summaryrefslogtreecommitdiffstats
path: root/xlators/nfs/lib/src/rpcsvc.c
blob: 930634d6553c35c0c0c81de12a1691f73d5616bd (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
/*
  Copyright (c) 2010 Gluster, Inc. <http://www.gluster.com>
  This file is part of GlusterFS.

  GlusterFS is free software; you can redistribute it and/or modify
  it under the terms of the GNU General Public License as published
  by the Free Software Foundation; either version 3 of the License,
  or (at your option) any later version.

  GlusterFS is distributed in the hope that it will be useful, but
  WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  General Public License for more details.

  You should have received a copy of the GNU General Public License
  along with this program.  If not, see
  <http://www.gnu.org/licenses/>.
*/

#ifndef _CONFIG_H
#define _CONFIG_H
#include "config.h"
#endif

#include "rpcsvc.h"
#include "rpc-socket.h"
#include "dict.h"
#include "logging.h"
#include "byte-order.h"
#include "common-utils.h"
#include "compat-errno.h"
#include "list.h"
#include "xdr-rpc.h"
#include "iobuf.h"
#include "globals.h"

#include <errno.h>
#include <pthread.h>
#include <stdlib.h>
#include <rpc/rpc.h>
#include <rpc/pmap_clnt.h>
#include <arpa/inet.h>
#include <rpc/xdr.h>
#include <fnmatch.h>
#include <stdarg.h>
#include <stdio.h>


#define nfs_rpcsvc_alloc_request(con, request)                          \
        do {                                                            \
                request = (rpcsvc_request_t *) mem_get ((con)->rxpool); \
                memset (request, 0, sizeof (rpcsvc_request_t));         \
        } while (0)                                                     \

/* The generic event handler for every stage */
void *
nfs_rpcsvc_stage_proc (void *arg)
{
        rpcsvc_stage_t          *stg = (rpcsvc_stage_t *)arg;

        if (!stg)
                return NULL;

        event_dispatch (stg->eventpool);
        return NULL;
}


rpcsvc_stage_t *
nfs_rpcsvc_stage_init (rpcsvc_t *svc)
{
        rpcsvc_stage_t          *stg = NULL;
        int                     ret = -1;
        size_t                  stacksize = RPCSVC_THREAD_STACK_SIZE;
        pthread_attr_t          stgattr;
        unsigned int            eventpoolsize = 0;

        if (!svc)
                return NULL;

        stg = GF_CALLOC (1, sizeof(*stg), gf_common_mt_rpcsvc_stage_t);
        if (!stg)
                return NULL;

        eventpoolsize = svc->memfactor * RPCSVC_EVENTPOOL_SIZE_MULT;
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "event pool size: %d", eventpoolsize);
        stg->eventpool = event_pool_new (eventpoolsize);
        if (!stg->eventpool)
                goto free_stg;

        pthread_attr_init (&stgattr);
        ret = pthread_attr_setstacksize (&stgattr, stacksize);
        if (ret == EINVAL)
                gf_log (GF_RPCSVC, GF_LOG_WARNING,
                                "Using default thread stack size");

        ret = pthread_create (&stg->tid, &stgattr, nfs_rpcsvc_stage_proc,
                              (void *)stg);
        if (ret != 0) {
                ret = -1;
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Stage creation failed");
                goto free_stg;
        }

        stg->svc = svc;
        ret = 0;
free_stg:
        if (ret == -1) {
                GF_FREE (stg);
                stg = NULL;
        }

        return stg;
}


int
nfs_rpcsvc_init_options (rpcsvc_t *svc, dict_t *options)
{
        svc->memfactor = RPCSVC_DEFAULT_MEMFACTOR;
        return 0;
}


/* The global RPC service initializer.
 * Starts up the stages and then waits for RPC program registrations
 * to come in.
 */
rpcsvc_t *
nfs_rpcsvc_init (glusterfs_ctx_t *ctx, dict_t *options)
{
        rpcsvc_t        *svc = NULL;
        int             ret = -1;

        if ((!ctx) || (!options))
                return NULL;

        svc = GF_CALLOC (1, sizeof (*svc), gf_common_mt_rpcsvc_t);
        if (!svc)
                return NULL;

        pthread_mutex_init (&svc->rpclock, NULL);
        INIT_LIST_HEAD (&svc->stages);
        INIT_LIST_HEAD (&svc->authschemes);

        ret = nfs_rpcsvc_init_options (svc, options);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to init options");
                goto free_svc;
        }

        ret = nfs_rpcsvc_auth_init (svc, options);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to init "
                        "authentication");
                goto free_svc;
        }

        ret = -1;
        svc->defaultstage = nfs_rpcsvc_stage_init (svc);
        if (!svc->defaultstage) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR,"RPC service init failed.");
                goto free_svc;
        }
        svc->options = options;
        svc->ctx = ctx;
        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "RPC service inited.");

        ret = 0;
free_svc:
        if (ret == -1) {
                GF_FREE (svc);
                svc = NULL;
        }

        return svc;
}


/* Once multi-threaded support is complete, we'll be able to round-robin
 * the various incoming connections over the many available stages. This
 * function selects one from among all the stages.
 */
rpcsvc_stage_t *
nfs_rpcsvc_select_stage (rpcsvc_t *rpcservice)
{
        if (!rpcservice)
                return NULL;

        return rpcservice->defaultstage;
}


int
nfs_rpcsvc_conn_peer_check_search (dict_t *options, char *pattern, char *clstr)
{
        int                     ret = -1;
        char                    *addrtok = NULL;
        char                    *addrstr = NULL;
        char                    *svptr = NULL;

        if ((!options) || (!clstr))
                return -1;

        if (!dict_get (options, pattern))
                return -1;

        ret = dict_get_str (options, pattern, &addrstr);
        if (ret < 0) {
                ret = -1;
                goto err;
        }

        if (!addrstr) {
                ret = -1;
                goto err;
        }

        addrtok = strtok_r (addrstr, ",", &svptr);
        while (addrtok) {

                /* CASEFOLD not present on Solaris */
#ifdef FNM_CASEFOLD
                ret = fnmatch (addrtok, clstr, FNM_CASEFOLD);
#else
                ret = fnmatch (addrtok, clstr, 0);
#endif
                if (ret == 0)
                        goto err;

                addrtok = strtok_r (NULL, ",", &svptr);
        }

        ret = -1;
err:

        return ret;
}


int
nfs_rpcsvc_conn_peer_check_allow (dict_t *options, char *volname, char *clstr)
{
        int     ret = RPCSVC_AUTH_DONTCARE;
        char    *srchstr = NULL;
        char    globalrule[] = "rpc-auth.addr.allow";

        if ((!options) || (!clstr))
                return ret;

        /* If volname is NULL, then we're searching for the general rule to
         * determine the current address in clstr is allowed or not for all
         * subvolumes.
         */
        if (volname) {
                ret = gf_asprintf (&srchstr, "rpc-auth.addr.%s.allow", volname);
                if (ret == -1) {
                        gf_log (GF_RPCSVC, GF_LOG_ERROR, "asprintf failed");
                        ret = RPCSVC_AUTH_DONTCARE;
                        goto out;
                }
        } else
                srchstr = globalrule;

        ret = nfs_rpcsvc_conn_peer_check_search (options, srchstr, clstr);
        if (volname)
                GF_FREE (srchstr);

        if (ret == 0)
                ret = RPCSVC_AUTH_ACCEPT;
        else
                ret = RPCSVC_AUTH_DONTCARE;
out:
        return ret;
}

int
nfs_rpcsvc_conn_peer_check_reject (dict_t *options, char *volname, char *clstr)
{
        int     ret = RPCSVC_AUTH_DONTCARE;
        char    *srchstr = NULL;
        char    generalrule[] = "rpc-auth.addr.reject";

        if ((!options) || (!clstr))
                return ret;

        if (volname) {
                ret = gf_asprintf (&srchstr, "rpc-auth.addr.%s.reject",
                                   volname);
                if (ret == -1) {
                        gf_log (GF_RPCSVC, GF_LOG_ERROR, "asprintf failed");
                        ret = RPCSVC_AUTH_REJECT;
                        goto out;
                }
        } else
                srchstr = generalrule;

        ret = nfs_rpcsvc_conn_peer_check_search (options, srchstr, clstr);
        if (volname)
                GF_FREE (srchstr);

        if (ret == 0)
                ret = RPCSVC_AUTH_REJECT;
        else
                ret = RPCSVC_AUTH_DONTCARE;
out:
        return ret;
}


/* This function tests the results of the allow rule and the reject rule to
 * combine them into a single result that can be used to determine if the
 * connection should be allowed to proceed.
 * Heres the test matrix we need to follow in this function.
 *
 * A -  Allow, the result of the allow test. Never returns R.
 * R - Reject, result of the reject test. Never returns A.
 * Both can return D or dont care if no rule was given.
 *
 * | @allow | @reject | Result |
 * |    A   |   R     | R      |
 * |    D   |   D     | D      |
 * |    A   |   D     | A      |
 * |    D   |   R     | R      |
 */
int
nfs_rpcsvc_combine_allow_reject_volume_check (int allow, int reject)
{
        int     final = RPCSVC_AUTH_REJECT;

        /* If allowed rule allows but reject rule rejects, we stay cautious
         * and reject. */
        if ((allow == RPCSVC_AUTH_ACCEPT) && (reject == RPCSVC_AUTH_REJECT))
                final = RPCSVC_AUTH_REJECT;
        /* if both are dont care, that is user did not specify for either allow
         * or reject, we leave it up to the general rule to apply, in the hope
         * that there is one.
         */
        else if ((allow == RPCSVC_AUTH_DONTCARE) &&
                 (reject == RPCSVC_AUTH_DONTCARE))
                final = RPCSVC_AUTH_DONTCARE;
        /* If one is dont care, the other one applies. */
        else if ((allow == RPCSVC_AUTH_ACCEPT) &&
                 (reject == RPCSVC_AUTH_DONTCARE))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((allow == RPCSVC_AUTH_DONTCARE) &&
                 (reject == RPCSVC_AUTH_REJECT))
                final = RPCSVC_AUTH_REJECT;

        return final;
}


/* Combines the result of the general rule test against, the specific rule
 * to determine final permission for the client's address.
 *
 * | @gen   | @spec   | Result |
 * |    A   |   A     | A      |
 * |    A   |   R     | R      |
 * |    A   |   D     | A      |
 * |    D   |   A     | A      |
 * |    D   |   R     | R      |
 * |    D   |   D     | D      |
 * |    R   |   A     | A      |
 * |    R   |   D     | R      |
 * |    R   |   R     | R      |
 */
int
nfs_rpcsvc_combine_gen_spec_addr_checks (int gen, int spec)
{
        int     final = RPCSVC_AUTH_REJECT;

        if ((gen == RPCSVC_AUTH_ACCEPT) && (spec == RPCSVC_AUTH_ACCEPT))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((gen == RPCSVC_AUTH_ACCEPT) && (spec == RPCSVC_AUTH_REJECT))
                final = RPCSVC_AUTH_REJECT;
        else if ((gen == RPCSVC_AUTH_ACCEPT) && (spec == RPCSVC_AUTH_DONTCARE))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((gen == RPCSVC_AUTH_DONTCARE) && (spec == RPCSVC_AUTH_ACCEPT))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((gen == RPCSVC_AUTH_DONTCARE) && (spec == RPCSVC_AUTH_REJECT))
                final = RPCSVC_AUTH_REJECT;
        else if ((gen == RPCSVC_AUTH_DONTCARE) && (spec== RPCSVC_AUTH_DONTCARE))
                final = RPCSVC_AUTH_DONTCARE;
        else if ((gen == RPCSVC_AUTH_REJECT) && (spec == RPCSVC_AUTH_ACCEPT))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((gen == RPCSVC_AUTH_REJECT) && (spec == RPCSVC_AUTH_DONTCARE))
                final = RPCSVC_AUTH_REJECT;
        else if ((gen == RPCSVC_AUTH_REJECT) && (spec == RPCSVC_AUTH_REJECT))
                final = RPCSVC_AUTH_REJECT;

        return final;
}



/* Combines the result of the general rule test against, the specific rule
 * to determine final test for the connection coming in for a given volume.
 *
 * | @gen   | @spec   | Result |
 * |    A   |   A     | A      |
 * |    A   |   R     | R      |
 * |    A   |   D     | A      |
 * |    D   |   A     | A      |
 * |    D   |   R     | R      |
 * |    D   |   D     | R      |, special case, we intentionally disallow this.
 * |    R   |   A     | A      |
 * |    R   |   D     | R      |
 * |    R   |   R     | R      |
 */
int
nfs_rpcsvc_combine_gen_spec_volume_checks (int gen, int spec)
{
        int     final = RPCSVC_AUTH_REJECT;

        if ((gen == RPCSVC_AUTH_ACCEPT) && (spec == RPCSVC_AUTH_ACCEPT))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((gen == RPCSVC_AUTH_ACCEPT) && (spec == RPCSVC_AUTH_REJECT))
                final = RPCSVC_AUTH_REJECT;
        else if ((gen == RPCSVC_AUTH_ACCEPT) && (spec == RPCSVC_AUTH_DONTCARE))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((gen == RPCSVC_AUTH_DONTCARE) && (spec == RPCSVC_AUTH_ACCEPT))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((gen == RPCSVC_AUTH_DONTCARE) && (spec == RPCSVC_AUTH_REJECT))
                final = RPCSVC_AUTH_REJECT;
        /* On no rule, we reject. */
        else if ((gen == RPCSVC_AUTH_DONTCARE) && (spec== RPCSVC_AUTH_DONTCARE))
                final = RPCSVC_AUTH_REJECT;
        else if ((gen == RPCSVC_AUTH_REJECT) && (spec == RPCSVC_AUTH_ACCEPT))
                final = RPCSVC_AUTH_ACCEPT;
        else if ((gen == RPCSVC_AUTH_REJECT) && (spec == RPCSVC_AUTH_DONTCARE))
                final = RPCSVC_AUTH_REJECT;
        else if ((gen == RPCSVC_AUTH_REJECT) && (spec == RPCSVC_AUTH_REJECT))
                final = RPCSVC_AUTH_REJECT;

        return final;
}


int
nfs_rpcsvc_conn_peer_check_name (dict_t *options, char *volname,
                                 rpcsvc_conn_t *conn)
{
        int     ret = RPCSVC_AUTH_REJECT;
        int     aret = RPCSVC_AUTH_REJECT;
        int     rjret = RPCSVC_AUTH_REJECT;
        char    clstr[RPCSVC_PEER_STRLEN];

        if (!conn)
                return ret;

        ret = nfs_rpcsvc_conn_peername (conn, clstr, RPCSVC_PEER_STRLEN);
        if (ret != 0) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to get remote addr: "
                        "%s", gai_strerror (ret));
                ret = RPCSVC_AUTH_REJECT;
                goto err;
        }

        aret = nfs_rpcsvc_conn_peer_check_allow (options, volname, clstr);
        rjret = nfs_rpcsvc_conn_peer_check_reject (options, volname, clstr);

        ret = nfs_rpcsvc_combine_allow_reject_volume_check (aret, rjret);

err:
        return ret;
}


int
nfs_rpcsvc_conn_peer_check_addr (dict_t *options, char *volname,
                                 rpcsvc_conn_t *conn)
{
        int     ret = RPCSVC_AUTH_REJECT;
        int     aret = RPCSVC_AUTH_DONTCARE;
        int     rjret = RPCSVC_AUTH_REJECT;
        char    clstr[RPCSVC_PEER_STRLEN];

        if (!conn)
                return ret;

        ret = nfs_rpcsvc_conn_peeraddr (conn, clstr, RPCSVC_PEER_STRLEN, NULL,
                                        0);
        if (ret != 0) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to get remote addr: "
                        "%s", gai_strerror (ret));
                ret = RPCSVC_AUTH_REJECT;
                goto err;
        }

        aret = nfs_rpcsvc_conn_peer_check_allow (options, volname, clstr);
        rjret = nfs_rpcsvc_conn_peer_check_reject (options, volname, clstr);

        ret = nfs_rpcsvc_combine_allow_reject_volume_check (aret, rjret);
err:
        return ret;
}


int
nfs_rpcsvc_conn_check_volume_specific (dict_t *options, char *volname,
                                       rpcsvc_conn_t *conn)
{
        int             namechk = RPCSVC_AUTH_REJECT;
        int             addrchk = RPCSVC_AUTH_REJECT;
        gf_boolean_t    namelookup = _gf_true;
        char            *namestr = NULL;
        int             ret = 0;

        if ((!options) || (!volname) || (!conn))
                return RPCSVC_AUTH_REJECT;

        /* Enabled by default */
        if ((dict_get (options, "rpc-auth.addr.namelookup"))) {
                ret = dict_get_str (options, "rpc-auth.addr.namelookup"
                                    , &namestr);
                if (ret == 0)
                        ret = gf_string2boolean (namestr, &namelookup);
        }

        /* We need two separate checks because the rules with addresses in them
         * can be network addresses which can be general and names can be
         * specific which will over-ride the network address rules.
         */
        if (namelookup)
                namechk = nfs_rpcsvc_conn_peer_check_name (options, volname,
                                                           conn);
        addrchk = nfs_rpcsvc_conn_peer_check_addr (options, volname, conn);

        if (namelookup)
                ret = nfs_rpcsvc_combine_gen_spec_addr_checks (addrchk,
                                                               namechk);
        else
                ret = addrchk;

        return ret;
}


int
nfs_rpcsvc_conn_check_volume_general (dict_t *options, rpcsvc_conn_t *conn)
{
        int             addrchk = RPCSVC_AUTH_REJECT;
        int             namechk = RPCSVC_AUTH_REJECT;
        gf_boolean_t    namelookup = _gf_true;
        char            *namestr = NULL;
        int             ret = 0;

        if ((!options) || (!conn))
                return RPCSVC_AUTH_REJECT;

        /* Enabled by default */
        if ((dict_get (options, "rpc-auth.addr.namelookup"))) {
                ret = dict_get_str (options, "rpc-auth.addr.namelookup"
                                    , &namestr);
                if (ret == 0)
                        ret = gf_string2boolean (namestr, &namelookup);
        }

        /* We need two separate checks because the rules with addresses in them
         * can be network addresses which can be general and names can be
         * specific which will over-ride the network address rules.
         */
        if (namelookup)
                namechk = nfs_rpcsvc_conn_peer_check_name (options, NULL, conn);
        addrchk = nfs_rpcsvc_conn_peer_check_addr (options, NULL, conn);

        if (namelookup)
                ret = nfs_rpcsvc_combine_gen_spec_addr_checks (addrchk,
                                                               namechk);
        else
                ret = addrchk;

        return ret;
}

int
nfs_rpcsvc_conn_peer_check (dict_t *options, char *volname, rpcsvc_conn_t *conn)
{
        int     general_chk = RPCSVC_AUTH_REJECT;
        int     specific_chk = RPCSVC_AUTH_REJECT;

        if ((!options) || (!volname) || (!conn))
                return RPCSVC_AUTH_REJECT;

        general_chk = nfs_rpcsvc_conn_check_volume_general (options, conn);
        specific_chk = nfs_rpcsvc_conn_check_volume_specific (options, volname,
                                                              conn);

        return nfs_rpcsvc_combine_gen_spec_volume_checks (general_chk,
                                                          specific_chk);
}


char *
nfs_rpcsvc_volume_allowed (dict_t *options, char *volname)
{
        char    globalrule[] = "rpc-auth.addr.allow";
        char    *srchstr = NULL;
        char    *addrstr = NULL;
        int     ret = -1;

        if ((!options) || (!volname))
                return NULL;

        ret = gf_asprintf (&srchstr, "rpc-auth.addr.%s.allow", volname);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "asprintf failed");
                goto out;
        }

        if (!dict_get (options, srchstr)) {
                GF_FREE (srchstr);
                srchstr = globalrule;
                ret = dict_get_str (options, srchstr, &addrstr);
        } else
                ret = dict_get_str (options, srchstr, &addrstr);

out:
        return addrstr;
}


/* Initialize the core of a connection */
rpcsvc_conn_t *
nfs_rpcsvc_conn_init (rpcsvc_t *svc, rpcsvc_program_t *prog, int sockfd)
{
        rpcsvc_conn_t  *conn = NULL;
        int             ret = -1;
        unsigned int    poolcount = 0;

        conn = GF_CALLOC (1, sizeof(*conn), gf_common_mt_rpcsvc_conn_t);
        if (!conn) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "memory allocation failed");
                return NULL;
        }

        conn->sockfd = sockfd;
        conn->program = (void *)prog;
        INIT_LIST_HEAD (&conn->txbufs);
        poolcount = RPCSVC_POOLCOUNT_MULT * svc->memfactor;
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "tx pool: %d", poolcount);
        conn->txpool = mem_pool_new (rpcsvc_txbuf_t, poolcount);
        if (!conn->txpool) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "mem pool allocation failed");
                goto free_conn;
        }

        gf_log (GF_RPCSVC, GF_LOG_TRACE, "rx pool: %d", poolcount);
        conn->rxpool = mem_pool_new (rpcsvc_request_t, poolcount);
        if (!conn->rxpool) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "mem pool allocation failed");
                goto free_txp;
        }

        /* Cannot consider a connection connected unless the user of this
         * connection decides it is ready to use. It is possible that we have
         * to free this connection soon after. That free will not happpen
         * unless the state is disconnected.
         */
        conn->connstate = RPCSVC_CONNSTATE_DISCONNECTED;
        pthread_mutex_init (&conn->connlock, NULL);
        conn->connref = 0;
        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "New connection inited: sockfd: %d",
                sockfd);

        ret = 0;
free_txp:
        if (ret == -1)
                mem_pool_destroy (conn->txpool);

free_conn:
        if (ret == -1) {
                GF_FREE (conn);
                conn = NULL;
        }

        return conn;
}


void
nfs_rpcsvc_conn_destroy (rpcsvc_conn_t *conn)
{
        mem_pool_destroy (conn->txpool);
        mem_pool_destroy (conn->rxpool);

        if (conn->program->conn_destroy)
                conn->program->conn_destroy (conn->program->private, conn);

        /* Need to destory record state, txlists etc. */
        GF_FREE (conn);
        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "Connection destroyed");
}


int
__nfs_rpcsvc_conn_unref (rpcsvc_conn_t *conn)
{
        --conn->connref;
        return conn->connref;
}


void
__nfs_rpcsvc_conn_deinit (rpcsvc_conn_t *conn)
{
        if (!conn)
                return;

        if ((conn->stage) && (conn->stage->eventpool)) {
                event_unregister (conn->stage->eventpool, conn->sockfd,
                                  conn->eventidx);
        }

        if (nfs_rpcsvc_conn_check_active (conn)) {
                gf_log (GF_RPCSVC, GF_LOG_DEBUG, "Connection de-activated:"
                        " sockfd: %d", conn->sockfd);
                conn->connstate = RPCSVC_CONNSTATE_DISCONNECTED;
        }

        if (conn->sockfd != -1) {
                close (conn->sockfd);
                conn->sockfd = -1;
        }
}


void
nfs_rpcsvc_conn_deinit (rpcsvc_conn_t *conn)
{
        int ref = 0;

        if (!conn)
                return;

        pthread_mutex_lock (&conn->connlock);
        {
                __nfs_rpcsvc_conn_deinit (conn);
                ref = __nfs_rpcsvc_conn_unref (conn);
        }
        pthread_mutex_unlock (&conn->connlock);

        if (ref == 0)
                nfs_rpcsvc_conn_destroy (conn);

        return;
}


void
nfs_rpcsvc_conn_unref (rpcsvc_conn_t *conn)
{
        int ref = 0;
        if (!conn)
                return;

        pthread_mutex_lock (&conn->connlock);
        {
                ref = __nfs_rpcsvc_conn_unref (conn);
        }
        pthread_mutex_unlock (&conn->connlock);

        if (ref == 0)
                nfs_rpcsvc_conn_destroy (conn);
}


int
nfs_rpcsvc_conn_active (rpcsvc_conn_t *conn)
{
        int     status = 0;

        if (!conn)
                return 0;

        pthread_mutex_lock (&conn->connlock);
        {
                status = nfs_rpcsvc_conn_check_active (conn);
        }
        pthread_mutex_unlock (&conn->connlock);

        return status;
}



void
nfs_rpcsvc_conn_ref (rpcsvc_conn_t *conn)
{
        if (!conn)
                return;

        pthread_mutex_lock (&conn->connlock);
        {
                ++conn->connref;
        }
        pthread_mutex_unlock (&conn->connlock);

        return;
}


void
nfs_rpcsvc_conn_state_init (rpcsvc_conn_t *conn)
{
        if (!conn)
                return;

        ++conn->connref;
        conn->connstate = RPCSVC_CONNSTATE_CONNECTED;
}

/* Builds a rpcsvc_conn_t with the aim of listening on it.
 */
rpcsvc_conn_t *
nfs_rpcsvc_conn_listen_init (rpcsvc_t *svc, rpcsvc_program_t *newprog)
{
        rpcsvc_conn_t  *conn = NULL;
        int             sock = -1;

        if (!newprog)
                return NULL;

        sock = nfs_rpcsvc_socket_listen (newprog->progaddrfamily,
                                         newprog->proghost, newprog->progport);
        if (sock == -1)
                goto err;

        conn = nfs_rpcsvc_conn_init (svc, newprog, sock);
        if (!conn)
                goto sock_close_err;

        nfs_rpcsvc_conn_state_init (conn);
sock_close_err:
        if (!conn)
                close (sock);

err:
        return conn;
}

void
nfs_rpcsvc_record_init (rpcsvc_record_state_t *rs, struct iobuf_pool *pool)
{
        if (!rs)
                return;

        rs->state = RPCSVC_READ_FRAGHDR;
        rs->vecstate = 0;
        rs->remainingfraghdr = RPCSVC_FRAGHDR_SIZE;
        rs->remainingfrag = 0;
        rs->fragsize = 0;
        rs->recordsize = 0;
        rs->islastfrag = 0;

        /* If the rs preserves a ref to the iob used by the previous request,
         * we must unref it here to prevent memory leak.
         * If program actor wanted to keep that memory around, it should've
         * refd it on entry into the actor.
         */
        if (rs->activeiob)
                iobuf_unref (rs->activeiob);

        if (rs->vectoriob) {
                iobuf_unref (rs->vectoriob);
                rs->vectoriob = NULL;
        }

        rs->activeiob = iobuf_get (pool);
        rs->fragcurrent = iobuf_ptr (rs->activeiob);

        memset (rs->fragheader, 0, RPCSVC_FRAGHDR_SIZE);
        rs->hdrcurrent = &rs->fragheader[0];

}


int
nfs_rpcsvc_conn_privport_check (rpcsvc_t *svc, char *volname,
                                rpcsvc_conn_t *conn)
{
        struct sockaddr_in      sa;
        int                     ret = RPCSVC_AUTH_REJECT;
        socklen_t               sasize = sizeof (sa);
        char                    *srchstr = NULL;
        char                    *valstr = NULL;
        int                     globalinsecure = RPCSVC_AUTH_REJECT;
        int                     exportinsecure = RPCSVC_AUTH_DONTCARE;
        uint16_t                port = 0;
        gf_boolean_t            insecure = _gf_false;

        if ((!svc) || (!volname) || (!conn))
                return ret;

        ret = nfs_rpcsvc_conn_peeraddr (conn, NULL, 0, (struct sockaddr *)&sa,
                                        sasize);
        if (ret != 0) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to get peer addr: %s",
                        gai_strerror (ret));
                ret = RPCSVC_AUTH_REJECT;
                goto err;
        }

        port = ntohs (sa.sin_port);
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Client port: %d", (int)port);
        /* If the port is already a privileged one, dont bother with checking
         * options.
         */
        if (port <= 1024) {
                ret = RPCSVC_AUTH_ACCEPT;
                goto err;
        }

        /* Disabled by default */
        if ((dict_get (svc->options, "rpc-auth.ports.insecure"))) {
                ret = dict_get_str (svc->options, "rpc-auth.ports.insecure"
                                    , &srchstr);
                if (ret == 0) {
                        ret = gf_string2boolean (srchstr, &insecure);
                        if (ret == 0) {
                                if (insecure == _gf_true)
                                        globalinsecure = RPCSVC_AUTH_ACCEPT;
                        } else
                                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to"
                                        " read rpc-auth.ports.insecure value");
                } else
                        gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to"
                                " read rpc-auth.ports.insecure value");
        }

        /* Disabled by default */
        ret = gf_asprintf (&srchstr, "rpc-auth.ports.%s.insecure", volname);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "asprintf failed");
                ret = RPCSVC_AUTH_REJECT;
                goto err;
        }

        if (dict_get (svc->options, srchstr)) {
                ret = dict_get_str (svc->options, srchstr, &valstr);
                if (ret == 0) {
                        ret = gf_string2boolean (srchstr, &insecure);
                        if (ret == 0) {
                                if (insecure == _gf_true)
                                       exportinsecure = RPCSVC_AUTH_ACCEPT;
                                else
                                        exportinsecure = RPCSVC_AUTH_REJECT;
                        } else
                                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to"
                                        " read rpc-auth.ports.insecure value");
                } else
                        gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to"
                                " read rpc-auth.ports.insecure value");
        }

        ret = nfs_rpcsvc_combine_gen_spec_volume_checks (globalinsecure,
                                                         exportinsecure);
        if (ret == RPCSVC_AUTH_ACCEPT)
                gf_log (GF_RPCSVC, GF_LOG_DEBUG, "Unprivileged port allowed");
        else
                gf_log (GF_RPCSVC, GF_LOG_DEBUG, "Unprivileged port not"
                        " allowed");

err:
        return ret;
}

/* Inits a rpcsvc_conn_t after accepting the connection.
 */
rpcsvc_conn_t *
nfs_rpcsvc_conn_accept_init (rpcsvc_t *svc, int listenfd,
                             rpcsvc_program_t *destprog)
{
        rpcsvc_conn_t   *newconn = NULL;
        int             sock = -1;
        int             ret = -1;

        sock = nfs_rpcsvc_socket_accept (listenfd);
        if (sock == -1)
                goto err;

        newconn = nfs_rpcsvc_conn_init (svc, destprog, sock);
        if (!newconn) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to init conn object");
                ret = -1;
                goto err;
        }

        nfs_rpcsvc_record_init (&newconn->rstate, svc->ctx->iobuf_pool);
        nfs_rpcsvc_conn_state_init (newconn);
        if (destprog->conn_init)
                destprog->conn_init (destprog->private, newconn);
        ret = 0;

err:
        if (ret == -1)
                close (sock);

        return newconn;
}


/* Once the connection has been created, we need to associate it with
 * a stage so that the selected stage will handle the event on this connection.
 * This function also allows the caller to decide which handler should
 * be executed in the context of the stage, and also which specific events
 * should be handed to the handler when running in this particular stage.
 */
int
nfs_rpcsvc_stage_conn_associate (rpcsvc_stage_t *stg, rpcsvc_conn_t *conn,
                                 event_handler_t handler, void *data)
{
        int     ret = -1;

        if ((!stg) || (!conn))
                return -1;

        conn->stage = stg;
        conn->eventidx = event_register (stg->eventpool, conn->sockfd, handler,
                                         data, 1, 0);
        if (conn->eventidx == -1)
                goto err;

        ret = 0;
err:
        return ret;
}


/* Depending on the state we're in, return the size of the next read request. */
size_t
nfs_rpcsvc_record_read_size (rpcsvc_record_state_t *rs)
{
        size_t  toread = -1;

        if (!rs)
                return -1;

        if (nfs_rpcsvc_record_readfraghdr (rs))
                toread = rs->remainingfraghdr;
        else if (nfs_rpcsvc_record_readfrag (rs))
                toread = rs->remainingfrag;
        else
                toread = RPCSVC_CONN_READ;

        return toread;
}


uint32_t
nfs_rpcsvc_record_extract_fraghdr (char *fraghdr)
{
        uint32_t        hdr = 0;
        if (!fraghdr)
                return 0;

        memcpy ((void *)&hdr, fraghdr, sizeof (hdr));

        hdr = ntohl (hdr);
        return hdr;
}


ssize_t
nfs_rpcsvc_record_read_complete_fraghdr (rpcsvc_record_state_t *rs,
                                         ssize_t dataread)
{
        uint32_t        remhdr = 0;
        char            *fraghdrstart = NULL;
        uint32_t        fraghdr = 0;

        fraghdrstart = &rs->fragheader[0];
        remhdr = rs->remainingfraghdr;
        fraghdr = nfs_rpcsvc_record_extract_fraghdr (fraghdrstart);
        rs->fragsize = RPCSVC_FRAGSIZE (fraghdr);
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Received fragment size: %d",
                rs->fragsize);
        if (nfs_rpcsvc_record_vectored (rs)) {
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC header,"
                        " remaining: %d", RPCSVC_BARERPC_MSGSZ);
                rs->remainingfrag = RPCSVC_BARERPC_MSGSZ;
        } else {
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Regular RPC header,"
                        " remaining: %d", rs->fragsize);
                rs->remainingfrag = rs->fragsize;
        }

        rs->state = RPCSVC_READ_FRAG;
        dataread -= remhdr;
        rs->remainingfraghdr -= remhdr;
        rs->islastfrag = RPCSVC_LASTFRAG (fraghdr);

        return dataread;
}


ssize_t
nfs_rpcsvc_record_read_partial_fraghdr (rpcsvc_record_state_t *rs,
                                        ssize_t dataread)
{

        /* In case we got less than even the remaining header size,
         * we need to consume it all and wait for remaining frag hdr
         * bytes to come in.
         */
         rs->remainingfraghdr -= dataread;
         nfs_rpcsvc_record_update_currenthdr (rs, dataread);
         dataread = 0;
         gf_log (GF_RPCSVC, GF_LOG_TRACE, "Fragment header remaining: %d",
                 rs->remainingfraghdr);

         return dataread;
}


ssize_t
nfs_rpcsvc_record_update_fraghdr (rpcsvc_record_state_t *rs, ssize_t dataread)
{
        if ((!rs) || (dataread <= 0))
                return -1;

        /* Why are we even here, we're not supposed to be in the fragment
         * header processing state.
         */
        if (!nfs_rpcsvc_record_readfraghdr(rs)) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "record state inconsistent"
                        ": request to update frag header when state is not"
                        "RPCSVC_READ_FRAGHDR");
                return -1;
        }

        /* Again, if header has been read then the state member above should've
         * been different, this is crazy. We should not be here.
         */
        if (rs->remainingfraghdr == 0) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "record state inconsistent"
                        ": request to update frag header when frag header"
                        "remaining is 0.");
                return -1;
        }

        /* We've definitely got the full header now and may be even more. */
        if (dataread >= rs->remainingfraghdr)
                dataread = nfs_rpcsvc_record_read_complete_fraghdr (rs,
                                                                    dataread);
        else
                dataread = nfs_rpcsvc_record_read_partial_fraghdr (rs,
                                                                   dataread);

        return dataread;
}

ssize_t
nfs_rpcsvc_record_read_complete_frag (rpcsvc_record_state_t *rs,
                                      ssize_t dataread)
{
        uint32_t        remfrag;

        /* Since the frag is now complete, change the state to the next
         * one, i.e. to read the header of the next fragment.
         */
        remfrag = rs->remainingfrag;
        rs->state = RPCSVC_READ_FRAGHDR;
        dataread -= remfrag;

        /* This will be 0 now. */
        rs->remainingfrag -= remfrag;

        /* Now that the fragment is complete, we must update the
         * record size. Recall that fragsize was got from the frag
         * header.
         */
        rs->recordsize += rs->fragsize;
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Fragment remaining: %d",
                rs->remainingfrag);

        return dataread;
}


ssize_t
nfs_rpcsvc_record_read_partial_frag (rpcsvc_record_state_t *rs,
                                     ssize_t dataread)
{
        /* Just take whatever has come through the current network buffer. */
        rs->remainingfrag -= dataread;

        nfs_rpcsvc_record_update_currentfrag (rs, dataread);
        /* Since we know we're consuming the whole buffer from dataread
         * simply setting to 0 zero is fine.
         */
        dataread = 0;
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Fragment remaining: %d",
                rs->remainingfrag);
        return dataread;
}


ssize_t
nfs_rpcsvc_record_update_frag (rpcsvc_record_state_t *rs, ssize_t dataread)
{
        if ((!rs) || (dataread <= 0))
                return -1;

        if (!nfs_rpcsvc_record_readfrag (rs)) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "record state inconsistent"
                        ": request to update fragment when record state is not"
                        "RPCSVC_READ_FRAG.");
                return -1;
        }

        if (rs->remainingfrag == 0) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "record state inconsistent"
                        ": request to update fragment when there is no fragment"
                        " data remaining to be read.");
                return -1;
        }

        /* We've read in more data than the current fragment requires. */
        if (dataread >= rs->remainingfrag)
                dataread = nfs_rpcsvc_record_read_complete_frag (rs, dataread);
        else
                dataread = nfs_rpcsvc_record_read_partial_frag (rs, dataread);

        return dataread;
}


/* This needs to change to returning errors, since
 * we need to return RPC specific error messages when some
 * of the pointers below are NULL.
 */
rpcsvc_actor_t *
nfs_rpcsvc_program_actor (rpcsvc_conn_t *conn, rpcsvc_request_t *req)
{
        rpcsvc_program_t        *program = NULL;
        int                     err = SYSTEM_ERR;
        rpcsvc_actor_t          *actor = NULL;

        if ((!conn) || (!req))
                goto err;

        program = (rpcsvc_program_t *)conn->program;
        if (!program)
                goto err;

        if (req->prognum != program->prognum) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "RPC program not available");
                err = PROG_UNAVAIL;
                goto err;
        }

        if (!program->actors) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "RPC System error");
                err = SYSTEM_ERR;
                goto err;
        }

        if (req->progver != program->progver) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "RPC program version not"
                        " available");
                err = PROG_MISMATCH;
                goto err;
        }

        if ((req->procnum < 0) || (req->procnum >= program->numactors)) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "RPC Program procedure not"
                        " available");
                err = PROC_UNAVAIL;
                goto err;
        }

        actor = &program->actors[req->procnum];
        if (!actor->actor) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "RPC Program procedure not"
                        " available");
                err = PROC_UNAVAIL;
                actor = NULL;
                goto err;
        }

        err = SUCCESS;
        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "Actor found: %s - %s",
                program->progname, actor->procname);
err:
        if (req)
                req->rpc_err = err;

        return actor;
}


rpcsvc_txbuf_t *
nfs_rpcsvc_init_txbuf (rpcsvc_conn_t *conn, struct iovec msg, struct iobuf *iob,
                       struct iobref *iobref, int txflags)
{
        rpcsvc_txbuf_t  *txbuf = NULL;

        txbuf = (rpcsvc_txbuf_t *) mem_get(conn->txpool);
        if (!txbuf) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to get txbuf");
                return NULL;
        }

        memset (txbuf, 0, sizeof (*txbuf));
        INIT_LIST_HEAD (&txbuf->txlist);
        txbuf->buf = msg;

        /* If it was required, this iob must've been ref'd already
         *  so I dont have to bother here.
         */
        txbuf->iob = iob;
        txbuf->iobref = iobref;
        txbuf->offset = 0;
        txbuf->txbehave = txflags;

        return txbuf;
}


int
nfs_rpcsvc_conn_append_txlist (rpcsvc_conn_t *conn, struct iovec msg,
                               struct iobuf *iob, int txflags)
{
        rpcsvc_txbuf_t          *txbuf = NULL;

        if ((!conn) || (!msg.iov_base) || (!iob))
                return -1;

        txbuf = nfs_rpcsvc_init_txbuf (conn, msg, iob, NULL, txflags);
        if (!txbuf)
                return -1;

        list_add_tail (&txbuf->txlist, &conn->txbufs);
        return 0;
}


void
nfs_rpcsvc_set_lastfrag (uint32_t *fragsize) {
        (*fragsize) |= 0x80000000U;
}

void
nfs_rpcsvc_set_frag_header_size (uint32_t size, char *haddr)
{
        size = htonl (size);
        memcpy (haddr, &size, sizeof (size));
}

void
nfs_rpcsvc_set_last_frag_header_size (uint32_t size, char *haddr)
{
        nfs_rpcsvc_set_lastfrag (&size);
        nfs_rpcsvc_set_frag_header_size (size, haddr);
}


/* Given the RPC reply structure and the payload handed by the RPC program,
 * encode the RPC record header into the buffer pointed by recordstart.
 */
struct iovec
nfs_rpcsvc_record_build_header (char *recordstart, size_t rlen,
                                struct rpc_msg reply, size_t payload)
{
        struct iovec    replyhdr;
        struct iovec    txrecord = {0, 0};
        size_t          fraglen = 0;
        int             ret = -1;

        /* After leaving aside the 4 bytes for the fragment header, lets
         * encode the RPC reply structure into the buffer given to us.
         */
        ret = nfs_rpc_reply_to_xdr (&reply,(recordstart + RPCSVC_FRAGHDR_SIZE),
                                    rlen, &replyhdr);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to create RPC reply");
                goto err;
        }

        fraglen = payload + replyhdr.iov_len;
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Reply fraglen %zu, payload: %zu, "
                "rpc hdr: %zu", fraglen, payload, replyhdr.iov_len);

        /* Since we're not spreading RPC records over mutiple fragments
         * we just set this fragment as the first and last fragment for this
         * record.
         */
        nfs_rpcsvc_set_last_frag_header_size (fraglen, recordstart);

        /* Even though the RPC record starts at recordstart+RPCSVC_FRAGHDR_SIZE
         * we need to transmit the record with the fragment header, which starts
         * at recordstart.
         */
        txrecord.iov_base = recordstart;

        /* Remember, this is only the vec for the RPC header and does not
         * include the payload above. We needed the payload only to calculate
         * the size of the full fragment. This size is sent in the fragment
         * header.
         */
        txrecord.iov_len = RPCSVC_FRAGHDR_SIZE + replyhdr.iov_len;

err:
        return txrecord;
}


int
nfs_rpcsvc_conn_submit (rpcsvc_conn_t *conn, struct iovec hdr,
                        struct iobuf *hdriob, struct iovec msgvec,
                        struct iobuf *msgiob)
{
        int     ret = -1;

        if ((!conn) || (!hdr.iov_base) || (!hdriob))
                return -1;

        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Tx Header: %zu, payload: %zu",
                hdr.iov_len, msgvec.iov_len);
        /* Now that we have both the RPC and Program buffers in xdr format
         * lets hand it to the transmission layer.
         */
        pthread_mutex_lock (&conn->connlock);
        {
                if (!nfs_rpcsvc_conn_check_active (conn)) {
                        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "Connection inactive");
                        goto unlock_err;
                }

                ret = nfs_rpcsvc_conn_append_txlist (conn, hdr, hdriob,
                                                     RPCSVC_TXB_FIRST);
                if (ret == -1) {
                        gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to append "
                                "header to transmission list");
                        goto unlock_err;
                }

                /* It is possible that this RPC reply is an error reply. In that
                 * case we might not have been handed a payload.
                 */
                ret = 0;
                if (msgiob)
                        ret = nfs_rpcsvc_conn_append_txlist (conn, msgvec,
                                                             msgiob,
                                                             RPCSVC_TXB_LAST);
                if (ret == -1) {
                        gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to append"
                                " payload to transmission list");
                        goto unlock_err;
                }
        }
unlock_err:
        pthread_mutex_unlock (&conn->connlock);

        if (ret == -1)
                goto err;

        /* Tell event pool, we're interested in poll_out to trigger flush
         * of our tx buffers.
         */
        conn->eventidx = event_select_on (conn->stage->eventpool, conn->sockfd,
                                          conn->eventidx, -1, 1);
        ret = 0;
err:

        return ret;
}


int
nfs_rpcsvc_fill_reply (rpcsvc_request_t *req, struct rpc_msg *reply)
{
        rpcsvc_program_t        *prog = NULL;
        if ((!req) || (!reply))
                return -1;

        prog = nfs_rpcsvc_request_program (req);
        nfs_rpc_fill_empty_reply (reply, req->xid);

        if (req->rpc_stat == MSG_DENIED)
                nfs_rpc_fill_denied_reply (reply, req->rpc_err, req->auth_err);
        else if (req->rpc_stat == MSG_ACCEPTED)
                nfs_rpc_fill_accepted_reply (reply, req->rpc_err,
                                             prog->proglowvers,
                                             prog->proghighvers,
                                             req->verf.flavour,
                                             req->verf.datalen,
                                             req->verf.authdata);
        else
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Invalid rpc_stat value");

        return 0;
}


/* Given a request and the reply payload, build a reply and encodes the reply
 * into a record header. This record header is encoded into the vector pointed
 * to be recbuf.
 * msgvec is the buffer that points to the payload of the RPC program.
 * This buffer can be NULL, if an RPC error reply is being constructed.
 * The only reason it is needed here is that in case the buffer is provided,
 * we should account for the length of that buffer in the RPC fragment header.
 */
struct iobuf *
nfs_rpcsvc_record_build_record (rpcsvc_request_t *req, size_t payload,
                                struct iovec *recbuf)
{
        struct rpc_msg          reply;
        struct iobuf            *replyiob = NULL;
        char                    *record = NULL;
        struct iovec            recordhdr = {0, };
        size_t                  pagesize = 0;
        rpcsvc_conn_t           *conn = NULL;
        rpcsvc_t                *svc = NULL;

        if ((!req) || (!req->conn) || (!recbuf))
                return NULL;

        /* First, try to get a pointer into the buffer which the RPC
         * layer can use.
         */
        conn = req->conn;
        svc = nfs_rpcsvc_conn_rpcsvc (conn);
        replyiob = iobuf_get (svc->ctx->iobuf_pool);
        pagesize = iobpool_pagesize ((struct iobuf_pool *)svc->ctx->iobuf_pool);
        if (!replyiob) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to get iobuf");
                goto err_exit;
        }

        record = iobuf_ptr (replyiob);  /* Now we have it. */

        /* Fill the rpc structure and XDR it into the buffer got above. */
        nfs_rpcsvc_fill_reply (req, &reply);
        recordhdr = nfs_rpcsvc_record_build_header (record, pagesize, reply,
                                                    payload);
        if (!recordhdr.iov_base) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to build record "
                        " header");
                iobuf_unref (replyiob);
                replyiob = NULL;
                recbuf->iov_base = NULL;
                goto err_exit;
        }

        recbuf->iov_base = recordhdr.iov_base;
        recbuf->iov_len = recordhdr.iov_len;
err_exit:
        return replyiob;
}


/*
 * The function to submit a program message to the RPC service.
 * This message is added to the transmission queue of the
 * conn.
 *
 * Program callers are not expected to use the msgvec->iov_base
 * address for anything else.
 * Nor are they expected to free it once this function returns.
 * Once the transmission of the buffer is completed by the RPC service,
 * the memory area as referenced through @msg will be unrefed.
 * If a higher layer does not want anything to do with this iobuf
 * after this function returns, it should call unref on it. For keeping
 * it around till the transmission is actually complete, rpcsvc also refs it.
 *  *
 * If this function returns an error by returning -1, the
 * higher layer programs should assume that a disconnection happened
 * and should know that the conn memory area as well as the req structure
 * has been freed internally.
 *
 * For now, this function assumes that a submit is always called
 * to send a new record. Later, if there is a situation where different
 * buffers for the same record come from different sources, then we'll
 * need to change this code to account for multiple submit calls adding
 * the buffers into a single record.
 */

int
nfs_rpcsvc_submit_generic (rpcsvc_request_t *req, struct iovec msgvec,
                           struct iobuf *msg)
{
        int                     ret = -1;
        struct iobuf            *replyiob = NULL;
        struct iovec            recordhdr = {0, };
        rpcsvc_conn_t           *conn = NULL;

        if ((!req) || (!req->conn))
                return -1;

        conn = req->conn;
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Tx message: %zu", msgvec.iov_len);
        /* Build the buffer containing the encoded RPC reply. */
        replyiob = nfs_rpcsvc_record_build_record (req, msgvec.iov_len,
                                                   &recordhdr);
        if (!replyiob) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR,"Reply record creation failed");
                goto disconnect_exit;
        }

        /* Must ref the iobuf got from higher layer so that the higher layer
         * can rest assured that it can unref it and leave the final freeing
         * of the buffer to us. Note msg can be NULL if an RPC-only message
         * was being sent. Happens when an RPC error reply is being sent.
         */
        if (msg)
                iobuf_ref (msg);
        ret = nfs_rpcsvc_conn_submit (conn, recordhdr, replyiob, msgvec, msg);
        mem_put (conn->rxpool, req);

        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to submit message");
                iobuf_unref (replyiob);
        }

disconnect_exit:
        /* Note that a unref is called everytime a reply is sent. This is in
         * response to the ref that is performed on the conn when a request is
         * handed to the RPC program.
         *
         * The catch, however, is that if the reply is an rpc error, we must
         * not unref. This is because the ref only contains
         * references for the actors to which the request was handed plus one
         * reference maintained by the RPC layer. By unrefing for a case where
         * no actor was called, we will be losing the ref held for the RPC
         * layer.
         */
        if ((nfs_rpcsvc_request_accepted (req)) &&
            (nfs_rpcsvc_request_accepted_success (req)))
                nfs_rpcsvc_conn_unref (conn);

        return ret;
}


int
nfs_rpcsvc_request_attach_vector (rpcsvc_request_t *req, struct iovec msgvec,
                                  struct iobuf *iob, struct iobref *iobref,
                                  int finalvector)
{
        rpcsvc_txbuf_t          *txb = NULL;
        int                     txflags = 0;

        if ((!req) || (!msgvec.iov_base))
                return -1;

        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Tx Vector: %zu", msgvec.iov_len);
        if (finalvector)
                txflags |= RPCSVC_TXB_LAST;
        /* We only let the user decide whether this is the last vector for the
         * record, since the first vector is always the RPC header.
         */
        txb = nfs_rpcsvc_init_txbuf (req->conn, msgvec, iob, iobref, txflags);
        if (!txb) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Could not init tx buf");
                return -1;
        }

        req->payloadsize += msgvec.iov_len;
        if (iob)
                iobuf_ref (iob);
        if (iobref)
                iobref_ref (iobref);
        list_add_tail (&txb->txlist, &req->txlist);

        return 0;
}


int
nfs_rpcsvc_request_attach_vectors (rpcsvc_request_t *req, struct iovec *payload,
                                   int vcount, struct iobref *piobref)
{
        int     c = 0;
        int     ret = -1;

        for (;c < (vcount-1); c++) {
                ret = nfs_rpcsvc_request_attach_vector (req, payload[c], NULL,
                                                        piobref, 0);
                if (ret < 0) {
                        gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to attach "
                                "vector");
                        goto out;
                }
        }

        ret = nfs_rpcsvc_request_attach_vector (req, payload[vcount-1], NULL,
                                                piobref, 1);
        if (ret < 0)
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to attach final vec");

out:
	return ret;
}


int
nfs_rpcsvc_submit_vectors (rpcsvc_request_t *req)
{
        int                     ret = -1;
        struct iobuf            *replyiob = NULL;
        struct iovec            recordhdr = {0, };
        rpcsvc_txbuf_t          *rpctxb = NULL;

        if ((!req) || (!req->conn))
                return -1;

        /* Build the buffer containing the encoded RPC reply. */
        replyiob = nfs_rpcsvc_record_build_record (req, req->payloadsize,
                                                   &recordhdr);
        if (!replyiob) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR,"Reply record creation failed");
                goto disconnect_exit;
        }

        rpctxb = nfs_rpcsvc_init_txbuf (req->conn, recordhdr, replyiob, NULL,
                                        RPCSVC_TXB_FIRST);
        if (!rpctxb) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to create tx buf");
                goto disconnect_exit;
        }

        pthread_mutex_lock (&req->conn->connlock);
        {
                list_splice_init (&req->txlist, &req->conn->txbufs);
                list_add (&rpctxb->txlist, &req->conn->txbufs);
        }
        pthread_mutex_unlock (&req->conn->connlock);

        ret = 0;
        req->conn->eventidx = event_select_on (req->conn->stage->eventpool,
                                               req->conn->sockfd,
                                               req->conn->eventidx, -1, 1);
disconnect_exit:
        /* Note that a unref is called everytime a reply is sent. This is in
         * response to the ref that is performed on the conn when a request is
         * handed to the RPC program.
         */
        nfs_rpcsvc_conn_unref (req->conn);
        if (ret == -1)
                iobuf_unref (replyiob);

        mem_put (req->conn->rxpool, req);
        return ret;
}


int
nfs_rpcsvc_error_reply (rpcsvc_request_t *req)
{
        struct iovec    dummyvec = {0, };

        if (!req)
                return -1;

        /* At this point the req should already have been filled with the
         * appropriate RPC error numbers.
         */
        return nfs_rpcsvc_submit_generic (req, dummyvec, NULL);
}


rpcsvc_request_t *
nfs_rpcsvc_request_init (rpcsvc_conn_t *conn, struct rpc_msg *callmsg,
                         struct iovec progmsg, rpcsvc_request_t *req)
{
        if ((!conn) || (!callmsg)|| (!req))
                return NULL;


        /* We start a RPC request as always denied. */
        req->rpc_stat = MSG_DENIED;
        req->xid = nfs_rpc_call_xid (callmsg);
        req->prognum = nfs_rpc_call_program (callmsg);
        req->progver = nfs_rpc_call_progver (callmsg);
        req->procnum = nfs_rpc_call_progproc (callmsg);
        req->conn = conn;
        req->msg = progmsg;
        req->recordiob = conn->rstate.activeiob;
        INIT_LIST_HEAD (&req->txlist);
        req->payloadsize = 0;

        /* By this time, the data bytes for the auth scheme would have already
         * been copied into the required sections of the req structure,
         * we just need to fill in the meta-data about it now.
         */
        req->cred.flavour = nfs_rpc_call_cred_flavour (callmsg);
        req->cred.datalen = nfs_rpc_call_cred_len (callmsg);
        req->verf.flavour = nfs_rpc_call_verf_flavour (callmsg);
        req->verf.datalen = nfs_rpc_call_verf_len (callmsg);

        /* AUTH */
        nfs_rpcsvc_auth_request_init (req);
        return req;
}


rpcsvc_request_t *
nfs_rpcsvc_request_create (rpcsvc_conn_t *conn)
{
        char                    *msgbuf = NULL;
        struct rpc_msg          rpcmsg;
        struct iovec            progmsg;        /* RPC Program payload */
        rpcsvc_request_t        *req = NULL;
        int                     ret = -1;

        if (!conn)
                return NULL;

        /* We need to allocate the request before actually calling
         * rpcsvc_request_init on the request so that we, can fill the auth
         * data directly into the request structure from the message iobuf.
         * This avoids a need to keep a temp buffer into which the auth data
         * would've been copied otherwise.
         */
        nfs_rpcsvc_alloc_request (conn, req);
        if (!req) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed to alloc request");
                goto err;
        }

        msgbuf = iobuf_ptr (conn->rstate.activeiob);
        ret = nfs_xdr_to_rpc_call (msgbuf, conn->rstate.recordsize, &rpcmsg,
                                   &progmsg, req->cred.authdata,
                                   req->verf.authdata);

        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "RPC call decoding failed");
                nfs_rpcsvc_request_seterr (req, GARBAGE_ARGS);
                goto err;
        }

        ret = -1;
        nfs_rpcsvc_request_init (conn, &rpcmsg, progmsg, req);

        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "RPC XID: %lx, Ver: %ld, Program: %ld,"
                " ProgVers: %ld, Proc: %ld", nfs_rpc_call_xid (&rpcmsg),
                nfs_rpc_call_rpcvers (&rpcmsg), nfs_rpc_call_program (&rpcmsg),
                nfs_rpc_call_progver (&rpcmsg),
                nfs_rpc_call_progproc (&rpcmsg));

        if (nfs_rpc_call_rpcvers (&rpcmsg) != 2) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "RPC version not supported");
                nfs_rpcsvc_request_seterr (req, RPC_MISMATCH);
                goto err;
        }

        ret = nfs_rpcsvc_authenticate (req);
        if (ret == RPCSVC_AUTH_REJECT) {
                /* No need to set auth_err, that is the responsibility of
                 * the authentication handler since only that know what exact
                 * error happened.
                 */
                nfs_rpcsvc_request_seterr (req, AUTH_ERROR);
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Failed authentication");
                ret = -1;
                goto err;
        }


        /* If the error is not RPC_MISMATCH, we consider the call as accepted
         * since we are not handling authentication failures for now.
         */
        req->rpc_stat = MSG_ACCEPTED;
        ret = 0;
err:
        if (ret == -1) {
                ret = nfs_rpcsvc_error_reply (req);
                req = NULL;
        }

        return req;
}


int
nfs_rpcsvc_handle_rpc_call (rpcsvc_conn_t *conn)
{
        rpcsvc_actor_t          *actor = NULL;
        rpcsvc_request_t        *req = NULL;
        int                     ret = -1;

        if (!conn)
                return -1;

        req = nfs_rpcsvc_request_create (conn);
        if (!req)
                goto err;

        if (!nfs_rpcsvc_request_accepted (req))
                goto err_reply;

        actor = nfs_rpcsvc_program_actor (conn, req);
        if (!actor)
                goto err_reply;

        if ((actor) && (actor->actor)) {
                nfs_rpcsvc_conn_ref (conn);
                ret = actor->actor (req);
        }

err_reply:
        if (ret == RPCSVC_ACTOR_ERROR)
                ret = nfs_rpcsvc_error_reply (req);

        /* No need to propagate error beyond this function since the reply
         * has now been queued. */
        ret = 0;
err:
        return ret;
}

#define nfs_rpc_call_cred_addr(rs) (iobuf_ptr ((rs)->activeiob) + RPCSVC_BARERPC_MSGSZ - 4)

uint32_t
nfs_rpcsvc_call_credlen (rpcsvc_record_state_t *rs)
{
        char                    *credaddr = NULL;
        uint32_t                credlen_nw = 0;
        uint32_t                credlen_host = 0;

        /* Position to the start of the credential length field. */
        credaddr = nfs_rpc_call_cred_addr (rs);
        credlen_nw = *(uint32_t *)credaddr;
        credlen_host = ntohl (credlen_nw);

        return credlen_host;
}

uint32_t
nfs_rpcsvc_call_verflen (rpcsvc_record_state_t *rs)
{
        char            *verfaddr = NULL;
        uint32_t        verflen_nw = 0;
        uint32_t        verflen_host = 0;
        uint32_t        credlen = 0;

        /* Position to the start of the verifier length field. */
        credlen = nfs_rpcsvc_call_credlen (rs);
        verfaddr = (nfs_rpc_call_cred_addr (rs) + 4 + credlen);
        verflen_nw = *(uint32_t *)verfaddr;
        verflen_host = ntohl (verflen_nw);

        return verflen_host;
}


void
nfs_rpcsvc_update_vectored_verf (rpcsvc_record_state_t *rs)
{
        if (!rs)
                return;

        rs->recordsize += nfs_rpcsvc_call_verflen (rs);
        return;
}


void
nfs_rpcsvc_handle_vectored_prep_rpc_call (rpcsvc_conn_t *conn)
{
        rpcsvc_actor_t          *actor = NULL;
        rpcsvc_request_t        *req = NULL;
        rpcsvc_record_state_t   *rs = NULL;
        rpcsvc_t                *svc = NULL;
        int                     ret = -1;
        ssize_t                 remfrag = RPCSVC_ACTOR_ERROR;
        int                     newbuf = 0;

        if (!conn)
                return;

        rs = &conn->rstate;

        /* In case one of the steps below fails, we need to make sure that the
         * remaining frag in the kernel's buffers are read-out so that the
         * requests that follow can be served.
         */
        rs->remainingfrag = rs->fragsize - rs->recordsize;
        rs->vecstate = RPCSVC_VECTOR_IGNORE;
        req = nfs_rpcsvc_request_create (conn);
        svc = nfs_rpcsvc_conn_rpcsvc (conn);
        if (!req)
                goto err;

        if (!nfs_rpcsvc_request_accepted (req))
                goto err_reply;

        actor = nfs_rpcsvc_program_actor (conn, req);
        if (!actor)
                goto err_reply;

        if (!actor->vector_sizer) {
                ret = -1;
                nfs_rpcsvc_request_seterr (req, PROC_UNAVAIL);
                goto err_reply;
        }

        nfs_rpcsvc_conn_ref (conn);
        ret = actor->vector_sizer (req, &remfrag, &newbuf);
        nfs_rpcsvc_conn_unref (conn);

        if (ret == RPCSVC_ACTOR_ERROR) {
                ret = -1;
                nfs_rpcsvc_request_seterr (req, SYSTEM_ERR);
                goto err_reply;
        }

        rs->remainingfrag = remfrag;
        rs->vecstate = RPCSVC_VECTOR_READPROCHDR;
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC proc header remaining:"
                " %d", rs->remainingfrag);
        conn->vectoredreq = req;

        /* Store the reference to the current frag pointer. This is where the
         * proc header will be read into.
         */
        req->msg.iov_base = rs->fragcurrent;
        req->msg.iov_len = rs->remainingfrag;
        ret = 0;

err_reply:
        if (ret == -1)
                ret = nfs_rpcsvc_error_reply (req);

        /* No need to propagate error beyond this function since the reply
         * has now been queued. */
        ret = 0;
err:
        return;
}


void
nfs_rpcsvc_update_vectored_verfsz (rpcsvc_conn_t *conn)
{
        rpcsvc_record_state_t   *rs = NULL;
        uint32_t                verflen = 0;

        if (!conn)
                return;

        rs = &conn->rstate;

        verflen = nfs_rpcsvc_call_verflen (rs);
        rs->recordsize += 8;
        if (verflen > 0) {
                rs->remainingfrag = verflen;
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC verf remaining: "
                        " %d", rs->remainingfrag);
                rs->vecstate = RPCSVC_VECTOR_READVERF;
        } else {
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC preparing call");
                nfs_rpcsvc_handle_vectored_prep_rpc_call (conn);
        }

        return;
}


void
nfs_rpcsvc_update_vectored_cred (rpcsvc_record_state_t *rs)
{
        uint32_t                credlen = 0;

        if (!rs)
                return;

        credlen = nfs_rpcsvc_call_credlen (rs);
        /* Update remainingfrag to read the 8 bytes needed for
         * reading verf flavour and verf len.
         */
        rs->remainingfrag = (2 * sizeof (uint32_t));
        rs->vecstate = RPCSVC_VECTOR_READVERFSZ;
        rs->recordsize += credlen;
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC verfsz remaining: %d",
                rs->remainingfrag);

        return;
}

void
nfs_rpcsvc_update_vectored_barerpc (rpcsvc_record_state_t *rs)
{
        uint32_t                credlen = 0;

        if (!rs)
                return;

        credlen = nfs_rpcsvc_call_credlen (rs);
        rs->recordsize = RPCSVC_BARERPC_MSGSZ;
        if (credlen == 0) {
                rs->remainingfrag = 8;
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC verfsz remaining"
                        ": %d", rs->remainingfrag);
                rs->vecstate = RPCSVC_VECTOR_READVERFSZ;
        } else {
                rs->remainingfrag = credlen;
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC cred remaining: "
                        "%d", rs->remainingfrag);
                rs->vecstate = RPCSVC_VECTOR_READCRED;
        }

        return;
}


void
nfs_rpcsvc_handle_vectored_rpc_call (rpcsvc_conn_t *conn)
{
        rpcsvc_actor_t          *actor = NULL;
        rpcsvc_request_t        *req = NULL;
        rpcsvc_record_state_t   *rs = NULL;
        rpcsvc_t                *svc = NULL;
        int                     ret = -1;
        ssize_t                 remfrag = -1;
        int                     newbuf = 0;

        if (!conn)
                return;

        rs = &conn->rstate;

        req = conn->vectoredreq;
        svc = nfs_rpcsvc_conn_rpcsvc (conn);

        if (!req)
                goto err;

        actor = nfs_rpcsvc_program_actor (conn, req);
        if (!actor)
                goto err_reply;

        if (!actor->vector_sizer) {
                ret = -1;
                nfs_rpcsvc_request_seterr (req, PROC_UNAVAIL);
                goto err_reply;
        }

        req->msg.iov_len = (unsigned long)((long)rs->fragcurrent - (long)req->msg.iov_base);
        nfs_rpcsvc_conn_ref (conn);
        ret = actor->vector_sizer (req, &remfrag, &newbuf);
        nfs_rpcsvc_conn_unref (conn);
        if (ret == RPCSVC_ACTOR_ERROR) {
                ret = -1;
                nfs_rpcsvc_request_seterr (req, SYSTEM_ERR);
                goto err_reply;
        }

        if (newbuf) {
                rs->vectoriob = iobuf_get (svc->ctx->iobuf_pool);
                rs->fragcurrent = iobuf_ptr (rs->vectoriob);
                rs->vecstate = RPCSVC_VECTOR_READVEC;
                rs->remainingfrag = remfrag;
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC buf remaining:"
                        " %d", rs->remainingfrag);
        } else {
                rs->remainingfrag = remfrag;
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC proc remaining:"
                        " %d", rs->remainingfrag);
        }

        ret = 0;
err_reply:
        if (ret == -1)
                ret = nfs_rpcsvc_error_reply (req);

        /* No need to propagate error beyond this function since the reply
         * has now been queued. */
        ret = 0;
err:
        return;
}



void
nfs_rpcsvc_record_vectored_call_actor (rpcsvc_conn_t *conn)
{
        rpcsvc_actor_t          *actor = NULL;
        rpcsvc_request_t        *req = NULL;
        rpcsvc_record_state_t   *rs = NULL;
        rpcsvc_t                *svc = NULL;
        int                     ret = -1;

        if (!conn)
                return;

        rs = &conn->rstate;
        req = conn->vectoredreq;
        svc = nfs_rpcsvc_conn_rpcsvc (conn);

        if (!req)
                goto err;

        actor = nfs_rpcsvc_program_actor (conn, req);
        if (!actor)
                goto err_reply;

        if (actor->vector_actor) {
                nfs_rpcsvc_conn_ref (conn);
                ret = actor->vector_actor (req, rs->vectoriob);
        } else {
                nfs_rpcsvc_request_seterr (req, PROC_UNAVAIL);
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "No vectored handler present");
                ret = RPCSVC_ACTOR_ERROR;
        }

err_reply:
        if (ret == RPCSVC_ACTOR_ERROR)
                ret = nfs_rpcsvc_error_reply (req);

        /* No need to propagate error beyond this function since the reply
         * has now been queued. */
        ret = 0;
err:
        return;
}



ssize_t
nfs_rpcsvc_update_vectored_state (rpcsvc_conn_t *conn)
{
        rpcsvc_record_state_t   *rs = NULL;
        rpcsvc_t                *svc = NULL;

        if (!conn)
                return 0;

        /* At this point, we can be confident that the activeiob contains
         * exactly the first RPCSVC_BARERPC_MSGSZ bytes needed in order to
         * determine the program and actor. So the next state will be
         * to read the credentials.
         *
         * But first, try to determine how many more bytes do we need from the
         * network to complete the RPC message including the credentials.
         */

        rs = &conn->rstate;
        if (nfs_rpcsvc_record_vectored_baremsg (rs))
                nfs_rpcsvc_update_vectored_barerpc (rs);
        else if (nfs_rpcsvc_record_vectored_cred (rs))
                nfs_rpcsvc_update_vectored_cred (rs);
        else if (nfs_rpcsvc_record_vectored_verfsz (rs))
                nfs_rpcsvc_update_vectored_verfsz (conn);
        else if (nfs_rpcsvc_record_vectored_verfread (rs)) {
                nfs_rpcsvc_update_vectored_verf (rs);
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC preparing call");
                nfs_rpcsvc_handle_vectored_prep_rpc_call (conn);
        } else if (nfs_rpcsvc_record_vectored_readprochdr (rs))
                nfs_rpcsvc_handle_vectored_rpc_call (conn);
        else if (nfs_rpcsvc_record_vectored_ignore (rs)) {
                svc = nfs_rpcsvc_conn_rpcsvc (conn);
                nfs_rpcsvc_record_init (rs, svc->ctx->iobuf_pool);
        } else if (nfs_rpcsvc_record_vectored_readvec (rs)) {
                svc = nfs_rpcsvc_conn_rpcsvc (conn);
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored RPC vector read");
                nfs_rpcsvc_record_vectored_call_actor (conn);
                nfs_rpcsvc_record_init (rs, svc->ctx->iobuf_pool);
        }

        return 0;
}


ssize_t
nfs_rpcsvc_record_read_partial_frag (rpcsvc_record_state_t *rs,
                                     ssize_t dataread);

ssize_t
nfs_rpcsvc_update_vectored_msg (rpcsvc_conn_t *conn, ssize_t dataread)
{

        if (!conn)
                return dataread;

        /* find out how much of the bare msg is pending and set that up to be
         * read into the updated fragcurrent along with the updated size into
         * remainingfrag.
         */


        /* Incidently, the logic needed here is similar to a regular partial
         * fragment read since we've already set the remainingfrag member in
         * rstate to be RPCSVC_BARERPC_MSGSZ for the purpose of a vectored
         * fragment.
         */
        return nfs_rpcsvc_record_read_partial_frag (&conn->rstate, dataread);
}

/* FIX: As a first version of vectored reading, I am assuming dataread will
 * always be equal to RPCSVC_BARERPC_MSGSZ for the sake of simplicity on the
 * belief that we're never actually reading more bytes than needed in each
 * poll_in.
 */
ssize_t
nfs_rpcsvc_handle_vectored_frag (rpcsvc_conn_t *conn, ssize_t dataread)
{
        if (!conn)
                return dataread;

        /* At this point we can be confident that only the frag size has been
         * read from the network. Now it is up to us to have the remaining RPC
         * fields given to us here.
         */

        /* Since the poll_in handler uses the remainingfrag field to determine
         * how much to read from the network, we'll hack this scheme to tell
         * the poll_in to read at most RPCSVC_BARERPC_MSGSZ bytes. This is done
         * to, as a first step, identify which (program, actor) we need to call.
         */

        dataread = nfs_rpcsvc_update_vectored_msg (conn, dataread);

        if (conn->rstate.remainingfrag == 0) {
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored frag complete");
                dataread = nfs_rpcsvc_update_vectored_state (conn);
        }

        return dataread;
}


int
nfs_rpcsvc_record_update_state (rpcsvc_conn_t *conn, ssize_t dataread)
{
        rpcsvc_record_state_t   *rs = NULL;
        rpcsvc_t                *svc = NULL;

        if (!conn)
                return -1;

        rs = &conn->rstate;
        /* At entry into this function, fragcurrent will be pointing to the\
         * start of the area into which dataread number of bytes were read.
         */

        if (nfs_rpcsvc_record_readfraghdr(rs))
                dataread = nfs_rpcsvc_record_update_fraghdr (rs, dataread);

        if (nfs_rpcsvc_record_readfrag(rs)) {
                /* The minimum needed for triggering the vectored handler is
                 * the frag size field. The fragsize member remains set to this
                 * size till this request is completely extracted from the
                 * network. Once all the data has been read from the network,
                 * the request structure would've been created. The point being
                 * that even if it takes multiple calls to network IO for
                 * getting the vectored fragment, we can continue to use this
                 * condition as the flag to tell us that this is a vectored
                 * fragment.
                 */
                if ((dataread > 0) && (nfs_rpcsvc_record_vectored (rs))) {
                        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Vectored frag");
                        dataread = nfs_rpcsvc_handle_vectored_frag (conn,
                                                                    dataread);
                } else if (dataread > 0) {
                        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Regular frag");
                        dataread = nfs_rpcsvc_record_update_frag (rs, dataread);
                }
        }

        /* This should not happen. We are never reading more than the current
         * fragment needs. Something is seriously wrong.
         */
        if (dataread > 0) {
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Data Left: %zd", dataread);
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Unwanted data read from "
                        " connection.");
        }

        /* If we're now supposed to wait for a new fragment header and if the
         * fragment that we just completed in the previous call to
         * rpcsvc_record_update_frag was the last fragment for the current
         * RPC record, then, it is time to perform the translation from
         * XDR formatted buffer in activeiob followed by the upcall to the
         * protocol actor.
         */
        if ((nfs_rpcsvc_record_readfraghdr(rs)) && (rs->islastfrag)) {
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "Full Record Received.");
                nfs_rpcsvc_handle_rpc_call (conn);
                svc = nfs_rpcsvc_conn_rpcsvc (conn);
                nfs_rpcsvc_record_init (rs, svc->ctx->iobuf_pool);
        }

        return 0;
}


char *
nfs_rpcsvc_record_read_addr (rpcsvc_record_state_t *rs)
{

        if (nfs_rpcsvc_record_readfraghdr (rs))
                return nfs_rpcsvc_record_currenthdr_addr (rs);
        else if (nfs_rpcsvc_record_readfrag (rs))
                return nfs_rpcsvc_record_currentfrag_addr (rs);

        return NULL;
}


int
nfs_rpcsvc_conn_data_poll_in (rpcsvc_conn_t *conn)
{
        ssize_t         dataread = -1;
        size_t          readsize = 0;
        char            *readaddr = NULL;
        int             ret = -1;

        readaddr = nfs_rpcsvc_record_read_addr (&conn->rstate);
        if (!readaddr)
                goto err;

        readsize = nfs_rpcsvc_record_read_size (&conn->rstate);
        if (readsize == -1)
                goto err;

        dataread = nfs_rpcsvc_socket_read (conn->sockfd, readaddr, readsize);
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "conn: 0x%lx, readsize: %zu, dataread:"
                "%zd", (long)conn, readsize, dataread);

        if (dataread > 0)
                ret = nfs_rpcsvc_record_update_state (conn, dataread);

err:
        return ret;
}


int
nfs_rpcsvc_conn_data_poll_err (rpcsvc_conn_t *conn)
{
        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Received error event");
        nfs_rpcsvc_conn_deinit (conn);
        return 0;
}


int
__nfs_rpcsvc_conn_data_poll_out (rpcsvc_conn_t *conn)
{
        rpcsvc_txbuf_t          *txbuf = NULL;
        rpcsvc_txbuf_t          *tmp = NULL;
        ssize_t                 written = -1;
        char                    *writeaddr = NULL;
        size_t                  writesize = -1;

        if (!conn)
                return -1;

        /* Attempt transmission of each of the pending buffers */
        list_for_each_entry_safe (txbuf, tmp, &conn->txbufs, txlist) {
tx_remaining:
                writeaddr = (char *)(txbuf->buf.iov_base + txbuf->offset);
                writesize = (txbuf->buf.iov_len - txbuf->offset);

                if (txbuf->txbehave & RPCSVC_TXB_FIRST) {
                        gf_log (GF_RPCSVC, GF_LOG_TRACE, "First Tx Buf");
                        nfs_rpcsvc_socket_block_tx (conn->sockfd);
                }

                written = nfs_rpcsvc_socket_write (conn->sockfd, writeaddr,
                                                   writesize);
                if (txbuf->txbehave & RPCSVC_TXB_LAST) {
                        gf_log (GF_RPCSVC, GF_LOG_TRACE, "Last Tx Buf");
                        nfs_rpcsvc_socket_unblock_tx (conn->sockfd);
                }
                gf_log (GF_RPCSVC, GF_LOG_TRACE, "conn: 0x%lx, Tx request: %zu,"
                        " Tx sent: %zd", (long)conn, writesize, written);

                /* There was an error transmitting this buffer */
                if (written == -1)
                        break;

                if (written >= 0)
                        txbuf->offset += written;

                /* If the current buffer has been completely transmitted,
                 * delete it from the list and move on to the next buffer.
                 */
                if (txbuf->offset == txbuf->buf.iov_len) {
                        /* It doesnt matter who ref'ed this iobuf, rpcsvc for
                         * its own header or a RPC program.
                         */
                        if (txbuf->iob)
                                iobuf_unref (txbuf->iob);
                        if (txbuf->iobref)
                                iobref_unref (txbuf->iobref);

                        list_del (&txbuf->txlist);
                        mem_put (conn->txpool, txbuf);
                } else
                        /* If the current buffer is incompletely tx'd, do not
                         * go to the head of the loop, since that moves us to
                         * the next buffer.
                         */
                        goto tx_remaining;
        }

        /* If we've broken out of the loop above then we must unblock
         * the transmission now.
         */
        nfs_rpcsvc_socket_unblock_tx (conn->sockfd);
        if (list_empty (&conn->txbufs))
                conn->eventidx = event_select_on (conn->stage->eventpool,
                                                  conn->sockfd, conn->eventidx,
                                                  -1, 0);

        return 0;
}


int
nfs_rpcsvc_conn_data_poll_out (rpcsvc_conn_t *conn)
{
        if (!conn)
                return -1;


        pthread_mutex_lock (&conn->connlock);
        {
                __nfs_rpcsvc_conn_data_poll_out (conn);
        }
        pthread_mutex_unlock (&conn->connlock);

        return 0;
}


int
nfs_rpcsvc_conn_data_handler (int fd, int idx, void *data, int poll_in,
                              int poll_out, int poll_err)
{
        rpcsvc_conn_t   *conn = NULL;
        int             ret = 0;

        if (!data)
                return 0;

        conn = (rpcsvc_conn_t *)data;

        if (poll_out)
                ret = nfs_rpcsvc_conn_data_poll_out (conn);

        if (poll_err) {
                ret = nfs_rpcsvc_conn_data_poll_err (conn);
                return 0;
        }

        if (poll_in) {
                ret = 0;
                ret = nfs_rpcsvc_conn_data_poll_in (conn);
        }

        if (ret == -1)
                nfs_rpcsvc_conn_data_poll_err (conn);

        return 0;
}


int
nfs_rpcsvc_conn_listening_handler (int fd, int idx, void *data, int poll_in,
                                   int poll_out, int poll_err)
{
        rpcsvc_conn_t           *newconn = NULL;
        rpcsvc_stage_t          *selectedstage = NULL;
        int                     ret = -1;
        rpcsvc_conn_t           *conn = NULL;
        rpcsvc_program_t        *prog = NULL;
        rpcsvc_t                *svc = NULL;

        if (!poll_in)
                return 0;

        conn = (rpcsvc_conn_t *)data;
        prog = (rpcsvc_program_t *)conn->program;
        svc = nfs_rpcsvc_conn_rpcsvc (conn);
        newconn = nfs_rpcsvc_conn_accept_init (svc, fd, prog);
        if (!newconn) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "failed to accept connection");
                goto err;
        }

        selectedstage = nfs_rpcsvc_select_stage (svc);
        if (!selectedstage)
                goto close_err;

        /* Now that we've accepted the connection, we need to associate
         * its events to a stage.
         */
        ret = nfs_rpcsvc_stage_conn_associate (selectedstage, newconn,
                                               nfs_rpcsvc_conn_data_handler,
                                               newconn);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "could not associated stage "
                        " with new connection");
                goto close_err;
        }
        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "New Connection: Program %s, Num: %d,"
                " Ver: %d, Port: %d", prog->progname, prog->prognum,
                prog->progver, prog->progport);
        ret = 0;
close_err:
        if (ret == -1)
                nfs_rpcsvc_conn_unref (newconn);

err:
        return ret;
}


/* Register the program with the local portmapper service. */
int
nfs_rpcsvc_program_register_portmap (rpcsvc_program_t *newprog)
{
        if (!newprog)
                return -1;

        if (!(pmap_set(newprog->prognum, newprog->progver, IPPROTO_TCP,
                       newprog->progport))) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Could not register with"
                        " portmap");
                return -1;
        }

        return 0;
}


int
nfs_rpcsvc_program_unregister_portmap (rpcsvc_program_t *prog)
{
        if (!prog)
                return -1;

        if (!(pmap_unset(prog->prognum, prog->progver))) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Could not unregister with"
                        " portmap");
                return -1;
        }

        return 0;
}


int
nfs_rpcsvc_stage_program_register (rpcsvc_stage_t *stg,
                                   rpcsvc_program_t *newprog)
{
        rpcsvc_conn_t           *newconn = NULL;
        rpcsvc_t                *svc = NULL;

        if ((!stg) || (!newprog))
                return -1;

        svc = nfs_rpcsvc_stage_service (stg);
        /* Create a listening socket */
        newconn = nfs_rpcsvc_conn_listen_init (svc, newprog);
        if (!newconn) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "could not create listening"
                        " connection");
                return -1;
        }

        if ((nfs_rpcsvc_stage_conn_associate (stg, newconn,
                                              nfs_rpcsvc_conn_listening_handler,
                                              newconn)) == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR,"could not associate stage with"
                        " listening connection");
                return -1;
        }

        return 0;
}


int
nfs_rpcsvc_program_register (rpcsvc_t *svc, rpcsvc_program_t program)
{
        rpcsvc_program_t        *newprog = NULL;
        rpcsvc_stage_t          *selectedstage = NULL;
        int                     ret = -1;

        if (!svc)
                return -1;

        newprog = GF_CALLOC (1, sizeof(*newprog),gf_common_mt_rpcsvc_program_t);
        if (!newprog)
                return -1;

        if (!program.actors)
                goto free_prog;

        memcpy (newprog, &program, sizeof (program));
        selectedstage = nfs_rpcsvc_select_stage (svc);

        ret = nfs_rpcsvc_stage_program_register (selectedstage, newprog);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "stage registration of program"
                        " failed");
                goto free_prog;
        }

        ret = nfs_rpcsvc_program_register_portmap (newprog);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "portmap registration of"
                        " program failed");
                goto free_prog;
        }

        ret = 0;
        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "New program registered: %s, Num: %d,"
                " Ver: %d, Port: %d", newprog->progname, newprog->prognum,
                newprog->progver, newprog->progport);

free_prog:
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Program registration failed:"
                        " %s, Num: %d, Ver: %d, Port: %d", newprog->progname,
                        newprog->prognum, newprog->progver, newprog->progport);
                GF_FREE (newprog);
        }

        return ret;
}

/* The only difference between the generic submit and this one is that the
 * generic submit is also used for submitting RPC error replies in where there
 * are no payloads so the msgvec and msgbuf can be NULL.
 * Since RPC programs should be using this function along with their payloads
 * we must perform NULL checks before calling the generic submit.
 */
int
nfs_rpcsvc_submit_message (rpcsvc_request_t *req, struct iovec msgvec,
                           struct iobuf *msg)
{
        if ((!req) || (!req->conn) || (!msg) || (!msgvec.iov_base))
                return -1;

        return nfs_rpcsvc_submit_generic (req, msgvec, msg);
}


int
nfs_rpcsvc_program_unregister (rpcsvc_t *svc, rpcsvc_program_t prog)
{
        int                     ret = -1;

        if (!svc)
                return -1;

        /* TODO: De-init the listening connection for this program. */
        ret = nfs_rpcsvc_program_unregister_portmap (&prog);
        if (ret == -1) {
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "portmap unregistration of"
                        " program failed");
                goto err;
        }

        ret = 0;
        gf_log (GF_RPCSVC, GF_LOG_DEBUG, "Program unregistered: %s, Num: %d,"
                " Ver: %d, Port: %d", prog.progname, prog.prognum,
                prog.progver, prog.progport);

err:
        if (ret == -1)
                gf_log (GF_RPCSVC, GF_LOG_ERROR, "Program unregistration failed"
                        ": %s, Num: %d, Ver: %d, Port: %d", prog.progname,
                        prog.prognum, prog.progver, prog.progport);

        return ret;
}


int
nfs_rpcsvc_conn_peername (rpcsvc_conn_t *conn, char *hostname, int hostlen)
{
        if (!conn)
                return -1;

        return nfs_rpcsvc_socket_peername (conn->sockfd, hostname, hostlen);
}


int
nfs_rpcsvc_conn_peeraddr (rpcsvc_conn_t *conn, char *addrstr, int addrlen,
                          struct sockaddr *sa, socklen_t sasize)
{
        if (!conn)
                return -1;

        return nfs_rpcsvc_socket_peeraddr (conn->sockfd, addrstr, addrlen, sa,
                                           sasize);
}