summaryrefslogtreecommitdiffstats
path: root/src/vnet/mpls/mpls.api
blob: c8a3ffb742203d3c7f019c207b82b6efa64310c1 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
/*
 * Copyright (c) 2015-2016 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

/** \brief Bind/Unbind an MPLS local label to an IP prefix. i.e. create
           a per-prefix label entry.
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param mb_mpls_table_id - The MPLS table-id the MPLS entry will be added in
    @param mb_label - The MPLS label value to bind
    @param mb_ip_table_id - The IP table-id of the IP prefix to bind to.
    @param mb_create_table_if_needed - Create either/both tables if required.
    @param mb_is_bind - Bind or unbind
    @param mb_is_ip4 - The prefix to bind to is IPv4
    @param mb_address_length - Length of IP prefix
    @param mb_address[16] - IP prefix/
*/
autoreply define mpls_ip_bind_unbind
{
  u32 client_index;
  u32 context;
  u32 mb_mpls_table_id;
  u32 mb_label;
  u32 mb_ip_table_id;
  u8 mb_create_table_if_needed;
  u8 mb_is_bind;
  u8 mb_is_ip4;
  u8 mb_address_length;
  u8 mb_address[16];
};

/** \brief MPLS tunnel Add / del route
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param mt_is_add - Is this a route add or delete
    @param mt_sw_if_index - The SW interface index of the tunnel to delete
    @param mt_is_multicast - Is the tunnel's underlying LSP multicast
    @param mt_next_hop_proto_is_ip4 - The next-hop is IPV4
    @param mt_next_hop_weight - The weight, for UCMP
    @param mt_next_hop[16] - the nextop address
    @param mt_next_hop_sw_if_index - the next-hop SW interface
    @param mt_next_hop_table_id - the next-hop table-id (if appropriate)
    @param mt_next_hop_n_out_labels - the number of next-hop output labels
    @param mt_next_hop_out_label_stack - the next-hop output label stack,  outer most first
*/
define mpls_tunnel_add_del
{
  u32 client_index;
  u32 context;
  u32 mt_sw_if_index;
  u8 mt_is_add;
  u8 mt_l2_only;
  u8 mt_is_multicast;
  u8 mt_next_hop_proto_is_ip4;
  u8 mt_next_hop_weight;
  u8 mt_next_hop[16];
  u8 mt_next_hop_n_out_labels;
  u32 mt_next_hop_sw_if_index;
  u32 mt_next_hop_table_id;
  u32 mt_next_hop_out_label_stack[mt_next_hop_n_out_labels];
};

/** \brief Reply for MPLS tunnel add / del request
    @param context - returned sender context, to match reply w/ request
    @param retval - return code
    @param sw_if_index - SW interface index of the tunnel created
*/
define mpls_tunnel_add_del_reply
{
  u32 context;
  i32 retval;
  u32 sw_if_index;
};

/** \brief Dump mpls eth tunnel table
    @param client_index - opaque cookie to identify the sender
    @param tunnel_index - eth tunnel identifier or -1 in case of all tunnels
*/
define mpls_tunnel_dump
{
  u32 client_index;
  u32 context;
  i32 tunnel_index;
};

/** \brief FIB path
    @param sw_if_index - index of the interface
    @param weight - The weight, for UCMP
    @param is_local - local if non-zero, else remote
    @param is_drop - Drop the packet
    @param is_unreach - Drop the packet and rate limit send ICMP unreachable
    @param is_prohibit - Drop the packet and rate limit send ICMP prohibited
    @param afi - the afi of the next hop, IP46_TYPE_IP4=1, IP46_TYPE_IP6=2
    @param next_hop[16] - the next hop address

    WARNING: this type is replicated, pending cleanup completion

*/
typeonly manual_print manual_endian define fib_path2
{
  u32 sw_if_index;
  u32 weight;
  u8 is_local;
  u8 is_drop;
  u8 is_unreach;
  u8 is_prohibit;
  u8 afi;
  u8 next_hop[16];
  u32 labels[16];
};

/** \brief mpls tunnel details
*/
manual_endian manual_print define mpls_tunnel_details
{
  u32 context;
  u8 mt_sw_if_index;
  u8 mt_tunnel_index;
  u8 mt_l2_only;
  u8 mt_is_multicast;
  u32 mt_count;
  vl_api_fib_path2_t mt_paths[mt_count];
};

/** \brief MPLS Route Add / del route
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param mr_label - The MPLS label value
    @param mr_eos - The End of stack bit
    @param mr_table_id - The MPLS table-id the route is added in
    @param mr_classify_table_index - If this is a classify route, 
                                     this is the classify table index
    @param  mr_create_table_if_needed - If the MPLS or IP tables do not exist,
                                        create them
    @param mr_is_add - Is this a route add or delete
    @param mr_is_classify - Is this route result a classify
    @param mr_is_multicast - Is this a multicast route
    @param mr_is_multipath - Is this route update a multipath - i.e. is this
                             a path addition to an existing route
    @param mr_is_resolve_host - Recurse resolution constraint via a host prefix
    @param mr_is_resolve_attached - Recurse resolution constraint via attached prefix
    @param mr_is_interface_rx - Interface Receive path
    @param mr_is_interface_rx - RPF-ID Receive path. The next-hop interface
                                is used as the RPF-ID
    @param mr_next_hop_proto_is_ip4 - The next-hop is IPV4
    @param mr_next_hop_weight - The weight, for UCMP
    @param mr_next_hop[16] - the nextop address
    @param mr_next_hop_sw_if_index - the next-hop SW interface
    @param mr_next_hop_table_id - the next-hop table-id (if appropriate)
    @param mr_next_hop_n_out_labels - the number of labels in the label stack
    @param mr_next_hop_out_label_stack - the next-hop output label stack, outer most first
    @param next_hop_via_label - The next-hop is a resolved via a local label
*/
autoreply define mpls_route_add_del
{
  u32 client_index;
  u32 context;
  u32 mr_label;
  u8 mr_eos;
  u32 mr_table_id;
  u32 mr_classify_table_index;
  u8 mr_create_table_if_needed;
  u8 mr_is_add;
  u8 mr_is_classify;
  u8 mr_is_multicast;
  u8 mr_is_multipath;
  u8 mr_is_resolve_host;
  u8 mr_is_resolve_attached;
  u8 mr_is_interface_rx;
  u8 mr_is_rpf_id;
  u8 mr_next_hop_proto_is_ip4;
  u8 mr_next_hop_weight;
  u8 mr_next_hop[16];
  u8 mr_next_hop_n_out_labels;
  u32 mr_next_hop_sw_if_index;
  u32 mr_next_hop_table_id;
  u32 mr_next_hop_via_label;
  u32 mr_next_hop_out_label_stack[mr_next_hop_n_out_labels];
};

/** \brief Dump MPLS fib table
    @param client_index - opaque cookie to identify the sender
*/
define mpls_fib_dump
{
  u32 client_index;
  u32 context;
};

/** \brief mpls FIB table response
    @param table_id - MPLS fib table id
    @param s_bit - End-of-stack bit
    @param label - MPLS label value
    @param count - the number of fib_path in path
    @param path  - array of of fib_path structures
*/
manual_endian manual_print define mpls_fib_details
{
  u32 context;
  u32 table_id;
  u8  eos_bit;
  u32 label;
  u32 count;
  vl_api_fib_path2_t path[count];
};

/*
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */
 
n724' href='#n724'>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
/*
 *------------------------------------------------------------------
 * vhost.c - vhost-user
 *
 * Copyright (c) 2014-2018 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 *------------------------------------------------------------------
 */

#include <fcntl.h>		/* for open */
#include <sys/ioctl.h>
#include <sys/socket.h>
#include <sys/un.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <sys/uio.h>		/* for iovec */
#include <netinet/in.h>
#include <sys/vfs.h>

#include <linux/if_arp.h>
#include <linux/if_tun.h>

#include <vlib/vlib.h>
#include <vlib/unix/unix.h>

#include <vnet/ethernet/ethernet.h>
#include <vnet/devices/devices.h>
#include <vnet/feature/feature.h>

#include <vnet/devices/virtio/vhost_user.h>
#include <vnet/devices/virtio/vhost_user_inline.h>

/**
 * @file
 * @brief vHost User Device Driver.
 *
 * This file contains the source code for vHost User interface.
 */


vlib_node_registration_t vhost_user_send_interrupt_node;

/* *INDENT-OFF* */
vhost_user_main_t vhost_user_main = {
  .mtu_bytes = 1518,
};

VNET_HW_INTERFACE_CLASS (vhost_interface_class, static) = {
  .name = "vhost-user",
};
/* *INDENT-ON* */

static long
get_huge_page_size (int fd)
{
  struct statfs s;
  fstatfs (fd, &s);
  return s.f_bsize;
}

static void
unmap_all_mem_regions (vhost_user_intf_t * vui)
{
  int i, r, q;
  vhost_user_vring_t *vq;

  for (i = 0; i < vui->nregions; i++)
    {
      if (vui->region_mmap_addr[i] != MAP_FAILED)
	{

	  long page_sz = get_huge_page_size (vui->region_mmap_fd[i]);

	  ssize_t map_sz = (vui->regions[i].memory_size +
			    vui->regions[i].mmap_offset +
			    page_sz - 1) & ~(page_sz - 1);

	  r =
	    munmap (vui->region_mmap_addr[i] - vui->regions[i].mmap_offset,
		    map_sz);

	  vu_log_debug (vui, "unmap memory region %d addr 0x%lx len 0x%lx "
			"page_sz 0x%x", i, vui->region_mmap_addr[i], map_sz,
			page_sz);

	  vui->region_mmap_addr[i] = MAP_FAILED;

	  if (r == -1)
	    {
	      vu_log_err (vui, "failed to unmap memory region (errno %d)",
			  errno);
	    }
	  close (vui->region_mmap_fd[i]);
	}
    }
  vui->nregions = 0;

  for (q = 0; q < vui->num_qid; q++)
    {
      vq = &vui->vrings[q];
      vq->avail = 0;
      vq->used = 0;
      vq->desc = 0;
    }
}

static_always_inline void
vhost_user_tx_thread_placement (vhost_user_intf_t * vui)
{
  //Let's try to assign one queue to each thread
  u32 qid;
  u32 thread_index = 0;

  vui->use_tx_spinlock = 0;
  while (1)
    {
      for (qid = 0; qid < vui->num_qid / 2; qid++)
	{
	  vhost_user_vring_t *rxvq = &vui->vrings[VHOST_VRING_IDX_RX (qid)];
	  if (!rxvq->started || !rxvq->enabled)
	    continue;

	  vui->per_cpu_tx_qid[thread_index] = qid;
	  thread_index++;
	  if (thread_index == vlib_get_thread_main ()->n_vlib_mains)
	    return;
	}
      //We need to loop, meaning the spinlock has to be used
      vui->use_tx_spinlock = 1;
      if (thread_index == 0)
	{
	  //Could not find a single valid one
	  for (thread_index = 0;
	       thread_index < vlib_get_thread_main ()->n_vlib_mains;
	       thread_index++)
	    {
	      vui->per_cpu_tx_qid[thread_index] = 0;
	    }
	  return;
	}
    }
}

/**
 * @brief Unassign existing interface/queue to thread mappings and re-assign
 * new interface/queue to thread mappings
 */
static_always_inline void
vhost_user_rx_thread_placement (vhost_user_intf_t * vui, u32 qid)
{
  vhost_user_vring_t *txvq = &vui->vrings[qid];
  vnet_main_t *vnm = vnet_get_main ();
  int rv;
  u32 q = qid >> 1;

  ASSERT ((qid & 1) == 1);	// should be odd
  // Assign new queue mappings for the interface
  vnet_hw_interface_set_input_node (vnm, vui->hw_if_index,
				    vhost_user_input_node.index);
  vnet_hw_interface_assign_rx_thread (vnm, vui->hw_if_index, q, ~0);
  if (txvq->mode == VNET_HW_IF_RX_MODE_UNKNOWN)
    /* Set polling as the default */
    txvq->mode = VNET_HW_IF_RX_MODE_POLLING;
  txvq->qid = q;
  rv = vnet_hw_interface_set_rx_mode (vnm, vui->hw_if_index, q, txvq->mode);
  if (rv)
    vu_log_warn (vui, "unable to set rx mode for interface %d, "
		 "queue %d: rc=%d", vui->hw_if_index, q, rv);
}

/** @brief Returns whether at least one TX and one RX vring are enabled */
static_always_inline int
vhost_user_intf_ready (vhost_user_intf_t * vui)
{
  int i, found[2] = { };	//RX + TX

  for (i = 0; i < vui->num_qid; i++)
    if (vui->vrings[i].started && vui->vrings[i].enabled)
      found[i & 1] = 1;

  return found[0] && found[1];
}

static_always_inline void
vhost_user_update_iface_state (vhost_user_intf_t * vui)
{
  /* if we have pointers to descriptor table, go up */
  int is_ready = vhost_user_intf_ready (vui);
  if (is_ready != vui->is_ready)
    {
      vu_log_debug (vui, "interface %d %s", vui->sw_if_index,
		    is_ready ? "ready" : "down");
      if (vui->admin_up)
	vnet_hw_interface_set_flags (vnet_get_main (), vui->hw_if_index,
				     is_ready ? VNET_HW_INTERFACE_FLAG_LINK_UP
				     : 0);
      vui->is_ready = is_ready;
    }
}

static void
vhost_user_set_interrupt_pending (vhost_user_intf_t * vui, u32 ifq)
{
  u32 qid;
  vnet_main_t *vnm = vnet_get_main ();

  qid = ifq & 0xff;
  if ((qid & 1) == 0)
    /* Only care about the odd number, or TX, virtqueue */
    return;

  if (vhost_user_intf_ready (vui))
    // qid >> 1 is to convert virtqueue number to vring queue index
    vnet_device_input_set_interrupt_pending (vnm, vui->hw_if_index, qid >> 1);
}

static clib_error_t *
vhost_user_callfd_read_ready (clib_file_t * uf)
{
  __attribute__ ((unused)) int n;
  u8 buff[8];

  n = read (uf->file_descriptor, ((char *) &buff), 8);

  return 0;
}

static_always_inline void
vhost_user_thread_placement (vhost_user_intf_t * vui, u32 qid)
{
  if (qid & 1)			// RX is odd, TX is even
    {
      if (vui->vrings[qid].qid == -1)
	vhost_user_rx_thread_placement (vui, qid);
    }
  else
    vhost_user_tx_thread_placement (vui);
}

static clib_error_t *
vhost_user_kickfd_read_ready (clib_file_t * uf)
{
  __attribute__ ((unused)) int n;
  u8 buff[8];
  vhost_user_intf_t *vui =
    pool_elt_at_index (vhost_user_main.vhost_user_interfaces,
		       uf->private_data >> 8);
  u32 qid = uf->private_data & 0xff;

  n = read (uf->file_descriptor, ((char *) &buff), 8);
  vu_log_debug (vui, "if %d KICK queue %d", vui->hw_if_index, qid);
  if (!vui->vrings[qid].started ||
      (vhost_user_intf_ready (vui) != vui->is_ready))
    {
      if (vui->vrings[qid].started == 0)
	{
	  vui->vrings[qid].started = 1;
	  vhost_user_thread_placement (vui, qid);
	  vhost_user_update_iface_state (vui);
	}
    }

  vhost_user_set_interrupt_pending (vui, uf->private_data);
  return 0;
}

static_always_inline void
vhost_user_vring_init (vhost_user_intf_t * vui, u32 qid)
{
  vhost_user_vring_t *vring = &vui->vrings[qid];

  clib_memset (vring, 0, sizeof (*vring));
  vring->kickfd_idx = ~0;
  vring->callfd_idx = ~0;
  vring->errfd = -1;
  vring->qid = -1;

  clib_spinlock_init (&vring->vring_lock);

  /*
   * We have a bug with some qemu 2.5, and this may be a fix.
   * Feel like interpretation holy text, but this is from vhost-user.txt.
   * "
   * One queue pair is enabled initially. More queues are enabled
   * dynamically, by sending message VHOST_USER_SET_VRING_ENABLE.
   * "
   * Don't know who's right, but this is what DPDK does.
   */
  if (qid == 0 || qid == 1)
    vring->enabled = 1;
}

static_always_inline void
vhost_user_vring_close (vhost_user_intf_t * vui, u32 qid)
{
  vhost_user_vring_t *vring = &vui->vrings[qid];

  if (vring->kickfd_idx != ~0)
    {
      clib_file_t *uf = pool_elt_at_index (file_main.file_pool,
					   vring->kickfd_idx);
      clib_file_del (&file_main, uf);
      vring->kickfd_idx = ~0;
    }
  if (vring->callfd_idx != ~0)
    {
      clib_file_t *uf = pool_elt_at_index (file_main.file_pool,
					   vring->callfd_idx);
      clib_file_del (&file_main, uf);
      vring->callfd_idx = ~0;
    }
  if (vring->errfd != -1)
    {
      close (vring->errfd);
      vring->errfd = -1;
    }

  clib_spinlock_free (&vring->vring_lock);

  // save the qid so that we don't need to unassign and assign_rx_thread
  // when the interface comes back up. They are expensive calls.
  u16 q = vui->vrings[qid].qid;
  vhost_user_vring_init (vui, qid);
  vui->vrings[qid].qid = q;
}

static_always_inline void
vhost_user_if_disconnect (vhost_user_intf_t * vui)
{
  vnet_main_t *vnm = vnet_get_main ();
  int q;

  vnet_hw_interface_set_flags (vnm, vui->hw_if_index, 0);

  if (vui->clib_file_index != ~0)
    {
      clib_file_del (&file_main, file_main.file_pool + vui->clib_file_index);
      vui->clib_file_index = ~0;
    }

  vui->is_ready = 0;

  for (q = 0; q < vui->num_qid; q++)
    vhost_user_vring_close (vui, q);

  unmap_all_mem_regions (vui);
  vu_log_debug (vui, "interface ifindex %d disconnected", vui->sw_if_index);
}

static clib_error_t *
vhost_user_socket_read (clib_file_t * uf)
{
  int n, i, j;
  int fd, number_of_fds = 0;
  int fds[VHOST_MEMORY_MAX_NREGIONS];
  vhost_user_msg_t msg;
  struct msghdr mh;
  struct iovec iov[1];
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui;
  struct cmsghdr *cmsg;
  u8 q;
  clib_file_t template = { 0 };
  vnet_main_t *vnm = vnet_get_main ();
  vlib_main_t *vm = vlib_get_main ();

  vui = pool_elt_at_index (vum->vhost_user_interfaces, uf->private_data);

  char control[CMSG_SPACE (VHOST_MEMORY_MAX_NREGIONS * sizeof (int))];

  clib_memset (&mh, 0, sizeof (mh));
  clib_memset (control, 0, sizeof (control));

  for (i = 0; i < VHOST_MEMORY_MAX_NREGIONS; i++)
    fds[i] = -1;

  /* set the payload */
  iov[0].iov_base = (void *) &msg;
  iov[0].iov_len = VHOST_USER_MSG_HDR_SZ;

  mh.msg_iov = iov;
  mh.msg_iovlen = 1;
  mh.msg_control = control;
  mh.msg_controllen = sizeof (control);

  n = recvmsg (uf->file_descriptor, &mh, 0);

  if (n != VHOST_USER_MSG_HDR_SZ)
    {
      if (n == -1)
	{
	  vu_log_debug (vui, "recvmsg returned error %d %s", errno,
			strerror (errno));
	}
      else
	{
	  vu_log_debug (vui, "n (%d) != VHOST_USER_MSG_HDR_SZ (%d)",
			n, VHOST_USER_MSG_HDR_SZ);
	}
      goto close_socket;
    }

  if (mh.msg_flags & MSG_CTRUNC)
    {
      vu_log_debug (vui, "MSG_CTRUNC is set");
      goto close_socket;
    }

  cmsg = CMSG_FIRSTHDR (&mh);

  if (cmsg && (cmsg->cmsg_len > 0) && (cmsg->cmsg_level == SOL_SOCKET) &&
      (cmsg->cmsg_type == SCM_RIGHTS) &&
      (cmsg->cmsg_len - CMSG_LEN (0) <=
       VHOST_MEMORY_MAX_NREGIONS * sizeof (int)))
    {
      number_of_fds = (cmsg->cmsg_len - CMSG_LEN (0)) / sizeof (int);
      clib_memcpy_fast (fds, CMSG_DATA (cmsg), number_of_fds * sizeof (int));
    }

  /* version 1, no reply bit set */
  if ((msg.flags & 7) != 1)
    {
      vu_log_debug (vui, "malformed message received. closing socket");
      goto close_socket;
    }

  {
    int rv;
    rv =
      read (uf->file_descriptor, ((char *) &msg) + VHOST_USER_MSG_HDR_SZ,
	    msg.size);
    if (rv < 0)
      {
	vu_log_debug (vui, "read failed %s", strerror (errno));
	goto close_socket;
      }
    else if (rv != msg.size)
      {
	vu_log_debug (vui, "message too short (read %dB should be %dB)", rv,
		      msg.size);
	goto close_socket;
      }
  }

  switch (msg.request)
    {
    case VHOST_USER_GET_FEATURES:
      msg.flags |= 4;
      msg.u64 = VIRTIO_FEATURE (VIRTIO_NET_F_MRG_RXBUF) |
	VIRTIO_FEATURE (VIRTIO_NET_F_CTRL_VQ) |
	VIRTIO_FEATURE (VIRTIO_F_ANY_LAYOUT) |
	VIRTIO_FEATURE (VIRTIO_RING_F_INDIRECT_DESC) |
	VIRTIO_FEATURE (VHOST_F_LOG_ALL) |
	VIRTIO_FEATURE (VIRTIO_NET_F_GUEST_ANNOUNCE) |
	VIRTIO_FEATURE (VIRTIO_NET_F_MQ) |
	VIRTIO_FEATURE (VHOST_USER_F_PROTOCOL_FEATURES) |
	VIRTIO_FEATURE (VIRTIO_F_VERSION_1);
      msg.u64 &= vui->feature_mask;

      if (vui->enable_event_idx)
	msg.u64 |= VIRTIO_FEATURE (VIRTIO_RING_F_EVENT_IDX);
      if (vui->enable_gso)
	msg.u64 |= FEATURE_VIRTIO_NET_F_HOST_GUEST_TSO_FEATURE_BITS;
      if (vui->enable_packed)
	msg.u64 |= VIRTIO_FEATURE (VIRTIO_F_RING_PACKED);

      msg.size = sizeof (msg.u64);
      vu_log_debug (vui, "if %d msg VHOST_USER_GET_FEATURES - reply "
		    "0x%016llx", vui->hw_if_index, msg.u64);
      n =
	send (uf->file_descriptor, &msg, VHOST_USER_MSG_HDR_SZ + msg.size, 0);
      if (n != (msg.size + VHOST_USER_MSG_HDR_SZ))
	{
	  vu_log_debug (vui, "could not send message response");
	  goto close_socket;
	}
      break;

    case VHOST_USER_SET_FEATURES:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_FEATURES features "
		    "0x%016llx", vui->hw_if_index, msg.u64);

      vui->features = msg.u64;

      if (vui->features &
	  (VIRTIO_FEATURE (VIRTIO_NET_F_MRG_RXBUF) |
	   VIRTIO_FEATURE (VIRTIO_F_VERSION_1)))
	vui->virtio_net_hdr_sz = 12;
      else
	vui->virtio_net_hdr_sz = 10;

      vui->is_any_layout =
	(vui->features & VIRTIO_FEATURE (VIRTIO_F_ANY_LAYOUT)) ? 1 : 0;

      ASSERT (vui->virtio_net_hdr_sz < VLIB_BUFFER_PRE_DATA_SIZE);
      vnet_hw_interface_t *hw = vnet_get_hw_interface (vnm, vui->hw_if_index);
      if (vui->enable_gso &&
	  ((vui->features & FEATURE_VIRTIO_NET_F_HOST_GUEST_TSO_FEATURE_BITS)
	   == FEATURE_VIRTIO_NET_F_HOST_GUEST_TSO_FEATURE_BITS))
	hw->flags |=
	  (VNET_HW_INTERFACE_FLAG_SUPPORTS_GSO |
	   VNET_HW_INTERFACE_FLAG_SUPPORTS_TX_L4_CKSUM_OFFLOAD);
      else
	hw->flags &= ~(VNET_HW_INTERFACE_FLAG_SUPPORTS_GSO |
		       VNET_HW_INTERFACE_FLAG_SUPPORTS_TX_L4_CKSUM_OFFLOAD);
      vnet_hw_interface_set_flags (vnm, vui->hw_if_index, 0);
      vui->is_ready = 0;
      vhost_user_update_iface_state (vui);
      break;

    case VHOST_USER_SET_MEM_TABLE:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_MEM_TABLE nregions %d",
		    vui->hw_if_index, msg.memory.nregions);

      if ((msg.memory.nregions < 1) ||
	  (msg.memory.nregions > VHOST_MEMORY_MAX_NREGIONS))
	{
	  vu_log_debug (vui, "number of mem regions must be between 1 and %i",
			VHOST_MEMORY_MAX_NREGIONS);
	  goto close_socket;
	}

      if (msg.memory.nregions != number_of_fds)
	{
	  vu_log_debug (vui, "each memory region must have FD");
	  goto close_socket;
	}

      /* Do the mmap without barrier sync */
      void *region_mmap_addr[VHOST_MEMORY_MAX_NREGIONS];
      for (i = 0; i < msg.memory.nregions; i++)
	{
	  long page_sz = get_huge_page_size (fds[i]);

	  /* align size to page */
	  ssize_t map_sz = (msg.memory.regions[i].memory_size +
			    msg.memory.regions[i].mmap_offset +
			    page_sz - 1) & ~(page_sz - 1);

	  region_mmap_addr[i] = mmap (0, map_sz, PROT_READ | PROT_WRITE,
				      MAP_SHARED, fds[i], 0);
	  if (region_mmap_addr[i] == MAP_FAILED)
	    {
	      vu_log_err (vui, "failed to map memory. errno is %d", errno);
	      for (j = 0; j < i; j++)
		munmap (region_mmap_addr[j], map_sz);
	      goto close_socket;
	    }
	  vu_log_debug (vui, "map memory region %d addr 0 len 0x%lx fd %d "
			"mapped 0x%lx page_sz 0x%x", i, map_sz, fds[i],
			region_mmap_addr[i], page_sz);
	}

      vlib_worker_thread_barrier_sync (vm);
      unmap_all_mem_regions (vui);
      for (i = 0; i < msg.memory.nregions; i++)
	{
	  clib_memcpy_fast (&(vui->regions[i]), &msg.memory.regions[i],
			    sizeof (vhost_user_memory_region_t));

	  vui->region_mmap_addr[i] = region_mmap_addr[i];
	  vui->region_guest_addr_lo[i] = vui->regions[i].guest_phys_addr;
	  vui->region_guest_addr_hi[i] = vui->regions[i].guest_phys_addr +
	    vui->regions[i].memory_size;

	  vui->region_mmap_addr[i] += vui->regions[i].mmap_offset;
	  vui->region_mmap_fd[i] = fds[i];

	  vui->nregions++;
	}

      /*
       * Re-compute desc, used, and avail descriptor table if vring address
       * is set.
       */
      for (q = 0; q < vui->num_qid; q++)
	{
	  if (vui->vrings[q].desc_user_addr &&
	      vui->vrings[q].used_user_addr && vui->vrings[q].avail_user_addr)
	    {
	      vui->vrings[q].desc =
		map_user_mem (vui, vui->vrings[q].desc_user_addr);
	      vui->vrings[q].used =
		map_user_mem (vui, vui->vrings[q].used_user_addr);
	      vui->vrings[q].avail =
		map_user_mem (vui, vui->vrings[q].avail_user_addr);
	    }
	}
      vlib_worker_thread_barrier_release (vm);
      break;

    case VHOST_USER_SET_VRING_NUM:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_VRING_NUM idx %d num %d",
		    vui->hw_if_index, msg.state.index, msg.state.num);

      if ((msg.state.num > 32768) ||	/* maximum ring size is 32768 */
	  (msg.state.num == 0) ||	/* it cannot be zero */
	  ((msg.state.num - 1) & msg.state.num) ||	/* must be power of 2 */
	  (msg.state.index >= vui->num_qid))
	{
	  vu_log_debug (vui, "invalid VHOST_USER_SET_VRING_NUM: msg.state.num"
			" %d, msg.state.index %d, curruent max q %d",
			msg.state.num, msg.state.index, vui->num_qid);
	  goto close_socket;
	}
      vui->vrings[msg.state.index].qsz_mask = msg.state.num - 1;
      break;

    case VHOST_USER_SET_VRING_ADDR:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_VRING_ADDR idx %d",
		    vui->hw_if_index, msg.state.index);

      if (msg.state.index >= vui->num_qid)
	{
	  vu_log_debug (vui, "invalid vring index VHOST_USER_SET_VRING_ADDR:"
			" %u >= %u", msg.state.index, vui->num_qid);
	  goto close_socket;
	}

      if (msg.size < sizeof (msg.addr))
	{
	  vu_log_debug (vui, "vhost message is too short (%d < %d)",
			msg.size, sizeof (msg.addr));
	  goto close_socket;
	}

      vring_desc_t *desc = map_user_mem (vui, msg.addr.desc_user_addr);
      vring_used_t *used = map_user_mem (vui, msg.addr.used_user_addr);
      vring_avail_t *avail = map_user_mem (vui, msg.addr.avail_user_addr);

      if ((desc == NULL) || (used == NULL) || (avail == NULL))
	{
	  vu_log_debug (vui, "failed to map user memory for hw_if_index %d",
			vui->hw_if_index);
	  goto close_socket;
	}

      vui->vrings[msg.state.index].desc_user_addr = msg.addr.desc_user_addr;
      vui->vrings[msg.state.index].used_user_addr = msg.addr.used_user_addr;
      vui->vrings[msg.state.index].avail_user_addr = msg.addr.avail_user_addr;

      vlib_worker_thread_barrier_sync (vm);
      vui->vrings[msg.state.index].desc = desc;
      vui->vrings[msg.state.index].used = used;
      vui->vrings[msg.state.index].avail = avail;

      vui->vrings[msg.state.index].log_guest_addr = msg.addr.log_guest_addr;
      vui->vrings[msg.state.index].log_used =
	(msg.addr.flags & (1 << VHOST_VRING_F_LOG)) ? 1 : 0;

      /* Spec says: If VHOST_USER_F_PROTOCOL_FEATURES has not been negotiated,
         the ring is initialized in an enabled state. */
      if (!(vui->features & VIRTIO_FEATURE (VHOST_USER_F_PROTOCOL_FEATURES)))
	vui->vrings[msg.state.index].enabled = 1;

      vui->vrings[msg.state.index].last_used_idx =
	vui->vrings[msg.state.index].last_avail_idx =
	vui->vrings[msg.state.index].used->idx;
      vui->vrings[msg.state.index].last_kick =
	vui->vrings[msg.state.index].last_used_idx;

      /* tell driver that we don't want interrupts */
      if (vhost_user_is_packed_ring_supported (vui))
	vui->vrings[msg.state.index].used_event->flags =
	  VRING_EVENT_F_DISABLE;
      else
	vui->vrings[msg.state.index].used->flags = VRING_USED_F_NO_NOTIFY;
      vlib_worker_thread_barrier_release (vm);
      vhost_user_update_iface_state (vui);
      break;

    case VHOST_USER_SET_OWNER:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_OWNER", vui->hw_if_index);
      break;

    case VHOST_USER_RESET_OWNER:
      vu_log_debug (vui, "if %d msg VHOST_USER_RESET_OWNER",
		    vui->hw_if_index);
      break;

    case VHOST_USER_SET_VRING_CALL:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_VRING_CALL %d",
		    vui->hw_if_index, msg.u64);

      q = (u8) (msg.u64 & 0xFF);
      if (vui->num_qid > q)
	{
	  /* if there is old fd, delete and close it */
	  if (vui->vrings[q].callfd_idx != ~0)
	    {
	      clib_file_t *uf = pool_elt_at_index (file_main.file_pool,
						   vui->vrings[q].callfd_idx);
	      clib_file_del (&file_main, uf);
	      vui->vrings[q].callfd_idx = ~0;
	    }
	}
      else if (vec_len (vui->vrings) > q)
	{
	  /* grow vrings by pair (RX + TX) */
	  vui->num_qid = (q & 1) ? (q + 1) : (q + 2);
	}
      else
	{
	  u32 i, new_max_q, old_max_q = vec_len (vui->vrings);

	  /*
	   * Double the array size if it is less than 64 entries.
	   * Slow down thereafter.
	   */
	  if (vec_len (vui->vrings) < (VHOST_VRING_INIT_MQ_PAIR_SZ << 3))
	    new_max_q = vec_len (vui->vrings) << 1;
	  else
	    new_max_q = vec_len (vui->vrings) +
	      (VHOST_VRING_INIT_MQ_PAIR_SZ << 2);
	  if (new_max_q > (VHOST_VRING_MAX_MQ_PAIR_SZ << 1))
	    new_max_q = (VHOST_VRING_MAX_MQ_PAIR_SZ << 1);

	  /* sync with the worker threads, vrings may move due to realloc */
	  vlib_worker_thread_barrier_sync (vm);
	  vec_validate_aligned (vui->vrings, new_max_q - 1,
				CLIB_CACHE_LINE_BYTES);
	  vlib_worker_thread_barrier_release (vm);

	  for (i = old_max_q; i < vec_len (vui->vrings); i++)
	    vhost_user_vring_init (vui, i);

	  /* grow vrings by pair (RX + TX) */
	  vui->num_qid = (q & 1) ? (q + 1) : (q + 2);
	}

      if (!(msg.u64 & VHOST_USER_VRING_NOFD_MASK))
	{
	  if (number_of_fds != 1)
	    {
	      vu_log_debug (vui, "More than one fd received !");
	      goto close_socket;
	    }

	  template.read_function = vhost_user_callfd_read_ready;
	  template.file_descriptor = fds[0];
	  template.private_data =
	    ((vui - vhost_user_main.vhost_user_interfaces) << 8) + q;
	  template.description = format (0, "vhost user");
	  vui->vrings[q].callfd_idx = clib_file_add (&file_main, &template);
	}
      else
	vui->vrings[q].callfd_idx = ~0;
      break;

    case VHOST_USER_SET_VRING_KICK:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_VRING_KICK %d",
		    vui->hw_if_index, msg.u64);

      q = (u8) (msg.u64 & 0xFF);
      if (q >= vui->num_qid)
	{
	  vu_log_debug (vui, "invalid vring index VHOST_USER_SET_VRING_KICK:"
			" %u >= %u", q, vui->num_qid);
	  goto close_socket;
	}

      if (vui->vrings[q].kickfd_idx != ~0)
	{
	  clib_file_t *uf = pool_elt_at_index (file_main.file_pool,
					       vui->vrings[q].kickfd_idx);
	  clib_file_del (&file_main, uf);
	  vui->vrings[q].kickfd_idx = ~0;
	}

      if (!(msg.u64 & VHOST_USER_VRING_NOFD_MASK))
	{
	  if (number_of_fds != 1)
	    {
	      vu_log_debug (vui, "More than one fd received !");
	      goto close_socket;
	    }

	  template.read_function = vhost_user_kickfd_read_ready;
	  template.file_descriptor = fds[0];
	  template.private_data =
	    (((uword) (vui - vhost_user_main.vhost_user_interfaces)) << 8) +
	    q;
	  vui->vrings[q].kickfd_idx = clib_file_add (&file_main, &template);
	}
      else
	{
	  //When no kickfd is set, the queue is initialized as started
	  vui->vrings[q].kickfd_idx = ~0;
	  vui->vrings[q].started = 1;
	  vhost_user_thread_placement (vui, q);
	}
      vhost_user_update_iface_state (vui);
      break;

    case VHOST_USER_SET_VRING_ERR:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_VRING_ERR %d",
		    vui->hw_if_index, msg.u64);

      q = (u8) (msg.u64 & 0xFF);
      if (q >= vui->num_qid)
	{
	  vu_log_debug (vui, "invalid vring index VHOST_USER_SET_VRING_ERR:"
			" %u >= %u", q, vui->num_qid);
	  goto close_socket;
	}

      if (vui->vrings[q].errfd != -1)
	close (vui->vrings[q].errfd);

      if (!(msg.u64 & VHOST_USER_VRING_NOFD_MASK))
	{
	  if (number_of_fds != 1)
	    goto close_socket;

	  vui->vrings[q].errfd = fds[0];
	}
      else
	vui->vrings[q].errfd = -1;
      break;

    case VHOST_USER_SET_VRING_BASE:
      vu_log_debug (vui,
		    "if %d msg VHOST_USER_SET_VRING_BASE idx %d num 0x%x",
		    vui->hw_if_index, msg.state.index, msg.state.num);
      if (msg.state.index >= vui->num_qid)
	{
	  vu_log_debug (vui, "invalid vring index VHOST_USER_SET_VRING_ADDR:"
			" %u >= %u", msg.state.index, vui->num_qid);
	  goto close_socket;
	}
      vlib_worker_thread_barrier_sync (vm);
      vui->vrings[msg.state.index].last_avail_idx = msg.state.num;
      if (vhost_user_is_packed_ring_supported (vui))
	{
	  /*
	   *  0                   1                   2                   3
	   *  0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
	   * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
	   * |    last avail idx           | |     last used idx           | |
	   * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
	   *                                ^                               ^
	   *                                |                               |
	   *                         avail wrap counter       used wrap counter
	   */
	  /* last avail idx at bit 0-14. */
	  vui->vrings[msg.state.index].last_avail_idx =
	    msg.state.num & 0x7fff;
	  /* avail wrap counter at bit 15 */
	  vui->vrings[msg.state.index].avail_wrap_counter =
	    ! !(msg.state.num & (1 << 15));

	  /*
	   * Although last_used_idx is passed in the upper 16 bits in qemu
	   * implementation, in practice, last_avail_idx and last_used_idx are
	   * usually the same. As a result, DPDK does not bother to pass us
	   * last_used_idx. The spec is not clear on thex coding. I figured it
	   * out by reading the qemu code. So let's just read last_avail_idx
	   * and set last_used_idx equals to last_avail_idx.
	   */
	  vui->vrings[msg.state.index].last_used_idx =
	    vui->vrings[msg.state.index].last_avail_idx;
	  vui->vrings[msg.state.index].last_kick =
	    vui->vrings[msg.state.index].last_used_idx;
	  vui->vrings[msg.state.index].used_wrap_counter =
	    vui->vrings[msg.state.index].avail_wrap_counter;

	  if (vui->vrings[msg.state.index].avail_wrap_counter == 1)
	    vui->vrings[msg.state.index].avail_wrap_counter =
	      VRING_DESC_F_AVAIL;
	}
      vlib_worker_thread_barrier_release (vm);
      break;

    case VHOST_USER_GET_VRING_BASE:
      if (msg.state.index >= vui->num_qid)
	{
	  vu_log_debug (vui, "invalid vring index VHOST_USER_GET_VRING_BASE:"
			" %u >= %u", msg.state.index, vui->num_qid);
	  goto close_socket;
	}

      /* protection is needed to prevent rx/tx from changing last_avail_idx */
      vlib_worker_thread_barrier_sync (vm);
      /*
       * Copy last_avail_idx from the vring before closing it because
       * closing the vring also initializes the vring last_avail_idx
       */
      msg.state.num = vui->vrings[msg.state.index].last_avail_idx;
      if (vhost_user_is_packed_ring_supported (vui))
	{
	  msg.state.num =
	    (vui->vrings[msg.state.index].last_avail_idx & 0x7fff) |
	    (! !vui->vrings[msg.state.index].avail_wrap_counter << 15);
	  msg.state.num |=
	    ((vui->vrings[msg.state.index].last_used_idx & 0x7fff) |
	     (! !vui->vrings[msg.state.index].used_wrap_counter << 15)) << 16;
	}
      msg.flags |= 4;
      msg.size = sizeof (msg.state);

      /*
       * Spec says: Client must [...] stop ring upon receiving
       * VHOST_USER_GET_VRING_BASE
       */
      vhost_user_vring_close (vui, msg.state.index);
      vlib_worker_thread_barrier_release (vm);
      vu_log_debug (vui,
		    "if %d msg VHOST_USER_GET_VRING_BASE idx %d num 0x%x",
		    vui->hw_if_index, msg.state.index, msg.state.num);
      n =
	send (uf->file_descriptor, &msg, VHOST_USER_MSG_HDR_SZ + msg.size, 0);
      if (n != (msg.size + VHOST_USER_MSG_HDR_SZ))
	{
	  vu_log_debug (vui, "could not send message response");
	  goto close_socket;
	}
      vhost_user_update_iface_state (vui);
      break;

    case VHOST_USER_NONE:
      vu_log_debug (vui, "if %d msg VHOST_USER_NONE", vui->hw_if_index);
      break;

    case VHOST_USER_SET_LOG_BASE:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_LOG_BASE",
		    vui->hw_if_index);

      if (msg.size != sizeof (msg.log))
	{
	  vu_log_debug (vui, "invalid msg size for VHOST_USER_SET_LOG_BASE:"
			" %d instead of %d", msg.size, sizeof (msg.log));
	  goto close_socket;
	}

      if (!(vui->protocol_features & (1 << VHOST_USER_PROTOCOL_F_LOG_SHMFD)))
	{
	  vu_log_debug (vui, "VHOST_USER_PROTOCOL_F_LOG_SHMFD not set but "
			"VHOST_USER_SET_LOG_BASE received");
	  goto close_socket;
	}

      fd = fds[0];
      /* align size to page */
      long page_sz = get_huge_page_size (fd);
      ssize_t map_sz =
	(msg.log.size + msg.log.offset + page_sz - 1) & ~(page_sz - 1);

      void *log_base_addr = mmap (0, map_sz, PROT_READ | PROT_WRITE,
				  MAP_SHARED, fd, 0);

      vu_log_debug (vui, "map log region addr 0 len 0x%lx off 0x%lx fd %d "
		    "mapped 0x%lx", map_sz, msg.log.offset, fd,
		    log_base_addr);

      if (log_base_addr == MAP_FAILED)
	{
	  vu_log_err (vui, "failed to map memory. errno is %d", errno);
	  goto close_socket;
	}

      vlib_worker_thread_barrier_sync (vm);
      vui->log_base_addr = log_base_addr;
      vui->log_base_addr += msg.log.offset;
      vui->log_size = msg.log.size;
      vlib_worker_thread_barrier_release (vm);

      msg.flags |= 4;
      msg.size = sizeof (msg.u64);
      n =
	send (uf->file_descriptor, &msg, VHOST_USER_MSG_HDR_SZ + msg.size, 0);
      if (n != (msg.size + VHOST_USER_MSG_HDR_SZ))
	{
	  vu_log_debug (vui, "could not send message response");
	  goto close_socket;
	}
      break;

    case VHOST_USER_SET_LOG_FD:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_LOG_FD", vui->hw_if_index);
      break;

    case VHOST_USER_GET_PROTOCOL_FEATURES:
      msg.flags |= 4;
      msg.u64 = (1 << VHOST_USER_PROTOCOL_F_LOG_SHMFD) |
	(1 << VHOST_USER_PROTOCOL_F_MQ);
      msg.size = sizeof (msg.u64);
      vu_log_debug (vui, "if %d msg VHOST_USER_GET_PROTOCOL_FEATURES - "
		    "reply 0x%016llx", vui->hw_if_index, msg.u64);
      n =
	send (uf->file_descriptor, &msg, VHOST_USER_MSG_HDR_SZ + msg.size, 0);
      if (n != (msg.size + VHOST_USER_MSG_HDR_SZ))
	{
	  vu_log_debug (vui, "could not send message response");
	  goto close_socket;
	}
      break;

    case VHOST_USER_SET_PROTOCOL_FEATURES:
      vu_log_debug (vui, "if %d msg VHOST_USER_SET_PROTOCOL_FEATURES "
		    "features 0x%016llx", vui->hw_if_index, msg.u64);
      vui->protocol_features = msg.u64;
      break;

    case VHOST_USER_GET_QUEUE_NUM:
      msg.flags |= 4;
      msg.u64 = VHOST_VRING_MAX_MQ_PAIR_SZ;
      msg.size = sizeof (msg.u64);
      vu_log_debug (vui, "if %d msg VHOST_USER_GET_QUEUE_NUM - reply %d",
		    vui->hw_if_index, msg.u64);
      n =
	send (uf->file_descriptor, &msg, VHOST_USER_MSG_HDR_SZ + msg.size, 0);
      if (n != (msg.size + VHOST_USER_MSG_HDR_SZ))
	{
	  vu_log_debug (vui, "could not send message response");
	  goto close_socket;
	}
      break;

    case VHOST_USER_SET_VRING_ENABLE:
      vu_log_debug (vui, "if %d VHOST_USER_SET_VRING_ENABLE: %s queue %d",
		    vui->hw_if_index, msg.state.num ? "enable" : "disable",
		    msg.state.index);
      if (msg.state.index >= vui->num_qid)
	{
	  vu_log_debug (vui, "invalid vring idx VHOST_USER_SET_VRING_ENABLE:"
			" %u >= %u", msg.state.index, vui->num_qid);
	  goto close_socket;
	}

      vui->vrings[msg.state.index].enabled = msg.state.num;
      vhost_user_thread_placement (vui, msg.state.index);
      vhost_user_update_iface_state (vui);
      break;

    default:
      vu_log_debug (vui, "unknown vhost-user message %d received. "
		    "closing socket", msg.request);
      goto close_socket;
    }

  return 0;

close_socket:
  vlib_worker_thread_barrier_sync (vm);
  vhost_user_if_disconnect (vui);
  vlib_worker_thread_barrier_release (vm);
  vhost_user_update_iface_state (vui);
  return 0;
}

static clib_error_t *
vhost_user_socket_error (clib_file_t * uf)
{
  vlib_main_t *vm = vlib_get_main ();
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui =
    pool_elt_at_index (vum->vhost_user_interfaces, uf->private_data);

  vu_log_debug (vui, "socket error on if %d", vui->sw_if_index);
  vlib_worker_thread_barrier_sync (vm);
  vhost_user_if_disconnect (vui);
  vlib_worker_thread_barrier_release (vm);
  return 0;
}

static clib_error_t *
vhost_user_socksvr_accept_ready (clib_file_t * uf)
{
  int client_fd, client_len;
  struct sockaddr_un client;
  clib_file_t template = { 0 };
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui;

  vui = pool_elt_at_index (vum->vhost_user_interfaces, uf->private_data);

  client_len = sizeof (client);
  client_fd = accept (uf->file_descriptor,
		      (struct sockaddr *) &client,
		      (socklen_t *) & client_len);

  if (client_fd < 0)
    return clib_error_return_unix (0, "accept");

  if (vui->clib_file_index != ~0)
    {
      vu_log_debug (vui, "Close client socket for vhost interface %d, fd %d",
		    vui->sw_if_index, UNIX_GET_FD (vui->clib_file_index));
      clib_file_del (&file_main, file_main.file_pool + vui->clib_file_index);
    }

  vu_log_debug (vui, "New client socket for vhost interface %d, fd %d",
		vui->sw_if_index, client_fd);
  template.read_function = vhost_user_socket_read;
  template.error_function = vhost_user_socket_error;
  template.file_descriptor = client_fd;
  template.private_data = vui - vhost_user_main.vhost_user_interfaces;
  template.description = format (0, "vhost interface %d", vui->sw_if_index);
  vui->clib_file_index = clib_file_add (&file_main, &template);
  vui->num_qid = 2;
  return 0;
}

static clib_error_t *
vhost_user_init (vlib_main_t * vm)
{
  vhost_user_main_t *vum = &vhost_user_main;
  vlib_thread_main_t *tm = vlib_get_thread_main ();

  vum->log_default = vlib_log_register_class ("vhost-user", 0);

  vum->coalesce_frames = 32;
  vum->coalesce_time = 1e-3;

  vec_validate (vum->cpus, tm->n_vlib_mains - 1);

  vhost_cpu_t *cpu;
  vec_foreach (cpu, vum->cpus)
  {
    /* This is actually not necessary as validate already zeroes it
     * Just keeping the loop here for later because I am lazy. */
    cpu->rx_buffers_len = 0;
  }

  vum->random = random_default_seed ();

  mhash_init_c_string (&vum->if_index_by_sock_name, sizeof (uword));

  return 0;
}

/* *INDENT-OFF* */
VLIB_INIT_FUNCTION (vhost_user_init) =
{
  .runs_after = VLIB_INITS("ip4_init"),
};
/* *INDENT-ON* */

static uword
vhost_user_send_interrupt_process (vlib_main_t * vm,
				   vlib_node_runtime_t * rt, vlib_frame_t * f)
{
  vhost_user_intf_t *vui;
  f64 timeout = 3153600000.0 /* 100 years */ ;
  uword event_type, *event_data = 0;
  vhost_user_main_t *vum = &vhost_user_main;
  u16 qid;
  f64 now, poll_time_remaining;
  f64 next_timeout;
  u8 stop_timer = 0;

  while (1)
    {
      poll_time_remaining =
	vlib_process_wait_for_event_or_clock (vm, timeout);
      event_type = vlib_process_get_events (vm, &event_data);
      vec_reset_length (event_data);

      /*
       * Use the remaining timeout if it is less than coalesce time to avoid
       * resetting the existing timer in the middle of expiration
       */
      timeout = poll_time_remaining;
      if (vlib_process_suspend_time_is_zero (timeout) ||
	  (timeout > vum->coalesce_time))
	timeout = vum->coalesce_time;

      now = vlib_time_now (vm);
      switch (event_type)
	{
	case VHOST_USER_EVENT_STOP_TIMER:
	  stop_timer = 1;
	  break;

	case VHOST_USER_EVENT_START_TIMER:
	  stop_timer = 0;
	  if (!vlib_process_suspend_time_is_zero (poll_time_remaining))
	    break;
	  /* fall through */

	case ~0:
	  /* *INDENT-OFF* */
	  pool_foreach (vui, vum->vhost_user_interfaces) {
	      next_timeout = timeout;
	      for (qid = 0; qid < vui->num_qid / 2; qid += 2)
		{
		  vhost_user_vring_t *rxvq = &vui->vrings[qid];
		  vhost_user_vring_t *txvq = &vui->vrings[qid + 1];

		  if (txvq->qid == -1)
		    continue;
		  if (txvq->n_since_last_int)
		    {
		      if (now >= txvq->int_deadline)
			vhost_user_send_call (vm, vui, txvq);
		      else
			next_timeout = txvq->int_deadline - now;
		    }

		  if (rxvq->n_since_last_int)
		    {
		      if (now >= rxvq->int_deadline)
			vhost_user_send_call (vm, vui, rxvq);
		      else
			next_timeout = rxvq->int_deadline - now;
		    }

		  if ((next_timeout < timeout) && (next_timeout > 0.0))
		    timeout = next_timeout;
		}
	  }
          /* *INDENT-ON* */
	  break;

	default:
	  clib_warning ("BUG: unhandled event type %d", event_type);
	  break;
	}
      /* No less than 1 millisecond */
      if (timeout < 1e-3)
	timeout = 1e-3;
      if (stop_timer)
	timeout = 3153600000.0;
    }
  return 0;
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (vhost_user_send_interrupt_node) = {
    .function = vhost_user_send_interrupt_process,
    .type = VLIB_NODE_TYPE_PROCESS,
    .name = "vhost-user-send-interrupt-process",
};
/* *INDENT-ON* */

static uword
vhost_user_process (vlib_main_t * vm,
		    vlib_node_runtime_t * rt, vlib_frame_t * f)
{
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui;
  struct sockaddr_un sun;
  int sockfd;
  clib_file_t template = { 0 };
  f64 timeout = 3153600000.0 /* 100 years */ ;
  uword *event_data = 0;

  sockfd = -1;
  sun.sun_family = AF_UNIX;
  template.read_function = vhost_user_socket_read;
  template.error_function = vhost_user_socket_error;
  template.description = format (0, "vhost user process");

  while (1)
    {
      vlib_process_wait_for_event_or_clock (vm, timeout);
      vlib_process_get_events (vm, &event_data);
      vec_reset_length (event_data);

      timeout = 3.0;

      /* *INDENT-OFF* */
      pool_foreach (vui, vum->vhost_user_interfaces) {

	  if (vui->unix_server_index == ~0) { //Nothing to do for server sockets
	      if (vui->clib_file_index == ~0)
		{
		  if ((sockfd < 0) &&
		      ((sockfd = socket (AF_UNIX, SOCK_STREAM, 0)) < 0))
		    {
		      /*
		       * 1st time error or new error for this interface,
		       * spit out the message and record the error
		       */
		      if (!vui->sock_errno || (vui->sock_errno != errno))
			{
			  clib_unix_warning
			    ("Error: Could not open unix socket for %s",
			     vui->sock_filename);
			  vui->sock_errno = errno;
			}
		      continue;
		    }

		  /* try to connect */
		  strncpy (sun.sun_path, (char *) vui->sock_filename,
			   sizeof (sun.sun_path) - 1);
		  sun.sun_path[sizeof (sun.sun_path) - 1] = 0;

		  /* Avoid hanging VPP if the other end does not accept */
		  if (fcntl(sockfd, F_SETFL, O_NONBLOCK) < 0)
                      clib_unix_warning ("fcntl");

		  if (connect (sockfd, (struct sockaddr *) &sun,
			       sizeof (struct sockaddr_un)) == 0)
		    {
		      /* Set the socket to blocking as it was before */
                      if (fcntl(sockfd, F_SETFL, 0) < 0)
                        clib_unix_warning ("fcntl2");

		      vui->sock_errno = 0;
		      template.file_descriptor = sockfd;
		      template.private_data =
			  vui - vhost_user_main.vhost_user_interfaces;
		      vui->clib_file_index = clib_file_add (&file_main, &template);
		      vui->num_qid = 2;

		      /* This sockfd is considered consumed */
		      sockfd = -1;
		    }
		  else
		    {
		      vui->sock_errno = errno;
		    }
		}
	      else
		{
		  /* check if socket is alive */
		  int error = 0;
		  socklen_t len = sizeof (error);
		  int fd = UNIX_GET_FD(vui->clib_file_index);
		  int retval =
		      getsockopt (fd, SOL_SOCKET, SO_ERROR, &error, &len);

		  if (retval)
		    {
		      vu_log_debug (vui, "getsockopt returned %d", retval);
		      vhost_user_if_disconnect (vui);
		    }
		}
	  }
      }
      /* *INDENT-ON* */
    }
  return 0;
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (vhost_user_process_node,static) = {
    .function = vhost_user_process,
    .type = VLIB_NODE_TYPE_PROCESS,
    .name = "vhost-user-process",
};
/* *INDENT-ON* */

/**
 * Disables and reset interface structure.
 * It can then be either init again, or removed from used interfaces.
 */
static void
vhost_user_term_if (vhost_user_intf_t * vui)
{
  int q;
  vhost_user_main_t *vum = &vhost_user_main;

  // disconnect interface sockets
  vhost_user_if_disconnect (vui);
  vhost_user_update_gso_interface_count (vui, 0 /* delete */ );
  vhost_user_update_iface_state (vui);

  for (q = 0; q < vui->num_qid; q++)
    {
      // Remove existing queue mapping for the interface
      if (q & 1)
	{
	  int rv;
	  vnet_main_t *vnm = vnet_get_main ();
	  vhost_user_vring_t *txvq = &vui->vrings[q];

	  if (txvq->qid != -1)
	    {
	      rv = vnet_hw_interface_unassign_rx_thread (vnm,
							 vui->hw_if_index,
							 q >> 1);
	      if (rv)
		vu_log_warn (vui, "unable to unassign interface %d, "
			     "queue %d: rc=%d", vui->hw_if_index, q >> 1, rv);
	    }
	}

      clib_spinlock_free (&vui->vrings[q].vring_lock);
    }

  if (vui->unix_server_index != ~0)
    {
      //Close server socket
      clib_file_t *uf = pool_elt_at_index (file_main.file_pool,
					   vui->unix_server_index);
      clib_file_del (&file_main, uf);
      vui->unix_server_index = ~0;
      unlink (vui->sock_filename);
    }

  mhash_unset (&vum->if_index_by_sock_name, vui->sock_filename,
	       &vui->if_index);
}

int
vhost_user_delete_if (vnet_main_t * vnm, vlib_main_t * vm, u32 sw_if_index)
{
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui;
  int rv = 0;
  vnet_hw_interface_t *hwif;
  u16 qid;

  if (!
      (hwif =
       vnet_get_sup_hw_interface_api_visible_or_null (vnm, sw_if_index))
      || hwif->dev_class_index != vhost_user_device_class.index)
    return VNET_API_ERROR_INVALID_SW_IF_INDEX;

  vui = pool_elt_at_index (vum->vhost_user_interfaces, hwif->dev_instance);

  vu_log_debug (vui, "Deleting vhost-user interface %s (instance %d)",
		hwif->name, hwif->dev_instance);

  for (qid = 1; qid < vui->num_qid / 2; qid += 2)
    {
      vhost_user_vring_t *txvq = &vui->vrings[qid];

      if (txvq->qid == -1)
	continue;
      if ((vum->ifq_count > 0) &&
	  ((txvq->mode == VNET_HW_IF_RX_MODE_INTERRUPT) ||
	   (txvq->mode == VNET_HW_IF_RX_MODE_ADAPTIVE)))
	{
	  vum->ifq_count--;
	  // Stop the timer if there is no more interrupt interface/queue
	  if ((vum->ifq_count == 0) &&
	      (vum->coalesce_time > 0.0) && (vum->coalesce_frames > 0))
	    {
	      vlib_process_signal_event (vm,
					 vhost_user_send_interrupt_node.index,
					 VHOST_USER_EVENT_STOP_TIMER, 0);
	      break;
	    }
	}
    }

  // Disable and reset interface
  vhost_user_term_if (vui);

  // Reset renumbered iface
  if (hwif->dev_instance <
      vec_len (vum->show_dev_instance_by_real_dev_instance))
    vum->show_dev_instance_by_real_dev_instance[hwif->dev_instance] = ~0;

  // Delete ethernet interface
  ethernet_delete_interface (vnm, vui->hw_if_index);

  // free vrings
  vec_free (vui->vrings);

  // Back to pool
  pool_put (vum->vhost_user_interfaces, vui);

  return rv;
}

static clib_error_t *
vhost_user_exit (vlib_main_t * vm)
{
  vnet_main_t *vnm = vnet_get_main ();
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui;

  vlib_worker_thread_barrier_sync (vlib_get_main ());
  /* *INDENT-OFF* */
  pool_foreach (vui, vum->vhost_user_interfaces) {
      vhost_user_delete_if (vnm, vm, vui->sw_if_index);
  }
  /* *INDENT-ON* */
  vlib_worker_thread_barrier_release (vlib_get_main ());
  return 0;
}

VLIB_MAIN_LOOP_EXIT_FUNCTION (vhost_user_exit);

/**
 * Open server unix socket on specified sock_filename.
 */
static int
vhost_user_init_server_sock (const char *sock_filename, int *sock_fd)
{
  int rv = 0;
  struct sockaddr_un un = { };
  int fd;
  /* create listening socket */
  if ((fd = socket (AF_UNIX, SOCK_STREAM, 0)) < 0)
    return VNET_API_ERROR_SYSCALL_ERROR_1;

  un.sun_family = AF_UNIX;
  strncpy ((char *) un.sun_path, (char *) sock_filename,
	   sizeof (un.sun_path) - 1);

  /* remove if exists */
  unlink ((char *) sock_filename);

  if (bind (fd, (struct sockaddr *) &un, sizeof (un)) == -1)
    {
      rv = VNET_API_ERROR_SYSCALL_ERROR_2;
      goto error;
    }

  if (listen (fd, 1) == -1)
    {
      rv = VNET_API_ERROR_SYSCALL_ERROR_3;
      goto error;
    }

  *sock_fd = fd;
  return 0;

error:
  close (fd);
  return rv;
}

/**
 * Create ethernet interface for vhost user interface.
 */
static void
vhost_user_create_ethernet (vnet_main_t * vnm, vlib_main_t * vm,
			    vhost_user_intf_t * vui, u8 * hwaddress)
{
  vhost_user_main_t *vum = &vhost_user_main;
  u8 hwaddr[6];
  clib_error_t *error;

  /* create hw and sw interface */
  if (hwaddress)
    {
      clib_memcpy (hwaddr, hwaddress, 6);
    }
  else
    {
      random_u32 (&vum->random);
      clib_memcpy (hwaddr + 2, &vum->random, sizeof (vum->random));
      hwaddr[0] = 2;
      hwaddr[1] = 0xfe;
    }

  error = ethernet_register_interface
    (vnm,
     vhost_user_device_class.index,
     vui - vum->vhost_user_interfaces /* device instance */ ,
     hwaddr /* ethernet address */ ,
     &vui->hw_if_index, 0 /* flag change */ );

  if (error)
    clib_error_report (error);
}

/*
 *  Initialize vui with specified attributes
 */
static void
vhost_user_vui_init (vnet_main_t * vnm, vhost_user_intf_t * vui,
		     int server_sock_fd, vhost_user_create_if_args_t * args,
		     u32 * sw_if_index)
{
  vnet_sw_interface_t *sw;
  int q;
  vhost_user_main_t *vum = &vhost_user_main;
  vnet_hw_interface_t *hw;

  hw = vnet_get_hw_interface (vnm, vui->hw_if_index);
  sw = vnet_get_hw_sw_interface (vnm, vui->hw_if_index);
  if (server_sock_fd != -1)
    {
      clib_file_t template = { 0 };
      template.read_function = vhost_user_socksvr_accept_ready;
      template.file_descriptor = server_sock_fd;
      template.private_data = vui - vum->vhost_user_interfaces;	//hw index
      template.description = format (0, "vhost user %d", sw);
      vui->unix_server_index = clib_file_add (&file_main, &template);
    }
  else
    {
      vui->unix_server_index = ~0;
    }

  vui->sw_if_index = sw->sw_if_index;
  strncpy (vui->sock_filename, args->sock_filename,
	   ARRAY_LEN (vui->sock_filename) - 1);
  vui->sock_errno = 0;
  vui->is_ready = 0;
  vui->feature_mask = args->feature_mask;
  vui->clib_file_index = ~0;
  vui->log_base_addr = 0;
  vui->if_index = vui - vum->vhost_user_interfaces;
  vui->enable_gso = args->enable_gso;
  vui->enable_event_idx = args->enable_event_idx;
  vui->enable_packed = args->enable_packed;
  /*
   * enable_gso takes precedence over configurable feature mask if there
   * is a clash.
   *   if feature mask disables gso, but enable_gso is configured,
   *     then gso is enable
   *   if feature mask enables gso, but enable_gso is not configured,
   *     then gso is enable
   *
   * if gso is enable via feature mask, it must enable both host and guest
   * gso feature mask, we don't support one sided GSO or partial GSO.
   */
  if ((vui->enable_gso == 0) &&
      ((args->feature_mask & FEATURE_VIRTIO_NET_F_HOST_GUEST_TSO_FEATURE_BITS)
       == (FEATURE_VIRTIO_NET_F_HOST_GUEST_TSO_FEATURE_BITS)))
    vui->enable_gso = 1;
  vhost_user_update_gso_interface_count (vui, 1 /* add */ );
  mhash_set_mem (&vum->if_index_by_sock_name, vui->sock_filename,
		 &vui->if_index, 0);

  vec_validate_aligned (vui->vrings, (VHOST_VRING_INIT_MQ_PAIR_SZ << 1) - 1,
			CLIB_CACHE_LINE_BYTES);
  vui->num_qid = 2;
  for (q = 0; q < vec_len (vui->vrings); q++)
    vhost_user_vring_init (vui, q);

  hw->flags |= VNET_HW_INTERFACE_FLAG_SUPPORTS_INT_MODE;
  vnet_hw_interface_set_flags (vnm, vui->hw_if_index, 0);

  if (sw_if_index)
    *sw_if_index = vui->sw_if_index;

  vec_validate (vui->per_cpu_tx_qid,
		vlib_get_thread_main ()->n_vlib_mains - 1);
  vhost_user_tx_thread_placement (vui);
}

int
vhost_user_create_if (vnet_main_t * vnm, vlib_main_t * vm,
		      vhost_user_create_if_args_t * args)
{
  vhost_user_intf_t *vui = NULL;
  u32 sw_if_idx = ~0;
  int rv = 0;
  int server_sock_fd = -1;
  vhost_user_main_t *vum = &vhost_user_main;
  uword *if_index;

  if (args->sock_filename == NULL || !(strlen (args->sock_filename) > 0))
    {
      return VNET_API_ERROR_INVALID_ARGUMENT;
    }

  if_index = mhash_get (&vum->if_index_by_sock_name,
			(void *) args->sock_filename);
  if (if_index)
    {
      vui = &vum->vhost_user_interfaces[*if_index];
      args->sw_if_index = vui->sw_if_index;
      return VNET_API_ERROR_IF_ALREADY_EXISTS;
    }

  if (args->is_server)
    {
      if ((rv =
	   vhost_user_init_server_sock (args->sock_filename,
					&server_sock_fd)) != 0)
	{
	  return rv;
	}
    }

  /* Protect the uninitialized vui from being dispatched by rx/tx */
  vlib_worker_thread_barrier_sync (vm);
  pool_get (vhost_user_main.vhost_user_interfaces, vui);
  vhost_user_create_ethernet (vnm, vm, vui, args->hwaddr);
  vlib_worker_thread_barrier_release (vm);

  vhost_user_vui_init (vnm, vui, server_sock_fd, args, &sw_if_idx);
  vnet_sw_interface_set_mtu (vnm, vui->sw_if_index, 9000);
  vhost_user_rx_thread_placement (vui, 1);

  if (args->renumber)
    vnet_interface_name_renumber (sw_if_idx, args->custom_dev_instance);

  args->sw_if_index = sw_if_idx;

  // Process node must connect
  vlib_process_signal_event (vm, vhost_user_process_node.index, 0, 0);

  return rv;
}

int
vhost_user_modify_if (vnet_main_t * vnm, vlib_main_t * vm,
		      vhost_user_create_if_args_t * args)
{
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui = NULL;
  u32 sw_if_idx = ~0;
  int server_sock_fd = -1;
  int rv = 0;
  vnet_hw_interface_t *hwif;
  uword *if_index;

  if (!(hwif = vnet_get_sup_hw_interface_api_visible_or_null (vnm,
							      args->sw_if_index))
      || hwif->dev_class_index != vhost_user_device_class.index)
    return VNET_API_ERROR_INVALID_SW_IF_INDEX;

  if (args->sock_filename == NULL || !(strlen (args->sock_filename) > 0))
    return VNET_API_ERROR_INVALID_ARGUMENT;

  vui = vec_elt_at_index (vum->vhost_user_interfaces, hwif->dev_instance);

  /*
   * Disallow changing the interface to have the same path name
   * as other interface
   */
  if_index = mhash_get (&vum->if_index_by_sock_name,
			(void *) args->sock_filename);
  if (if_index && (*if_index != vui->if_index))
    return VNET_API_ERROR_IF_ALREADY_EXISTS;

  // First try to open server socket
  if (args->is_server)
    if ((rv = vhost_user_init_server_sock (args->sock_filename,
					   &server_sock_fd)) != 0)
      return rv;

  vhost_user_term_if (vui);
  vhost_user_vui_init (vnm, vui, server_sock_fd, args, &sw_if_idx);

  if (args->renumber)
    vnet_interface_name_renumber (sw_if_idx, args->custom_dev_instance);

  // Process node must connect
  vlib_process_signal_event (vm, vhost_user_process_node.index, 0, 0);

  return rv;
}

clib_error_t *
vhost_user_connect_command_fn (vlib_main_t * vm,
			       unformat_input_t * input,
			       vlib_cli_command_t * cmd)
{
  vnet_main_t *vnm = vnet_get_main ();
  unformat_input_t _line_input, *line_input = &_line_input;
  clib_error_t *error = NULL;
  vhost_user_create_if_args_t args = { 0 };
  int rv;

  /* Get a line of input. */
  if (!unformat_user (input, unformat_line_input, line_input))
    return 0;

  args.feature_mask = (u64) ~ (0ULL);
  args.custom_dev_instance = ~0;
  /* GSO feature is disable by default */
  args.feature_mask &= ~FEATURE_VIRTIO_NET_F_HOST_GUEST_TSO_FEATURE_BITS;
  /* packed-ring feature is disable by default */
  args.feature_mask &= ~VIRTIO_FEATURE (VIRTIO_F_RING_PACKED);
  /* event_idx feature is disable by default */
  args.feature_mask &= ~VIRTIO_FEATURE (VIRTIO_RING_F_EVENT_IDX);

  while (unformat_check_input (line_input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (line_input, "socket %s", &args.sock_filename))
	;
      else if (unformat (line_input, "server"))
	args.is_server = 1;
      else if (unformat (line_input, "gso"))
	args.enable_gso = 1;
      else if (unformat (line_input, "packed"))
	args.enable_packed = 1;
      else if (unformat (line_input, "event-idx"))
	args.enable_event_idx = 1;
      else if (unformat (line_input, "feature-mask 0x%llx",
			 &args.feature_mask))
	;
      else if (unformat (line_input, "hwaddr %U", unformat_ethernet_address,
			 args.hwaddr))
	;
      else if (unformat (line_input, "renumber %d",
			 &args.custom_dev_instance))
	args.renumber = 1;
      else
	{
	  error = clib_error_return (0, "unknown input `%U'",
				     format_unformat_error, line_input);
	  goto done;
	}
    }

  if ((rv = vhost_user_create_if (vnm, vm, &args)))
    {
      error = clib_error_return (0, "vhost_user_create_if returned %d", rv);
      goto done;
    }

  vlib_cli_output (vm, "%U\n", format_vnet_sw_if_index_name, vnm,
		   args.sw_if_index);

done:
  vec_free (args.sock_filename);
  unformat_free (line_input);

  return error;
}

clib_error_t *
vhost_user_delete_command_fn (vlib_main_t * vm,
			      unformat_input_t * input,
			      vlib_cli_command_t * cmd)
{
  unformat_input_t _line_input, *line_input = &_line_input;
  u32 sw_if_index = ~0;
  vnet_main_t *vnm = vnet_get_main ();
  clib_error_t *error = NULL;

  /* Get a line of input. */
  if (!unformat_user (input, unformat_line_input, line_input))
    return 0;

  while (unformat_check_input (line_input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (line_input, "sw_if_index %d", &sw_if_index))
	;
      else if (unformat
	       (line_input, "%U", unformat_vnet_sw_interface, vnm,
		&sw_if_index))
	{
	  vnet_hw_interface_t *hwif =
	    vnet_get_sup_hw_interface_api_visible_or_null (vnm, sw_if_index);
	  if (hwif == NULL ||
	      vhost_user_device_class.index != hwif->dev_class_index)
	    {
	      error = clib_error_return (0, "Not a vhost interface");
	      goto done;
	    }
	}
      else
	{
	  error = clib_error_return (0, "unknown input `%U'",
				     format_unformat_error, line_input);
	  goto done;
	}
    }

  vhost_user_delete_if (vnm, vm, sw_if_index);

done:
  unformat_free (line_input);

  return error;
}

int
vhost_user_dump_ifs (vnet_main_t * vnm, vlib_main_t * vm,
		     vhost_user_intf_details_t ** out_vuids)
{
  int rv = 0;
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui;
  vhost_user_intf_details_t *r_vuids = NULL;
  vhost_user_intf_details_t *vuid = NULL;
  u32 *hw_if_indices = 0;
  vnet_hw_interface_t *hi;
  int i;

  if (!out_vuids)
    return -1;

  pool_foreach (vui, vum->vhost_user_interfaces)
    vec_add1 (hw_if_indices, vui->hw_if_index);

  for (i = 0; i < vec_len (hw_if_indices); i++)
    {
      hi = vnet_get_hw_interface (vnm, hw_if_indices[i]);
      vui = pool_elt_at_index (vum->vhost_user_interfaces, hi->dev_instance);

      vec_add2 (r_vuids, vuid, 1);
      vuid->sw_if_index = vui->sw_if_index;
      vuid->virtio_net_hdr_sz = vui->virtio_net_hdr_sz;
      vuid->features = vui->features;
      vuid->num_regions = vui->nregions;
      vuid->is_server = vui->unix_server_index != ~0;
      vuid->sock_errno = vui->sock_errno;
      snprintf ((char *) vuid->sock_filename, sizeof (vuid->sock_filename),
		"%s", vui->sock_filename);
      memcpy_s (vuid->if_name, sizeof (vuid->if_name), hi->name,
		clib_min (vec_len (hi->name), sizeof (vuid->if_name) - 1));
      vuid->if_name[sizeof (vuid->if_name) - 1] = 0;
    }

  vec_free (hw_if_indices);

  *out_vuids = r_vuids;

  return rv;
}

static u8 *
format_vhost_user_desc (u8 * s, va_list * args)
{
  char *fmt = va_arg (*args, char *);
  vhost_user_intf_t *vui = va_arg (*args, vhost_user_intf_t *);
  vring_desc_t *desc_table = va_arg (*args, vring_desc_t *);
  int idx = va_arg (*args, int);
  u32 *mem_hint = va_arg (*args, u32 *);

  s = format (s, fmt, idx, desc_table[idx].addr, desc_table[idx].len,
	      desc_table[idx].flags, desc_table[idx].next,
	      pointer_to_uword (map_guest_mem (vui, desc_table[idx].addr,
					       mem_hint)));
  return s;
}

static void
vhost_user_show_fds (vlib_main_t * vm, vhost_user_vring_t * vq)
{
  int kickfd = UNIX_GET_FD (vq->kickfd_idx);
  int callfd = UNIX_GET_FD (vq->callfd_idx);

  vlib_cli_output (vm, "  kickfd %d callfd %d errfd %d\n", kickfd, callfd,
		   vq->errfd);
}

static void
vhost_user_show_desc (vlib_main_t * vm, vhost_user_intf_t * vui, int q,
		      int show_descr, int show_verbose)
{
  int j;
  u32 mem_hint = 0;
  u32 idx;
  u32 n_entries;
  vring_desc_t *desc_table;
  vhost_user_vring_t *vq = &vui->vrings[q];

  if (vq->avail && vq->used)
    vlib_cli_output (vm, "  avail.flags %x avail event idx %u avail.idx %d "
		     "used event idx %u used.idx %d\n", vq->avail->flags,
		     vhost_user_avail_event_idx (vq), vq->avail->idx,
		     vhost_user_used_event_idx (vq), vq->used->idx);

  vhost_user_show_fds (vm, vq);

  if (show_descr)
    {
      vlib_cli_output (vm, "\n  descriptor table:\n");
      vlib_cli_output (vm,
		       "  slot         addr         len  flags  next      "
		       "user_addr\n");
      vlib_cli_output (vm,
		       "  ===== ================== ===== ====== ===== "
		       "==================\n");
      for (j = 0; j < vq->qsz_mask + 1; j++)
	{
	  desc_table = vq->desc;
	  vlib_cli_output (vm, "%U", format_vhost_user_desc,
			   "  %-5d 0x%016lx %-5d 0x%04x %-5d 0x%016lx\n", vui,
			   desc_table, j, &mem_hint);
	  if (show_verbose && (desc_table[j].flags & VRING_DESC_F_INDIRECT))
	    {
	      n_entries = desc_table[j].len / sizeof (vring_desc_t);
	      desc_table = map_guest_mem (vui, desc_table[j].addr, &mem_hint);
	      if (desc_table)
		{
		  for (idx = 0; idx < clib_min (20, n_entries); idx++)
		    {
		      vlib_cli_output
			(vm, "%U", format_vhost_user_desc,
			 ">  %-4u 0x%016lx %-5u 0x%04x %-5u 0x%016lx\n", vui,
			 desc_table, idx, &mem_hint);
		    }
		  if (n_entries >= 20)
		    vlib_cli_output (vm, "Skip displaying entries 20...%u\n",
				     n_entries);
		}
	    }
	}
    }
}

static u8 *
format_vhost_user_packed_desc (u8 * s, va_list * args)
{
  char *fmt = va_arg (*args, char *);
  vhost_user_intf_t *vui = va_arg (*args, vhost_user_intf_t *);
  vring_packed_desc_t *desc_table = va_arg (*args, vring_packed_desc_t *);
  int idx = va_arg (*args, int);
  u32 *mem_hint = va_arg (*args, u32 *);

  s = format (s, fmt, idx, desc_table[idx].addr, desc_table[idx].len,
	      desc_table[idx].flags, desc_table[idx].id,
	      pointer_to_uword (map_guest_mem (vui, desc_table[idx].addr,
					       mem_hint)));
  return s;
}

static u8 *
format_vhost_user_event_idx_flags (u8 * s, va_list * args)
{
  u32 flags = va_arg (*args, u32);
  typedef struct
  {
    u8 value;
    char *str;
  } event_idx_flags;
  static event_idx_flags event_idx_array[] = {
#define _(s,v) { .str = #s, .value = v, },
    foreach_virtio_event_idx_flags
#undef _
  };
  u32 num_entries = sizeof (event_idx_array) / sizeof (event_idx_flags);

  if (flags < num_entries)
    s = format (s, "%s", event_idx_array[flags].str);
  else
    s = format (s, "%u", flags);
  return s;
}

static void
vhost_user_show_desc_packed (vlib_main_t * vm, vhost_user_intf_t * vui, int q,
			     int show_descr, int show_verbose)
{
  int j;
  u32 mem_hint = 0;
  u32 idx;
  u32 n_entries;
  vring_packed_desc_t *desc_table;
  vhost_user_vring_t *vq = &vui->vrings[q];
  u16 off_wrap, event_idx;

  off_wrap = vq->avail_event->off_wrap;
  event_idx = off_wrap & 0x7fff;
  vlib_cli_output (vm, "  avail_event.flags %U avail_event.off_wrap %u "
		   "avail event idx %u\n", format_vhost_user_event_idx_flags,
		   (u32) vq->avail_event->flags, off_wrap, event_idx);

  off_wrap = vq->used_event->off_wrap;
  event_idx = off_wrap & 0x7fff;
  vlib_cli_output (vm, "  used_event.flags %U used_event.off_wrap %u "
		   "used event idx %u\n", format_vhost_user_event_idx_flags,
		   (u32) vq->used_event->flags, off_wrap, event_idx);

  vlib_cli_output (vm, "  avail wrap counter %u, used wrap counter %u\n",
		   vq->avail_wrap_counter, vq->used_wrap_counter);

  vhost_user_show_fds (vm, vq);

  if (show_descr)
    {
      vlib_cli_output (vm, "\n  descriptor table:\n");
      vlib_cli_output (vm,
		       "  slot         addr         len  flags  id    "
		       "user_addr\n");
      vlib_cli_output (vm,
		       "  ===== ================== ===== ====== ===== "
		       "==================\n");
      for (j = 0; j < vq->qsz_mask + 1; j++)
	{
	  desc_table = vq->packed_desc;
	  vlib_cli_output (vm, "%U", format_vhost_user_packed_desc,
			   "  %-5u 0x%016lx %-5u 0x%04x %-5u 0x%016lx\n", vui,
			   desc_table, j, &mem_hint);
	  if (show_verbose && (desc_table[j].flags & VRING_DESC_F_INDIRECT))
	    {
	      n_entries = desc_table[j].len >> 4;
	      desc_table = map_guest_mem (vui, desc_table[j].addr, &mem_hint);
	      if (desc_table)
		{
		  for (idx = 0; idx < clib_min (20, n_entries); idx++)
		    {
		      vlib_cli_output
			(vm, "%U", format_vhost_user_packed_desc,
			 ">  %-4u 0x%016lx %-5u 0x%04x %-5u 0x%016lx\n", vui,
			 desc_table, idx, &mem_hint);
		    }
		  if (n_entries >= 20)
		    vlib_cli_output (vm, "Skip displaying entries 20...%u\n",
				     n_entries);
		}
	    }
	}
    }
}

clib_error_t *
show_vhost_user_command_fn (vlib_main_t * vm,
			    unformat_input_t * input,
			    vlib_cli_command_t * cmd)
{
  clib_error_t *error = 0;
  vnet_main_t *vnm = vnet_get_main ();
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui;
  u32 hw_if_index, *hw_if_indices = 0;
  vnet_hw_interface_t *hi;
  u16 qid;
  u32 ci;
  int i, j, q;
  int show_descr = 0;
  int show_verbose = 0;
  struct feat_struct
  {
    u8 bit;
    char *str;
  };
  struct feat_struct *feat_entry;

  static struct feat_struct feat_array[] = {
#define _(s,b) { .str = #s, .bit = b, },
    foreach_virtio_net_features
#undef _
    {.str = NULL}
  };

#define foreach_protocol_feature \
  _(VHOST_USER_PROTOCOL_F_MQ) \
  _(VHOST_USER_PROTOCOL_F_LOG_SHMFD)

  static struct feat_struct proto_feat_array[] = {
#define _(s) { .str = #s, .bit = s},
    foreach_protocol_feature
#undef _
    {.str = NULL}
  };

  while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat
	  (input, "%U", unformat_vnet_hw_interface, vnm, &hw_if_index))
	{
	  hi = vnet_get_hw_interface (vnm, hw_if_index);
	  if (vhost_user_device_class.index != hi->dev_class_index)
	    {
	      error = clib_error_return (0, "unknown input `%U'",
					 format_unformat_error, input);
	      goto done;
	    }
	  vec_add1 (hw_if_indices, hw_if_index);
	}
      else if (unformat (input, "descriptors") || unformat (input, "desc"))
	show_descr = 1;
      else if (unformat (input, "verbose"))
	show_verbose = 1;
      else
	{
	  error = clib_error_return (0, "unknown input `%U'",
				     format_unformat_error, input);
	  goto done;
	}
    }
  if (vec_len (hw_if_indices) == 0)
    {
      pool_foreach (vui, vum->vhost_user_interfaces)
	vec_add1 (hw_if_indices, vui->hw_if_index);
    }
  vlib_cli_output (vm, "Virtio vhost-user interfaces");
  vlib_cli_output (vm, "Global:\n  coalesce frames %d time %e",
		   vum->coalesce_frames, vum->coalesce_time);
  vlib_cli_output (vm, "  Number of rx virtqueues in interrupt mode: %d",
		   vum->ifq_count);
  vlib_cli_output (vm, "  Number of GSO interfaces: %d", vum->gso_count);

  for (i = 0; i < vec_len (hw_if_indices); i++)
    {
      hi = vnet_get_hw_interface (vnm, hw_if_indices[i]);
      vui = pool_elt_at_index (vum->vhost_user_interfaces, hi->dev_instance);
      vlib_cli_output (vm, "Interface: %U (ifindex %d)",
		       format_vnet_hw_if_index_name, vnm, hw_if_indices[i],
		       hw_if_indices[i]);
      vlib_cli_output (vm, "  Number of qids %u", vui->num_qid);
      if (vui->enable_gso)
	vlib_cli_output (vm, "  GSO enable");
      if (vui->enable_packed)
	vlib_cli_output (vm, "  Packed ring enable");
      if (vui->enable_event_idx)
	vlib_cli_output (vm, "  Event index enable");

      vlib_cli_output (vm, "virtio_net_hdr_sz %d\n"
		       " features mask (0x%llx): \n"
		       " features (0x%llx): \n",
		       vui->virtio_net_hdr_sz, vui->feature_mask,
		       vui->features);

      feat_entry = (struct feat_struct *) &feat_array;
      while (feat_entry->str)
	{
	  if (vui->features & (1ULL << feat_entry->bit))
	    vlib_cli_output (vm, "   %s (%d)", feat_entry->str,
			     feat_entry->bit);
	  feat_entry++;
	}

      vlib_cli_output (vm, "  protocol features (0x%llx)",
		       vui->protocol_features);
      feat_entry = (struct feat_struct *) &proto_feat_array;
      while (feat_entry->str)
	{
	  if (vui->protocol_features & (1ULL << feat_entry->bit))
	    vlib_cli_output (vm, "   %s (%d)", feat_entry->str,
			     feat_entry->bit);
	  feat_entry++;
	}

      vlib_cli_output (vm, "\n");

      vlib_cli_output (vm, " socket filename %s type %s errno \"%s\"\n\n",
		       vui->sock_filename,
		       (vui->unix_server_index != ~0) ? "server" : "client",
		       strerror (vui->sock_errno));

      vlib_cli_output (vm, " rx placement: ");

      for (qid = 1; qid < vui->num_qid / 2; qid += 2)
	{
	  vnet_main_t *vnm = vnet_get_main ();
	  uword thread_index;
	  vnet_hw_if_rx_mode mode;
	  vhost_user_vring_t *txvq = &vui->vrings[qid];

	  if (txvq->qid == -1)
	    continue;
	  thread_index =
	    vnet_get_device_input_thread_index (vnm, vui->hw_if_index,
						qid >> 1);
	  vnet_hw_interface_get_rx_mode (vnm, vui->hw_if_index, qid >> 1,
					 &mode);
	  vlib_cli_output (vm, "   thread %d on vring %d, %U\n",
			   thread_index, qid,
			   format_vnet_hw_if_rx_mode, mode);
	}

      vlib_cli_output (vm, " tx placement: %s\n",
		       vui->use_tx_spinlock ? "spin-lock" : "lock-free");

      vec_foreach_index (ci, vui->per_cpu_tx_qid)
      {
	vlib_cli_output (vm, "   thread %d on vring %d\n", ci,
			 VHOST_VRING_IDX_RX (vui->per_cpu_tx_qid[ci]));
      }

      vlib_cli_output (vm, "\n");

      vlib_cli_output (vm, " Memory regions (total %d)\n", vui->nregions);

      if (vui->nregions)
	{
	  vlib_cli_output (vm,
			   " region fd    guest_phys_addr    memory_size        userspace_addr     mmap_offset        mmap_addr\n");
	  vlib_cli_output (vm,
			   " ====== ===== ================== ================== ================== ================== ==================\n");
	}
      for (j = 0; j < vui->nregions; j++)
	{
	  vlib_cli_output (vm,
			   "  %d     %-5d 0x%016lx 0x%016lx 0x%016lx 0x%016lx 0x%016lx\n",
			   j, vui->region_mmap_fd[j],
			   vui->regions[j].guest_phys_addr,
			   vui->regions[j].memory_size,
			   vui->regions[j].userspace_addr,
			   vui->regions[j].mmap_offset,
			   pointer_to_uword (vui->region_mmap_addr[j]));
	}
      for (q = 0; q < vui->num_qid; q++)
	{
	  if (!vui->vrings[q].started)
	    continue;

	  vlib_cli_output (vm, "\n Virtqueue %d (%s%s)\n", q,
			   (q & 1) ? "RX" : "TX",
			   vui->vrings[q].enabled ? "" : " disabled");

	  vlib_cli_output (vm,
			   "  qsz %d last_avail_idx %d last_used_idx %d"
			   " last_kick %u\n",
			   vui->vrings[q].qsz_mask + 1,
			   vui->vrings[q].last_avail_idx,
			   vui->vrings[q].last_used_idx,
			   vui->vrings[q].last_kick);

	  if (vhost_user_is_packed_ring_supported (vui))
	    vhost_user_show_desc_packed (vm, vui, q, show_descr,
					 show_verbose);
	  else
	    vhost_user_show_desc (vm, vui, q, show_descr, show_verbose);
	}
      vlib_cli_output (vm, "\n");
    }
done:
  vec_free (hw_if_indices);
  return error;
}

/*
 * CLI functions
 */

/*?
 * Create a vHost User interface. Once created, a new virtual interface
 * will exist with the name '<em>VirtualEthernet0/0/x</em>', where '<em>x</em>'
 * is the next free index.
 *
 * There are several parameters associated with a vHost interface:
 *
 * - <b>socket <socket-filename></b> - Name of the linux socket used by hypervisor
 * and VPP to manage the vHost interface. If in '<em>server</em>' mode, VPP will
 * create the socket if it does not already exist. If in '<em>client</em>' mode,
 * hypervisor will create the socket if it does not already exist. The VPP code
 * is indifferent to the file location. However, if SELinux is enabled, then the
 * socket needs to be created in '<em>/var/run/vpp/</em>'.
 *
 * - <b>server</b> - Optional flag to indicate that VPP should be the server for
 * the linux socket. If not provided, VPP will be the client. In '<em>server</em>'
 *  mode, the VM can be reset without tearing down the vHost Interface. In
 * '<em>client</em>' mode, VPP can be reset without bringing down the VM and
 * tearing down the vHost Interface.
 *
 * - <b>feature-mask <hex></b> - Optional virtio/vhost feature set negotiated at
 * startup. <b>This is intended for degugging only.</b> It is recommended that this
 * parameter not be used except by experienced users. By default, all supported
 * features will be advertised. Otherwise, provide the set of features desired.
 *   - 0x000008000 (15) - VIRTIO_NET_F_MRG_RXBUF
 *   - 0x000020000 (17) - VIRTIO_NET_F_CTRL_VQ
 *   - 0x000200000 (21) - VIRTIO_NET_F_GUEST_ANNOUNCE
 *   - 0x000400000 (22) - VIRTIO_NET_F_MQ
 *   - 0x004000000 (26) - VHOST_F_LOG_ALL
 *   - 0x008000000 (27) - VIRTIO_F_ANY_LAYOUT
 *   - 0x010000000 (28) - VIRTIO_F_INDIRECT_DESC
 *   - 0x040000000 (30) - VHOST_USER_F_PROTOCOL_FEATURES
 *   - 0x100000000 (32) - VIRTIO_F_VERSION_1
 *
 * - <b>hwaddr <mac-addr></b> - Optional ethernet address, can be in either
 * X:X:X:X:X:X unix or X.X.X cisco format.
 *
 * - <b>renumber <dev_instance></b> - Optional parameter which allows the instance
 * in the name to be specified. If instance already exists, name will be used
 * anyway and multiple instances will have the same name. Use with caution.
 *
 * @cliexpar
 * Example of how to create a vhost interface with VPP as the client and all features enabled:
 * @cliexstart{create vhost-user socket /var/run/vpp/vhost1.sock}
 * VirtualEthernet0/0/0
 * @cliexend
 * Example of how to create a vhost interface with VPP as the server and with just
 * multiple queues enabled:
 * @cliexstart{create vhost-user socket /var/run/vpp/vhost2.sock server feature-mask 0x40400000}
 * VirtualEthernet0/0/1
 * @cliexend
 * Once the vHost interface is created, enable the interface using:
 * @cliexcmd{set interface state VirtualEthernet0/0/0 up}
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (vhost_user_connect_command, static) = {
    .path = "create vhost-user",
    .short_help = "create vhost-user socket <socket-filename> [server] "
    "[feature-mask <hex>] [hwaddr <mac-addr>] [renumber <dev_instance>] [gso] "
    "[packed] [event-idx]",
    .function = vhost_user_connect_command_fn,
    .is_mp_safe = 1,
};
/* *INDENT-ON* */

/*?
 * Delete a vHost User interface using the interface name or the
 * software interface index. Use the '<em>show interface</em>'
 * command to determine the software interface index. On deletion,
 * the linux socket will not be deleted.
 *
 * @cliexpar
 * Example of how to delete a vhost interface by name:
 * @cliexcmd{delete vhost-user VirtualEthernet0/0/1}
 * Example of how to delete a vhost interface by software interface index:
 * @cliexcmd{delete vhost-user sw_if_index 1}
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (vhost_user_delete_command, static) = {
    .path = "delete vhost-user",
    .short_help = "delete vhost-user {<interface> | sw_if_index <sw_idx>}",
    .function = vhost_user_delete_command_fn,
};

/*?
 * Display the attributes of a single vHost User interface (provide interface
 * name), multiple vHost User interfaces (provide a list of interface names seperated
 * by spaces) or all Vhost User interfaces (omit an interface name to display all
 * vHost interfaces).
 *
 * @cliexpar
 * @parblock
 * Example of how to display a vhost interface:
 * @cliexstart{show vhost-user VirtualEthernet0/0/0}
 * Virtio vhost-user interfaces
 * Global:
 *   coalesce frames 32 time 1e-3
 * Interface: VirtualEthernet0/0/0 (ifindex 1)
 * virtio_net_hdr_sz 12
 *  features mask (0xffffffffffffffff):
 *  features (0x50408000):
 *    VIRTIO_NET_F_MRG_RXBUF (15)
 *    VIRTIO_NET_F_MQ (22)
 *    VIRTIO_F_INDIRECT_DESC (28)
 *    VHOST_USER_F_PROTOCOL_FEATURES (30)
 *   protocol features (0x3)
 *    VHOST_USER_PROTOCOL_F_MQ (0)
 *    VHOST_USER_PROTOCOL_F_LOG_SHMFD (1)
 *
 *  socket filename /var/run/vpp/vhost1.sock type client errno "Success"
 *
 * rx placement:
 *    thread 1 on vring 1
 *    thread 1 on vring 5
 *    thread 2 on vring 3
 *    thread 2 on vring 7
 *  tx placement: spin-lock
 *    thread 0 on vring 0
 *    thread 1 on vring 2
 *    thread 2 on vring 0
 *
 * Memory regions (total 2)
 * region fd    guest_phys_addr    memory_size        userspace_addr     mmap_offset        mmap_addr
 * ====== ===== ================== ================== ================== ================== ==================
 *   0     60    0x0000000000000000 0x00000000000a0000 0x00002aaaaac00000 0x0000000000000000 0x00002aab2b400000
 *   1     61    0x00000000000c0000 0x000000003ff40000 0x00002aaaaacc0000 0x00000000000c0000 0x00002aababcc0000
 *
 *  Virtqueue 0 (TX)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *   avail.flags 1 avail.idx 128 used.flags 1 used.idx 0
 *   kickfd 62 callfd 64 errfd -1
 *
 *  Virtqueue 1 (RX)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *   avail.flags 1 avail.idx 0 used.flags 1 used.idx 0
 *   kickfd 65 callfd 66 errfd -1
 *
 *  Virtqueue 2 (TX)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *   avail.flags 1 avail.idx 128 used.flags 1 used.idx 0
 *   kickfd 63 callfd 70 errfd -1
 *
 *  Virtqueue 3 (RX)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *   avail.flags 1 avail.idx 0 used.flags 1 used.idx 0
 *   kickfd 72 callfd 74 errfd -1
 *
 *  Virtqueue 4 (TX disabled)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *   avail.flags 1 avail.idx 0 used.flags 1 used.idx 0
 *   kickfd 76 callfd 78 errfd -1
 *
 *  Virtqueue 5 (RX disabled)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *   avail.flags 1 avail.idx 0 used.flags 1 used.idx 0
 *   kickfd 80 callfd 82 errfd -1
 *
 *  Virtqueue 6 (TX disabled)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *  avail.flags 1 avail.idx 0 used.flags 1 used.idx 0
 *   kickfd 84 callfd 86 errfd -1
 *
 *  Virtqueue 7 (RX disabled)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *   avail.flags 1 avail.idx 0 used.flags 1 used.idx 0
 *   kickfd 88 callfd 90 errfd -1
 *
 * @cliexend
 *
 * The optional '<em>descriptors</em>' parameter will display the same output as
 * the previous example but will include the descriptor table for each queue.
 * The output is truncated below:
 * @cliexstart{show vhost-user VirtualEthernet0/0/0 descriptors}
 * Virtio vhost-user interfaces
 * Global:
 *   coalesce frames 32 time 1e-3
 * Interface: VirtualEthernet0/0/0 (ifindex 1)
 * virtio_net_hdr_sz 12
 *  features mask (0xffffffffffffffff):
 *  features (0x50408000):
 *    VIRTIO_NET_F_MRG_RXBUF (15)
 *    VIRTIO_NET_F_MQ (22)
 * :
 *  Virtqueue 0 (TX)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 *   avail.flags 1 avail.idx 128 used.flags 1 used.idx 0
 *   kickfd 62 callfd 64 errfd -1
 *
 *   descriptor table:
 *    id          addr         len  flags  next      user_addr
 *   ===== ================== ===== ====== ===== ==================
 *   0     0x0000000010b6e974 2060  0x0002 1     0x00002aabbc76e974
 *   1     0x0000000010b6e034 2060  0x0002 2     0x00002aabbc76e034
 *   2     0x0000000010b6d6f4 2060  0x0002 3     0x00002aabbc76d6f4
 *   3     0x0000000010b6cdb4 2060  0x0002 4     0x00002aabbc76cdb4
 *   4     0x0000000010b6c474 2060  0x0002 5     0x00002aabbc76c474
 *   5     0x0000000010b6bb34 2060  0x0002 6     0x00002aabbc76bb34
 *   6     0x0000000010b6b1f4 2060  0x0002 7     0x00002aabbc76b1f4
 *   7     0x0000000010b6a8b4 2060  0x0002 8     0x00002aabbc76a8b4
 *   8     0x0000000010b69f74 2060  0x0002 9     0x00002aabbc769f74
 *   9     0x0000000010b69634 2060  0x0002 10    0x00002aabbc769634
 *   10    0x0000000010b68cf4 2060  0x0002 11    0x00002aabbc768cf4
 * :
 *   249   0x0000000000000000 0     0x0000 250   0x00002aab2b400000
 *   250   0x0000000000000000 0     0x0000 251   0x00002aab2b400000
 *   251   0x0000000000000000 0     0x0000 252   0x00002aab2b400000
 *   252   0x0000000000000000 0     0x0000 253   0x00002aab2b400000
 *   253   0x0000000000000000 0     0x0000 254   0x00002aab2b400000
 *   254   0x0000000000000000 0     0x0000 255   0x00002aab2b400000
 *   255   0x0000000000000000 0     0x0000 32768 0x00002aab2b400000
 *
 *  Virtqueue 1 (RX)
 *   qsz 256 last_avail_idx 0 last_used_idx 0
 * :
 * @cliexend
 * @endparblock
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (show_vhost_user_command, static) = {
    .path = "show vhost-user",
    .short_help = "show vhost-user [<interface> [<interface> [..]]] "
    "[[descriptors] [verbose]]",
    .function = show_vhost_user_command_fn,
};
/* *INDENT-ON* */


static clib_error_t *
vhost_user_config (vlib_main_t * vm, unformat_input_t * input)
{
  vhost_user_main_t *vum = &vhost_user_main;

  while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (input, "coalesce-frames %d", &vum->coalesce_frames))
	;
      else if (unformat (input, "coalesce-time %f", &vum->coalesce_time))
	;
      else if (unformat (input, "dont-dump-memory"))
	vum->dont_dump_vhost_user_memory = 1;
      else
	return clib_error_return (0, "unknown input `%U'",
				  format_unformat_error, input);
    }

  return 0;
}

/* vhost-user { ... } configuration. */
VLIB_CONFIG_FUNCTION (vhost_user_config, "vhost-user");

void
vhost_user_unmap_all (void)
{
  vhost_user_main_t *vum = &vhost_user_main;
  vhost_user_intf_t *vui;

  if (vum->dont_dump_vhost_user_memory)
    {
      pool_foreach (vui, vum->vhost_user_interfaces)
	unmap_all_mem_regions (vui);
    }
}

/*
 * fd.io coding-style-patch-verification: ON
 *
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */