aboutsummaryrefslogtreecommitdiff
path: root/sys/dev/virtio/network/if_vtnet.c
blob: 7cbbe56e6c6b5e95dffb114c11c3b4e97eae5b93 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
/*-
 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
 *
 * Copyright (c) 2011, Bryan Venteicher <bryanv@FreeBSD.org>
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice unmodified, this list of conditions, and the following
 *    disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

/* Driver for VirtIO network devices. */

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

#include <sys/param.h>
#include <sys/eventhandler.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/sockio.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/module.h>
#include <sys/msan.h>
#include <sys/socket.h>
#include <sys/sysctl.h>
#include <sys/random.h>
#include <sys/sglist.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/taskqueue.h>
#include <sys/smp.h>
#include <machine/smp.h>

#include <vm/uma.h>

#include <net/debugnet.h>
#include <net/ethernet.h>
#include <net/pfil.h>
#include <net/if.h>
#include <net/if_var.h>
#include <net/if_arp.h>
#include <net/if_dl.h>
#include <net/if_types.h>
#include <net/if_media.h>
#include <net/if_vlan_var.h>

#include <net/bpf.h>

#include <netinet/in_systm.h>
#include <netinet/in.h>
#include <netinet/ip.h>
#include <netinet/ip6.h>
#include <netinet6/ip6_var.h>
#include <netinet/udp.h>
#include <netinet/tcp.h>
#include <netinet/tcp_lro.h>

#include <machine/bus.h>
#include <machine/resource.h>
#include <sys/bus.h>
#include <sys/rman.h>

#include <dev/virtio/virtio.h>
#include <dev/virtio/virtqueue.h>
#include <dev/virtio/network/virtio_net.h>
#include <dev/virtio/network/if_vtnetvar.h>
#include "virtio_if.h"

#include "opt_inet.h"
#include "opt_inet6.h"

#if defined(INET) || defined(INET6)
#include <machine/in_cksum.h>
#endif

static int	vtnet_modevent(module_t, int, void *);

static int	vtnet_probe(device_t);
static int	vtnet_attach(device_t);
static int	vtnet_detach(device_t);
static int	vtnet_suspend(device_t);
static int	vtnet_resume(device_t);
static int	vtnet_shutdown(device_t);
static int	vtnet_attach_completed(device_t);
static int	vtnet_config_change(device_t);

static int	vtnet_negotiate_features(struct vtnet_softc *);
static int	vtnet_setup_features(struct vtnet_softc *);
static int	vtnet_init_rxq(struct vtnet_softc *, int);
static int	vtnet_init_txq(struct vtnet_softc *, int);
static int	vtnet_alloc_rxtx_queues(struct vtnet_softc *);
static void	vtnet_free_rxtx_queues(struct vtnet_softc *);
static int	vtnet_alloc_rx_filters(struct vtnet_softc *);
static void	vtnet_free_rx_filters(struct vtnet_softc *);
static int	vtnet_alloc_virtqueues(struct vtnet_softc *);
static int	vtnet_alloc_interface(struct vtnet_softc *);
static int	vtnet_setup_interface(struct vtnet_softc *);
static int	vtnet_ioctl_mtu(struct vtnet_softc *, u_int);
static int	vtnet_ioctl_ifflags(struct vtnet_softc *);
static int	vtnet_ioctl_multi(struct vtnet_softc *);
static int	vtnet_ioctl_ifcap(struct vtnet_softc *, struct ifreq *);
static int	vtnet_ioctl(struct ifnet *, u_long, caddr_t);
static uint64_t	vtnet_get_counter(struct ifnet *, ift_counter);

static int	vtnet_rxq_populate(struct vtnet_rxq *);
static void	vtnet_rxq_free_mbufs(struct vtnet_rxq *);
static struct mbuf *
		vtnet_rx_alloc_buf(struct vtnet_softc *, int , struct mbuf **);
static int	vtnet_rxq_replace_lro_nomrg_buf(struct vtnet_rxq *,
		    struct mbuf *, int);
static int	vtnet_rxq_replace_buf(struct vtnet_rxq *, struct mbuf *, int);
static int	vtnet_rxq_enqueue_buf(struct vtnet_rxq *, struct mbuf *);
static int	vtnet_rxq_new_buf(struct vtnet_rxq *);
static int	vtnet_rxq_csum_needs_csum(struct vtnet_rxq *, struct mbuf *,
		     uint16_t, int, struct virtio_net_hdr *);
static int	vtnet_rxq_csum_data_valid(struct vtnet_rxq *, struct mbuf *,
		     uint16_t, int, struct virtio_net_hdr *);
static int	vtnet_rxq_csum(struct vtnet_rxq *, struct mbuf *,
		     struct virtio_net_hdr *);
static void	vtnet_rxq_discard_merged_bufs(struct vtnet_rxq *, int);
static void	vtnet_rxq_discard_buf(struct vtnet_rxq *, struct mbuf *);
static int	vtnet_rxq_merged_eof(struct vtnet_rxq *, struct mbuf *, int);
static void	vtnet_rxq_input(struct vtnet_rxq *, struct mbuf *,
		    struct virtio_net_hdr *);
static int	vtnet_rxq_eof(struct vtnet_rxq *);
static void	vtnet_rx_vq_process(struct vtnet_rxq *rxq, int tries);
static void	vtnet_rx_vq_intr(void *);
static void	vtnet_rxq_tq_intr(void *, int);

static int	vtnet_txq_intr_threshold(struct vtnet_txq *);
static int	vtnet_txq_below_threshold(struct vtnet_txq *);
static int	vtnet_txq_notify(struct vtnet_txq *);
static void	vtnet_txq_free_mbufs(struct vtnet_txq *);
static int	vtnet_txq_offload_ctx(struct vtnet_txq *, struct mbuf *,
		    int *, int *, int *);
static int	vtnet_txq_offload_tso(struct vtnet_txq *, struct mbuf *, int,
		    int, struct virtio_net_hdr *);
static struct mbuf *
		vtnet_txq_offload(struct vtnet_txq *, struct mbuf *,
		    struct virtio_net_hdr *);
static int	vtnet_txq_enqueue_buf(struct vtnet_txq *, struct mbuf **,
		    struct vtnet_tx_header *);
static int	vtnet_txq_encap(struct vtnet_txq *, struct mbuf **, int);
#ifdef VTNET_LEGACY_TX
static void	vtnet_start_locked(struct vtnet_txq *, struct ifnet *);
static void	vtnet_start(struct ifnet *);
#else
static int	vtnet_txq_mq_start_locked(struct vtnet_txq *, struct mbuf *);
static int	vtnet_txq_mq_start(struct ifnet *, struct mbuf *);
static void	vtnet_txq_tq_deferred(void *, int);
#endif
static void	vtnet_txq_start(struct vtnet_txq *);
static void	vtnet_txq_tq_intr(void *, int);
static int	vtnet_txq_eof(struct vtnet_txq *);
static void	vtnet_tx_vq_intr(void *);
static void	vtnet_tx_start_all(struct vtnet_softc *);

#ifndef VTNET_LEGACY_TX
static void	vtnet_qflush(struct ifnet *);
#endif

static int	vtnet_watchdog(struct vtnet_txq *);
static void	vtnet_accum_stats(struct vtnet_softc *,
		    struct vtnet_rxq_stats *, struct vtnet_txq_stats *);
static void	vtnet_tick(void *);

static void	vtnet_start_taskqueues(struct vtnet_softc *);
static void	vtnet_free_taskqueues(struct vtnet_softc *);
static void	vtnet_drain_taskqueues(struct vtnet_softc *);

static void	vtnet_drain_rxtx_queues(struct vtnet_softc *);
static void	vtnet_stop_rendezvous(struct vtnet_softc *);
static void	vtnet_stop(struct vtnet_softc *);
static int	vtnet_virtio_reinit(struct vtnet_softc *);
static void	vtnet_init_rx_filters(struct vtnet_softc *);
static int	vtnet_init_rx_queues(struct vtnet_softc *);
static int	vtnet_init_tx_queues(struct vtnet_softc *);
static int	vtnet_init_rxtx_queues(struct vtnet_softc *);
static void	vtnet_set_active_vq_pairs(struct vtnet_softc *);
static void	vtnet_update_rx_offloads(struct vtnet_softc *);
static int	vtnet_reinit(struct vtnet_softc *);
static void	vtnet_init_locked(struct vtnet_softc *, int);
static void	vtnet_init(void *);

static void	vtnet_free_ctrl_vq(struct vtnet_softc *);
static void	vtnet_exec_ctrl_cmd(struct vtnet_softc *, void *,
		    struct sglist *, int, int);
static int	vtnet_ctrl_mac_cmd(struct vtnet_softc *, uint8_t *);
static int	vtnet_ctrl_guest_offloads(struct vtnet_softc *, uint64_t);
static int	vtnet_ctrl_mq_cmd(struct vtnet_softc *, uint16_t);
static int	vtnet_ctrl_rx_cmd(struct vtnet_softc *, uint8_t, bool);
static int	vtnet_set_promisc(struct vtnet_softc *, bool);
static int	vtnet_set_allmulti(struct vtnet_softc *, bool);
static void	vtnet_rx_filter(struct vtnet_softc *);
static void	vtnet_rx_filter_mac(struct vtnet_softc *);
static int	vtnet_exec_vlan_filter(struct vtnet_softc *, int, uint16_t);
static void	vtnet_rx_filter_vlan(struct vtnet_softc *);
static void	vtnet_update_vlan_filter(struct vtnet_softc *, int, uint16_t);
static void	vtnet_register_vlan(void *, struct ifnet *, uint16_t);
static void	vtnet_unregister_vlan(void *, struct ifnet *, uint16_t);

static void	vtnet_update_speed_duplex(struct vtnet_softc *);
static int	vtnet_is_link_up(struct vtnet_softc *);
static void	vtnet_update_link_status(struct vtnet_softc *);
static int	vtnet_ifmedia_upd(struct ifnet *);
static void	vtnet_ifmedia_sts(struct ifnet *, struct ifmediareq *);
static void	vtnet_get_macaddr(struct vtnet_softc *);
static void	vtnet_set_macaddr(struct vtnet_softc *);
static void	vtnet_attached_set_macaddr(struct vtnet_softc *);
static void	vtnet_vlan_tag_remove(struct mbuf *);
static void	vtnet_set_rx_process_limit(struct vtnet_softc *);

static void	vtnet_setup_rxq_sysctl(struct sysctl_ctx_list *,
		    struct sysctl_oid_list *, struct vtnet_rxq *);
static void	vtnet_setup_txq_sysctl(struct sysctl_ctx_list *,
		    struct sysctl_oid_list *, struct vtnet_txq *);
static void	vtnet_setup_queue_sysctl(struct vtnet_softc *);
static void	vtnet_load_tunables(struct vtnet_softc *);
static void	vtnet_setup_sysctl(struct vtnet_softc *);

static int	vtnet_rxq_enable_intr(struct vtnet_rxq *);
static void	vtnet_rxq_disable_intr(struct vtnet_rxq *);
static int	vtnet_txq_enable_intr(struct vtnet_txq *);
static void	vtnet_txq_disable_intr(struct vtnet_txq *);
static void	vtnet_enable_rx_interrupts(struct vtnet_softc *);
static void	vtnet_enable_tx_interrupts(struct vtnet_softc *);
static void	vtnet_enable_interrupts(struct vtnet_softc *);
static void	vtnet_disable_rx_interrupts(struct vtnet_softc *);
static void	vtnet_disable_tx_interrupts(struct vtnet_softc *);
static void	vtnet_disable_interrupts(struct vtnet_softc *);

static int	vtnet_tunable_int(struct vtnet_softc *, const char *, int);

DEBUGNET_DEFINE(vtnet);

#define vtnet_htog16(_sc, _val)	virtio_htog16(vtnet_modern(_sc), _val)
#define vtnet_htog32(_sc, _val)	virtio_htog32(vtnet_modern(_sc), _val)
#define vtnet_htog64(_sc, _val)	virtio_htog64(vtnet_modern(_sc), _val)
#define vtnet_gtoh16(_sc, _val)	virtio_gtoh16(vtnet_modern(_sc), _val)
#define vtnet_gtoh32(_sc, _val)	virtio_gtoh32(vtnet_modern(_sc), _val)
#define vtnet_gtoh64(_sc, _val)	virtio_gtoh64(vtnet_modern(_sc), _val)

/* Tunables. */
static SYSCTL_NODE(_hw, OID_AUTO, vtnet, CTLFLAG_RD | CTLFLAG_MPSAFE, 0,
    "VirtIO Net driver parameters");

static int vtnet_csum_disable = 0;
SYSCTL_INT(_hw_vtnet, OID_AUTO, csum_disable, CTLFLAG_RDTUN,
    &vtnet_csum_disable, 0, "Disables receive and send checksum offload");

static int vtnet_fixup_needs_csum = 0;
SYSCTL_INT(_hw_vtnet, OID_AUTO, fixup_needs_csum, CTLFLAG_RDTUN,
    &vtnet_fixup_needs_csum, 0,
    "Calculate valid checksum for NEEDS_CSUM packets");

static int vtnet_tso_disable = 0;
SYSCTL_INT(_hw_vtnet, OID_AUTO, tso_disable, CTLFLAG_RDTUN,
    &vtnet_tso_disable, 0, "Disables TSO");

static int vtnet_lro_disable = 0;
SYSCTL_INT(_hw_vtnet, OID_AUTO, lro_disable, CTLFLAG_RDTUN,
    &vtnet_lro_disable, 0, "Disables hardware LRO");

static int vtnet_mq_disable = 0;
SYSCTL_INT(_hw_vtnet, OID_AUTO, mq_disable, CTLFLAG_RDTUN,
    &vtnet_mq_disable, 0, "Disables multiqueue support");

static int vtnet_mq_max_pairs = VTNET_MAX_QUEUE_PAIRS;
SYSCTL_INT(_hw_vtnet, OID_AUTO, mq_max_pairs, CTLFLAG_RDTUN,
    &vtnet_mq_max_pairs, 0, "Maximum number of multiqueue pairs");

static int vtnet_tso_maxlen = IP_MAXPACKET;
SYSCTL_INT(_hw_vtnet, OID_AUTO, tso_maxlen, CTLFLAG_RDTUN,
    &vtnet_tso_maxlen, 0, "TSO burst limit");

static int vtnet_rx_process_limit = 1024;
SYSCTL_INT(_hw_vtnet, OID_AUTO, rx_process_limit, CTLFLAG_RDTUN,
    &vtnet_rx_process_limit, 0,
    "Number of RX segments processed in one pass");

static int vtnet_lro_entry_count = 128;
SYSCTL_INT(_hw_vtnet, OID_AUTO, lro_entry_count, CTLFLAG_RDTUN,
    &vtnet_lro_entry_count, 0, "Software LRO entry count");

/* Enable sorted LRO, and the depth of the mbuf queue. */
static int vtnet_lro_mbufq_depth = 0;
SYSCTL_UINT(_hw_vtnet, OID_AUTO, lro_mbufq_depth, CTLFLAG_RDTUN,
    &vtnet_lro_mbufq_depth, 0, "Depth of software LRO mbuf queue");

static uma_zone_t vtnet_tx_header_zone;

static struct virtio_feature_desc vtnet_feature_desc[] = {
	{ VIRTIO_NET_F_CSUM,			"TxChecksum"		},
	{ VIRTIO_NET_F_GUEST_CSUM,		"RxChecksum"		},
	{ VIRTIO_NET_F_CTRL_GUEST_OFFLOADS,	"CtrlRxOffloads"	},
	{ VIRTIO_NET_F_MAC,			"MAC"			},
	{ VIRTIO_NET_F_GSO,			"TxGSO"			},
	{ VIRTIO_NET_F_GUEST_TSO4,		"RxLROv4"		},
	{ VIRTIO_NET_F_GUEST_TSO6,		"RxLROv6"		},
	{ VIRTIO_NET_F_GUEST_ECN,		"RxLROECN"		},
	{ VIRTIO_NET_F_GUEST_UFO,		"RxUFO"			},
	{ VIRTIO_NET_F_HOST_TSO4,		"TxTSOv4"		},
	{ VIRTIO_NET_F_HOST_TSO6,		"TxTSOv6"		},
	{ VIRTIO_NET_F_HOST_ECN,		"TxTSOECN"		},
	{ VIRTIO_NET_F_HOST_UFO,		"TxUFO"			},
	{ VIRTIO_NET_F_MRG_RXBUF,		"MrgRxBuf"		},
	{ VIRTIO_NET_F_STATUS,			"Status"		},
	{ VIRTIO_NET_F_CTRL_VQ,			"CtrlVq"		},
	{ VIRTIO_NET_F_CTRL_RX,			"CtrlRxMode"		},
	{ VIRTIO_NET_F_CTRL_VLAN,		"CtrlVLANFilter"	},
	{ VIRTIO_NET_F_CTRL_RX_EXTRA,		"CtrlRxModeExtra"	},
	{ VIRTIO_NET_F_GUEST_ANNOUNCE,		"GuestAnnounce"		},
	{ VIRTIO_NET_F_MQ,			"Multiqueue"		},
	{ VIRTIO_NET_F_CTRL_MAC_ADDR,		"CtrlMacAddr"		},
	{ VIRTIO_NET_F_SPEED_DUPLEX,		"SpeedDuplex"		},

	{ 0, NULL }
};

static device_method_t vtnet_methods[] = {
	/* Device methods. */
	DEVMETHOD(device_probe,			vtnet_probe),
	DEVMETHOD(device_attach,		vtnet_attach),
	DEVMETHOD(device_detach,		vtnet_detach),
	DEVMETHOD(device_suspend,		vtnet_suspend),
	DEVMETHOD(device_resume,		vtnet_resume),
	DEVMETHOD(device_shutdown,		vtnet_shutdown),

	/* VirtIO methods. */
	DEVMETHOD(virtio_attach_completed,	vtnet_attach_completed),
	DEVMETHOD(virtio_config_change,		vtnet_config_change),

	DEVMETHOD_END
};

#ifdef DEV_NETMAP
#include <dev/netmap/if_vtnet_netmap.h>
#endif

static driver_t vtnet_driver = {
    .name = "vtnet",
    .methods = vtnet_methods,
    .size = sizeof(struct vtnet_softc)
};
static devclass_t vtnet_devclass;

VIRTIO_DRIVER_MODULE(vtnet, vtnet_driver, vtnet_devclass,
    vtnet_modevent, 0);
MODULE_VERSION(vtnet, 1);
MODULE_DEPEND(vtnet, virtio, 1, 1, 1);
#ifdef DEV_NETMAP
MODULE_DEPEND(vtnet, netmap, 1, 1, 1);
#endif

VIRTIO_SIMPLE_PNPINFO(vtnet, VIRTIO_ID_NETWORK, "VirtIO Networking Adapter");

static int
vtnet_modevent(module_t mod __unused, int type, void *unused __unused)
{
	int error = 0;
	static int loaded = 0;

	switch (type) {
	case MOD_LOAD:
		if (loaded++ == 0) {
			vtnet_tx_header_zone = uma_zcreate("vtnet_tx_hdr",
				sizeof(struct vtnet_tx_header),
				NULL, NULL, NULL, NULL, 0, 0);
#ifdef DEBUGNET
			/*
			 * We need to allocate from this zone in the transmit path, so ensure
			 * that we have at least one item per header available.
			 * XXX add a separate zone like we do for mbufs? otherwise we may alloc
			 * buckets
			 */
			uma_zone_reserve(vtnet_tx_header_zone, DEBUGNET_MAX_IN_FLIGHT * 2);
			uma_prealloc(vtnet_tx_header_zone, DEBUGNET_MAX_IN_FLIGHT * 2);
#endif
		}
		break;
	case MOD_QUIESCE:
		if (uma_zone_get_cur(vtnet_tx_header_zone) > 0)
			error = EBUSY;
		break;
	case MOD_UNLOAD:
		if (--loaded == 0) {
			uma_zdestroy(vtnet_tx_header_zone);
			vtnet_tx_header_zone = NULL;
		}
		break;
	case MOD_SHUTDOWN:
		break;
	default:
		error = EOPNOTSUPP;
		break;
	}

	return (error);
}

static int
vtnet_probe(device_t dev)
{
	return (VIRTIO_SIMPLE_PROBE(dev, vtnet));
}

static int
vtnet_attach(device_t dev)
{
	struct vtnet_softc *sc;
	int error;

	sc = device_get_softc(dev);
	sc->vtnet_dev = dev;
	virtio_set_feature_desc(dev, vtnet_feature_desc);

	VTNET_CORE_LOCK_INIT(sc);
	callout_init_mtx(&sc->vtnet_tick_ch, VTNET_CORE_MTX(sc), 0);
	vtnet_load_tunables(sc);

	error = vtnet_alloc_interface(sc);
	if (error) {
		device_printf(dev, "cannot allocate interface\n");
		goto fail;
	}

	vtnet_setup_sysctl(sc);

	error = vtnet_setup_features(sc);
	if (error) {
		device_printf(dev, "cannot setup features\n");
		goto fail;
	}

	error = vtnet_alloc_rx_filters(sc);
	if (error) {
		device_printf(dev, "cannot allocate Rx filters\n");
		goto fail;
	}

	error = vtnet_alloc_rxtx_queues(sc);
	if (error) {
		device_printf(dev, "cannot allocate queues\n");
		goto fail;
	}

	error = vtnet_alloc_virtqueues(sc);
	if (error) {
		device_printf(dev, "cannot allocate virtqueues\n");
		goto fail;
	}

	error = vtnet_setup_interface(sc);
	if (error) {
		device_printf(dev, "cannot setup interface\n");
		goto fail;
	}

	error = virtio_setup_intr(dev, INTR_TYPE_NET);
	if (error) {
		device_printf(dev, "cannot setup interrupts\n");
		ether_ifdetach(sc->vtnet_ifp);
		goto fail;
	}

#ifdef DEV_NETMAP
	vtnet_netmap_attach(sc);
#endif
	vtnet_start_taskqueues(sc);

fail:
	if (error)
		vtnet_detach(dev);

	return (error);
}

static int
vtnet_detach(device_t dev)
{
	struct vtnet_softc *sc;
	struct ifnet *ifp;

	sc = device_get_softc(dev);
	ifp = sc->vtnet_ifp;

	if (device_is_attached(dev)) {
		VTNET_CORE_LOCK(sc);
		vtnet_stop(sc);
		VTNET_CORE_UNLOCK(sc);

		callout_drain(&sc->vtnet_tick_ch);
		vtnet_drain_taskqueues(sc);

		ether_ifdetach(ifp);
	}

#ifdef DEV_NETMAP
	netmap_detach(ifp);
#endif

	if (sc->vtnet_pfil != NULL) {
		pfil_head_unregister(sc->vtnet_pfil);
		sc->vtnet_pfil = NULL;
	}

	vtnet_free_taskqueues(sc);

	if (sc->vtnet_vlan_attach != NULL) {
		EVENTHANDLER_DEREGISTER(vlan_config, sc->vtnet_vlan_attach);
		sc->vtnet_vlan_attach = NULL;
	}
	if (sc->vtnet_vlan_detach != NULL) {
		EVENTHANDLER_DEREGISTER(vlan_unconfig, sc->vtnet_vlan_detach);
		sc->vtnet_vlan_detach = NULL;
	}

	ifmedia_removeall(&sc->vtnet_media);

	if (ifp != NULL) {
		if_free(ifp);
		sc->vtnet_ifp = NULL;
	}

	vtnet_free_rxtx_queues(sc);
	vtnet_free_rx_filters(sc);

	if (sc->vtnet_ctrl_vq != NULL)
		vtnet_free_ctrl_vq(sc);

	VTNET_CORE_LOCK_DESTROY(sc);

	return (0);
}

static int
vtnet_suspend(device_t dev)
{
	struct vtnet_softc *sc;

	sc = device_get_softc(dev);

	VTNET_CORE_LOCK(sc);
	vtnet_stop(sc);
	sc->vtnet_flags |= VTNET_FLAG_SUSPENDED;
	VTNET_CORE_UNLOCK(sc);

	return (0);
}

static int
vtnet_resume(device_t dev)
{
	struct vtnet_softc *sc;
	struct ifnet *ifp;

	sc = device_get_softc(dev);
	ifp = sc->vtnet_ifp;

	VTNET_CORE_LOCK(sc);
	if (ifp->if_flags & IFF_UP)
		vtnet_init_locked(sc, 0);
	sc->vtnet_flags &= ~VTNET_FLAG_SUSPENDED;
	VTNET_CORE_UNLOCK(sc);

	return (0);
}

static int
vtnet_shutdown(device_t dev)
{
	/*
	 * Suspend already does all of what we need to
	 * do here; we just never expect to be resumed.
	 */
	return (vtnet_suspend(dev));
}

static int
vtnet_attach_completed(device_t dev)
{
	struct vtnet_softc *sc;

	sc = device_get_softc(dev);

	VTNET_CORE_LOCK(sc);
	vtnet_attached_set_macaddr(sc);
	VTNET_CORE_UNLOCK(sc);

	return (0);
}

static int
vtnet_config_change(device_t dev)
{
	struct vtnet_softc *sc;

	sc = device_get_softc(dev);

	VTNET_CORE_LOCK(sc);
	vtnet_update_link_status(sc);
	if (sc->vtnet_link_active != 0)
		vtnet_tx_start_all(sc);
	VTNET_CORE_UNLOCK(sc);

	return (0);
}

static int
vtnet_negotiate_features(struct vtnet_softc *sc)
{
	device_t dev;
	uint64_t features, negotiated_features;
	int no_csum;

	dev = sc->vtnet_dev;
	features = virtio_bus_is_modern(dev) ? VTNET_MODERN_FEATURES :
	    VTNET_LEGACY_FEATURES;

	/*
	 * TSO and LRO are only available when their corresponding checksum
	 * offload feature is also negotiated.
	 */
	no_csum = vtnet_tunable_int(sc, "csum_disable", vtnet_csum_disable);
	if (no_csum)
		features &= ~(VIRTIO_NET_F_CSUM | VIRTIO_NET_F_GUEST_CSUM);
	if (no_csum || vtnet_tunable_int(sc, "tso_disable", vtnet_tso_disable))
		features &= ~VTNET_TSO_FEATURES;
	if (no_csum || vtnet_tunable_int(sc, "lro_disable", vtnet_lro_disable))
		features &= ~VTNET_LRO_FEATURES;

#ifndef VTNET_LEGACY_TX
	if (vtnet_tunable_int(sc, "mq_disable", vtnet_mq_disable))
		features &= ~VIRTIO_NET_F_MQ;
#else
	features &= ~VIRTIO_NET_F_MQ;
#endif

	negotiated_features = virtio_negotiate_features(dev, features);

	if (virtio_with_feature(dev, VIRTIO_NET_F_MTU)) {
		uint16_t mtu;

		mtu = virtio_read_dev_config_2(dev,
		    offsetof(struct virtio_net_config, mtu));
		if (mtu < VTNET_MIN_MTU /* || mtu > VTNET_MAX_MTU */) {
			device_printf(dev, "Invalid MTU value: %d. "
			    "MTU feature disabled.\n", mtu);
			features &= ~VIRTIO_NET_F_MTU;
			negotiated_features =
			    virtio_negotiate_features(dev, features);
		}
	}

	if (virtio_with_feature(dev, VIRTIO_NET_F_MQ)) {
		uint16_t npairs;

		npairs = virtio_read_dev_config_2(dev,
		    offsetof(struct virtio_net_config, max_virtqueue_pairs));
		if (npairs < VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MIN ||
		    npairs > VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MAX) {
			device_printf(dev, "Invalid max_virtqueue_pairs value: "
			    "%d. Multiqueue feature disabled.\n", npairs);
			features &= ~VIRTIO_NET_F_MQ;
			negotiated_features =
			    virtio_negotiate_features(dev, features);
		}
	}

	if (virtio_with_feature(dev, VTNET_LRO_FEATURES) &&
	    virtio_with_feature(dev, VIRTIO_NET_F_MRG_RXBUF) == 0) {
		/*
		 * LRO without mergeable buffers requires special care. This
		 * is not ideal because every receive buffer must be large
		 * enough to hold the maximum TCP packet, the Ethernet header,
		 * and the header. This requires up to 34 descriptors with
		 * MCLBYTES clusters. If we do not have indirect descriptors,
		 * LRO is disabled since the virtqueue will not contain very
		 * many receive buffers.
		 */
		if (!virtio_with_feature(dev, VIRTIO_RING_F_INDIRECT_DESC)) {
			device_printf(dev,
			    "Host LRO disabled since both mergeable buffers "
			    "and indirect descriptors were not negotiated\n");
			features &= ~VTNET_LRO_FEATURES;
			negotiated_features =
			    virtio_negotiate_features(dev, features);
		} else
			sc->vtnet_flags |= VTNET_FLAG_LRO_NOMRG;
	}

	sc->vtnet_features = negotiated_features;
	sc->vtnet_negotiated_features = negotiated_features;

	return (virtio_finalize_features(dev));
}

static int
vtnet_setup_features(struct vtnet_softc *sc)
{
	device_t dev;
	int error;

	dev = sc->vtnet_dev;

	error = vtnet_negotiate_features(sc);
	if (error)
		return (error);

	if (virtio_with_feature(dev, VIRTIO_F_VERSION_1))
		sc->vtnet_flags |= VTNET_FLAG_MODERN;
	if (virtio_with_feature(dev, VIRTIO_RING_F_INDIRECT_DESC))
		sc->vtnet_flags |= VTNET_FLAG_INDIRECT;
	if (virtio_with_feature(dev, VIRTIO_RING_F_EVENT_IDX))
		sc->vtnet_flags |= VTNET_FLAG_EVENT_IDX;

	if (virtio_with_feature(dev, VIRTIO_NET_F_MAC)) {
		/* This feature should always be negotiated. */
		sc->vtnet_flags |= VTNET_FLAG_MAC;
	}

	if (virtio_with_feature(dev, VIRTIO_NET_F_MTU)) {
		sc->vtnet_max_mtu = virtio_read_dev_config_2(dev,
		    offsetof(struct virtio_net_config, mtu));
	} else
		sc->vtnet_max_mtu = VTNET_MAX_MTU;

	if (virtio_with_feature(dev, VIRTIO_NET_F_MRG_RXBUF)) {
		sc->vtnet_flags |= VTNET_FLAG_MRG_RXBUFS;
		sc->vtnet_hdr_size = sizeof(struct virtio_net_hdr_mrg_rxbuf);
	} else if (vtnet_modern(sc)) {
		/* This is identical to the mergeable header. */
		sc->vtnet_hdr_size = sizeof(struct virtio_net_hdr_v1);
	} else
		sc->vtnet_hdr_size = sizeof(struct virtio_net_hdr);

	if (vtnet_modern(sc) || sc->vtnet_flags & VTNET_FLAG_MRG_RXBUFS)
		sc->vtnet_rx_nsegs = VTNET_RX_SEGS_HDR_INLINE;
	else if (sc->vtnet_flags & VTNET_FLAG_LRO_NOMRG)
		sc->vtnet_rx_nsegs = VTNET_RX_SEGS_LRO_NOMRG;
	else
		sc->vtnet_rx_nsegs = VTNET_RX_SEGS_HDR_SEPARATE;

	/*
	 * Favor "hardware" LRO if negotiated, but support software LRO as
	 * a fallback; there is usually little benefit (or worse) with both.
	 */
	if (virtio_with_feature(dev, VIRTIO_NET_F_GUEST_TSO4) == 0 &&
	    virtio_with_feature(dev, VIRTIO_NET_F_GUEST_TSO6) == 0)
		sc->vtnet_flags |= VTNET_FLAG_SW_LRO;

	if (virtio_with_feature(dev, VIRTIO_NET_F_GSO) ||
	    virtio_with_feature(dev, VIRTIO_NET_F_HOST_TSO4) ||
	    virtio_with_feature(dev, VIRTIO_NET_F_HOST_TSO6))
		sc->vtnet_tx_nsegs = VTNET_TX_SEGS_MAX;
	else
		sc->vtnet_tx_nsegs = VTNET_TX_SEGS_MIN;

	sc->vtnet_req_vq_pairs = 1;
	sc->vtnet_max_vq_pairs = 1;

	if (virtio_with_feature(dev, VIRTIO_NET_F_CTRL_VQ)) {
		sc->vtnet_flags |= VTNET_FLAG_CTRL_VQ;

		if (virtio_with_feature(dev, VIRTIO_NET_F_CTRL_RX))
			sc->vtnet_flags |= VTNET_FLAG_CTRL_RX;
		if (virtio_with_feature(dev, VIRTIO_NET_F_CTRL_VLAN))
			sc->vtnet_flags |= VTNET_FLAG_VLAN_FILTER;
		if (virtio_with_feature(dev, VIRTIO_NET_F_CTRL_MAC_ADDR))
			sc->vtnet_flags |= VTNET_FLAG_CTRL_MAC;

		if (virtio_with_feature(dev, VIRTIO_NET_F_MQ)) {
			sc->vtnet_max_vq_pairs = virtio_read_dev_config_2(dev,
			    offsetof(struct virtio_net_config,
			    max_virtqueue_pairs));
		}
	}

	if (sc->vtnet_max_vq_pairs > 1) {
		int req;

		/*
		 * Limit the maximum number of requested queue pairs to the
		 * number of CPUs and the configured maximum.
		 */
		req = vtnet_tunable_int(sc, "mq_max_pairs", vtnet_mq_max_pairs);
		if (req < 0)
			req = 1;
		if (req == 0)
			req = mp_ncpus;
		if (req > sc->vtnet_max_vq_pairs)
			req = sc->vtnet_max_vq_pairs;
		if (req > mp_ncpus)
			req = mp_ncpus;
		if (req > 1) {
			sc->vtnet_req_vq_pairs = req;
			sc->vtnet_flags |= VTNET_FLAG_MQ;
		}
	}

	return (0);
}

static int
vtnet_init_rxq(struct vtnet_softc *sc, int id)
{
	struct vtnet_rxq *rxq;

	rxq = &sc->vtnet_rxqs[id];

	snprintf(rxq->vtnrx_name, sizeof(rxq->vtnrx_name), "%s-rx%d",
	    device_get_nameunit(sc->vtnet_dev), id);
	mtx_init(&rxq->vtnrx_mtx, rxq->vtnrx_name, NULL, MTX_DEF);

	rxq->vtnrx_sc = sc;
	rxq->vtnrx_id = id;

	rxq->vtnrx_sg = sglist_alloc(sc->vtnet_rx_nsegs, M_NOWAIT);
	if (rxq->vtnrx_sg == NULL)
		return (ENOMEM);

#if defined(INET) || defined(INET6)
	if (vtnet_software_lro(sc)) {
		if (tcp_lro_init_args(&rxq->vtnrx_lro, sc->vtnet_ifp,
		    sc->vtnet_lro_entry_count, sc->vtnet_lro_mbufq_depth) != 0)
			return (ENOMEM);
	}
#endif

	NET_TASK_INIT(&rxq->vtnrx_intrtask, 0, vtnet_rxq_tq_intr, rxq);
	rxq->vtnrx_tq = taskqueue_create(rxq->vtnrx_name, M_NOWAIT,
	    taskqueue_thread_enqueue, &rxq->vtnrx_tq);

	return (rxq->vtnrx_tq == NULL ? ENOMEM : 0);
}

static int
vtnet_init_txq(struct vtnet_softc *sc, int id)
{
	struct vtnet_txq *txq;

	txq = &sc->vtnet_txqs[id];

	snprintf(txq->vtntx_name, sizeof(txq->vtntx_name), "%s-tx%d",
	    device_get_nameunit(sc->vtnet_dev), id);
	mtx_init(&txq->vtntx_mtx, txq->vtntx_name, NULL, MTX_DEF);

	txq->vtntx_sc = sc;
	txq->vtntx_id = id;

	txq->vtntx_sg = sglist_alloc(sc->vtnet_tx_nsegs, M_NOWAIT);
	if (txq->vtntx_sg == NULL)
		return (ENOMEM);

#ifndef VTNET_LEGACY_TX
	txq->vtntx_br = buf_ring_alloc(VTNET_DEFAULT_BUFRING_SIZE, M_DEVBUF,
	    M_NOWAIT, &txq->vtntx_mtx);
	if (txq->vtntx_br == NULL)
		return (ENOMEM);

	TASK_INIT(&txq->vtntx_defrtask, 0, vtnet_txq_tq_deferred, txq);
#endif
	TASK_INIT(&txq->vtntx_intrtask, 0, vtnet_txq_tq_intr, txq);
	txq->vtntx_tq = taskqueue_create(txq->vtntx_name, M_NOWAIT,
	    taskqueue_thread_enqueue, &txq->vtntx_tq);
	if (txq->vtntx_tq == NULL)
		return (ENOMEM);

	return (0);
}

static int
vtnet_alloc_rxtx_queues(struct vtnet_softc *sc)
{
	int i, npairs, error;

	npairs = sc->vtnet_max_vq_pairs;

	sc->vtnet_rxqs = malloc(sizeof(struct vtnet_rxq) * npairs, M_DEVBUF,
	    M_NOWAIT | M_ZERO);
	sc->vtnet_txqs = malloc(sizeof(struct vtnet_txq) * npairs, M_DEVBUF,
	    M_NOWAIT | M_ZERO);
	if (sc->vtnet_rxqs == NULL || sc->vtnet_txqs == NULL)
		return (ENOMEM);

	for (i = 0; i < npairs; i++) {
		error = vtnet_init_rxq(sc, i);
		if (error)
			return (error);
		error = vtnet_init_txq(sc, i);
		if (error)
			return (error);
	}

	vtnet_set_rx_process_limit(sc);
	vtnet_setup_queue_sysctl(sc);

	return (0);
}

static void
vtnet_destroy_rxq(struct vtnet_rxq *rxq)
{

	rxq->vtnrx_sc = NULL;
	rxq->vtnrx_id = -1;

#if defined(INET) || defined(INET6)
	tcp_lro_free(&rxq->vtnrx_lro);
#endif

	if (rxq->vtnrx_sg != NULL) {
		sglist_free(rxq->vtnrx_sg);
		rxq->vtnrx_sg = NULL;
	}

	if (mtx_initialized(&rxq->vtnrx_mtx) != 0)
		mtx_destroy(&rxq->vtnrx_mtx);
}

static void
vtnet_destroy_txq(struct vtnet_txq *txq)
{

	txq->vtntx_sc = NULL;
	txq->vtntx_id = -1;

	if (txq->vtntx_sg != NULL) {
		sglist_free(txq->vtntx_sg);
		txq->vtntx_sg = NULL;
	}

#ifndef VTNET_LEGACY_TX
	if (txq->vtntx_br != NULL) {
		buf_ring_free(txq->vtntx_br, M_DEVBUF);
		txq->vtntx_br = NULL;
	}
#endif

	if (mtx_initialized(&txq->vtntx_mtx) != 0)
		mtx_destroy(&txq->vtntx_mtx);
}

static void
vtnet_free_rxtx_queues(struct vtnet_softc *sc)
{
	int i;

	if (sc->vtnet_rxqs != NULL) {
		for (i = 0; i < sc->vtnet_max_vq_pairs; i++)
			vtnet_destroy_rxq(&sc->vtnet_rxqs[i]);
		free(sc->vtnet_rxqs, M_DEVBUF);
		sc->vtnet_rxqs = NULL;
	}

	if (sc->vtnet_txqs != NULL) {
		for (i = 0; i < sc->vtnet_max_vq_pairs; i++)
			vtnet_destroy_txq(&sc->vtnet_txqs[i]);
		free(sc->vtnet_txqs, M_DEVBUF);
		sc->vtnet_txqs = NULL;
	}
}

static int
vtnet_alloc_rx_filters(struct vtnet_softc *sc)
{

	if (sc->vtnet_flags & VTNET_FLAG_CTRL_RX) {
		sc->vtnet_mac_filter = malloc(sizeof(struct vtnet_mac_filter),
		    M_DEVBUF, M_NOWAIT | M_ZERO);
		if (sc->vtnet_mac_filter == NULL)
			return (ENOMEM);
	}

	if (sc->vtnet_flags & VTNET_FLAG_VLAN_FILTER) {
		sc->vtnet_vlan_filter = malloc(sizeof(uint32_t) *
		    VTNET_VLAN_FILTER_NWORDS, M_DEVBUF, M_NOWAIT | M_ZERO);
		if (sc->vtnet_vlan_filter == NULL)
			return (ENOMEM);
	}

	return (0);
}

static void
vtnet_free_rx_filters(struct vtnet_softc *sc)
{

	if (sc->vtnet_mac_filter != NULL) {
		free(sc->vtnet_mac_filter, M_DEVBUF);
		sc->vtnet_mac_filter = NULL;
	}

	if (sc->vtnet_vlan_filter != NULL) {
		free(sc->vtnet_vlan_filter, M_DEVBUF);
		sc->vtnet_vlan_filter = NULL;
	}
}

static int
vtnet_alloc_virtqueues(struct vtnet_softc *sc)
{
	device_t dev;
	struct vq_alloc_info *info;
	struct vtnet_rxq *rxq;
	struct vtnet_txq *txq;
	int i, idx, flags, nvqs, error;

	dev = sc->vtnet_dev;
	flags = 0;

	nvqs = sc->vtnet_max_vq_pairs * 2;
	if (sc->vtnet_flags & VTNET_FLAG_CTRL_VQ)
		nvqs++;

	info = malloc(sizeof(struct vq_alloc_info) * nvqs, M_TEMP, M_NOWAIT);
	if (info == NULL)
		return (ENOMEM);

	for (i = 0, idx = 0; i < sc->vtnet_req_vq_pairs; i++, idx += 2) {
		rxq = &sc->vtnet_rxqs[i];
		VQ_ALLOC_INFO_INIT(&info[idx], sc->vtnet_rx_nsegs,
		    vtnet_rx_vq_intr, rxq, &rxq->vtnrx_vq,
		    "%s-rx%d", device_get_nameunit(dev), rxq->vtnrx_id);

		txq = &sc->vtnet_txqs[i];
		VQ_ALLOC_INFO_INIT(&info[idx+1], sc->vtnet_tx_nsegs,
		    vtnet_tx_vq_intr, txq, &txq->vtntx_vq,
		    "%s-tx%d", device_get_nameunit(dev), txq->vtntx_id);
	}

	/* These queues will not be used so allocate the minimum resources. */
	for (/**/; i < sc->vtnet_max_vq_pairs; i++, idx += 2) {
		rxq = &sc->vtnet_rxqs[i];
		VQ_ALLOC_INFO_INIT(&info[idx], 0, NULL, rxq, &rxq->vtnrx_vq,
		    "%s-rx%d", device_get_nameunit(dev), rxq->vtnrx_id);

		txq = &sc->vtnet_txqs[i];
		VQ_ALLOC_INFO_INIT(&info[idx+1], 0, NULL, txq, &txq->vtntx_vq,
		    "%s-tx%d", device_get_nameunit(dev), txq->vtntx_id);
	}

	if (sc->vtnet_flags & VTNET_FLAG_CTRL_VQ) {
		VQ_ALLOC_INFO_INIT(&info[idx], 0, NULL, NULL,
		    &sc->vtnet_ctrl_vq, "%s ctrl", device_get_nameunit(dev));
	}

	/*
	 * TODO: Enable interrupt binding if this is multiqueue. This will
	 * only matter when per-virtqueue MSIX is available.
	 */
	if (sc->vtnet_flags & VTNET_FLAG_MQ)
		flags |= 0;

	error = virtio_alloc_virtqueues(dev, flags, nvqs, info);
	free(info, M_TEMP);

	return (error);
}

static int
vtnet_alloc_interface(struct vtnet_softc *sc)
{
	device_t dev;
	struct ifnet *ifp;

	dev = sc->vtnet_dev;

	ifp = if_alloc(IFT_ETHER);
	if (ifp == NULL)
		return (ENOMEM);

	sc->vtnet_ifp = ifp;
	ifp->if_softc = sc;
	if_initname(ifp, device_get_name(dev), device_get_unit(dev));

	return (0);
}

static int
vtnet_setup_interface(struct vtnet_softc *sc)
{
	device_t dev;
	struct pfil_head_args pa;
	struct ifnet *ifp;

	dev = sc->vtnet_dev;
	ifp = sc->vtnet_ifp;

	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST |
	    IFF_KNOWSEPOCH;
	ifp->if_baudrate = IF_Gbps(10);
	ifp->if_init = vtnet_init;
	ifp->if_ioctl = vtnet_ioctl;
	ifp->if_get_counter = vtnet_get_counter;
#ifndef VTNET_LEGACY_TX
	ifp->if_transmit = vtnet_txq_mq_start;
	ifp->if_qflush = vtnet_qflush;
#else
	struct virtqueue *vq = sc->vtnet_txqs[0].vtntx_vq;
	ifp->if_start = vtnet_start;
	IFQ_SET_MAXLEN(&ifp->if_snd, virtqueue_size(vq) - 1);
	ifp->if_snd.ifq_drv_maxlen = virtqueue_size(vq) - 1;
	IFQ_SET_READY(&ifp->if_snd);
#endif

	vtnet_get_macaddr(sc);

	if (virtio_with_feature(dev, VIRTIO_NET_F_STATUS))
		ifp->if_capabilities |= IFCAP_LINKSTATE;

	ifmedia_init(&sc->vtnet_media, 0, vtnet_ifmedia_upd, vtnet_ifmedia_sts);
	ifmedia_add(&sc->vtnet_media, IFM_ETHER | IFM_AUTO, 0, NULL);
	ifmedia_set(&sc->vtnet_media, IFM_ETHER | IFM_AUTO);

	if (virtio_with_feature(dev, VIRTIO_NET_F_CSUM)) {
		int gso;

		ifp->if_capabilities |= IFCAP_TXCSUM | IFCAP_TXCSUM_IPV6;

		gso = virtio_with_feature(dev, VIRTIO_NET_F_GSO);
		if (gso || virtio_with_feature(dev, VIRTIO_NET_F_HOST_TSO4))
			ifp->if_capabilities |= IFCAP_TSO4;
		if (gso || virtio_with_feature(dev, VIRTIO_NET_F_HOST_TSO6))
			ifp->if_capabilities |= IFCAP_TSO6;
		if (gso || virtio_with_feature(dev, VIRTIO_NET_F_HOST_ECN))
			sc->vtnet_flags |= VTNET_FLAG_TSO_ECN;

		if (ifp->if_capabilities & (IFCAP_TSO4 | IFCAP_TSO6)) {
			int tso_maxlen;

			ifp->if_capabilities |= IFCAP_VLAN_HWTSO;

			tso_maxlen = vtnet_tunable_int(sc, "tso_maxlen",
			    vtnet_tso_maxlen);
			ifp->if_hw_tsomax = tso_maxlen -
			    (ETHER_HDR_LEN + ETHER_VLAN_ENCAP_LEN);
			ifp->if_hw_tsomaxsegcount = sc->vtnet_tx_nsegs - 1;
			ifp->if_hw_tsomaxsegsize = PAGE_SIZE;
		}
	}

	if (virtio_with_feature(dev, VIRTIO_NET_F_GUEST_CSUM)) {
		ifp->if_capabilities |= IFCAP_RXCSUM;
#ifdef notyet
		/* BMV: Rx checksums not distinguished between IPv4 and IPv6. */
		ifp->if_capabilities |= IFCAP_RXCSUM_IPV6;
#endif

		if (vtnet_tunable_int(sc, "fixup_needs_csum",
		    vtnet_fixup_needs_csum) != 0)
			sc->vtnet_flags |= VTNET_FLAG_FIXUP_NEEDS_CSUM;

		/* Support either "hardware" or software LRO. */
		ifp->if_capabilities |= IFCAP_LRO;
	}

	if (ifp->if_capabilities & (IFCAP_HWCSUM | IFCAP_HWCSUM_IPV6)) {
		/*
		 * VirtIO does not support VLAN tagging, but we can fake
		 * it by inserting and removing the 802.1Q header during
		 * transmit and receive. We are then able to do checksum
		 * offloading of VLAN frames.
		 */
		ifp->if_capabilities |=
		    IFCAP_VLAN_HWTAGGING | IFCAP_VLAN_HWCSUM;
	}

	if (sc->vtnet_max_mtu >= ETHERMTU_JUMBO)
		ifp->if_capabilities |= IFCAP_JUMBO_MTU;
	ifp->if_capabilities |= IFCAP_VLAN_MTU;

	/*
	 * Capabilities after here are not enabled by default.
	 */
	ifp->if_capenable = ifp->if_capabilities;

	if (sc->vtnet_flags & VTNET_FLAG_VLAN_FILTER) {
		ifp->if_capabilities |= IFCAP_VLAN_HWFILTER;

		sc->vtnet_vlan_attach = EVENTHANDLER_REGISTER(vlan_config,
		    vtnet_register_vlan, sc, EVENTHANDLER_PRI_FIRST);
		sc->vtnet_vlan_detach = EVENTHANDLER_REGISTER(vlan_unconfig,
		    vtnet_unregister_vlan, sc, EVENTHANDLER_PRI_FIRST);
	}

	ether_ifattach(ifp, sc->vtnet_hwaddr);

	/* Tell the upper layer(s) we support long frames. */
	ifp->if_hdrlen = sizeof(struct ether_vlan_header);

	DEBUGNET_SET(ifp, vtnet);

	pa.pa_version = PFIL_VERSION;
	pa.pa_flags = PFIL_IN;
	pa.pa_type = PFIL_TYPE_ETHERNET;
	pa.pa_headname = ifp->if_xname;
	sc->vtnet_pfil = pfil_head_register(&pa);

	return (0);
}

static int
vtnet_rx_cluster_size(struct vtnet_softc *sc, int mtu)
{
	int framesz;

	if (sc->vtnet_flags & VTNET_FLAG_MRG_RXBUFS)
		return (MJUMPAGESIZE);
	else if (sc->vtnet_flags & VTNET_FLAG_LRO_NOMRG)
		return (MCLBYTES);

	/*
	 * Try to scale the receive mbuf cluster size from the MTU. We
	 * could also use the VQ size to influence the selected size,
	 * but that would only matter for very small queues.
	 */
	if (vtnet_modern(sc)) {
		MPASS(sc->vtnet_hdr_size == sizeof(struct virtio_net_hdr_v1));
		framesz = sizeof(struct virtio_net_hdr_v1);
	} else
		framesz = sizeof(struct vtnet_rx_header);
	framesz += sizeof(struct ether_vlan_header) + mtu;

	if (framesz <= MCLBYTES)
		return (MCLBYTES);
	else if (framesz <= MJUMPAGESIZE)
		return (MJUMPAGESIZE);
	else if (framesz <= MJUM9BYTES)
		return (MJUM9BYTES);

	/* Sane default; avoid 16KB clusters. */
	return (MCLBYTES);
}

static int
vtnet_ioctl_mtu(struct vtnet_softc *sc, u_int mtu)
{
	struct ifnet *ifp;
	int clustersz;

	ifp = sc->vtnet_ifp;
	VTNET_CORE_LOCK_ASSERT(sc);

	if (ifp->if_mtu == mtu)
		return (0);
	else if (mtu < ETHERMIN || mtu > sc->vtnet_max_mtu)
		return (EINVAL);

	ifp->if_mtu = mtu;
	clustersz = vtnet_rx_cluster_size(sc, mtu);

	if (clustersz != sc->vtnet_rx_clustersz &&
	    ifp->if_drv_flags & IFF_DRV_RUNNING) {
		ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
		vtnet_init_locked(sc, 0);
	}

	return (0);
}

static int
vtnet_ioctl_ifflags(struct vtnet_softc *sc)
{
	struct ifnet *ifp;
	int drv_running;

	ifp = sc->vtnet_ifp;
	drv_running = (ifp->if_drv_flags & IFF_DRV_RUNNING) != 0;

	VTNET_CORE_LOCK_ASSERT(sc);

	if ((ifp->if_flags & IFF_UP) == 0) {
		if (drv_running)
			vtnet_stop(sc);
		goto out;
	}

	if (!drv_running) {
		vtnet_init_locked(sc, 0);
		goto out;
	}

	if ((ifp->if_flags ^ sc->vtnet_if_flags) &
	    (IFF_PROMISC | IFF_ALLMULTI)) {
		if ((sc->vtnet_flags & VTNET_FLAG_CTRL_RX) == 0)
			return (ENOTSUP);
		vtnet_rx_filter(sc);
	}

out:
	sc->vtnet_if_flags = ifp->if_flags;
	return (0);
}

static int
vtnet_ioctl_multi(struct vtnet_softc *sc)
{
	struct ifnet *ifp;

	ifp = sc->vtnet_ifp;

	VTNET_CORE_LOCK_ASSERT(sc);

	if (sc->vtnet_flags & VTNET_FLAG_CTRL_RX &&
	    ifp->if_drv_flags & IFF_DRV_RUNNING)
		vtnet_rx_filter_mac(sc);

	return (0);
}

static int
vtnet_ioctl_ifcap(struct vtnet_softc *sc, struct ifreq *ifr)
{
	struct ifnet *ifp;
	int mask, reinit, update;

	ifp = sc->vtnet_ifp;
	mask = (ifr->ifr_reqcap & ifp->if_capabilities) ^ ifp->if_capenable;
	reinit = update = 0;

	VTNET_CORE_LOCK_ASSERT(sc);

	if (mask & IFCAP_TXCSUM)
		ifp->if_capenable ^= IFCAP_TXCSUM;
	if (mask & IFCAP_TXCSUM_IPV6)
		ifp->if_capenable ^= IFCAP_TXCSUM_IPV6;
	if (mask & IFCAP_TSO4)
		ifp->if_capenable ^= IFCAP_TSO4;
	if (mask & IFCAP_TSO6)
		ifp->if_capenable ^= IFCAP_TSO6;

	if (mask & (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6 | IFCAP_LRO)) {
		/*
		 * These Rx features require the negotiated features to
		 * be updated. Avoid a full reinit if possible.
		 */
		if (sc->vtnet_features & VIRTIO_NET_F_CTRL_GUEST_OFFLOADS)
			update = 1;
		else
			reinit = 1;

		/* BMV: Avoid needless renegotiation for just software LRO. */
		if ((mask & (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6 | IFCAP_LRO)) ==
		    IFCAP_LRO && vtnet_software_lro(sc))
			reinit = update = 0;

		if (mask & IFCAP_RXCSUM)
			ifp->if_capenable ^= IFCAP_RXCSUM;
		if (mask & IFCAP_RXCSUM_IPV6)
			ifp->if_capenable ^= IFCAP_RXCSUM_IPV6;
		if (mask & IFCAP_LRO)
			ifp->if_capenable ^= IFCAP_LRO;

		/*
		 * VirtIO does not distinguish between IPv4 and IPv6 checksums
		 * so treat them as a pair. Guest TSO (LRO) requires receive
		 * checksums.
		 */
		if (ifp->if_capenable & (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6)) {
			ifp->if_capenable |= IFCAP_RXCSUM;
#ifdef notyet
			ifp->if_capenable |= IFCAP_RXCSUM_IPV6;
#endif
		} else
			ifp->if_capenable &=
			    ~(IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6 | IFCAP_LRO);
	}

	if (mask & IFCAP_VLAN_HWFILTER) {
		/* These Rx features require renegotiation. */
		reinit = 1;

		if (mask & IFCAP_VLAN_HWFILTER)
			ifp->if_capenable ^= IFCAP_VLAN_HWFILTER;
	}

	if (mask & IFCAP_VLAN_HWTSO)
		ifp->if_capenable ^= IFCAP_VLAN_HWTSO;
	if (mask & IFCAP_VLAN_HWTAGGING)
		ifp->if_capenable ^= IFCAP_VLAN_HWTAGGING;

	if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
		if (reinit) {
			ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
			vtnet_init_locked(sc, 0);
		} else if (update)
			vtnet_update_rx_offloads(sc);
	}

	return (0);
}

static int
vtnet_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
{
	struct vtnet_softc *sc;
	struct ifreq *ifr;
	int error;

	sc = ifp->if_softc;
	ifr = (struct ifreq *) data;
	error = 0;

	switch (cmd) {
	case SIOCSIFMTU:
		VTNET_CORE_LOCK(sc);
		error = vtnet_ioctl_mtu(sc, ifr->ifr_mtu);
		VTNET_CORE_UNLOCK(sc);
		break;

	case SIOCSIFFLAGS:
		VTNET_CORE_LOCK(sc);
		error = vtnet_ioctl_ifflags(sc);
		VTNET_CORE_UNLOCK(sc);
		break;

	case SIOCADDMULTI:
	case SIOCDELMULTI:
		VTNET_CORE_LOCK(sc);
		error = vtnet_ioctl_multi(sc);
		VTNET_CORE_UNLOCK(sc);
		break;

	case SIOCSIFMEDIA:
	case SIOCGIFMEDIA:
		error = ifmedia_ioctl(ifp, ifr, &sc->vtnet_media, cmd);
		break;

	case SIOCSIFCAP:
		VTNET_CORE_LOCK(sc);
		error = vtnet_ioctl_ifcap(sc, ifr);
		VTNET_CORE_UNLOCK(sc);
		VLAN_CAPABILITIES(ifp);
		break;

	default:
		error = ether_ioctl(ifp, cmd, data);
		break;
	}

	VTNET_CORE_LOCK_ASSERT_NOTOWNED(sc);

	return (error);
}

static int
vtnet_rxq_populate(struct vtnet_rxq *rxq)
{
	struct virtqueue *vq;
	int nbufs, error;

#ifdef DEV_NETMAP
	error = vtnet_netmap_rxq_populate(rxq);
	if (error >= 0)
		return (error);
#endif  /* DEV_NETMAP */

	vq = rxq->vtnrx_vq;
	error = ENOSPC;

	for (nbufs = 0; !virtqueue_full(vq); nbufs++) {
		error = vtnet_rxq_new_buf(rxq);
		if (error)
			break;
	}

	if (nbufs > 0) {
		virtqueue_notify(vq);
		/*
		 * EMSGSIZE signifies the virtqueue did not have enough
		 * entries available to hold the last mbuf. This is not
		 * an error.
		 */
		if (error == EMSGSIZE)
			error = 0;
	}

	return (error);
}

static void
vtnet_rxq_free_mbufs(struct vtnet_rxq *rxq)
{
	struct virtqueue *vq;
	struct mbuf *m;
	int last;
#ifdef DEV_NETMAP
	struct netmap_kring *kring = netmap_kring_on(NA(rxq->vtnrx_sc->vtnet_ifp),
							rxq->vtnrx_id, NR_RX);
#else  /* !DEV_NETMAP */
	void *kring = NULL;
#endif /* !DEV_NETMAP */

	vq = rxq->vtnrx_vq;
	last = 0;

	while ((m = virtqueue_drain(vq, &last)) != NULL) {
		if (kring == NULL)
			m_freem(m);
	}

	KASSERT(virtqueue_empty(vq),
	    ("%s: mbufs remaining in rx queue %p", __func__, rxq));
}

static struct mbuf *
vtnet_rx_alloc_buf(struct vtnet_softc *sc, int nbufs, struct mbuf **m_tailp)
{
	struct mbuf *m_head, *m_tail, *m;
	int i, size;

	m_head = NULL;
	size = sc->vtnet_rx_clustersz;

	KASSERT(nbufs == 1 || sc->vtnet_flags & VTNET_FLAG_LRO_NOMRG,
	    ("%s: mbuf %d chain requested without LRO_NOMRG", __func__, nbufs));

	for (i = 0; i < nbufs; i++) {
		m = m_getjcl(M_NOWAIT, MT_DATA, i == 0 ? M_PKTHDR : 0, size);
		if (m == NULL) {
			sc->vtnet_stats.mbuf_alloc_failed++;
			m_freem(m_head);
			return (NULL);
		}

		m->m_len = size;
		if (m_head != NULL) {
			m_tail->m_next = m;
			m_tail = m;
		} else
			m_head = m_tail = m;
	}

	if (m_tailp != NULL)
		*m_tailp = m_tail;

	return (m_head);
}

/*
 * Slow path for when LRO without mergeable buffers is negotiated.
 */
static int
vtnet_rxq_replace_lro_nomrg_buf(struct vtnet_rxq *rxq, struct mbuf *m0,
    int len0)
{
	struct vtnet_softc *sc;
	struct mbuf *m, *m_prev, *m_new, *m_tail;
	int len, clustersz, nreplace, error;

	sc = rxq->vtnrx_sc;
	clustersz = sc->vtnet_rx_clustersz;

	m_prev = NULL;
	m_tail = NULL;
	nreplace = 0;

	m = m0;
	len = len0;

	/*
	 * Since these mbuf chains are so large, avoid allocating a complete
	 * replacement when the received frame did not consume the entire
	 * chain. Unused mbufs are moved to the tail of the replacement mbuf.
	 */
	while (len > 0) {
		if (m == NULL) {
			sc->vtnet_stats.rx_frame_too_large++;
			return (EMSGSIZE);
		}

		/*
		 * Every mbuf should have the expected cluster size since that
		 * is also used to allocate the replacements.
		 */
		KASSERT(m->m_len == clustersz,
		    ("%s: mbuf size %d not expected cluster size %d", __func__,
		    m->m_len, clustersz));

		m->m_len = MIN(m->m_len, len);
		len -= m->m_len;

		m_prev = m;
		m = m->m_next;
		nreplace++;
	}

	KASSERT(nreplace > 0 && nreplace <= sc->vtnet_rx_nmbufs,
	    ("%s: invalid replacement mbuf count %d max %d", __func__,
	    nreplace, sc->vtnet_rx_nmbufs));

	m_new = vtnet_rx_alloc_buf(sc, nreplace, &m_tail);
	if (m_new == NULL) {
		m_prev->m_len = clustersz;
		return (ENOBUFS);
	}

	/*
	 * Move any unused mbufs from the received mbuf chain onto the
	 * end of the replacement chain.
	 */
	if (m_prev->m_next != NULL) {
		m_tail->m_next = m_prev->m_next;
		m_prev->m_next = NULL;
	}

	error = vtnet_rxq_enqueue_buf(rxq, m_new);
	if (error) {
		/*
		 * The replacement is suppose to be an copy of the one
		 * dequeued so this is a very unexpected error.
		 *
		 * Restore the m0 chain to the original state if it was
		 * modified so we can then discard it.
		 */
		if (m_tail->m_next != NULL) {
			m_prev->m_next = m_tail->m_next;
			m_tail->m_next = NULL;
		}
		m_prev->m_len = clustersz;
		sc->vtnet_stats.rx_enq_replacement_failed++;
		m_freem(m_new);
	}

	return (error);
}

static int
vtnet_rxq_replace_buf(struct vtnet_rxq *rxq, struct mbuf *m, int len)
{
	struct vtnet_softc *sc;
	struct mbuf *m_new;
	int error;

	sc = rxq->vtnrx_sc;

	if (sc->vtnet_flags & VTNET_FLAG_LRO_NOMRG)
		return (vtnet_rxq_replace_lro_nomrg_buf(rxq, m, len));

	MPASS(m->m_next == NULL);
	if (m->m_len < len)
		return (EMSGSIZE);

	m_new = vtnet_rx_alloc_buf(sc, 1, NULL);
	if (m_new == NULL)
		return (ENOBUFS);

	error = vtnet_rxq_enqueue_buf(rxq, m_new);
	if (error) {
		sc->vtnet_stats.rx_enq_replacement_failed++;
		m_freem(m_new);
	} else
		m->m_len = len;

	return (error);
}

static int
vtnet_rxq_enqueue_buf(struct vtnet_rxq *rxq, struct mbuf *m)
{
	struct vtnet_softc *sc;
	struct sglist *sg;
	int header_inlined, error;

	sc = rxq->vtnrx_sc;
	sg = rxq->vtnrx_sg;

	KASSERT(m->m_next == NULL || sc->vtnet_flags & VTNET_FLAG_LRO_NOMRG,
	    ("%s: mbuf chain without LRO_NOMRG", __func__));
	VTNET_RXQ_LOCK_ASSERT(rxq);

	sglist_reset(sg);
	header_inlined = vtnet_modern(sc) ||
	    (sc->vtnet_flags & VTNET_FLAG_MRG_RXBUFS) != 0; /* TODO: ANY_LAYOUT */

	if (header_inlined)
		error = sglist_append_mbuf(sg, m);
	else {
		struct vtnet_rx_header *rxhdr =
		    mtod(m, struct vtnet_rx_header *);
		MPASS(sc->vtnet_hdr_size == sizeof(struct virtio_net_hdr));

		/* Append the header and remaining mbuf data. */
		error = sglist_append(sg, &rxhdr->vrh_hdr, sc->vtnet_hdr_size);
		if (error)
			return (error);
		error = sglist_append(sg, &rxhdr[1],
		    m->m_len - sizeof(struct vtnet_rx_header));
		if (error)
			return (error);

		if (m->m_next != NULL)
			error = sglist_append_mbuf(sg, m->m_next);
	}

	if (error)
		return (error);

	return (virtqueue_enqueue(rxq->vtnrx_vq, m, sg, 0, sg->sg_nseg));
}

static int
vtnet_rxq_new_buf(struct vtnet_rxq *rxq)
{
	struct vtnet_softc *sc;
	struct mbuf *m;
	int error;

	sc = rxq->vtnrx_sc;

	m = vtnet_rx_alloc_buf(sc, sc->vtnet_rx_nmbufs, NULL);
	if (m == NULL)
		return (ENOBUFS);

	error = vtnet_rxq_enqueue_buf(rxq, m);
	if (error)
		m_freem(m);

	return (error);
}

static int
vtnet_rxq_csum_needs_csum(struct vtnet_rxq *rxq, struct mbuf *m, uint16_t etype,
    int hoff, struct virtio_net_hdr *hdr)
{
	struct vtnet_softc *sc;
	int error;

	sc = rxq->vtnrx_sc;

	/*
	 * NEEDS_CSUM corresponds to Linux's CHECKSUM_PARTIAL, but FreeBSD does
	 * not have an analogous CSUM flag. The checksum has been validated,
	 * but is incomplete (TCP/UDP pseudo header).
	 *
	 * The packet is likely from another VM on the same host that itself
	 * performed checksum offloading so Tx/Rx is basically a memcpy and
	 * the checksum has little value.
	 *
	 * Default to receiving the packet as-is for performance reasons, but
	 * this can cause issues if the packet is to be forwarded because it
	 * does not contain a valid checksum. This patch may be helpful:
	 * https://reviews.freebsd.org/D6611. In the meantime, have the driver
	 * compute the checksum if requested.
	 *
	 * BMV: Need to add an CSUM_PARTIAL flag?
	 */
	if ((sc->vtnet_flags & VTNET_FLAG_FIXUP_NEEDS_CSUM) == 0) {
		error = vtnet_rxq_csum_data_valid(rxq, m, etype, hoff, hdr);
		return (error);
	}

	/*
	 * Compute the checksum in the driver so the packet will contain a
	 * valid checksum. The checksum is at csum_offset from csum_start.
	 */
	switch (etype) {
#if defined(INET) || defined(INET6)
	case ETHERTYPE_IP:
	case ETHERTYPE_IPV6: {
		int csum_off, csum_end;
		uint16_t csum;

		csum_off = hdr->csum_start + hdr->csum_offset;
		csum_end = csum_off + sizeof(uint16_t);

		/* Assume checksum will be in the first mbuf. */
		if (m->m_len < csum_end || m->m_pkthdr.len < csum_end)
			return (1);

		/*
		 * Like in_delayed_cksum()/in6_delayed_cksum(), compute the
		 * checksum and write it at the specified offset. We could
		 * try to verify the packet: csum_start should probably
		 * correspond to the start of the TCP/UDP header.
		 *
		 * BMV: Need to properly handle UDP with zero checksum. Is
		 * the IPv4 header checksum implicitly validated?
		 */
		csum = in_cksum_skip(m, m->m_pkthdr.len, hdr->csum_start);
		*(uint16_t *)(mtodo(m, csum_off)) = csum;
		m->m_pkthdr.csum_flags |= CSUM_DATA_VALID | CSUM_PSEUDO_HDR;
		m->m_pkthdr.csum_data = 0xFFFF;
		break;
	}
#endif
	default:
		sc->vtnet_stats.rx_csum_bad_ethtype++;
		return (1);
	}

	return (0);
}

static int
vtnet_rxq_csum_data_valid(struct vtnet_rxq *rxq, struct mbuf *m,
    uint16_t etype, int hoff, struct virtio_net_hdr *hdr __unused)
{
#if 0
	struct vtnet_softc *sc;
#endif
	int protocol;

#if 0
	sc = rxq->vtnrx_sc;
#endif

	switch (etype) {
#if defined(INET)
	case ETHERTYPE_IP:
		if (__predict_false(m->m_len < hoff + sizeof(struct ip)))
			protocol = IPPROTO_DONE;
		else {
			struct ip *ip = (struct ip *)(m->m_data + hoff);
			protocol = ip->ip_p;
		}
		break;
#endif
#if defined(INET6)
	case ETHERTYPE_IPV6:
		if (__predict_false(m->m_len < hoff + sizeof(struct ip6_hdr))
		    || ip6_lasthdr(m, hoff, IPPROTO_IPV6, &protocol) < 0)
			protocol = IPPROTO_DONE;
		break;
#endif
	default:
		protocol = IPPROTO_DONE;
		break;
	}

	switch (protocol) {
	case IPPROTO_TCP:
	case IPPROTO_UDP:
		m->m_pkthdr.csum_flags |= CSUM_DATA_VALID | CSUM_PSEUDO_HDR;
		m->m_pkthdr.csum_data = 0xFFFF;
		break;
	default:
		/*
		 * FreeBSD does not support checksum offloading of this
		 * protocol. Let the stack re-verify the checksum later
		 * if the protocol is supported.
		 */
#if 0
		if_printf(sc->vtnet_ifp,
		    "%s: checksum offload of unsupported protocol "
		    "etype=%#x protocol=%d csum_start=%d csum_offset=%d\n",
		    __func__, etype, protocol, hdr->csum_start,
		    hdr->csum_offset);
#endif
		break;
	}

	return (0);
}

static int
vtnet_rxq_csum(struct vtnet_rxq *rxq, struct mbuf *m,
    struct virtio_net_hdr *hdr)
{
	const struct ether_header *eh;
	int hoff;
	uint16_t etype;

	eh = mtod(m, const struct ether_header *);
	etype = ntohs(eh->ether_type);
	if (etype == ETHERTYPE_VLAN) {
		/* TODO BMV: Handle QinQ. */
		const struct ether_vlan_header *evh =
		    mtod(m, const struct ether_vlan_header *);
		etype = ntohs(evh->evl_proto);
		hoff = sizeof(struct ether_vlan_header);
	} else
		hoff = sizeof(struct ether_header);

	if (hdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM)
		return (vtnet_rxq_csum_needs_csum(rxq, m, etype, hoff, hdr));
	else /* VIRTIO_NET_HDR_F_DATA_VALID */
		return (vtnet_rxq_csum_data_valid(rxq, m, etype, hoff, hdr));
}

static void
vtnet_rxq_discard_merged_bufs(struct vtnet_rxq *rxq, int nbufs)
{
	struct mbuf *m;

	while (--nbufs > 0) {
		m = virtqueue_dequeue(rxq->vtnrx_vq, NULL);
		if (m == NULL)
			break;
		vtnet_rxq_discard_buf(rxq, m);
	}
}

static void
vtnet_rxq_discard_buf(struct vtnet_rxq *rxq, struct mbuf *m)
{
	int error __diagused;

	/*
	 * Requeue the discarded mbuf. This should always be successful
	 * since it was just dequeued.
	 */
	error = vtnet_rxq_enqueue_buf(rxq, m);
	KASSERT(error == 0,
	    ("%s: cannot requeue discarded mbuf %d", __func__, error));
}

static int
vtnet_rxq_merged_eof(struct vtnet_rxq *rxq, struct mbuf *m_head, int nbufs)
{
	struct vtnet_softc *sc;
	struct virtqueue *vq;
	struct mbuf *m_tail;

	sc = rxq->vtnrx_sc;
	vq = rxq->vtnrx_vq;
	m_tail = m_head;

	while (--nbufs > 0) {
		struct mbuf *m;
		uint32_t len;

		m = virtqueue_dequeue(vq, &len);
		if (m == NULL) {
			rxq->vtnrx_stats.vrxs_ierrors++;
			goto fail;
		}

		if (vtnet_rxq_new_buf(rxq) != 0) {
			rxq->vtnrx_stats.vrxs_iqdrops++;
			vtnet_rxq_discard_buf(rxq, m);
			if (nbufs > 1)
				vtnet_rxq_discard_merged_bufs(rxq, nbufs);
			goto fail;
		}

		if (m->m_len < len)
			len = m->m_len;

		m->m_len = len;
		m->m_flags &= ~M_PKTHDR;

		m_head->m_pkthdr.len += len;
		m_tail->m_next = m;
		m_tail = m;
	}

	return (0);

fail:
	sc->vtnet_stats.rx_mergeable_failed++;
	m_freem(m_head);

	return (1);
}

#if defined(INET) || defined(INET6)
static int
vtnet_lro_rx(struct vtnet_rxq *rxq, struct mbuf *m)
{
	struct lro_ctrl *lro;

	lro = &rxq->vtnrx_lro;

	if (lro->lro_mbuf_max != 0) {
		tcp_lro_queue_mbuf(lro, m);
		return (0);
	}

	return (tcp_lro_rx(lro, m, 0));
}
#endif

static void
vtnet_rxq_input(struct vtnet_rxq *rxq, struct mbuf *m,
    struct virtio_net_hdr *hdr)
{
	struct vtnet_softc *sc;
	struct ifnet *ifp;

	sc = rxq->vtnrx_sc;
	ifp = sc->vtnet_ifp;

	if (ifp->if_capenable & IFCAP_VLAN_HWTAGGING) {
		struct ether_header *eh = mtod(m, struct ether_header *);
		if (eh->ether_type == htons(ETHERTYPE_VLAN)) {
			vtnet_vlan_tag_remove(m);
			/*
			 * With the 802.1Q header removed, update the
			 * checksum starting location accordingly.
			 */
			if (hdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM)
				hdr->csum_start -= ETHER_VLAN_ENCAP_LEN;
		}
	}

	m->m_pkthdr.flowid = rxq->vtnrx_id;
	M_HASHTYPE_SET(m, M_HASHTYPE_OPAQUE);

	if (hdr->flags &
	    (VIRTIO_NET_HDR_F_NEEDS_CSUM | VIRTIO_NET_HDR_F_DATA_VALID)) {
		if (vtnet_rxq_csum(rxq, m, hdr) == 0)
			rxq->vtnrx_stats.vrxs_csum++;
		else
			rxq->vtnrx_stats.vrxs_csum_failed++;
	}

	if (hdr->gso_size != 0) {
		switch (hdr->gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
		case VIRTIO_NET_HDR_GSO_TCPV4:
		case VIRTIO_NET_HDR_GSO_TCPV6:
			m->m_pkthdr.lro_nsegs =
			    howmany(m->m_pkthdr.len, hdr->gso_size);
			rxq->vtnrx_stats.vrxs_host_lro++;
			break;
		}
	}

	rxq->vtnrx_stats.vrxs_ipackets++;
	rxq->vtnrx_stats.vrxs_ibytes += m->m_pkthdr.len;

#if defined(INET) || defined(INET6)
	if (vtnet_software_lro(sc) && ifp->if_capenable & IFCAP_LRO) {
		if (vtnet_lro_rx(rxq, m) == 0)
			return;
	}
#endif

	(*ifp->if_input)(ifp, m);
}

static int
vtnet_rxq_eof(struct vtnet_rxq *rxq)
{
	struct virtio_net_hdr lhdr, *hdr;
	struct vtnet_softc *sc;
	struct ifnet *ifp;
	struct virtqueue *vq;
	int deq, count;

	sc = rxq->vtnrx_sc;
	vq = rxq->vtnrx_vq;
	ifp = sc->vtnet_ifp;
	deq = 0;
	count = sc->vtnet_rx_process_limit;

	VTNET_RXQ_LOCK_ASSERT(rxq);

	while (count-- > 0) {
		struct mbuf *m;
		uint32_t len, nbufs, adjsz;

		m = virtqueue_dequeue(vq, &len);
		if (m == NULL)
			break;
		deq++;

		if (len < sc->vtnet_hdr_size + ETHER_HDR_LEN) {
			rxq->vtnrx_stats.vrxs_ierrors++;
			vtnet_rxq_discard_buf(rxq, m);
			continue;
		}

		if (sc->vtnet_flags & VTNET_FLAG_MRG_RXBUFS) {
			struct virtio_net_hdr_mrg_rxbuf *mhdr =
			    mtod(m, struct virtio_net_hdr_mrg_rxbuf *);
			kmsan_mark(mhdr, sizeof(*mhdr), KMSAN_STATE_INITED);
			nbufs = vtnet_htog16(sc, mhdr->num_buffers);
			adjsz = sizeof(struct virtio_net_hdr_mrg_rxbuf);
		} else if (vtnet_modern(sc)) {
			nbufs = 1; /* num_buffers is always 1 */
			adjsz = sizeof(struct virtio_net_hdr_v1);
		} else {
			nbufs = 1;
			adjsz = sizeof(struct vtnet_rx_header);
			/*
			 * Account for our gap between the header and start of
			 * data to keep the segments separated.
			 */
			len += VTNET_RX_HEADER_PAD;
		}

		if (vtnet_rxq_replace_buf(rxq, m, len) != 0) {
			rxq->vtnrx_stats.vrxs_iqdrops++;
			vtnet_rxq_discard_buf(rxq, m);
			if (nbufs > 1)
				vtnet_rxq_discard_merged_bufs(rxq, nbufs);
			continue;
		}

		m->m_pkthdr.len = len;
		m->m_pkthdr.rcvif = ifp;
		m->m_pkthdr.csum_flags = 0;

		if (nbufs > 1) {
			/* Dequeue the rest of chain. */
			if (vtnet_rxq_merged_eof(rxq, m, nbufs) != 0)
				continue;
		}

		kmsan_mark_mbuf(m, KMSAN_STATE_INITED);

		/*
		 * Save an endian swapped version of the header prior to it
		 * being stripped. The header is always at the start of the
		 * mbuf data. num_buffers was already saved (and not needed)
		 * so use the standard header.
		 */
		hdr = mtod(m, struct virtio_net_hdr *);
		lhdr.flags = hdr->flags;
		lhdr.gso_type = hdr->gso_type;
		lhdr.hdr_len = vtnet_htog16(sc, hdr->hdr_len);
		lhdr.gso_size = vtnet_htog16(sc, hdr->gso_size);
		lhdr.csum_start = vtnet_htog16(sc, hdr->csum_start);
		lhdr.csum_offset = vtnet_htog16(sc, hdr->csum_offset);
		m_adj(m, adjsz);

		if (PFIL_HOOKED_IN(sc->vtnet_pfil)) {
			pfil_return_t pfil;

			pfil = pfil_run_hooks(sc->vtnet_pfil, &m, ifp, PFIL_IN,
			    NULL);
			switch (pfil) {
			case PFIL_REALLOCED:
				m = pfil_mem2mbuf(m->m_data);
				break;
			case PFIL_DROPPED:
			case PFIL_CONSUMED:
				continue;
			default:
				KASSERT(pfil == PFIL_PASS,
				    ("Filter returned %d!", pfil));
			}
		}

		vtnet_rxq_input(rxq, m, &lhdr);
	}

	if (deq > 0) {
#if defined(INET) || defined(INET6)
		if (vtnet_software_lro(sc))
			tcp_lro_flush_all(&rxq->vtnrx_lro);
#endif
		virtqueue_notify(vq);
	}

	return (count > 0 ? 0 : EAGAIN);
}

static void
vtnet_rx_vq_process(struct vtnet_rxq *rxq, int tries)
{
	struct vtnet_softc *sc;
	struct ifnet *ifp;
	u_int more;
#ifdef DEV_NETMAP
	int nmirq;
#endif /* DEV_NETMAP */

	sc = rxq->vtnrx_sc;
	ifp = sc->vtnet_ifp;

	if (__predict_false(rxq->vtnrx_id >= sc->vtnet_act_vq_pairs)) {
		/*
		 * Ignore this interrupt. Either this is a spurious interrupt
		 * or multiqueue without per-VQ MSIX so every queue needs to
		 * be polled (a brain dead configuration we could try harder
		 * to avoid).
		 */
		vtnet_rxq_disable_intr(rxq);
		return;
	}

	VTNET_RXQ_LOCK(rxq);

#ifdef DEV_NETMAP
	/*
	 * We call netmap_rx_irq() under lock to prevent concurrent calls.
	 * This is not necessary to serialize the access to the RX vq, but
	 * rather to avoid races that may happen if this interface is
	 * attached to a VALE switch, which would cause received packets
	 * to stall in the RX queue (nm_kr_tryget() could find the kring
	 * busy when called from netmap_bwrap_intr_notify()).
	 */
	nmirq = netmap_rx_irq(ifp, rxq->vtnrx_id, &more);
	if (nmirq != NM_IRQ_PASS) {
		VTNET_RXQ_UNLOCK(rxq);
		if (nmirq == NM_IRQ_RESCHED) {
			taskqueue_enqueue(rxq->vtnrx_tq, &rxq->vtnrx_intrtask);
		}
		return;
	}
#endif /* DEV_NETMAP */

again:
	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) {
		VTNET_RXQ_UNLOCK(rxq);
		return;
	}

	more = vtnet_rxq_eof(rxq);
	if (more || vtnet_rxq_enable_intr(rxq) != 0) {
		if (!more)
			vtnet_rxq_disable_intr(rxq);
		/*
		 * This is an occasional condition or race (when !more),
		 * so retry a few times before scheduling the taskqueue.
		 */
		if (tries-- > 0)
			goto again;

		rxq->vtnrx_stats.vrxs_rescheduled++;
		VTNET_RXQ_UNLOCK(rxq);
		taskqueue_enqueue(rxq->vtnrx_tq, &rxq->vtnrx_intrtask);
	} else
		VTNET_RXQ_UNLOCK(rxq);
}

static void
vtnet_rx_vq_intr(void *xrxq)
{
	struct vtnet_rxq *rxq;

	rxq = xrxq;
	vtnet_rx_vq_process(rxq, VTNET_INTR_DISABLE_RETRIES);
}

static void
vtnet_rxq_tq_intr(void *xrxq, int pending __unused)
{
	struct vtnet_rxq *rxq;

	rxq = xrxq;
	vtnet_rx_vq_process(rxq, 0);
}

static int
vtnet_txq_intr_threshold(struct vtnet_txq *txq)
{
	struct vtnet_softc *sc;
	int threshold;

	sc = txq->vtntx_sc;

	/*
	 * The Tx interrupt is disabled until the queue free count falls
	 * below our threshold. Completed frames are drained from the Tx
	 * virtqueue before transmitting new frames and in the watchdog
	 * callout, so the frequency of Tx interrupts is greatly reduced,
	 * at the cost of not freeing mbufs as quickly as they otherwise
	 * would be.
	 */
	threshold = virtqueue_size(txq->vtntx_vq) / 4;

	/*
	 * Without indirect descriptors, leave enough room for the most
	 * segments we handle.
	 */
	if ((sc->vtnet_flags & VTNET_FLAG_INDIRECT) == 0 &&
	    threshold < sc->vtnet_tx_nsegs)
		threshold = sc->vtnet_tx_nsegs;

	return (threshold);
}

static int
vtnet_txq_below_threshold(struct vtnet_txq *txq)
{
	struct virtqueue *vq;

	vq = txq->vtntx_vq;

	return (virtqueue_nfree(vq) <= txq->vtntx_intr_threshold);
}

static int
vtnet_txq_notify(struct vtnet_txq *txq)
{
	struct virtqueue *vq;

	vq = txq->vtntx_vq;

	txq->vtntx_watchdog = VTNET_TX_TIMEOUT;
	virtqueue_notify(vq);

	if (vtnet_txq_enable_intr(txq) == 0)
		return (0);

	/*
	 * Drain frames that were completed since last checked. If this
	 * causes the queue to go above the threshold, the caller should
	 * continue transmitting.
	 */
	if (vtnet_txq_eof(txq) != 0 && vtnet_txq_below_threshold(txq) == 0) {
		virtqueue_disable_intr(vq);
		return (1);
	}

	return (0);
}

static void
vtnet_txq_free_mbufs(struct vtnet_txq *txq)
{
	struct virtqueue *vq;
	struct vtnet_tx_header *txhdr;
	int last;
#ifdef DEV_NETMAP
	struct netmap_kring *kring = netmap_kring_on(NA(txq->vtntx_sc->vtnet_ifp),
							txq->vtntx_id, NR_TX);
#else  /* !DEV_NETMAP */
	void *kring = NULL;
#endif /* !DEV_NETMAP */

	vq = txq->vtntx_vq;
	last = 0;

	while ((txhdr = virtqueue_drain(vq, &last)) != NULL) {
		if (kring == NULL) {
			m_freem(txhdr->vth_mbuf);
			uma_zfree(vtnet_tx_header_zone, txhdr);
		}
	}

	KASSERT(virtqueue_empty(vq),
	    ("%s: mbufs remaining in tx queue %p", __func__, txq));
}

/*
 * BMV: This can go away once we finally have offsets in the mbuf header.
 */
static int
vtnet_txq_offload_ctx(struct vtnet_txq *txq, struct mbuf *m, int *etype,
    int *proto, int *start)
{
	struct vtnet_softc *sc;
	struct ether_vlan_header *evh;
	int offset;

	sc = txq->vtntx_sc;

	evh = mtod(m, struct ether_vlan_header *);
	if (evh->evl_encap_proto == htons(ETHERTYPE_VLAN)) {
		/* BMV: We should handle nested VLAN tags too. */
		*etype = ntohs(evh->evl_proto);
		offset = sizeof(struct ether_vlan_header);
	} else {
		*etype = ntohs(evh->evl_encap_proto);
		offset = sizeof(struct ether_header);
	}

	switch (*etype) {
#if defined(INET)
	case ETHERTYPE_IP: {
		struct ip *ip, iphdr;
		if (__predict_false(m->m_len < offset + sizeof(struct ip))) {
			m_copydata(m, offset, sizeof(struct ip),
			    (caddr_t) &iphdr);
			ip = &iphdr;
		} else
			ip = (struct ip *)(m->m_data + offset);
		*proto = ip->ip_p;
		*start = offset + (ip->ip_hl << 2);
		break;
	}
#endif
#if defined(INET6)
	case ETHERTYPE_IPV6:
		*proto = -1;
		*start = ip6_lasthdr(m, offset, IPPROTO_IPV6, proto);
		/* Assert the network stack sent us a valid packet. */
		KASSERT(*start > offset,
		    ("%s: mbuf %p start %d offset %d proto %d", __func__, m,
		    *start, offset, *proto));
		break;
#endif
	default:
		sc->vtnet_stats.tx_csum_unknown_ethtype++;
		return (EINVAL);
	}

	return (0);
}

static int
vtnet_txq_offload_tso(struct vtnet_txq *txq, struct mbuf *m, int eth_type,
    int offset, struct virtio_net_hdr *hdr)
{
	static struct timeval lastecn;
	static int curecn;
	struct vtnet_softc *sc;
	struct tcphdr *tcp, tcphdr;

	sc = txq->vtntx_sc;

	if (__predict_false(m->m_len < offset + sizeof(struct tcphdr))) {
		m_copydata(m, offset, sizeof(struct tcphdr), (caddr_t) &tcphdr);
		tcp = &tcphdr;
	} else
		tcp = (struct tcphdr *)(m->m_data + offset);

	hdr->hdr_len = vtnet_gtoh16(sc, offset + (tcp->th_off << 2));
	hdr->gso_size = vtnet_gtoh16(sc, m->m_pkthdr.tso_segsz);
	hdr->gso_type = eth_type == ETHERTYPE_IP ? VIRTIO_NET_HDR_GSO_TCPV4 :
	    VIRTIO_NET_HDR_GSO_TCPV6;

	if (__predict_false(tcp->th_flags & TH_CWR)) {
		/*
		 * Drop if VIRTIO_NET_F_HOST_ECN was not negotiated. In
		 * FreeBSD, ECN support is not on a per-interface basis,
		 * but globally via the net.inet.tcp.ecn.enable sysctl
		 * knob. The default is off.
		 */
		if ((sc->vtnet_flags & VTNET_FLAG_TSO_ECN) == 0) {
			if (ppsratecheck(&lastecn, &curecn, 1))
				if_printf(sc->vtnet_ifp,
				    "TSO with ECN not negotiated with host\n");
			return (ENOTSUP);
		}
		hdr->gso_type |= VIRTIO_NET_HDR_GSO_ECN;
	}

	txq->vtntx_stats.vtxs_tso++;

	return (0);
}

static struct mbuf *
vtnet_txq_offload(struct vtnet_txq *txq, struct mbuf *m,
    struct virtio_net_hdr *hdr)
{
	struct vtnet_softc *sc;
	int flags, etype, csum_start, proto, error;

	sc = txq->vtntx_sc;
	flags = m->m_pkthdr.csum_flags;

	error = vtnet_txq_offload_ctx(txq, m, &etype, &proto, &csum_start);
	if (error)
		goto drop;

	if (flags & (VTNET_CSUM_OFFLOAD | VTNET_CSUM_OFFLOAD_IPV6)) {
		/* Sanity check the parsed mbuf matches the offload flags. */
		if (__predict_false((flags & VTNET_CSUM_OFFLOAD &&
		    etype != ETHERTYPE_IP) || (flags & VTNET_CSUM_OFFLOAD_IPV6
		    && etype != ETHERTYPE_IPV6))) {
			sc->vtnet_stats.tx_csum_proto_mismatch++;
			goto drop;
		}

		hdr->flags |= VIRTIO_NET_HDR_F_NEEDS_CSUM;
		hdr->csum_start = vtnet_gtoh16(sc, csum_start);
		hdr->csum_offset = vtnet_gtoh16(sc, m->m_pkthdr.csum_data);
		txq->vtntx_stats.vtxs_csum++;
	}

	if (flags & (CSUM_IP_TSO | CSUM_IP6_TSO)) {
		/*
		 * Sanity check the parsed mbuf IP protocol is TCP, and
		 * VirtIO TSO reqires the checksum offloading above.
		 */
		if (__predict_false(proto != IPPROTO_TCP)) {
			sc->vtnet_stats.tx_tso_not_tcp++;
			goto drop;
		} else if (__predict_false((hdr->flags &
		    VIRTIO_NET_HDR_F_NEEDS_CSUM) == 0)) {
			sc->vtnet_stats.tx_tso_without_csum++;
			goto drop;
		}

		error = vtnet_txq_offload_tso(txq, m, etype, csum_start, hdr);
		if (error)
			goto drop;
	}

	return (m);

drop:
	m_freem(m);
	return (NULL);
}

static int
vtnet_txq_enqueue_buf(struct vtnet_txq *txq, struct mbuf **m_head,
    struct vtnet_tx_header *txhdr)
{
	struct vtnet_softc *sc;
	struct virtqueue *vq;
	struct sglist *sg;
	struct mbuf *m;
	int error;

	sc = txq->vtntx_sc;
	vq = txq->vtntx_vq;
	sg = txq->vtntx_sg;
	m = *m_head;

	sglist_reset(sg);
	error = sglist_append(sg, &txhdr->vth_uhdr, sc->vtnet_hdr_size);
	if (error != 0 || sg->sg_nseg != 1) {
		KASSERT(0, ("%s: cannot add header to sglist error %d nseg %d",
		    __func__, error, sg->sg_nseg));
		goto fail;
	}

	error = sglist_append_mbuf(sg, m);
	if (error) {
		m = m_defrag(m, M_NOWAIT);
		if (m == NULL)
			goto fail;

		*m_head = m;
		sc->vtnet_stats.tx_defragged++;

		error = sglist_append_mbuf(sg, m);
		if (error)
			goto fail;
	}

	txhdr->vth_mbuf = m;
	error = virtqueue_enqueue(vq, txhdr, sg, sg->sg_nseg, 0);

	return (error);

fail:
	sc->vtnet_stats.tx_defrag_failed++;
	m_freem(*m_head);
	*m_head = NULL;

	return (ENOBUFS);
}

static int
vtnet_txq_encap(struct vtnet_txq *txq, struct mbuf **m_head, int flags)
{
	struct vtnet_tx_header *txhdr;
	struct virtio_net_hdr *hdr;
	struct mbuf *m;
	int error;

	m = *m_head;
	M_ASSERTPKTHDR(m);

	txhdr = uma_zalloc(vtnet_tx_header_zone, flags | M_ZERO);
	if (txhdr == NULL) {
		m_freem(m);
		*m_head = NULL;
		return (ENOMEM);
	}

	/*
	 * Always use the non-mergeable header, regardless if mergable headers
	 * were negotiated, because for transmit num_buffers is always zero.
	 * The vtnet_hdr_size is used to enqueue the right header size segment.
	 */
	hdr = &txhdr->vth_uhdr.hdr;

	if (m->m_flags & M_VLANTAG) {
		m = ether_vlanencap(m, m->m_pkthdr.ether_vtag);
		if ((*m_head = m) == NULL) {
			error = ENOBUFS;
			goto fail;
		}
		m->m_flags &= ~M_VLANTAG;
	}

	if (m->m_pkthdr.csum_flags & VTNET_CSUM_ALL_OFFLOAD) {
		m = vtnet_txq_offload(txq, m, hdr);
		if ((*m_head = m) == NULL) {
			error = ENOBUFS;
			goto fail;
		}
	}

	error = vtnet_txq_enqueue_buf(txq, m_head, txhdr);
fail:
	if (error)
		uma_zfree(vtnet_tx_header_zone, txhdr);

	return (error);
}

#ifdef VTNET_LEGACY_TX

static void
vtnet_start_locked(struct vtnet_txq *txq, struct ifnet *ifp)
{
	struct vtnet_softc *sc;
	struct virtqueue *vq;
	struct mbuf *m0;
	int tries, enq;

	sc = txq->vtntx_sc;
	vq = txq->vtntx_vq;
	tries = 0;

	VTNET_TXQ_LOCK_ASSERT(txq);

	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0 ||
	    sc->vtnet_link_active == 0)
		return;

	vtnet_txq_eof(txq);

again:
	enq = 0;

	while (!IFQ_DRV_IS_EMPTY(&ifp->if_snd)) {
		if (virtqueue_full(vq))
			break;

		IFQ_DRV_DEQUEUE(&ifp->if_snd, m0);
		if (m0 == NULL)
			break;

		if (vtnet_txq_encap(txq, &m0, M_NOWAIT) != 0) {
			if (m0 != NULL)
				IFQ_DRV_PREPEND(&ifp->if_snd, m0);
			break;
		}

		enq++;
		ETHER_BPF_MTAP(ifp, m0);
	}

	if (enq > 0 && vtnet_txq_notify(txq) != 0) {
		if (tries++ < VTNET_NOTIFY_RETRIES)
			goto again;

		txq->vtntx_stats.vtxs_rescheduled++;
		taskqueue_enqueue(txq->vtntx_tq, &txq->vtntx_intrtask);
	}
}

static void
vtnet_start(struct ifnet *ifp)
{
	struct vtnet_softc *sc;
	struct vtnet_txq *txq;

	sc = ifp->if_softc;
	txq = &sc->vtnet_txqs[0];

	VTNET_TXQ_LOCK(txq);
	vtnet_start_locked(txq, ifp);
	VTNET_TXQ_UNLOCK(txq);
}

#else /* !VTNET_LEGACY_TX */

static int
vtnet_txq_mq_start_locked(struct vtnet_txq *txq, struct mbuf *m)
{
	struct vtnet_softc *sc;
	struct virtqueue *vq;
	struct buf_ring *br;
	struct ifnet *ifp;
	int enq, tries, error;

	sc = txq->vtntx_sc;
	vq = txq->vtntx_vq;
	br = txq->vtntx_br;
	ifp = sc->vtnet_ifp;
	tries = 0;
	error = 0;

	VTNET_TXQ_LOCK_ASSERT(txq);

	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0 ||
	    sc->vtnet_link_active == 0) {
		if (m != NULL)
			error = drbr_enqueue(ifp, br, m);
		return (error);
	}

	if (m != NULL) {
		error = drbr_enqueue(ifp, br, m);
		if (error)
			return (error);
	}

	vtnet_txq_eof(txq);

again:
	enq = 0;

	while ((m = drbr_peek(ifp, br)) != NULL) {
		if (virtqueue_full(vq)) {
			drbr_putback(ifp, br, m);
			break;
		}

		if (vtnet_txq_encap(txq, &m, M_NOWAIT) != 0) {
			if (m != NULL)
				drbr_putback(ifp, br, m);
			else
				drbr_advance(ifp, br);
			break;
		}
		drbr_advance(ifp, br);

		enq++;
		ETHER_BPF_MTAP(ifp, m);
	}

	if (enq > 0 && vtnet_txq_notify(txq) != 0) {
		if (tries++ < VTNET_NOTIFY_RETRIES)
			goto again;

		txq->vtntx_stats.vtxs_rescheduled++;
		taskqueue_enqueue(txq->vtntx_tq, &txq->vtntx_intrtask);
	}

	return (0);
}

static int
vtnet_txq_mq_start(struct ifnet *ifp, struct mbuf *m)
{
	struct vtnet_softc *sc;
	struct vtnet_txq *txq;
	int i, npairs, error;

	sc = ifp->if_softc;
	npairs = sc->vtnet_act_vq_pairs;

	if (M_HASHTYPE_GET(m) != M_HASHTYPE_NONE)
		i = m->m_pkthdr.flowid % npairs;
	else
		i = curcpu % npairs;

	txq = &sc->vtnet_txqs[i];

	if (VTNET_TXQ_TRYLOCK(txq) != 0) {
		error = vtnet_txq_mq_start_locked(txq, m);
		VTNET_TXQ_UNLOCK(txq);
	} else {
		error = drbr_enqueue(ifp, txq->vtntx_br, m);
		taskqueue_enqueue(txq->vtntx_tq, &txq->vtntx_defrtask);
	}

	return (error);
}

static void
vtnet_txq_tq_deferred(void *xtxq, int pending __unused)
{
	struct vtnet_softc *sc;
	struct vtnet_txq *txq;

	txq = xtxq;
	sc = txq->vtntx_sc;

	VTNET_TXQ_LOCK(txq);
	if (!drbr_empty(sc->vtnet_ifp, txq->vtntx_br))
		vtnet_txq_mq_start_locked(txq, NULL);
	VTNET_TXQ_UNLOCK(txq);
}

#endif /* VTNET_LEGACY_TX */

static void
vtnet_txq_start(struct vtnet_txq *txq)
{
	struct vtnet_softc *sc;
	struct ifnet *ifp;

	sc = txq->vtntx_sc;
	ifp = sc->vtnet_ifp;

#ifdef VTNET_LEGACY_TX
	if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
		vtnet_start_locked(txq, ifp);
#else
	if (!drbr_empty(ifp, txq->vtntx_br))
		vtnet_txq_mq_start_locked(txq, NULL);
#endif
}

static void
vtnet_txq_tq_intr(void *xtxq, int pending __unused)
{
	struct vtnet_softc *sc;
	struct vtnet_txq *txq;
	struct ifnet *ifp;

	txq = xtxq;
	sc = txq->vtntx_sc;
	ifp = sc->vtnet_ifp;

	VTNET_TXQ_LOCK(txq);

	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) {
		VTNET_TXQ_UNLOCK(txq);
		return;
	}

	vtnet_txq_eof(txq);
	vtnet_txq_start(txq);

	VTNET_TXQ_UNLOCK(txq);
}

static int
vtnet_txq_eof(struct vtnet_txq *txq)
{
	struct virtqueue *vq;
	struct vtnet_tx_header *txhdr;
	struct mbuf *m;
	int deq;

	vq = txq->vtntx_vq;
	deq = 0;
	VTNET_TXQ_LOCK_ASSERT(txq);

	while ((txhdr = virtqueue_dequeue(vq, NULL)) != NULL) {
		m = txhdr->vth_mbuf;
		deq++;

		txq->vtntx_stats.vtxs_opackets++;
		txq->vtntx_stats.vtxs_obytes += m->m_pkthdr.len;
		if (m->m_flags & M_MCAST)
			txq->vtntx_stats.vtxs_omcasts++;

		m_freem(m);
		uma_zfree(vtnet_tx_header_zone, txhdr);
	}

	if (virtqueue_empty(vq))
		txq->vtntx_watchdog = 0;

	return (deq);
}

static void
vtnet_tx_vq_intr(void *xtxq)
{
	struct vtnet_softc *sc;
	struct vtnet_txq *txq;
	struct ifnet *ifp;

	txq = xtxq;
	sc = txq->vtntx_sc;
	ifp = sc->vtnet_ifp;

	if (__predict_false(txq->vtntx_id >= sc->vtnet_act_vq_pairs)) {
		/*
		 * Ignore this interrupt. Either this is a spurious interrupt
		 * or multiqueue without per-VQ MSIX so every queue needs to
		 * be polled (a brain dead configuration we could try harder
		 * to avoid).
		 */
		vtnet_txq_disable_intr(txq);
		return;
	}

#ifdef DEV_NETMAP
	if (netmap_tx_irq(ifp, txq->vtntx_id) != NM_IRQ_PASS)
		return;
#endif /* DEV_NETMAP */

	VTNET_TXQ_LOCK(txq);

	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) {
		VTNET_TXQ_UNLOCK(txq);
		return;
	}

	vtnet_txq_eof(txq);
	vtnet_txq_start(txq);

	VTNET_TXQ_UNLOCK(txq);
}

static void
vtnet_tx_start_all(struct vtnet_softc *sc)
{
	struct vtnet_txq *txq;
	int i;

	VTNET_CORE_LOCK_ASSERT(sc);

	for (i = 0; i < sc->vtnet_act_vq_pairs; i++) {
		txq = &sc->vtnet_txqs[i];

		VTNET_TXQ_LOCK(txq);
		vtnet_txq_start(txq);
		VTNET_TXQ_UNLOCK(txq);
	}
}

#ifndef VTNET_LEGACY_TX
static void
vtnet_qflush(struct ifnet *ifp)
{
	struct vtnet_softc *sc;
	struct vtnet_txq *txq;
	struct mbuf *m;
	int i;

	sc = ifp->if_softc;

	for (i = 0; i < sc->vtnet_act_vq_pairs; i++) {
		txq = &sc->vtnet_txqs[i];

		VTNET_TXQ_LOCK(txq);
		while ((m = buf_ring_dequeue_sc(txq->vtntx_br)) != NULL)
			m_freem(m);
		VTNET_TXQ_UNLOCK(txq);
	}

	if_qflush(ifp);
}
#endif

static int
vtnet_watchdog(struct vtnet_txq *txq)
{
	struct ifnet *ifp;

	ifp = txq->vtntx_sc->vtnet_ifp;

	VTNET_TXQ_LOCK(txq);
	if (txq->vtntx_watchdog == 1) {
		/*
		 * Only drain completed frames if the watchdog is about to
		 * expire. If any frames were drained, there may be enough
		 * free descriptors now available to transmit queued frames.
		 * In that case, the timer will immediately be decremented
		 * below, but the timeout is generous enough that should not
		 * be a problem.
		 */
		if (vtnet_txq_eof(txq) != 0)
			vtnet_txq_start(txq);
	}

	if (txq->vtntx_watchdog == 0 || --txq->vtntx_watchdog) {
		VTNET_TXQ_UNLOCK(txq);
		return (0);
	}
	VTNET_TXQ_UNLOCK(txq);

	if_printf(ifp, "watchdog timeout on queue %d\n", txq->vtntx_id);
	return (1);
}

static void
vtnet_accum_stats(struct vtnet_softc *sc, struct vtnet_rxq_stats *rxacc,
    struct vtnet_txq_stats *txacc)
{

	bzero(rxacc, sizeof(struct vtnet_rxq_stats));
	bzero(txacc, sizeof(struct vtnet_txq_stats));

	for (int i = 0; i < sc->vtnet_max_vq_pairs; i++) {
		struct vtnet_rxq_stats *rxst;
		struct vtnet_txq_stats *txst;

		rxst = &sc->vtnet_rxqs[i].vtnrx_stats;
		rxacc->vrxs_ipackets += rxst->vrxs_ipackets;
		rxacc->vrxs_ibytes += rxst->vrxs_ibytes;
		rxacc->vrxs_iqdrops += rxst->vrxs_iqdrops;
		rxacc->vrxs_csum += rxst->vrxs_csum;
		rxacc->vrxs_csum_failed += rxst->vrxs_csum_failed;
		rxacc->vrxs_rescheduled += rxst->vrxs_rescheduled;

		txst = &sc->vtnet_txqs[i].vtntx_stats;
		txacc->vtxs_opackets += txst->vtxs_opackets;
		txacc->vtxs_obytes += txst->vtxs_obytes;
		txacc->vtxs_csum += txst->vtxs_csum;
		txacc->vtxs_tso += txst->vtxs_tso;
		txacc->vtxs_rescheduled += txst->vtxs_rescheduled;
	}
}

static uint64_t
vtnet_get_counter(if_t ifp, ift_counter cnt)
{
	struct vtnet_softc *sc;
	struct vtnet_rxq_stats rxaccum;
	struct vtnet_txq_stats txaccum;

	sc = if_getsoftc(ifp);
	vtnet_accum_stats(sc, &rxaccum, &txaccum);

	switch (cnt) {
	case IFCOUNTER_IPACKETS:
		return (rxaccum.vrxs_ipackets);
	case IFCOUNTER_IQDROPS:
		return (rxaccum.vrxs_iqdrops);
	case IFCOUNTER_IERRORS:
		return (rxaccum.vrxs_ierrors);
	case IFCOUNTER_OPACKETS:
		return (txaccum.vtxs_opackets);
#ifndef VTNET_LEGACY_TX
	case IFCOUNTER_OBYTES:
		return (txaccum.vtxs_obytes);
	case IFCOUNTER_OMCASTS:
		return (txaccum.vtxs_omcasts);
#endif
	default:
		return (if_get_counter_default(ifp, cnt));
	}
}

static void
vtnet_tick(void *xsc)
{
	struct vtnet_softc *sc;
	struct ifnet *ifp;
	int i, timedout;

	sc = xsc;
	ifp = sc->vtnet_ifp;
	timedout = 0;

	VTNET_CORE_LOCK_ASSERT(sc);

	for (i = 0; i < sc->vtnet_act_vq_pairs; i++)
		timedout |= vtnet_watchdog(&sc->vtnet_txqs[i]);

	if (timedout != 0) {
		ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
		vtnet_init_locked(sc, 0);
	} else
		callout_schedule(&sc->vtnet_tick_ch, hz);
}

static void
vtnet_start_taskqueues(struct vtnet_softc *sc)
{
	device_t dev;
	struct vtnet_rxq *rxq;
	struct vtnet_txq *txq;
	int i, error;

	dev = sc->vtnet_dev;

	/*
	 * Errors here are very difficult to recover from - we cannot
	 * easily fail because, if this is during boot, we will hang
	 * when freeing any successfully started taskqueues because
	 * the scheduler isn't up yet.
	 *
	 * Most drivers just ignore the return value - it only fails
	 * with ENOMEM so an error is not likely.
	 */
	for (i = 0; i < sc->vtnet_req_vq_pairs; i++) {
		rxq = &sc->vtnet_rxqs[i];
		error = taskqueue_start_threads(&rxq->vtnrx_tq, 1, PI_NET,
		    "%s rxq %d", device_get_nameunit(dev), rxq->vtnrx_id);
		if (error) {
			device_printf(dev, "failed to start rx taskq %d\n",
			    rxq->vtnrx_id);
		}

		txq = &sc->vtnet_txqs[i];
		error = taskqueue_start_threads(&txq->vtntx_tq, 1, PI_NET,
		    "%s txq %d", device_get_nameunit(dev), txq->vtntx_id);
		if (error) {
			device_printf(dev, "failed to start tx taskq %d\n",
			    txq->vtntx_id);
		}
	}
}

static void
vtnet_free_taskqueues(struct vtnet_softc *sc)
{
	struct vtnet_rxq *rxq;
	struct vtnet_txq *txq;
	int i;

	for (i = 0; i < sc->vtnet_max_vq_pairs; i++) {
		rxq = &sc->vtnet_rxqs[i];
		if (rxq->vtnrx_tq != NULL) {
			taskqueue_free(rxq->vtnrx_tq);
			rxq->vtnrx_tq = NULL;
		}

		txq = &sc->vtnet_txqs[i];
		if (txq->vtntx_tq != NULL) {
			taskqueue_free(txq->vtntx_tq);
			txq->vtntx_tq = NULL;
		}
	}
}

static void
vtnet_drain_taskqueues(struct vtnet_softc *sc)
{
	struct vtnet_rxq *rxq;
	struct vtnet_txq *txq;
	int i;

	for (i = 0; i < sc->vtnet_max_vq_pairs; i++) {
		rxq = &sc->vtnet_rxqs[i];
		if (rxq->vtnrx_tq != NULL)
			taskqueue_drain(rxq->vtnrx_tq, &rxq->vtnrx_intrtask);

		txq = &sc->vtnet_txqs[i];
		if (txq->vtntx_tq != NULL) {
			taskqueue_drain(txq->vtntx_tq, &txq->vtntx_intrtask);
#ifndef VTNET_LEGACY_TX
			taskqueue_drain(txq->vtntx_tq, &txq->vtntx_defrtask);
#endif
		}
	}
}

static void
vtnet_drain_rxtx_queues(struct vtnet_softc *sc)
{
	struct vtnet_rxq *rxq;
	struct vtnet_txq *txq;
	int i;

	for (i = 0; i < sc->vtnet_max_vq_pairs; i++) {
		rxq = &sc->vtnet_rxqs[i];
		vtnet_rxq_free_mbufs(rxq);

		txq = &sc->vtnet_txqs[i];
		vtnet_txq_free_mbufs(txq);
	}
}

static void
vtnet_stop_rendezvous(struct vtnet_softc *sc)
{
	struct vtnet_rxq *rxq;
	struct vtnet_txq *txq;
	int i;

	VTNET_CORE_LOCK_ASSERT(sc);

	/*
	 * Lock and unlock the per-queue mutex so we known the stop
	 * state is visible. Doing only the active queues should be
	 * sufficient, but it does not cost much extra to do all the
	 * queues.
	 */
	for (i = 0; i < sc->vtnet_max_vq_pairs; i++) {
		rxq = &sc->vtnet_rxqs[i];
		VTNET_RXQ_LOCK(rxq);
		VTNET_RXQ_UNLOCK(rxq);

		txq = &sc->vtnet_txqs[i];
		VTNET_TXQ_LOCK(txq);
		VTNET_TXQ_UNLOCK(txq);
	}
}

static void
vtnet_stop(struct vtnet_softc *sc)
{
	device_t dev;
	struct ifnet *ifp;

	dev = sc->vtnet_dev;
	ifp = sc->vtnet_ifp;

	VTNET_CORE_LOCK_ASSERT(sc);

	ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
	sc->vtnet_link_active = 0;
	callout_stop(&sc->vtnet_tick_ch);

	/* Only advisory. */
	vtnet_disable_interrupts(sc);

#ifdef DEV_NETMAP
	/* Stop any pending txsync/rxsync and disable them. */
	netmap_disable_all_rings(ifp);
#endif /* DEV_NETMAP */

	/*
	 * Stop the host adapter. This resets it to the pre-initialized
	 * state. It will not generate any interrupts until after it is
	 * reinitialized.
	 */
	virtio_stop(dev);
	vtnet_stop_rendezvous(sc);

	vtnet_drain_rxtx_queues(sc);
	sc->vtnet_act_vq_pairs = 1;
}

static int
vtnet_virtio_reinit(struct vtnet_softc *sc)
{
	device_t dev;
	struct ifnet *ifp;
	uint64_t features;
	int error;

	dev = sc->vtnet_dev;
	ifp = sc->vtnet_ifp;
	features = sc->vtnet_negotiated_features;

	/*
	 * Re-negotiate with the host, removing any disabled receive
	 * features. Transmit features are disabled only on our side
	 * via if_capenable and if_hwassist.
	 */

	if ((ifp->if_capenable & (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6)) == 0)
		features &= ~(VIRTIO_NET_F_GUEST_CSUM | VTNET_LRO_FEATURES);

	if ((ifp->if_capenable & IFCAP_LRO) == 0)
		features &= ~VTNET_LRO_FEATURES;

	if ((ifp->if_capenable & IFCAP_VLAN_HWFILTER) == 0)
		features &= ~VIRTIO_NET_F_CTRL_VLAN;

	error = virtio_reinit(dev, features);
	if (error) {
		device_printf(dev, "virtio reinit error %d\n", error);
		return (error);
	}

	sc->vtnet_features = features;
	virtio_reinit_complete(dev);

	return (0);
}

static void
vtnet_init_rx_filters(struct vtnet_softc *sc)
{
	struct ifnet *ifp;

	ifp = sc->vtnet_ifp;

	if (sc->vtnet_flags & VTNET_FLAG_CTRL_RX) {
		vtnet_rx_filter(sc);
		vtnet_rx_filter_mac(sc);
	}

	if (ifp->if_capenable & IFCAP_VLAN_HWFILTER)
		vtnet_rx_filter_vlan(sc);
}

static int
vtnet_init_rx_queues(struct vtnet_softc *sc)
{
	device_t dev;
	struct ifnet *ifp;
	struct vtnet_rxq *rxq;
	int i, clustersz, error;

	dev = sc->vtnet_dev;
	ifp = sc->vtnet_ifp;

	clustersz = vtnet_rx_cluster_size(sc, ifp->if_mtu);
	sc->vtnet_rx_clustersz = clustersz;

	if (sc->vtnet_flags & VTNET_FLAG_LRO_NOMRG) {
		sc->vtnet_rx_nmbufs = howmany(sizeof(struct vtnet_rx_header) +
		    VTNET_MAX_RX_SIZE, clustersz);
		KASSERT(sc->vtnet_rx_nmbufs < sc->vtnet_rx_nsegs,
		    ("%s: too many rx mbufs %d for %d segments", __func__,
		    sc->vtnet_rx_nmbufs, sc->vtnet_rx_nsegs));
	} else
		sc->vtnet_rx_nmbufs = 1;

	for (i = 0; i < sc->vtnet_act_vq_pairs; i++) {
		rxq = &sc->vtnet_rxqs[i];

		/* Hold the lock to satisfy asserts. */
		VTNET_RXQ_LOCK(rxq);
		error = vtnet_rxq_populate(rxq);
		VTNET_RXQ_UNLOCK(rxq);

		if (error) {
			device_printf(dev, "cannot populate Rx queue %d\n", i);
			return (error);
		}
	}

	return (0);
}

static int
vtnet_init_tx_queues(struct vtnet_softc *sc)
{
	struct vtnet_txq *txq;
	int i;

	for (i = 0; i < sc->vtnet_act_vq_pairs; i++) {
		txq = &sc->vtnet_txqs[i];
		txq->vtntx_watchdog = 0;
		txq->vtntx_intr_threshold = vtnet_txq_intr_threshold(txq);
#ifdef DEV_NETMAP
		netmap_reset(NA(sc->vtnet_ifp), NR_TX, i, 0);
#endif /* DEV_NETMAP */
	}

	return (0);
}

static int
vtnet_init_rxtx_queues(struct vtnet_softc *sc)
{
	int error;

	error = vtnet_init_rx_queues(sc);
	if (error)
		return (error);

	error = vtnet_init_tx_queues(sc);
	if (error)
		return (error);

	return (0);
}

static void
vtnet_set_active_vq_pairs(struct vtnet_softc *sc)
{
	device_t dev;
	int npairs;

	dev = sc->vtnet_dev;

	if ((sc->vtnet_flags & VTNET_FLAG_MQ) == 0) {
		sc->vtnet_act_vq_pairs = 1;
		return;
	}

	npairs = sc->vtnet_req_vq_pairs;

	if (vtnet_ctrl_mq_cmd(sc, npairs) != 0) {
		device_printf(dev, "cannot set active queue pairs to %d, "
		    "falling back to 1 queue pair\n", npairs);
		npairs = 1;
	}

	sc->vtnet_act_vq_pairs = npairs;
}

static void
vtnet_update_rx_offloads(struct vtnet_softc *sc)
{
	struct ifnet *ifp;
	uint64_t features;
	int error;

	ifp = sc->vtnet_ifp;
	features = sc->vtnet_features;

	VTNET_CORE_LOCK_ASSERT(sc);

	if (ifp->if_capabilities & (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6)) {
		if (ifp->if_capenable & (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6))
			features |= VIRTIO_NET_F_GUEST_CSUM;
		else
			features &= ~VIRTIO_NET_F_GUEST_CSUM;
	}

	if (ifp->if_capabilities & IFCAP_LRO && !vtnet_software_lro(sc)) {
		if (ifp->if_capenable & IFCAP_LRO)
			features |= VTNET_LRO_FEATURES;
		else
			features &= ~VTNET_LRO_FEATURES;
	}

	error = vtnet_ctrl_guest_offloads(sc,
	    features & (VIRTIO_NET_F_GUEST_CSUM | VIRTIO_NET_F_GUEST_TSO4 |
		        VIRTIO_NET_F_GUEST_TSO6 | VIRTIO_NET_F_GUEST_ECN  |
			VIRTIO_NET_F_GUEST_UFO));
	if (error) {
		device_printf(sc->vtnet_dev,
		    "%s: cannot update Rx features\n", __func__);
		if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
			ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
			vtnet_init_locked(sc, 0);
		}
	} else
		sc->vtnet_features = features;
}

static int
vtnet_reinit(struct vtnet_softc *sc)
{
	struct ifnet *ifp;
	int error;

	ifp = sc->vtnet_ifp;

	bcopy(IF_LLADDR(ifp), sc->vtnet_hwaddr, ETHER_ADDR_LEN);

	error = vtnet_virtio_reinit(sc);
	if (error)
		return (error);

	vtnet_set_macaddr(sc);
	vtnet_set_active_vq_pairs(sc);

	if (sc->vtnet_flags & VTNET_FLAG_CTRL_VQ)
		vtnet_init_rx_filters(sc);

	ifp->if_hwassist = 0;
	if (ifp->if_capenable & IFCAP_TXCSUM)
		ifp->if_hwassist |= VTNET_CSUM_OFFLOAD;
	if (ifp->if_capenable & IFCAP_TXCSUM_IPV6)
		ifp->if_hwassist |= VTNET_CSUM_OFFLOAD_IPV6;
	if (ifp->if_capenable & IFCAP_TSO4)
		ifp->if_hwassist |= CSUM_IP_TSO;
	if (ifp->if_capenable & IFCAP_TSO6)
		ifp->if_hwassist |= CSUM_IP6_TSO;

	error = vtnet_init_rxtx_queues(sc);
	if (error)
		return (error);

	return (0);
}

static void
vtnet_init_locked(struct vtnet_softc *sc, int init_mode)
{
	struct ifnet *ifp;

	ifp = sc->vtnet_ifp;

	VTNET_CORE_LOCK_ASSERT(sc);

	if (ifp->if_drv_flags & IFF_DRV_RUNNING)
		return;

	vtnet_stop(sc);

#ifdef DEV_NETMAP
	/* Once stopped we can update the netmap flags, if necessary. */
	switch (init_mode) {
	case VTNET_INIT_NETMAP_ENTER:
		nm_set_native_flags(NA(ifp));
		break;
	case VTNET_INIT_NETMAP_EXIT:
		nm_clear_native_flags(NA(ifp));
		break;
	}
#endif /* DEV_NETMAP */

	if (vtnet_reinit(sc) != 0) {
		vtnet_stop(sc);
		return;
	}

	ifp->if_drv_flags |= IFF_DRV_RUNNING;
	vtnet_update_link_status(sc);
	vtnet_enable_interrupts(sc);
	callout_reset(&sc->vtnet_tick_ch, hz, vtnet_tick, sc);

#ifdef DEV_NETMAP
	/* Re-enable txsync/rxsync. */
	netmap_enable_all_rings(ifp);
#endif /* DEV_NETMAP */
}

static void
vtnet_init(void *xsc)
{
	struct vtnet_softc *sc;

	sc = xsc;

	VTNET_CORE_LOCK(sc);
	vtnet_init_locked(sc, 0);
	VTNET_CORE_UNLOCK(sc);
}

static void
vtnet_free_ctrl_vq(struct vtnet_softc *sc)
{

	/*
	 * The control virtqueue is only polled and therefore it should
	 * already be empty.
	 */
	KASSERT(virtqueue_empty(sc->vtnet_ctrl_vq),
	    ("%s: ctrl vq %p not empty", __func__, sc->vtnet_ctrl_vq));
}

static void
vtnet_exec_ctrl_cmd(struct vtnet_softc *sc, void *cookie,
    struct sglist *sg, int readable, int writable)
{
	struct virtqueue *vq;

	vq = sc->vtnet_ctrl_vq;

	MPASS(sc->vtnet_flags & VTNET_FLAG_CTRL_VQ);
	VTNET_CORE_LOCK_ASSERT(sc);

	if (!virtqueue_empty(vq))
		return;

	/*
	 * Poll for the response, but the command is likely completed before
	 * returning from the notify.
	 */
	if (virtqueue_enqueue(vq, cookie, sg, readable, writable) == 0)  {
		virtqueue_notify(vq);
		virtqueue_poll(vq, NULL);
	}
}

static int
vtnet_ctrl_mac_cmd(struct vtnet_softc *sc, uint8_t *hwaddr)
{
	struct sglist_seg segs[3];
	struct sglist sg;
	struct {
		struct virtio_net_ctrl_hdr hdr __aligned(2);
		uint8_t pad1;
		uint8_t addr[ETHER_ADDR_LEN] __aligned(8);
		uint8_t pad2;
		uint8_t ack;
	} s;
	int error;

	error = 0;
	MPASS(sc->vtnet_flags & VTNET_FLAG_CTRL_MAC);

	s.hdr.class = VIRTIO_NET_CTRL_MAC;
	s.hdr.cmd = VIRTIO_NET_CTRL_MAC_ADDR_SET;
	bcopy(hwaddr, &s.addr[0], ETHER_ADDR_LEN);
	s.ack = VIRTIO_NET_ERR;

	sglist_init(&sg, nitems(segs), segs);
	error |= sglist_append(&sg, &s.hdr, sizeof(struct virtio_net_ctrl_hdr));
	error |= sglist_append(&sg, &s.addr[0], ETHER_ADDR_LEN);
	error |= sglist_append(&sg, &s.ack, sizeof(uint8_t));
	MPASS(error == 0 && sg.sg_nseg == nitems(segs));

	if (error == 0)
		vtnet_exec_ctrl_cmd(sc, &s.ack, &sg, sg.sg_nseg - 1, 1);

	return (s.ack == VIRTIO_NET_OK ? 0 : EIO);
}

static int
vtnet_ctrl_guest_offloads(struct vtnet_softc *sc, uint64_t offloads)
{
	struct sglist_seg segs[3];
	struct sglist sg;
	struct {
		struct virtio_net_ctrl_hdr hdr __aligned(2);
		uint8_t pad1;
		uint64_t offloads __aligned(8);
		uint8_t pad2;
		uint8_t ack;
	} s;
	int error;

	error = 0;
	MPASS(sc->vtnet_features & VIRTIO_NET_F_CTRL_GUEST_OFFLOADS);

	s.hdr.class = VIRTIO_NET_CTRL_GUEST_OFFLOADS;
	s.hdr.cmd = VIRTIO_NET_CTRL_GUEST_OFFLOADS_SET;
	s.offloads = vtnet_gtoh64(sc, offloads);
	s.ack = VIRTIO_NET_ERR;

	sglist_init(&sg, nitems(segs), segs);
	error |= sglist_append(&sg, &s.hdr, sizeof(struct virtio_net_ctrl_hdr));
	error |= sglist_append(&sg, &s.offloads, sizeof(uint64_t));
	error |= sglist_append(&sg, &s.ack, sizeof(uint8_t));
	MPASS(error == 0 && sg.sg_nseg == nitems(segs));

	if (error == 0)
		vtnet_exec_ctrl_cmd(sc, &s.ack, &sg, sg.sg_nseg - 1, 1);

	return (s.ack == VIRTIO_NET_OK ? 0 : EIO);
}

static int
vtnet_ctrl_mq_cmd(struct vtnet_softc *sc, uint16_t npairs)
{
	struct sglist_seg segs[3];
	struct sglist sg;
	struct {
		struct virtio_net_ctrl_hdr hdr __aligned(2);
		uint8_t pad1;
		struct virtio_net_ctrl_mq mq __aligned(2);
		uint8_t pad2;
		uint8_t ack;
	} s;
	int error;

	error = 0;
	MPASS(sc->vtnet_flags & VTNET_FLAG_MQ);

	s.hdr.class = VIRTIO_NET_CTRL_MQ;
	s.hdr.cmd = VIRTIO_NET_CTRL_MQ_VQ_PAIRS_SET;
	s.mq.virtqueue_pairs = vtnet_gtoh16(sc, npairs);
	s.ack = VIRTIO_NET_ERR;

	sglist_init(&sg, nitems(segs), segs);
	error |= sglist_append(&sg, &s.hdr, sizeof(struct virtio_net_ctrl_hdr));
	error |= sglist_append(&sg, &s.mq, sizeof(struct virtio_net_ctrl_mq));
	error |= sglist_append(&sg, &s.ack, sizeof(uint8_t));
	MPASS(error == 0 && sg.sg_nseg == nitems(segs));

	if (error == 0)
		vtnet_exec_ctrl_cmd(sc, &s.ack, &sg, sg.sg_nseg - 1, 1);

	return (s.ack == VIRTIO_NET_OK ? 0 : EIO);
}

static int
vtnet_ctrl_rx_cmd(struct vtnet_softc *sc, uint8_t cmd, bool on)
{
	struct sglist_seg segs[3];
	struct sglist sg;
	struct {
		struct virtio_net_ctrl_hdr hdr __aligned(2);
		uint8_t pad1;
		uint8_t onoff;
		uint8_t pad2;
		uint8_t ack;
	} s;
	int error;

	error = 0;
	MPASS(sc->vtnet_flags & VTNET_FLAG_CTRL_RX);

	s.hdr.class = VIRTIO_NET_CTRL_RX;
	s.hdr.cmd = cmd;
	s.onoff = on;
	s.ack = VIRTIO_NET_ERR;

	sglist_init(&sg, nitems(segs), segs);
	error |= sglist_append(&sg, &s.hdr, sizeof(struct virtio_net_ctrl_hdr));
	error |= sglist_append(&sg, &s.onoff, sizeof(uint8_t));
	error |= sglist_append(&sg, &s.ack, sizeof(uint8_t));
	MPASS(error == 0 && sg.sg_nseg == nitems(segs));

	if (error == 0)
		vtnet_exec_ctrl_cmd(sc, &s.ack, &sg, sg.sg_nseg - 1, 1);

	return (s.ack == VIRTIO_NET_OK ? 0 : EIO);
}

static int
vtnet_set_promisc(struct vtnet_softc *sc, bool on)
{
	return (vtnet_ctrl_rx_cmd(sc, VIRTIO_NET_CTRL_RX_PROMISC, on));
}

static int
vtnet_set_allmulti(struct vtnet_softc *sc, bool on)
{
	return (vtnet_ctrl_rx_cmd(sc, VIRTIO_NET_CTRL_RX_ALLMULTI, on));
}

static void
vtnet_rx_filter(struct vtnet_softc *sc)
{
	device_t dev;
	struct ifnet *ifp;

	dev = sc->vtnet_dev;
	ifp = sc->vtnet_ifp;

	VTNET_CORE_LOCK_ASSERT(sc);

	if (vtnet_set_promisc(sc, ifp->if_flags & IFF_PROMISC) != 0) {
		device_printf(dev, "cannot %s promiscuous mode\n",
		    ifp->if_flags & IFF_PROMISC ? "enable" : "disable");
	}

	if (vtnet_set_allmulti(sc, ifp->if_flags & IFF_ALLMULTI) != 0) {
		device_printf(dev, "cannot %s all-multicast mode\n",
		    ifp->if_flags & IFF_ALLMULTI ? "enable" : "disable");
	}
}

static u_int
vtnet_copy_ifaddr(void *arg, struct sockaddr_dl *sdl, u_int ucnt)
{
	struct vtnet_softc *sc = arg;

	if (memcmp(LLADDR(sdl), sc->vtnet_hwaddr, ETHER_ADDR_LEN) == 0)
		return (0);

	if (ucnt < VTNET_MAX_MAC_ENTRIES)
		bcopy(LLADDR(sdl),
		    &sc->vtnet_mac_filter->vmf_unicast.macs[ucnt],
		    ETHER_ADDR_LEN);

	return (1);
}

static u_int
vtnet_copy_maddr(void *arg, struct sockaddr_dl *sdl, u_int mcnt)
{
	struct vtnet_mac_filter *filter = arg;

	if (mcnt < VTNET_MAX_MAC_ENTRIES)
		bcopy(LLADDR(sdl), &filter->vmf_multicast.macs[mcnt],
		    ETHER_ADDR_LEN);

	return (1);
}

static void
vtnet_rx_filter_mac(struct vtnet_softc *sc)
{
	struct virtio_net_ctrl_hdr hdr __aligned(2);
	struct vtnet_mac_filter *filter;
	struct sglist_seg segs[4];
	struct sglist sg;
	struct ifnet *ifp;
	bool promisc, allmulti;
	u_int ucnt, mcnt;
	int error;
	uint8_t ack;

	ifp = sc->vtnet_ifp;
	filter = sc->vtnet_mac_filter;
	error = 0;

	MPASS(sc->vtnet_flags & VTNET_FLAG_CTRL_RX);
	VTNET_CORE_LOCK_ASSERT(sc);

	/* Unicast MAC addresses: */
	ucnt = if_foreach_lladdr(ifp, vtnet_copy_ifaddr, sc);
	promisc = (ucnt > VTNET_MAX_MAC_ENTRIES);

	if (promisc) {
		ucnt = 0;
		if_printf(ifp, "more than %d MAC addresses assigned, "
		    "falling back to promiscuous mode\n",
		    VTNET_MAX_MAC_ENTRIES);
	}

	/* Multicast MAC addresses: */
	mcnt = if_foreach_llmaddr(ifp, vtnet_copy_maddr, filter);
	allmulti = (mcnt > VTNET_MAX_MAC_ENTRIES);

	if (allmulti) {
		mcnt = 0;
		if_printf(ifp, "more than %d multicast MAC addresses "
		    "assigned, falling back to all-multicast mode\n",
		    VTNET_MAX_MAC_ENTRIES);
	}

	if (promisc && allmulti)
		goto out;

	filter->vmf_unicast.nentries = vtnet_gtoh32(sc, ucnt);
	filter->vmf_multicast.nentries = vtnet_gtoh32(sc, mcnt);

	hdr.class = VIRTIO_NET_CTRL_MAC;
	hdr.cmd = VIRTIO_NET_CTRL_MAC_TABLE_SET;
	ack = VIRTIO_NET_ERR;

	sglist_init(&sg, nitems(segs), segs);
	error |= sglist_append(&sg, &hdr, sizeof(struct virtio_net_ctrl_hdr));
	error |= sglist_append(&sg, &filter->vmf_unicast,
	    sizeof(uint32_t) + ucnt * ETHER_ADDR_LEN);
	error |= sglist_append(&sg, &filter->vmf_multicast,
	    sizeof(uint32_t) + mcnt * ETHER_ADDR_LEN);
	error |= sglist_append(&sg, &ack, sizeof(uint8_t));
	MPASS(error == 0 && sg.sg_nseg == nitems(segs));

	if (error == 0)
		vtnet_exec_ctrl_cmd(sc, &ack, &sg, sg.sg_nseg - 1, 1);
	if (ack != VIRTIO_NET_OK)
		if_printf(ifp, "error setting host MAC filter table\n");

out:
	if (promisc != 0 && vtnet_set_promisc(sc, true) != 0)
		if_printf(ifp, "cannot enable promiscuous mode\n");
	if (allmulti != 0 && vtnet_set_allmulti(sc, true) != 0)
		if_printf(ifp, "cannot enable all-multicast mode\n");
}

static int
vtnet_exec_vlan_filter(struct vtnet_softc *sc, int add, uint16_t tag)
{
	struct sglist_seg segs[3];
	struct sglist sg;
	struct {
		struct virtio_net_ctrl_hdr hdr __aligned(2);
		uint8_t pad1;
		uint16_t tag __aligned(2);
		uint8_t pad2;
		uint8_t ack;
	} s;
	int error;

	error = 0;
	MPASS(sc->vtnet_flags & VTNET_FLAG_VLAN_FILTER);

	s.hdr.class = VIRTIO_NET_CTRL_VLAN;
	s.hdr.cmd = add ? VIRTIO_NET_CTRL_VLAN_ADD : VIRTIO_NET_CTRL_VLAN_DEL;
	s.tag = vtnet_gtoh16(sc, tag);
	s.ack = VIRTIO_NET_ERR;

	sglist_init(&sg, nitems(segs), segs);
	error |= sglist_append(&sg, &s.hdr, sizeof(struct virtio_net_ctrl_hdr));
	error |= sglist_append(&sg, &s.tag, sizeof(uint16_t));
	error |= sglist_append(&sg, &s.ack, sizeof(uint8_t));
	MPASS(error == 0 && sg.sg_nseg == nitems(segs));

	if (error == 0)
		vtnet_exec_ctrl_cmd(sc, &s.ack, &sg, sg.sg_nseg - 1, 1);

	return (s.ack == VIRTIO_NET_OK ? 0 : EIO);
}

static void
vtnet_rx_filter_vlan(struct vtnet_softc *sc)
{
	int i, bit;
	uint32_t w;
	uint16_t tag;

	MPASS(sc->vtnet_flags & VTNET_FLAG_VLAN_FILTER);
	VTNET_CORE_LOCK_ASSERT(sc);

	/* Enable the filter for each configured VLAN. */
	for (i = 0; i < VTNET_VLAN_FILTER_NWORDS; i++) {
		w = sc->vtnet_vlan_filter[i];

		while ((bit = ffs(w) - 1) != -1) {
			w &= ~(1 << bit);
			tag = sizeof(w) * CHAR_BIT * i + bit;

			if (vtnet_exec_vlan_filter(sc, 1, tag) != 0) {
				device_printf(sc->vtnet_dev,
				    "cannot enable VLAN %d filter\n", tag);
			}
		}
	}
}

static void
vtnet_update_vlan_filter(struct vtnet_softc *sc, int add, uint16_t tag)
{
	struct ifnet *ifp;
	int idx, bit;

	ifp = sc->vtnet_ifp;
	idx = (tag >> 5) & 0x7F;
	bit = tag & 0x1F;

	if (tag == 0 || tag > 4095)
		return;

	VTNET_CORE_LOCK(sc);

	if (add)
		sc->vtnet_vlan_filter[idx] |= (1 << bit);
	else
		sc->vtnet_vlan_filter[idx] &= ~(1 << bit);

	if (ifp->if_capenable & IFCAP_VLAN_HWFILTER &&
	    ifp->if_drv_flags & IFF_DRV_RUNNING &&
	    vtnet_exec_vlan_filter(sc, add, tag) != 0) {
		device_printf(sc->vtnet_dev,
		    "cannot %s VLAN %d %s the host filter table\n",
		    add ? "add" : "remove", tag, add ? "to" : "from");
	}

	VTNET_CORE_UNLOCK(sc);
}

static void
vtnet_register_vlan(void *arg, struct ifnet *ifp, uint16_t tag)
{

	if (ifp->if_softc != arg)
		return;

	vtnet_update_vlan_filter(arg, 1, tag);
}

static void
vtnet_unregister_vlan(void *arg, struct ifnet *ifp, uint16_t tag)
{

	if (ifp->if_softc != arg)
		return;

	vtnet_update_vlan_filter(arg, 0, tag);
}

static void
vtnet_update_speed_duplex(struct vtnet_softc *sc)
{
	struct ifnet *ifp;
	uint32_t speed;

	ifp = sc->vtnet_ifp;

	if ((sc->vtnet_features & VIRTIO_NET_F_SPEED_DUPLEX) == 0)
		return;

	/* BMV: Ignore duplex. */
	speed = virtio_read_dev_config_4(sc->vtnet_dev,
	    offsetof(struct virtio_net_config, speed));
	if (speed != UINT32_MAX)
		ifp->if_baudrate = IF_Mbps(speed);
}

static int
vtnet_is_link_up(struct vtnet_softc *sc)
{
	uint16_t status;

	if ((sc->vtnet_features & VIRTIO_NET_F_STATUS) == 0)
		return (1);

	status = virtio_read_dev_config_2(sc->vtnet_dev,
	    offsetof(struct virtio_net_config, status));

	return ((status & VIRTIO_NET_S_LINK_UP) != 0);
}

static void
vtnet_update_link_status(struct vtnet_softc *sc)
{
	struct ifnet *ifp;
	int link;

	ifp = sc->vtnet_ifp;
	VTNET_CORE_LOCK_ASSERT(sc);
	link = vtnet_is_link_up(sc);

	/* Notify if the link status has changed. */
	if (link != 0 && sc->vtnet_link_active == 0) {
		vtnet_update_speed_duplex(sc);
		sc->vtnet_link_active = 1;
		if_link_state_change(ifp, LINK_STATE_UP);
	} else if (link == 0 && sc->vtnet_link_active != 0) {
		sc->vtnet_link_active = 0;
		if_link_state_change(ifp, LINK_STATE_DOWN);
	}
}

static int
vtnet_ifmedia_upd(struct ifnet *ifp __unused)
{
	return (EOPNOTSUPP);
}

static void
vtnet_ifmedia_sts(struct ifnet *ifp, struct ifmediareq *ifmr)
{
	struct vtnet_softc *sc;

	sc = ifp->if_softc;

	ifmr->ifm_status = IFM_AVALID;
	ifmr->ifm_active = IFM_ETHER;

	VTNET_CORE_LOCK(sc);
	if (vtnet_is_link_up(sc) != 0) {
		ifmr->ifm_status |= IFM_ACTIVE;
		ifmr->ifm_active |= IFM_10G_T | IFM_FDX;
	} else
		ifmr->ifm_active |= IFM_NONE;
	VTNET_CORE_UNLOCK(sc);
}

static void
vtnet_get_macaddr(struct vtnet_softc *sc)
{

	if (sc->vtnet_flags & VTNET_FLAG_MAC) {
		virtio_read_device_config_array(sc->vtnet_dev,
		    offsetof(struct virtio_net_config, mac),
		    &sc->vtnet_hwaddr[0], sizeof(uint8_t), ETHER_ADDR_LEN);
	} else {
		/* Generate a random locally administered unicast address. */
		sc->vtnet_hwaddr[0] = 0xB2;
		arc4rand(&sc->vtnet_hwaddr[1], ETHER_ADDR_LEN - 1, 0);
	}
}

static void
vtnet_set_macaddr(struct vtnet_softc *sc)
{
	device_t dev;
	int error;

	dev = sc->vtnet_dev;

	if (sc->vtnet_flags & VTNET_FLAG_CTRL_MAC) {
		error = vtnet_ctrl_mac_cmd(sc, sc->vtnet_hwaddr);
		if (error)
			device_printf(dev, "unable to set MAC address\n");
		return;
	}

	/* MAC in config is read-only in modern VirtIO. */
	if (!vtnet_modern(sc) && sc->vtnet_flags & VTNET_FLAG_MAC) {
		for (int i = 0; i < ETHER_ADDR_LEN; i++) {
			virtio_write_dev_config_1(dev,
			    offsetof(struct virtio_net_config, mac) + i,
			    sc->vtnet_hwaddr[i]);
		}
	}
}

static void
vtnet_attached_set_macaddr(struct vtnet_softc *sc)
{

	/* Assign MAC address if it was generated. */
	if ((sc->vtnet_flags & VTNET_FLAG_MAC) == 0)
		vtnet_set_macaddr(sc);
}

static void
vtnet_vlan_tag_remove(struct mbuf *m)
{
	struct ether_vlan_header *evh;

	evh = mtod(m, struct ether_vlan_header *);
	m->m_pkthdr.ether_vtag = ntohs(evh->evl_tag);
	m->m_flags |= M_VLANTAG;

	/* Strip the 802.1Q header. */
	bcopy((char *) evh, (char *) evh + ETHER_VLAN_ENCAP_LEN,
	    ETHER_HDR_LEN - ETHER_TYPE_LEN);
	m_adj(m, ETHER_VLAN_ENCAP_LEN);
}

static void
vtnet_set_rx_process_limit(struct vtnet_softc *sc)
{
	int limit;

	limit = vtnet_tunable_int(sc, "rx_process_limit",
	    vtnet_rx_process_limit);
	if (limit < 0)
		limit = INT_MAX;
	sc->vtnet_rx_process_limit = limit;
}

static void
vtnet_setup_rxq_sysctl(struct sysctl_ctx_list *ctx,
    struct sysctl_oid_list *child, struct vtnet_rxq *rxq)
{
	struct sysctl_oid *node;
	struct sysctl_oid_list *list;
	struct vtnet_rxq_stats *stats;
	char namebuf[16];

	snprintf(namebuf, sizeof(namebuf), "rxq%d", rxq->vtnrx_id);
	node = SYSCTL_ADD_NODE(ctx, child, OID_AUTO, namebuf,
	    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "Receive Queue");
	list = SYSCTL_CHILDREN(node);

	stats = &rxq->vtnrx_stats;

	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "ipackets", CTLFLAG_RD,
	    &stats->vrxs_ipackets, "Receive packets");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "ibytes", CTLFLAG_RD,
	    &stats->vrxs_ibytes, "Receive bytes");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "iqdrops", CTLFLAG_RD,
	    &stats->vrxs_iqdrops, "Receive drops");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "ierrors", CTLFLAG_RD,
	    &stats->vrxs_ierrors, "Receive errors");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "csum", CTLFLAG_RD,
	    &stats->vrxs_csum, "Receive checksum offloaded");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "csum_failed", CTLFLAG_RD,
	    &stats->vrxs_csum_failed, "Receive checksum offload failed");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "host_lro", CTLFLAG_RD,
	    &stats->vrxs_host_lro, "Receive host segmentation offloaded");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "rescheduled", CTLFLAG_RD,
	    &stats->vrxs_rescheduled,
	    "Receive interrupt handler rescheduled");
}

static void
vtnet_setup_txq_sysctl(struct sysctl_ctx_list *ctx,
    struct sysctl_oid_list *child, struct vtnet_txq *txq)
{
	struct sysctl_oid *node;
	struct sysctl_oid_list *list;
	struct vtnet_txq_stats *stats;
	char namebuf[16];

	snprintf(namebuf, sizeof(namebuf), "txq%d", txq->vtntx_id);
	node = SYSCTL_ADD_NODE(ctx, child, OID_AUTO, namebuf,
	    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "Transmit Queue");
	list = SYSCTL_CHILDREN(node);

	stats = &txq->vtntx_stats;

	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "opackets", CTLFLAG_RD,
	    &stats->vtxs_opackets, "Transmit packets");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "obytes", CTLFLAG_RD,
	    &stats->vtxs_obytes, "Transmit bytes");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "omcasts", CTLFLAG_RD,
	    &stats->vtxs_omcasts, "Transmit multicasts");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "csum", CTLFLAG_RD,
	    &stats->vtxs_csum, "Transmit checksum offloaded");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "tso", CTLFLAG_RD,
	    &stats->vtxs_tso, "Transmit TCP segmentation offloaded");
	SYSCTL_ADD_UQUAD(ctx, list, OID_AUTO, "rescheduled", CTLFLAG_RD,
	    &stats->vtxs_rescheduled,
	    "Transmit interrupt handler rescheduled");
}

static void
vtnet_setup_queue_sysctl(struct vtnet_softc *sc)
{
	device_t dev;
	struct sysctl_ctx_list *ctx;
	struct sysctl_oid *tree;
	struct sysctl_oid_list *child;
	int i;

	dev = sc->vtnet_dev;
	ctx = device_get_sysctl_ctx(dev);
	tree = device_get_sysctl_tree(dev);
	child = SYSCTL_CHILDREN(tree);

	for (i = 0; i < sc->vtnet_req_vq_pairs; i++) {
		vtnet_setup_rxq_sysctl(ctx, child, &sc->vtnet_rxqs[i]);
		vtnet_setup_txq_sysctl(ctx, child, &sc->vtnet_txqs[i]);
	}
}

static void
vtnet_setup_stat_sysctl(struct sysctl_ctx_list *ctx,
    struct sysctl_oid_list *child, struct vtnet_softc *sc)
{
	struct vtnet_statistics *stats;
	struct vtnet_rxq_stats rxaccum;
	struct vtnet_txq_stats txaccum;

	vtnet_accum_stats(sc, &rxaccum, &txaccum);

	stats = &sc->vtnet_stats;
	stats->rx_csum_offloaded = rxaccum.vrxs_csum;
	stats->rx_csum_failed = rxaccum.vrxs_csum_failed;
	stats->rx_task_rescheduled = rxaccum.vrxs_rescheduled;
	stats->tx_csum_offloaded = txaccum.vtxs_csum;
	stats->tx_tso_offloaded = txaccum.vtxs_tso;
	stats->tx_task_rescheduled = txaccum.vtxs_rescheduled;

	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "mbuf_alloc_failed",
	    CTLFLAG_RD, &stats->mbuf_alloc_failed,
	    "Mbuf cluster allocation failures");

	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_frame_too_large",
	    CTLFLAG_RD, &stats->rx_frame_too_large,
	    "Received frame larger than the mbuf chain");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_enq_replacement_failed",
	    CTLFLAG_RD, &stats->rx_enq_replacement_failed,
	    "Enqueuing the replacement receive mbuf failed");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_mergeable_failed",
	    CTLFLAG_RD, &stats->rx_mergeable_failed,
	    "Mergeable buffers receive failures");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_csum_bad_ethtype",
	    CTLFLAG_RD, &stats->rx_csum_bad_ethtype,
	    "Received checksum offloaded buffer with unsupported "
	    "Ethernet type");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_csum_bad_ipproto",
	    CTLFLAG_RD, &stats->rx_csum_bad_ipproto,
	    "Received checksum offloaded buffer with incorrect IP protocol");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_csum_bad_offset",
	    CTLFLAG_RD, &stats->rx_csum_bad_offset,
	    "Received checksum offloaded buffer with incorrect offset");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_csum_bad_proto",
	    CTLFLAG_RD, &stats->rx_csum_bad_proto,
	    "Received checksum offloaded buffer with incorrect protocol");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_csum_failed",
	    CTLFLAG_RD, &stats->rx_csum_failed,
	    "Received buffer checksum offload failed");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_csum_offloaded",
	    CTLFLAG_RD, &stats->rx_csum_offloaded,
	    "Received buffer checksum offload succeeded");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "rx_task_rescheduled",
	    CTLFLAG_RD, &stats->rx_task_rescheduled,
	    "Times the receive interrupt task rescheduled itself");

	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_csum_unknown_ethtype",
	    CTLFLAG_RD, &stats->tx_csum_unknown_ethtype,
	    "Aborted transmit of checksum offloaded buffer with unknown "
	    "Ethernet type");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_csum_proto_mismatch",
	    CTLFLAG_RD, &stats->tx_csum_proto_mismatch,
	    "Aborted transmit of checksum offloaded buffer because mismatched "
	    "protocols");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_tso_not_tcp",
	    CTLFLAG_RD, &stats->tx_tso_not_tcp,
	    "Aborted transmit of TSO buffer with non TCP protocol");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_tso_without_csum",
	    CTLFLAG_RD, &stats->tx_tso_without_csum,
	    "Aborted transmit of TSO buffer without TCP checksum offload");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_defragged",
	    CTLFLAG_RD, &stats->tx_defragged,
	    "Transmit mbufs defragged");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_defrag_failed",
	    CTLFLAG_RD, &stats->tx_defrag_failed,
	    "Aborted transmit of buffer because defrag failed");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_csum_offloaded",
	    CTLFLAG_RD, &stats->tx_csum_offloaded,
	    "Offloaded checksum of transmitted buffer");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_tso_offloaded",
	    CTLFLAG_RD, &stats->tx_tso_offloaded,
	    "Segmentation offload of transmitted buffer");
	SYSCTL_ADD_UQUAD(ctx, child, OID_AUTO, "tx_task_rescheduled",
	    CTLFLAG_RD, &stats->tx_task_rescheduled,
	    "Times the transmit interrupt task rescheduled itself");
}

static void
vtnet_setup_sysctl(struct vtnet_softc *sc)
{
	device_t dev;
	struct sysctl_ctx_list *ctx;
	struct sysctl_oid *tree;
	struct sysctl_oid_list *child;

	dev = sc->vtnet_dev;
	ctx = device_get_sysctl_ctx(dev);
	tree = device_get_sysctl_tree(dev);
	child = SYSCTL_CHILDREN(tree);

	SYSCTL_ADD_INT(ctx, child, OID_AUTO, "max_vq_pairs",
	    CTLFLAG_RD, &sc->vtnet_max_vq_pairs, 0,
	    "Number of maximum supported virtqueue pairs");
	SYSCTL_ADD_INT(ctx, child, OID_AUTO, "req_vq_pairs",
	    CTLFLAG_RD, &sc->vtnet_req_vq_pairs, 0,
	    "Number of requested virtqueue pairs");
	SYSCTL_ADD_INT(ctx, child, OID_AUTO, "act_vq_pairs",
	    CTLFLAG_RD, &sc->vtnet_act_vq_pairs, 0,
	    "Number of active virtqueue pairs");

	vtnet_setup_stat_sysctl(ctx, child, sc);
}

static void
vtnet_load_tunables(struct vtnet_softc *sc)
{

	sc->vtnet_lro_entry_count = vtnet_tunable_int(sc,
	    "lro_entry_count", vtnet_lro_entry_count);
	if (sc->vtnet_lro_entry_count < TCP_LRO_ENTRIES)
		sc->vtnet_lro_entry_count = TCP_LRO_ENTRIES;

	sc->vtnet_lro_mbufq_depth = vtnet_tunable_int(sc,
	    "lro_mbufq_depth", vtnet_lro_mbufq_depth);
}

static int
vtnet_rxq_enable_intr(struct vtnet_rxq *rxq)
{

	return (virtqueue_enable_intr(rxq->vtnrx_vq));
}

static void
vtnet_rxq_disable_intr(struct vtnet_rxq *rxq)
{

	virtqueue_disable_intr(rxq->vtnrx_vq);
}

static int
vtnet_txq_enable_intr(struct vtnet_txq *txq)
{
	struct virtqueue *vq;

	vq = txq->vtntx_vq;

	if (vtnet_txq_below_threshold(txq) != 0)
		return (virtqueue_postpone_intr(vq, VQ_POSTPONE_LONG));

	/*
	 * The free count is above our threshold. Keep the Tx interrupt
	 * disabled until the queue is fuller.
	 */
	return (0);
}

static void
vtnet_txq_disable_intr(struct vtnet_txq *txq)
{

	virtqueue_disable_intr(txq->vtntx_vq);
}

static void
vtnet_enable_rx_interrupts(struct vtnet_softc *sc)
{
	struct vtnet_rxq *rxq;
	int i;

	for (i = 0; i < sc->vtnet_act_vq_pairs; i++) {
		rxq = &sc->vtnet_rxqs[i];
		if (vtnet_rxq_enable_intr(rxq) != 0)
			taskqueue_enqueue(rxq->vtnrx_tq, &rxq->vtnrx_intrtask);
	}
}

static void
vtnet_enable_tx_interrupts(struct vtnet_softc *sc)
{
	int i;

	for (i = 0; i < sc->vtnet_act_vq_pairs; i++)
		vtnet_txq_enable_intr(&sc->vtnet_txqs[i]);
}

static void
vtnet_enable_interrupts(struct vtnet_softc *sc)
{

	vtnet_enable_rx_interrupts(sc);
	vtnet_enable_tx_interrupts(sc);
}

static void
vtnet_disable_rx_interrupts(struct vtnet_softc *sc)
{
	int i;

	for (i = 0; i < sc->vtnet_max_vq_pairs; i++)
		vtnet_rxq_disable_intr(&sc->vtnet_rxqs[i]);
}

static void
vtnet_disable_tx_interrupts(struct vtnet_softc *sc)
{
	int i;

	for (i = 0; i < sc->vtnet_max_vq_pairs; i++)
		vtnet_txq_disable_intr(&sc->vtnet_txqs[i]);
}

static void
vtnet_disable_interrupts(struct vtnet_softc *sc)
{

	vtnet_disable_rx_interrupts(sc);
	vtnet_disable_tx_interrupts(sc);
}

static int
vtnet_tunable_int(struct vtnet_softc *sc, const char *knob, int def)
{
	char path[64];

	snprintf(path, sizeof(path),
	    "hw.vtnet.%d.%s", device_get_unit(sc->vtnet_dev), knob);
	TUNABLE_INT_FETCH(path, &def);

	return (def);
}

#ifdef DEBUGNET
static void
vtnet_debugnet_init(struct ifnet *ifp, int *nrxr, int *ncl, int *clsize)
{
	struct vtnet_softc *sc;

	sc = if_getsoftc(ifp);

	VTNET_CORE_LOCK(sc);
	*nrxr = sc->vtnet_req_vq_pairs;
	*ncl = DEBUGNET_MAX_IN_FLIGHT;
	*clsize = sc->vtnet_rx_clustersz;
	VTNET_CORE_UNLOCK(sc);
}

static void
vtnet_debugnet_event(struct ifnet *ifp __unused, enum debugnet_ev event __unused)
{
}

static int
vtnet_debugnet_transmit(struct ifnet *ifp, struct mbuf *m)
{
	struct vtnet_softc *sc;
	struct vtnet_txq *txq;
	int error;

	sc = if_getsoftc(ifp);
	if ((if_getdrvflags(ifp) & (IFF_DRV_RUNNING | IFF_DRV_OACTIVE)) !=
	    IFF_DRV_RUNNING)
		return (EBUSY);

	txq = &sc->vtnet_txqs[0];
	error = vtnet_txq_encap(txq, &m, M_NOWAIT | M_USE_RESERVE);
	if (error == 0)
		(void)vtnet_txq_notify(txq);
	return (error);
}

static int
vtnet_debugnet_poll(struct ifnet *ifp, int count)
{
	struct vtnet_softc *sc;
	int i;

	sc = if_getsoftc(ifp);
	if ((if_getdrvflags(ifp) & (IFF_DRV_RUNNING | IFF_DRV_OACTIVE)) !=
	    IFF_DRV_RUNNING)
		return (EBUSY);

	(void)vtnet_txq_eof(&sc->vtnet_txqs[0]);
	for (i = 0; i < sc->vtnet_act_vq_pairs; i++)
		(void)vtnet_rxq_eof(&sc->vtnet_rxqs[i]);
	return (0);
}
#endif /* DEBUGNET */