void gen6_sema_emit_breadcrumb(struct drm_i915_gem_request *req, u32 *cs)
{
	return i9xx_emit_breadcrumb(req,
				    req->engine->semaphore.signal(req, cs));
}

static void gen8_render_emit_breadcrumb(struct drm_i915_gem_request *req,
					u32 *cs)
{
	struct intel_engine_cs *engine = req->engine;

	if (engine->semaphore.signal)
		cs = engine->semaphore.signal(req, cs);

	*cs++ = GFX_OP_PIPE_CONTROL(6);
	*cs++ = PIPE_CONTROL_GLOBAL_GTT_IVB | PIPE_CONTROL_CS_STALL |
		PIPE_CONTROL_QW_WRITE;
	*cs++ = intel_hws_seqno_address(engine);
	*cs++ = 0;
	*cs++ = req->global_seqno;
	/* We're thrashing one dword of HWS. */
	*cs++ = 0;
	*cs++ = MI_USER_INTERRUPT;
	*cs++ = MI_NOOP;

	req->tail = intel_ring_offset(req, cs);
	assert_ring_tail_valid(req->ring, req->tail);
}

static const int gen8_render_emit_breadcrumb_sz = 8;

/**
 * intel_ring_sync - sync the waiter to the signaller on seqno
 *
 * @waiter - ring that is waiting
 * @signaller - ring which has, or will signal
 * @seqno - seqno which the waiter will block on
 */

static int
gen8_ring_sync_to(struct drm_i915_gem_request *req,
		  struct drm_i915_gem_request *signal)
{
	struct drm_i915_private *dev_priv = req->i915;
	u64 offset = GEN8_WAIT_OFFSET(req->engine, signal->engine->id);
	struct i915_hw_ppgtt *ppgtt;
	u32 *cs;

	cs = intel_ring_begin(req, 4);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = MI_SEMAPHORE_WAIT | MI_SEMAPHORE_GLOBAL_GTT |
		MI_SEMAPHORE_SAD_GTE_SDD;
	*cs++ = signal->global_seqno;
	*cs++ = lower_32_bits(offset);
	*cs++ = upper_32_bits(offset);
	intel_ring_advance(req, cs);

	/* When the !RCS engines idle waiting upon a semaphore, they lose their
	 * pagetables and we must reload them before executing the batch.
	 * We do this on the i915_switch_context() following the wait and
	 * before the dispatch.
	 */
	ppgtt = req->ctx->ppgtt;
	if (ppgtt && req->engine->id != RCS)
		ppgtt->pd_dirty_rings |= intel_engine_flag(req->engine);
	return 0;
}

static int
gen6_ring_sync_to(struct drm_i915_gem_request *req,
		  struct drm_i915_gem_request *signal)
{
	u32 dw1 = MI_SEMAPHORE_MBOX |
		  MI_SEMAPHORE_COMPARE |
		  MI_SEMAPHORE_REGISTER;
	u32 wait_mbox = signal->engine->semaphore.mbox.wait[req->engine->hw_id];
	u32 *cs;

	WARN_ON(wait_mbox == MI_SEMAPHORE_SYNC_INVALID);

	cs = intel_ring_begin(req, 4);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = dw1 | wait_mbox;
	/* Throughout all of the GEM code, seqno passed implies our current
	 * seqno is >= the last seqno executed. However for hardware the
	 * comparison is strictly greater than.
	 */
	*cs++ = signal->global_seqno - 1;
	*cs++ = 0;
	*cs++ = MI_NOOP;
	intel_ring_advance(req, cs);

	return 0;
}

static void
gen5_seqno_barrier(struct intel_engine_cs *engine)
{
	/* MI_STORE are internally buffered by the GPU and not flushed
	 * either by MI_FLUSH or SyncFlush or any other combination of
	 * MI commands.
	 *
	 * "Only the submission of the store operation is guaranteed.
	 * The write result will be complete (coherent) some time later
	 * (this is practically a finite period but there is no guaranteed
	 * latency)."
	 *
	 * Empirically, we observe that we need a delay of at least 75us to
	 * be sure that the seqno write is visible by the CPU.
	 */
	usleep_range(125, 250);
}

static void
gen6_seqno_barrier(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	/* Workaround to force correct ordering between irq and seqno writes on
	 * ivb (and maybe also on snb) by reading from a CS register (like
	 * ACTHD) before reading the status page.
	 *
	 * Note that this effectively stalls the read by the time it takes to
	 * do a memory transaction, which more or less ensures that the write
	 * from the GPU has sufficient time to invalidate the CPU cacheline.
	 * Alternatively we could delay the interrupt from the CS ring to give
	 * the write time to land, but that would incur a delay after every
	 * batch i.e. much more frequent than a delay when waiting for the
	 * interrupt (with the same net latency).
	 *
	 * Also note that to prevent whole machine hangs on gen7, we have to
	 * take the spinlock to guard against concurrent cacheline access.
	 */
	spin_lock_irq(&dev_priv->uncore.lock);
	POSTING_READ_FW(RING_ACTHD(engine->mmio_base));
	spin_unlock_irq(&dev_priv->uncore.lock);
}

static void
gen5_irq_enable(struct intel_engine_cs *engine)
{
	gen5_enable_gt_irq(engine->i915, engine->irq_enable_mask);
}

static void
gen5_irq_disable(struct intel_engine_cs *engine)
{
	gen5_disable_gt_irq(engine->i915, engine->irq_enable_mask);
}

static void
i9xx_irq_enable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	dev_priv->irq_mask &= ~engine->irq_enable_mask;
	I915_WRITE(IMR, dev_priv->irq_mask);
	POSTING_READ_FW(RING_IMR(engine->mmio_base));
}

static void
i9xx_irq_disable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	dev_priv->irq_mask |= engine->irq_enable_mask;
	I915_WRITE(IMR, dev_priv->irq_mask);
}

static void
i8xx_irq_enable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	dev_priv->irq_mask &= ~engine->irq_enable_mask;
	I915_WRITE16(IMR, dev_priv->irq_mask);
	POSTING_READ16(RING_IMR(engine->mmio_base));
}

static void
i8xx_irq_disable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	dev_priv->irq_mask |= engine->irq_enable_mask;
	I915_WRITE16(IMR, dev_priv->irq_mask);
}

static int
bsd_ring_flush(struct drm_i915_gem_request *req, u32 mode)
{
	u32 *cs;

	cs = intel_ring_begin(req, 2);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = MI_FLUSH;
	*cs++ = MI_NOOP;
	intel_ring_advance(req, cs);
	return 0;
}

static void
gen6_irq_enable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	I915_WRITE_IMR(engine,
		       ~(engine->irq_enable_mask |
			 engine->irq_keep_mask));
	gen5_enable_gt_irq(dev_priv, engine->irq_enable_mask);
}

static void
gen6_irq_disable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	I915_WRITE_IMR(engine, ~engine->irq_keep_mask);
	gen5_disable_gt_irq(dev_priv, engine->irq_enable_mask);
}

static void
hsw_vebox_irq_enable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	I915_WRITE_IMR(engine, ~engine->irq_enable_mask);
	gen6_unmask_pm_irq(dev_priv, engine->irq_enable_mask);
}

static void
hsw_vebox_irq_disable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	I915_WRITE_IMR(engine, ~0);
	gen6_mask_pm_irq(dev_priv, engine->irq_enable_mask);
}

static void
gen8_irq_enable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	I915_WRITE_IMR(engine,
		       ~(engine->irq_enable_mask |
			 engine->irq_keep_mask));
	POSTING_READ_FW(RING_IMR(engine->mmio_base));
}

static void
gen8_irq_disable(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	I915_WRITE_IMR(engine, ~engine->irq_keep_mask);
}

static int
i965_emit_bb_start(struct drm_i915_gem_request *req,
		   u64 offset, u32 length,
		   unsigned int dispatch_flags)
{
	u32 *cs;

	cs = intel_ring_begin(req, 2);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT | (dispatch_flags &
		I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE_I965);
	*cs++ = offset;
	intel_ring_advance(req, cs);

	return 0;
}

/* Just userspace ABI convention to limit the wa batch bo to a resonable size */
#define I830_BATCH_LIMIT (256*1024)
#define I830_TLB_ENTRIES (2)
#define I830_WA_SIZE max(I830_TLB_ENTRIES*4096, I830_BATCH_LIMIT)
static int
i830_emit_bb_start(struct drm_i915_gem_request *req,
		   u64 offset, u32 len,
		   unsigned int dispatch_flags)
{
	u32 *cs, cs_offset = i915_ggtt_offset(req->engine->scratch);

	cs = intel_ring_begin(req, 6);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	/* Evict the invalid PTE TLBs */
	*cs++ = COLOR_BLT_CMD | BLT_WRITE_RGBA;
	*cs++ = BLT_DEPTH_32 | BLT_ROP_COLOR_COPY | 4096;
	*cs++ = I830_TLB_ENTRIES << 16 | 4; /* load each page */
	*cs++ = cs_offset;
	*cs++ = 0xdeadbeef;
	*cs++ = MI_NOOP;
	intel_ring_advance(req, cs);

	if ((dispatch_flags & I915_DISPATCH_PINNED) == 0) {
		if (len > I830_BATCH_LIMIT)
			return -ENOSPC;

		cs = intel_ring_begin(req, 6 + 2);
		if (IS_ERR(cs))
			return PTR_ERR(cs);

		/* Blit the batch (which has now all relocs applied) to the
		 * stable batch scratch bo area (so that the CS never
		 * stumbles over its tlb invalidation bug) ...
		 */
		*cs++ = SRC_COPY_BLT_CMD | BLT_WRITE_RGBA;
		*cs++ = BLT_DEPTH_32 | BLT_ROP_SRC_COPY | 4096;
		*cs++ = DIV_ROUND_UP(len, 4096) << 16 | 4096;
		*cs++ = cs_offset;
		*cs++ = 4096;
		*cs++ = offset;

		*cs++ = MI_FLUSH;
		*cs++ = MI_NOOP;
		intel_ring_advance(req, cs);

		/* ... and execute it. */
		offset = cs_offset;
	}

	cs = intel_ring_begin(req, 2);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
	*cs++ = offset | (dispatch_flags & I915_DISPATCH_SECURE ? 0 :
		MI_BATCH_NON_SECURE);
	intel_ring_advance(req, cs);

	return 0;
}

static int
i915_emit_bb_start(struct drm_i915_gem_request *req,
		   u64 offset, u32 len,
		   unsigned int dispatch_flags)
{
	u32 *cs;

	cs = intel_ring_begin(req, 2);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
	*cs++ = offset | (dispatch_flags & I915_DISPATCH_SECURE ? 0 :
		MI_BATCH_NON_SECURE);
	intel_ring_advance(req, cs);

	return 0;
}



int intel_ring_pin(struct intel_ring *ring,
		   struct drm_i915_private *i915,
		   unsigned int offset_bias)
{
	enum i915_map_type map = HAS_LLC(i915) ? I915_MAP_WB : I915_MAP_WC;
	struct i915_vma *vma = ring->vma;
	unsigned int flags;
	void *addr;
	int ret;

	GEM_BUG_ON(ring->vaddr);


	flags = PIN_GLOBAL;
	if (offset_bias)
		flags |= PIN_OFFSET_BIAS | offset_bias;
	if (vma->obj->stolen)
		flags |= PIN_MAPPABLE;

	if (!(vma->flags & I915_VMA_GLOBAL_BIND)) {
		if (flags & PIN_MAPPABLE || map == I915_MAP_WC)
			ret = i915_gem_object_set_to_gtt_domain(vma->obj, true);
		else
			ret = i915_gem_object_set_to_cpu_domain(vma->obj, true);
		if (unlikely(ret))
			return ret;
	}

	ret = i915_vma_pin(vma, 0, PAGE_SIZE, flags);
	if (unlikely(ret))
		return ret;

	if (i915_vma_is_map_and_fenceable(vma))
		addr = (void __force *)i915_vma_pin_iomap(vma);
	else
		addr = i915_gem_object_pin_map(vma->obj, map);
	if (IS_ERR(addr))
		goto err;

	vma->obj->pin_global++;

	ring->vaddr = addr;
	return 0;

err:
	i915_vma_unpin(vma);
	return PTR_ERR(addr);
}

void intel_ring_reset(struct intel_ring *ring, u32 tail)
{
	GEM_BUG_ON(!list_empty(&ring->request_list));
	ring->tail = tail;
	ring->head = tail;
	ring->emit = tail;
	intel_ring_update_space(ring);
}

void intel_ring_unpin(struct intel_ring *ring)
{
	GEM_BUG_ON(!ring->vma);
	GEM_BUG_ON(!ring->vaddr);

	/* Discard any unused bytes beyond that submitted to hw. */
	intel_ring_reset(ring, ring->tail);

	if (i915_vma_is_map_and_fenceable(ring->vma))
		i915_vma_unpin_iomap(ring->vma);
	else
		i915_gem_object_unpin_map(ring->vma->obj);
	ring->vaddr = NULL;

	ring->vma->obj->pin_global--;
	i915_vma_unpin(ring->vma);
}

static struct i915_vma *
intel_ring_create_vma(struct drm_i915_private *dev_priv, int size)
{
	struct drm_i915_gem_object *obj;
	struct i915_vma *vma;

	obj = i915_gem_object_create_stolen(dev_priv, size);
	if (!obj)
		obj = i915_gem_object_create_internal(dev_priv, size);
	if (IS_ERR(obj))
		return ERR_CAST(obj);

	/* mark ring buffers as read-only from GPU side by default */
	obj->gt_ro = 1;

	vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL);
	if (IS_ERR(vma))
		goto err;

	return vma;

err:
	i915_gem_object_put(obj);
	return vma;
}

struct intel_ring *
intel_engine_create_ring(struct intel_engine_cs *engine, int size)
{
	struct intel_ring *ring;
	struct i915_vma *vma;

	GEM_BUG_ON(!is_power_of_2(size));
	GEM_BUG_ON(RING_CTL_SIZE(size) & ~RING_NR_PAGES);

	ring = kzalloc(sizeof(*ring), GFP_KERNEL);
	if (!ring)
		return ERR_PTR(-ENOMEM);

	INIT_LIST_HEAD(&ring->request_list);

	ring->size = size;
	/* Workaround an erratum on the i830 which causes a hang if
	 * the TAIL pointer points to within the last 2 cachelines
	 * of the buffer.
	 */
	ring->effective_size = size;
	if (IS_I830(engine->i915) || IS_I845G(engine->i915))
		ring->effective_size -= 2 * CACHELINE_BYTES;

	intel_ring_update_space(ring);

	vma = intel_ring_create_vma(engine->i915, size);
	if (IS_ERR(vma)) {
		kfree(ring);
		return ERR_CAST(vma);
	}
	ring->vma = vma;

	return ring;
}

void
intel_ring_free(struct intel_ring *ring)
{
	struct drm_i915_gem_object *obj = ring->vma->obj;

	i915_vma_close(ring->vma);
	__i915_gem_object_release_unless_active(obj);

	kfree(ring);
}

static int context_pin(struct i915_gem_context *ctx)
{
	struct i915_vma *vma = ctx->engine[RCS].state;
	int ret;

	/* Clear this page out of any CPU caches for coherent swap-in/out.
	 * We only want to do this on the first bind so that we do not stall
	 * on an active context (which by nature is already on the GPU).
	 */
	if (!(vma->flags & I915_VMA_GLOBAL_BIND)) {
		ret = i915_gem_object_set_to_gtt_domain(vma->obj, false);
		if (ret)
			return ret;
	}

	return i915_vma_pin(vma, 0, I915_GTT_MIN_ALIGNMENT,
			    PIN_GLOBAL | PIN_HIGH);
}

static struct i915_vma *
alloc_context_vma(struct intel_engine_cs *engine)
{
	struct drm_i915_private *i915 = engine->i915;
	struct drm_i915_gem_object *obj;
	struct i915_vma *vma;

	obj = i915_gem_object_create(i915, engine->context_size);
	if (IS_ERR(obj))
		return ERR_CAST(obj);

	/*
	 * Try to make the context utilize L3 as well as LLC.
	 *
	 * On VLV we don't have L3 controls in the PTEs so we
	 * shouldn't touch the cache level, especially as that
	 * would make the object snooped which might have a
	 * negative performance impact.
	 *
	 * Snooping is required on non-llc platforms in execlist
	 * mode, but since all GGTT accesses use PAT entry 0 we
	 * get snooping anyway regardless of cache_level.
	 *
	 * This is only applicable for Ivy Bridge devices since
	 * later platforms don't have L3 control bits in the PTE.
	 */
	if (IS_IVYBRIDGE(i915)) {
		/* Ignore any error, regard it as a simple optimisation */
		i915_gem_object_set_cache_level(obj, I915_CACHE_L3_LLC);
	}

	vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
	if (IS_ERR(vma))
		i915_gem_object_put(obj);

	return vma;
}

static struct intel_ring *
intel_ring_context_pin(struct intel_engine_cs *engine,
		       struct i915_gem_context *ctx)
{
	struct intel_context *ce = &ctx->engine[engine->id];
	int ret;

	lockdep_assert_held(&ctx->i915->drm.struct_mutex);

	if (likely(ce->pin_count++))
		goto out;
	GEM_BUG_ON(!ce->pin_count); /* no overflow please! */

	if (!ce->state && engine->context_size) {
		struct i915_vma *vma;

		vma = alloc_context_vma(engine);
		if (IS_ERR(vma)) {
			ret = PTR_ERR(vma);
			goto err;
		}

		ce->state = vma;
	}

	if (ce->state) {
		ret = context_pin(ctx);
		if (ret)
			goto err;

		ce->state->obj->mm.dirty = true;
		ce->state->obj->pin_global++;
	}

	/* The kernel context is only used as a placeholder for flushing the
	 * active context. It is never used for submitting user rendering and
	 * as such never requires the golden render context, and so we can skip
	 * emitting it when we switch to the kernel context. This is required
	 * as during eviction we cannot allocate and pin the renderstate in
	 * order to initialise the context.
	 */
	if (i915_gem_context_is_kernel(ctx))
		ce->initialised = true;

	i915_gem_context_get(ctx);

out:
	/* One ringbuffer to rule them all */
	return engine->buffer;

err:
	ce->pin_count = 0;
	return ERR_PTR(ret);
}

static void intel_ring_context_unpin(struct intel_engine_cs *engine,
				     struct i915_gem_context *ctx)
{
	struct intel_context *ce = &ctx->engine[engine->id];

	lockdep_assert_held(&ctx->i915->drm.struct_mutex);
	GEM_BUG_ON(ce->pin_count == 0);

	if (--ce->pin_count)
		return;

	if (ce->state) {
		ce->state->obj->pin_global--;
		i915_vma_unpin(ce->state);
	}

	i915_gem_context_put(ctx);
}

static int intel_init_ring_buffer(struct intel_engine_cs *engine)
{
	struct intel_ring *ring;
	int err;

	intel_engine_setup_common(engine);

	err = intel_engine_init_common(engine);
	if (err)
		goto err;

	ring = intel_engine_create_ring(engine, 32 * PAGE_SIZE);
	if (IS_ERR(ring)) {
		err = PTR_ERR(ring);
		goto err;
	}

	/* Ring wraparound at offset 0 sometimes hangs. No idea why. */
	err = intel_ring_pin(ring, engine->i915, I915_GTT_PAGE_SIZE);
	if (err)
		goto err_ring;

	GEM_BUG_ON(engine->buffer);
	engine->buffer = ring;

	return 0;

err_ring:
	intel_ring_free(ring);
err:
	intel_engine_cleanup_common(engine);
	return err;
}

void intel_engine_cleanup(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	WARN_ON(INTEL_GEN(dev_priv) > 2 &&
		(I915_READ_MODE(engine) & MODE_IDLE) == 0);

	intel_ring_unpin(engine->buffer);
	intel_ring_free(engine->buffer);

	if (engine->cleanup)
		engine->cleanup(engine);

	intel_engine_cleanup_common(engine);

	dev_priv->engine[engine->id] = NULL;
	kfree(engine);
}

void intel_legacy_submission_resume(struct drm_i915_private *dev_priv)
{
	struct intel_engine_cs *engine;
	enum intel_engine_id id;

	/* Restart from the beginning of the rings for convenience */
	for_each_engine(engine, dev_priv, id)
		intel_ring_reset(engine->buffer, 0);
}

static int ring_request_alloc(struct drm_i915_gem_request *request)
{
	u32 *cs;

	GEM_BUG_ON(!request->ctx->engine[request->engine->id].pin_count);

	/* Flush enough space to reduce the likelihood of waiting after
	 * we start building the request - in which case we will just
	 * have to repeat work.
	 */
	request->reserved_space += LEGACY_REQUEST_SIZE;

	cs = intel_ring_begin(request, 0);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	request->reserved_space -= LEGACY_REQUEST_SIZE;
	return 0;
}

static noinline int wait_for_space(struct drm_i915_gem_request *req,
				   unsigned int bytes)
{
	struct intel_ring *ring = req->ring;
	struct drm_i915_gem_request *target;
	long timeout;

	lockdep_assert_held(&req->i915->drm.struct_mutex);

	if (intel_ring_update_space(ring) >= bytes)
		return 0;

	/*
	 * Space is reserved in the ringbuffer for finalising the request,
	 * as that cannot be allowed to fail. During request finalisation,
	 * reserved_space is set to 0 to stop the overallocation and the
	 * assumption is that then we never need to wait (which has the
	 * risk of failing with EINTR).
	 *
	 * See also i915_gem_request_alloc() and i915_add_request().
	 */
	GEM_BUG_ON(!req->reserved_space);

	list_for_each_entry(target, &ring->request_list, ring_link) {
		/* Would completion of this request free enough space? */
		if (bytes <= __intel_ring_space(target->postfix,
						ring->emit, ring->size))
			break;
	}

	if (WARN_ON(&target->ring_link == &ring->request_list))
		return -ENOSPC;

	timeout = i915_wait_request(target,
				    I915_WAIT_INTERRUPTIBLE | I915_WAIT_LOCKED,
				    MAX_SCHEDULE_TIMEOUT);
	if (timeout < 0)
		return timeout;

	i915_gem_request_retire_upto(target);

	intel_ring_update_space(ring);
	GEM_BUG_ON(ring->space < bytes);
	return 0;
}

u32 *intel_ring_begin(struct drm_i915_gem_request *req,
		      unsigned int num_dwords)
{
	struct intel_ring *ring = req->ring;
	const unsigned int remain_usable = ring->effective_size - ring->emit;
	const unsigned int bytes = num_dwords * sizeof(u32);
	unsigned int need_wrap = 0;
	unsigned int total_bytes;
	u32 *cs;

	/* Packets must be qword aligned. */
	GEM_BUG_ON(num_dwords & 1);

	total_bytes = bytes + req->reserved_space;
	GEM_BUG_ON(total_bytes > ring->effective_size);

	if (unlikely(total_bytes > remain_usable)) {
		const int remain_actual = ring->size - ring->emit;

		if (bytes > remain_usable) {
			/*
			 * Not enough space for the basic request. So need to
			 * flush out the remainder and then wait for
			 * base + reserved.
			 */
			total_bytes += remain_actual;
			need_wrap = remain_actual | 1;
		} else  {
			/*
			 * The base request will fit but the reserved space
			 * falls off the end. So we don't need an immediate
			 * wrap and only need to effectively wait for the
			 * reserved size from the start of ringbuffer.
			 */
			total_bytes = req->reserved_space + remain_actual;
		}
	}

	if (unlikely(total_bytes > ring->space)) {
		int ret = wait_for_space(req, total_bytes);
		if (unlikely(ret))
			return ERR_PTR(ret);
	}

	if (unlikely(need_wrap)) {
		need_wrap &= ~1;
		GEM_BUG_ON(need_wrap > ring->space);
		GEM_BUG_ON(ring->emit + need_wrap > ring->size);

		/* Fill the tail with MI_NOOP */
		memset(ring->vaddr + ring->emit, 0, need_wrap);
		ring->emit = 0;
		ring->space -= need_wrap;
	}

	GEM_BUG_ON(ring->emit > ring->size - bytes);
	GEM_BUG_ON(ring->space < bytes);
	cs = ring->vaddr + ring->emit;
	GEM_DEBUG_EXEC(memset(cs, POISON_INUSE, bytes));
	ring->emit += bytes;
	ring->space -= bytes;

	return cs;
}

/* Align the ring tail to a cacheline boundary */
int intel_ring_cacheline_align(struct drm_i915_gem_request *req)
{
	int num_dwords =
		(req->ring->emit & (CACHELINE_BYTES - 1)) / sizeof(uint32_t);
	u32 *cs;

	if (num_dwords == 0)
		return 0;

	num_dwords = CACHELINE_BYTES / sizeof(uint32_t) - num_dwords;
	cs = intel_ring_begin(req, num_dwords);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	while (num_dwords--)
		*cs++ = MI_NOOP;

	intel_ring_advance(req, cs);

	return 0;
}

static void gen6_bsd_submit_request(struct drm_i915_gem_request *request)
{
	struct drm_i915_private *dev_priv = request->i915;

	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);

       /* Every tail move must follow the sequence below */

	/* Disable notification that the ring is IDLE. The GT
	 * will then assume that it is busy and bring it out of rc6.
	 */
	I915_WRITE_FW(GEN6_BSD_SLEEP_PSMI_CONTROL,
		      _MASKED_BIT_ENABLE(GEN6_BSD_SLEEP_MSG_DISABLE));

	/* Clear the context id. Here be magic! */
	I915_WRITE64_FW(GEN6_BSD_RNCID, 0x0);

	/* Wait for the ring not to be idle, i.e. for it to wake up. */
	if (__intel_wait_for_register_fw(dev_priv,
					 GEN6_BSD_SLEEP_PSMI_CONTROL,
					 GEN6_BSD_SLEEP_INDICATOR,
					 0,
					 1000, 0, NULL))
		DRM_ERROR("timed out waiting for the BSD ring to wake up\n");

	/* Now that the ring is fully powered up, update the tail */
	i9xx_submit_request(request);

	/* Let the ring send IDLE messages to the GT again,
	 * and so let it sleep to conserve power when idle.
	 */
	I915_WRITE_FW(GEN6_BSD_SLEEP_PSMI_CONTROL,
		      _MASKED_BIT_DISABLE(GEN6_BSD_SLEEP_MSG_DISABLE));

	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
}

static int gen6_bsd_ring_flush(struct drm_i915_gem_request *req, u32 mode)
{
	u32 cmd, *cs;

	cs = intel_ring_begin(req, 4);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	cmd = MI_FLUSH_DW;
	if (INTEL_GEN(req->i915) >= 8)
		cmd += 1;

	/* We always require a command barrier so that subsequent
	 * commands, such as breadcrumb interrupts, are strictly ordered
	 * wrt the contents of the write cache being flushed to memory
	 * (and thus being coherent from the CPU).
	 */
	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;

	/*
	 * Bspec vol 1c.5 - video engine command streamer:
	 * "If ENABLED, all TLBs will be invalidated once the flush
	 * operation is complete. This bit is only valid when the
	 * Post-Sync Operation field is a value of 1h or 3h."
	 */
	if (mode & EMIT_INVALIDATE)
		cmd |= MI_INVALIDATE_TLB | MI_INVALIDATE_BSD;

	*cs++ = cmd;
	*cs++ = I915_GEM_HWS_SCRATCH_ADDR | MI_FLUSH_DW_USE_GTT;
	if (INTEL_GEN(req->i915) >= 8) {
		*cs++ = 0; /* upper addr */
		*cs++ = 0; /* value */
	} else  {
		*cs++ = 0;
		*cs++ = MI_NOOP;
	}
	intel_ring_advance(req, cs);
	return 0;
}

static int
gen8_emit_bb_start(struct drm_i915_gem_request *req,
		   u64 offset, u32 len,
		   unsigned int dispatch_flags)
{
	bool ppgtt = USES_PPGTT(req->i915) &&
			!(dispatch_flags & I915_DISPATCH_SECURE);
	u32 *cs;

	cs = intel_ring_begin(req, 4);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	/* FIXME(BDW): Address space and security selectors. */
	*cs++ = MI_BATCH_BUFFER_START_GEN8 | (ppgtt << 8) | (dispatch_flags &
		I915_DISPATCH_RS ? MI_BATCH_RESOURCE_STREAMER : 0);
	*cs++ = lower_32_bits(offset);
	*cs++ = upper_32_bits(offset);
	*cs++ = MI_NOOP;
	intel_ring_advance(req, cs);

	return 0;
}

static int
hsw_emit_bb_start(struct drm_i915_gem_request *req,
		  u64 offset, u32 len,
		  unsigned int dispatch_flags)
{
	u32 *cs;

	cs = intel_ring_begin(req, 2);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = MI_BATCH_BUFFER_START | (dispatch_flags & I915_DISPATCH_SECURE ?
		0 : MI_BATCH_PPGTT_HSW | MI_BATCH_NON_SECURE_HSW) |
		(dispatch_flags & I915_DISPATCH_RS ?
		MI_BATCH_RESOURCE_STREAMER : 0);
	/* bit0-7 is the length on GEN6+ */
	*cs++ = offset;
	intel_ring_advance(req, cs);

	return 0;
}

static int
gen6_emit_bb_start(struct drm_i915_gem_request *req,
		   u64 offset, u32 len,
		   unsigned int dispatch_flags)
{
	u32 *cs;

	cs = intel_ring_begin(req, 2);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	*cs++ = MI_BATCH_BUFFER_START | (dispatch_flags & I915_DISPATCH_SECURE ?
		0 : MI_BATCH_NON_SECURE_I965);
	/* bit0-7 is the length on GEN6+ */
	*cs++ = offset;
	intel_ring_advance(req, cs);

	return 0;
}

/* Blitter support (SandyBridge+) */

static int gen6_ring_flush(struct drm_i915_gem_request *req, u32 mode)
{
	u32 cmd, *cs;

	cs = intel_ring_begin(req, 4);
	if (IS_ERR(cs))
		return PTR_ERR(cs);

	cmd = MI_FLUSH_DW;
	if (INTEL_GEN(req->i915) >= 8)
		cmd += 1;

	/* We always require a command barrier so that subsequent
	 * commands, such as breadcrumb interrupts, are strictly ordered
	 * wrt the contents of the write cache being flushed to memory
	 * (and thus being coherent from the CPU).
	 */
	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;

	/*
	 * Bspec vol 1c.3 - blitter engine command streamer:
	 * "If ENABLED, all TLBs will be invalidated once the flush
	 * operation is complete. This bit is only valid when the
	 * Post-Sync Operation field is a value of 1h or 3h."
	 */
	if (mode & EMIT_INVALIDATE)
		cmd |= MI_INVALIDATE_TLB;
	*cs++ = cmd;
	*cs++ = I915_GEM_HWS_SCRATCH_ADDR | MI_FLUSH_DW_USE_GTT;
	if (INTEL_GEN(req->i915) >= 8) {
		*cs++ = 0; /* upper addr */
		*cs++ = 0; /* value */
	} else  {
		*cs++ = 0;
		*cs++ = MI_NOOP;
	}
	intel_ring_advance(req, cs);

	return 0;
}

static void intel_ring_init_semaphores(struct drm_i915_private *dev_priv,
				       struct intel_engine_cs *engine)
{
	struct drm_i915_gem_object *obj;
	int ret, i;

	if (!i915_modparams.semaphores)
		return;

	if (INTEL_GEN(dev_priv) >= 8 && !dev_priv->semaphore) {
		struct i915_vma *vma;

		obj = i915_gem_object_create(dev_priv, PAGE_SIZE);
		if (IS_ERR(obj))
			goto err;

		vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL);
		if (IS_ERR(vma))
			goto err_obj;

		ret = i915_gem_object_set_to_gtt_domain(obj, false);
		if (ret)
			goto err_obj;

		ret = i915_vma_pin(vma, 0, 0, PIN_GLOBAL | PIN_HIGH);
		if (ret)
			goto err_obj;

		dev_priv->semaphore = vma;
	}

	if (INTEL_GEN(dev_priv) >= 8) {
		u32 offset = i915_ggtt_offset(dev_priv->semaphore);

		engine->semaphore.sync_to = gen8_ring_sync_to;
		engine->semaphore.signal = gen8_xcs_signal;

		for (i = 0; i < I915_NUM_ENGINES; i++) {
			u32 ring_offset;

			if (i != engine->id)
				ring_offset = offset + GEN8_SEMAPHORE_OFFSET(engine->id, i);
			else
				ring_offset = MI_SEMAPHORE_SYNC_INVALID;

			engine->semaphore.signal_ggtt[i] = ring_offset;
		}
	} else if (INTEL_GEN(dev_priv) >= 6) {
		engine->semaphore.sync_to = gen6_ring_sync_to;
		engine->semaphore.signal = gen6_signal;

		/*
		 * The current semaphore is only applied on pre-gen8
		 * platform.  And there is no VCS2 ring on the pre-gen8
		 * platform. So the semaphore between RCS and VCS2 is
		 * initialized as INVALID.  Gen8 will initialize the
		 * sema between VCS2 and RCS later.
		 */
		for (i = 0; i < GEN6_NUM_SEMAPHORES; i++) {
			static const struct {
				u32 wait_mbox;
				i915_reg_t mbox_reg;
			} sem_data[GEN6_NUM_SEMAPHORES][GEN6_NUM_SEMAPHORES] = {
				[RCS_HW] = {
					[VCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_RV,  .mbox_reg = GEN6_VRSYNC },
					[BCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_RB,  .mbox_reg = GEN6_BRSYNC },
					[VECS_HW] = { .wait_mbox = MI_SEMAPHORE_SYNC_RVE, .mbox_reg = GEN6_VERSYNC },
				},
				[VCS_HW] = {
					[RCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VR,  .mbox_reg = GEN6_RVSYNC },
					[BCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VB,  .mbox_reg = GEN6_BVSYNC },
					[VECS_HW] = { .wait_mbox = MI_SEMAPHORE_SYNC_VVE, .mbox_reg = GEN6_VEVSYNC },
				},
				[BCS_HW] = {
					[RCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_BR,  .mbox_reg = GEN6_RBSYNC },
					[VCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_BV,  .mbox_reg = GEN6_VBSYNC },
					[VECS_HW] = { .wait_mbox = MI_SEMAPHORE_SYNC_BVE, .mbox_reg = GEN6_VEBSYNC },
				},
				[VECS_HW] = {
					[RCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VER, .mbox_reg = GEN6_RVESYNC },
					[VCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VEV, .mbox_reg = GEN6_VVESYNC },
					[BCS_HW] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VEB, .mbox_reg = GEN6_BVESYNC },
				},
			};
			u32 wait_mbox;
			i915_reg_t mbox_reg;

			if (i == engine->hw_id) {
				wait_mbox = MI_SEMAPHORE_SYNC_INVALID;
				mbox_reg = GEN6_NOSYNC;
			} else {
				wait_mbox = sem_data[engine->hw_id][i].wait_mbox;
				mbox_reg = sem_data[engine->hw_id][i].mbox_reg;
			}

			engine->semaphore.mbox.wait[i] = wait_mbox;
			engine->semaphore.mbox.signal[i] = mbox_reg;
		}
	}

	return;

err_obj:
	i915_gem_object_put(obj);
err:
	DRM_DEBUG_DRIVER("Failed to allocate space for semaphores, disabling\n");
	i915_modparams.semaphores = 0;
}

static void intel_ring_init_irq(struct drm_i915_private *dev_priv,
				struct intel_engine_cs *engine)
{
	engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << engine->irq_shift;

	if (INTEL_GEN(dev_priv) >= 8) {
		engine->irq_enable = gen8_irq_enable;
		engine->irq_disable = gen8_irq_disable;
		engine->irq_seqno_barrier = gen6_seqno_barrier;
	} else if (INTEL_GEN(dev_priv) >= 6) {
		engine->irq_enable = gen6_irq_enable;
		engine->irq_disable = gen6_irq_disable;
		engine->irq_seqno_barrier = gen6_seqno_barrier;
	} else if (INTEL_GEN(dev_priv) >= 5) {
		engine->irq_enable = gen5_irq_enable;
		engine->irq_disable = gen5_irq_disable;
		engine->irq_seqno_barrier = gen5_seqno_barrier;
	} else if (INTEL_GEN(dev_priv) >= 3) {
		engine->irq_enable = i9xx_irq_enable;
		engine->irq_disable = i9xx_irq_disable;
	} else {
		engine->irq_enable = i8xx_irq_enable;
		engine->irq_disable = i8xx_irq_disable;
	}
}

static void i9xx_set_default_submission(struct intel_engine_cs *engine)
{
	engine->submit_request = i9xx_submit_request;
	engine->cancel_requests = cancel_requests;
}

static void gen6_bsd_set_default_submission(struct intel_engine_cs *engine)
{
	engine->submit_request = gen6_bsd_submit_request;
	engine->cancel_requests = cancel_requests;
}

static void intel_ring_default_vfuncs(struct drm_i915_private *dev_priv,
				      struct intel_engine_cs *engine)
{
	intel_ring_init_irq(dev_priv, engine);
	intel_ring_init_semaphores(dev_priv, engine);

	engine->init_hw = init_ring_common;
	engine->reset_hw = reset_ring_common;

	engine->context_pin = intel_ring_context_pin;
	engine->context_unpin = intel_ring_context_unpin;

	engine->request_alloc = ring_request_alloc;

	engine->emit_breadcrumb = i9xx_emit_breadcrumb;
	engine->emit_breadcrumb_sz = i9xx_emit_breadcrumb_sz;
	if (i915_modparams.semaphores) {
		int num_rings;

		engine->emit_breadcrumb = gen6_sema_emit_breadcrumb;

		num_rings = INTEL_INFO(dev_priv)->num_rings - 1;
		if (INTEL_GEN(dev_priv) >= 8) {
			engine->emit_breadcrumb_sz += num_rings * 6;
		} else {
			engine->emit_breadcrumb_sz += num_rings * 3;
			if (num_rings & 1)
				engine->emit_breadcrumb_sz++;
		}
	}

	engine->set_default_submission = i9xx_set_default_submission;

	if (INTEL_GEN(dev_priv) >= 8)
		engine->emit_bb_start = gen8_emit_bb_start;
	else if (INTEL_GEN(dev_priv) >= 6)
		engine->emit_bb_start = gen6_emit_bb_start;
	else if (INTEL_GEN(dev_priv) >= 4)
		engine->emit_bb_start = i965_emit_bb_start;
	else if (IS_I830(dev_priv) || IS_I845G(dev_priv))
		engine->emit_bb_start = i830_emit_bb_start;
	else
		engine->emit_bb_start = i915_emit_bb_start;
}

int intel_init_render_ring_buffer(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;
	int ret;

	intel_ring_default_vfuncs(dev_priv, engine);

	if (HAS_L3_DPF(dev_priv))
		engine->irq_keep_mask = GT_RENDER_L3_PARITY_ERROR_INTERRUPT;

	if (INTEL_GEN(dev_priv) >= 8) {
		engine->init_context = intel_rcs_ctx_init;
		engine->emit_breadcrumb = gen8_render_emit_breadcrumb;
		engine->emit_breadcrumb_sz = gen8_render_emit_breadcrumb_sz;
		engine->emit_flush = gen8_render_ring_flush;
		if (i915_modparams.semaphores) {
			int num_rings;

			engine->semaphore.signal = gen8_rcs_signal;

			num_rings = INTEL_INFO(dev_priv)->num_rings - 1;
			engine->emit_breadcrumb_sz += num_rings * 8;
		}
	} else if (INTEL_GEN(dev_priv) >= 6) {
		engine->init_context = intel_rcs_ctx_init;
		engine->emit_flush = gen7_render_ring_flush;
		if (IS_GEN6(dev_priv))
			engine->emit_flush = gen6_render_ring_flush;
	} else if (IS_GEN5(dev_priv)) {
		engine->emit_flush = gen4_render_ring_flush;
	} else {
		if (INTEL_GEN(dev_priv) < 4)
			engine->emit_flush = gen2_render_ring_flush;
		else
			engine->emit_flush = gen4_render_ring_flush;
		engine->irq_enable_mask = I915_USER_INTERRUPT;
	}

	if (IS_HASWELL(dev_priv))
		engine->emit_bb_start = hsw_emit_bb_start;

	engine->init_hw = init_render_ring;
	engine->cleanup = render_ring_cleanup;

	ret = intel_init_ring_buffer(engine);
	if (ret)
		return ret;

	if (INTEL_GEN(dev_priv) >= 6) {
		ret = intel_engine_create_scratch(engine, PAGE_SIZE);
		if (ret)
			return ret;
	} else if (HAS_BROKEN_CS_TLB(dev_priv)) {
		ret = intel_engine_create_scratch(engine, I830_WA_SIZE);
		if (ret)
			return ret;
	}

	return 0;
}

int intel_init_bsd_ring_buffer(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	intel_ring_default_vfuncs(dev_priv, engine);

	if (INTEL_GEN(dev_priv) >= 6) {
		/* gen6 bsd needs a special wa for tail updates */
		if (IS_GEN6(dev_priv))
			engine->set_default_submission = gen6_bsd_set_default_submission;
		engine->emit_flush = gen6_bsd_ring_flush;
		if (INTEL_GEN(dev_priv) < 8)
			engine->irq_enable_mask = GT_BSD_USER_INTERRUPT;
	} else {
		engine->mmio_base = BSD_RING_BASE;
		engine->emit_flush = bsd_ring_flush;
		if (IS_GEN5(dev_priv))
			engine->irq_enable_mask = ILK_BSD_USER_INTERRUPT;
		else
			engine->irq_enable_mask = I915_BSD_USER_INTERRUPT;
	}

	return intel_init_ring_buffer(engine);
}

int intel_init_blt_ring_buffer(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	intel_ring_default_vfuncs(dev_priv, engine);

	engine->emit_flush = gen6_ring_flush;
	if (INTEL_GEN(dev_priv) < 8)
		engine->irq_enable_mask = GT_BLT_USER_INTERRUPT;

	return intel_init_ring_buffer(engine);
}

int intel_init_vebox_ring_buffer(struct intel_engine_cs *engine)
{
	struct drm_i915_private *dev_priv = engine->i915;

	intel_ring_default_vfuncs(dev_priv, engine);

	engine->emit_flush = gen6_ring_flush;

	if (INTEL_GEN(dev_priv) < 8) {
		engine->irq_enable_mask = PM_VEBOX_USER_INTERRUPT;
		engine->irq_enable = hsw_vebox_irq_enable;
		engine->irq_disable = hsw_vebox_irq_disable;
	}

	return intel_init_ring_buffer(engine);
}