summaryrefslogtreecommitdiff
path: root/drivers/net/ethernet/sfc/tc.c
blob: 834f000ba1c49e6b192a2d34068e1d299e5ff6a8 (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
// SPDX-License-Identifier: GPL-2.0-only
/****************************************************************************
 * Driver for Solarflare network controllers and boards
 * Copyright 2019 Solarflare Communications Inc.
 * Copyright 2020-2022 Xilinx Inc.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License version 2 as published
 * by the Free Software Foundation, incorporated herein by reference.
 */

#include <net/pkt_cls.h>
#include <net/vxlan.h>
#include <net/geneve.h>
#include <net/tc_act/tc_ct.h>
#include "tc.h"
#include "tc_bindings.h"
#include "tc_encap_actions.h"
#include "tc_conntrack.h"
#include "mae.h"
#include "ef100_rep.h"
#include "efx.h"

enum efx_encap_type efx_tc_indr_netdev_type(struct net_device *net_dev)
{
	if (netif_is_vxlan(net_dev))
		return EFX_ENCAP_TYPE_VXLAN;
	if (netif_is_geneve(net_dev))
		return EFX_ENCAP_TYPE_GENEVE;

	return EFX_ENCAP_TYPE_NONE;
}

#define EFX_TC_HDR_TYPE_TTL_MASK ((u32)0xff)
/* Hoplimit is stored in the most significant byte in the pedit ipv6 header action */
#define EFX_TC_HDR_TYPE_HLIMIT_MASK ~((u32)0xff000000)
#define EFX_EFV_PF	NULL
/* Look up the representor information (efv) for a device.
 * May return NULL for the PF (us), or an error pointer for a device that
 * isn't supported as a TC offload endpoint
 */
struct efx_rep *efx_tc_flower_lookup_efv(struct efx_nic *efx,
					 struct net_device *dev)
{
	struct efx_rep *efv;

	if (!dev)
		return ERR_PTR(-EOPNOTSUPP);
	/* Is it us (the PF)? */
	if (dev == efx->net_dev)
		return EFX_EFV_PF;
	/* Is it an efx vfrep at all? */
	if (dev->netdev_ops != &efx_ef100_rep_netdev_ops)
		return ERR_PTR(-EOPNOTSUPP);
	/* Is it ours?  We don't support TC rules that include another
	 * EF100's netdevices (not even on another port of the same NIC).
	 */
	efv = netdev_priv(dev);
	if (efv->parent != efx)
		return ERR_PTR(-EOPNOTSUPP);
	return efv;
}

/* Convert a driver-internal vport ID into an internal device (PF or VF) */
static s64 efx_tc_flower_internal_mport(struct efx_nic *efx, struct efx_rep *efv)
{
	u32 mport;

	if (IS_ERR(efv))
		return PTR_ERR(efv);
	if (!efv) /* device is PF (us) */
		efx_mae_mport_uplink(efx, &mport);
	else /* device is repr */
		efx_mae_mport_mport(efx, efv->mport, &mport);
	return mport;
}

/* Convert a driver-internal vport ID into an external device (wire or VF) */
s64 efx_tc_flower_external_mport(struct efx_nic *efx, struct efx_rep *efv)
{
	u32 mport;

	if (IS_ERR(efv))
		return PTR_ERR(efv);
	if (!efv) /* device is PF (us) */
		efx_mae_mport_wire(efx, &mport);
	else /* device is repr */
		efx_mae_mport_mport(efx, efv->mport, &mport);
	return mport;
}

static const struct rhashtable_params efx_tc_mac_ht_params = {
	.key_len	= offsetofend(struct efx_tc_mac_pedit_action, h_addr),
	.key_offset	= 0,
	.head_offset	= offsetof(struct efx_tc_mac_pedit_action, linkage),
};

static const struct rhashtable_params efx_tc_encap_match_ht_params = {
	.key_len	= offsetof(struct efx_tc_encap_match, linkage),
	.key_offset	= 0,
	.head_offset	= offsetof(struct efx_tc_encap_match, linkage),
};

static const struct rhashtable_params efx_tc_match_action_ht_params = {
	.key_len	= sizeof(unsigned long),
	.key_offset	= offsetof(struct efx_tc_flow_rule, cookie),
	.head_offset	= offsetof(struct efx_tc_flow_rule, linkage),
};

static const struct rhashtable_params efx_tc_lhs_rule_ht_params = {
	.key_len	= sizeof(unsigned long),
	.key_offset	= offsetof(struct efx_tc_lhs_rule, cookie),
	.head_offset	= offsetof(struct efx_tc_lhs_rule, linkage),
};

static const struct rhashtable_params efx_tc_recirc_ht_params = {
	.key_len	= offsetof(struct efx_tc_recirc_id, linkage),
	.key_offset	= 0,
	.head_offset	= offsetof(struct efx_tc_recirc_id, linkage),
};

static struct efx_tc_mac_pedit_action *efx_tc_flower_get_mac(struct efx_nic *efx,
							     unsigned char h_addr[ETH_ALEN],
							     struct netlink_ext_ack *extack)
{
	struct efx_tc_mac_pedit_action *ped, *old;
	int rc;

	ped = kzalloc(sizeof(*ped), GFP_USER);
	if (!ped)
		return ERR_PTR(-ENOMEM);
	memcpy(ped->h_addr, h_addr, ETH_ALEN);
	old = rhashtable_lookup_get_insert_fast(&efx->tc->mac_ht,
						&ped->linkage,
						efx_tc_mac_ht_params);
	if (old) {
		/* don't need our new entry */
		kfree(ped);
		if (IS_ERR(old)) /* oh dear, it's actually an error */
			return ERR_CAST(old);
		if (!refcount_inc_not_zero(&old->ref))
			return ERR_PTR(-EAGAIN);
		/* existing entry found, ref taken */
		return old;
	}

	rc = efx_mae_allocate_pedit_mac(efx, ped);
	if (rc < 0) {
		NL_SET_ERR_MSG_MOD(extack, "Failed to store pedit MAC address in hw");
		goto out_remove;
	}

	/* ref and return */
	refcount_set(&ped->ref, 1);
	return ped;
out_remove:
	rhashtable_remove_fast(&efx->tc->mac_ht, &ped->linkage,
			       efx_tc_mac_ht_params);
	kfree(ped);
	return ERR_PTR(rc);
}

static void efx_tc_flower_put_mac(struct efx_nic *efx,
				  struct efx_tc_mac_pedit_action *ped)
{
	if (!refcount_dec_and_test(&ped->ref))
		return; /* still in use */
	rhashtable_remove_fast(&efx->tc->mac_ht, &ped->linkage,
			       efx_tc_mac_ht_params);
	efx_mae_free_pedit_mac(efx, ped);
	kfree(ped);
}

static void efx_tc_free_action_set(struct efx_nic *efx,
				   struct efx_tc_action_set *act, bool in_hw)
{
	/* Failure paths calling this on the 'cursor' action set in_hw=false,
	 * because if the alloc had succeeded we'd've put it in acts.list and
	 * not still have it in act.
	 */
	if (in_hw) {
		efx_mae_free_action_set(efx, act->fw_id);
		/* in_hw is true iff we are on an acts.list; make sure to
		 * remove ourselves from that list before we are freed.
		 */
		list_del(&act->list);
	}
	if (act->count) {
		spin_lock_bh(&act->count->cnt->lock);
		if (!list_empty(&act->count_user))
			list_del(&act->count_user);
		spin_unlock_bh(&act->count->cnt->lock);
		efx_tc_flower_put_counter_index(efx, act->count);
	}
	if (act->encap_md) {
		list_del(&act->encap_user);
		efx_tc_flower_release_encap_md(efx, act->encap_md);
	}
	if (act->src_mac)
		efx_tc_flower_put_mac(efx, act->src_mac);
	if (act->dst_mac)
		efx_tc_flower_put_mac(efx, act->dst_mac);
	kfree(act);
}

static void efx_tc_free_action_set_list(struct efx_nic *efx,
					struct efx_tc_action_set_list *acts,
					bool in_hw)
{
	struct efx_tc_action_set *act, *next;

	/* Failure paths set in_hw=false, because usually the acts didn't get
	 * to efx_mae_alloc_action_set_list(); if they did, the failure tree
	 * has a separate efx_mae_free_action_set_list() before calling us.
	 */
	if (in_hw)
		efx_mae_free_action_set_list(efx, acts);
	/* Any act that's on the list will be in_hw even if the list isn't */
	list_for_each_entry_safe(act, next, &acts->list, list)
		efx_tc_free_action_set(efx, act, true);
	/* Don't kfree, as acts is embedded inside a struct efx_tc_flow_rule */
}

/* Boilerplate for the simple 'copy a field' cases */
#define _MAP_KEY_AND_MASK(_name, _type, _tcget, _tcfield, _field)	\
if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_##_name)) {		\
	struct flow_match_##_type fm;					\
									\
	flow_rule_match_##_tcget(rule, &fm);				\
	match->value._field = fm.key->_tcfield;				\
	match->mask._field = fm.mask->_tcfield;				\
}
#define MAP_KEY_AND_MASK(_name, _type, _tcfield, _field)	\
	_MAP_KEY_AND_MASK(_name, _type, _type, _tcfield, _field)
#define MAP_ENC_KEY_AND_MASK(_name, _type, _tcget, _tcfield, _field)	\
	_MAP_KEY_AND_MASK(ENC_##_name, _type, _tcget, _tcfield, _field)

static int efx_tc_flower_parse_match(struct efx_nic *efx,
				     struct flow_rule *rule,
				     struct efx_tc_match *match,
				     struct netlink_ext_ack *extack)
{
	struct flow_dissector *dissector = rule->match.dissector;
	unsigned char ipv = 0;

	/* Owing to internal TC infelicities, the IPV6_ADDRS key might be set
	 * even on IPv4 filters; so rather than relying on dissector->used_keys
	 * we check the addr_type in the CONTROL key.  If we don't find it (or
	 * it's masked, which should never happen), we treat both IPV4_ADDRS
	 * and IPV6_ADDRS as absent.
	 */
	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_CONTROL)) {
		struct flow_match_control fm;

		flow_rule_match_control(rule, &fm);
		if (IS_ALL_ONES(fm.mask->addr_type))
			switch (fm.key->addr_type) {
			case FLOW_DISSECTOR_KEY_IPV4_ADDRS:
				ipv = 4;
				break;
			case FLOW_DISSECTOR_KEY_IPV6_ADDRS:
				ipv = 6;
				break;
			default:
				break;
			}

		if (fm.mask->flags & FLOW_DIS_IS_FRAGMENT) {
			match->value.ip_frag = fm.key->flags & FLOW_DIS_IS_FRAGMENT;
			match->mask.ip_frag = true;
		}
		if (fm.mask->flags & FLOW_DIS_FIRST_FRAG) {
			match->value.ip_firstfrag = fm.key->flags & FLOW_DIS_FIRST_FRAG;
			match->mask.ip_firstfrag = true;
		}
		if (fm.mask->flags & ~(FLOW_DIS_IS_FRAGMENT | FLOW_DIS_FIRST_FRAG)) {
			NL_SET_ERR_MSG_FMT_MOD(extack, "Unsupported match on control.flags %#x",
					       fm.mask->flags);
			return -EOPNOTSUPP;
		}
	}
	if (dissector->used_keys &
	    ~(BIT_ULL(FLOW_DISSECTOR_KEY_CONTROL) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_BASIC) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_ETH_ADDRS) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_VLAN) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_CVLAN) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_IPV4_ADDRS) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_IPV6_ADDRS) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_PORTS) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_ENC_KEYID) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_ENC_IP) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_ENC_PORTS) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_ENC_CONTROL) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_CT) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_TCP) |
	      BIT_ULL(FLOW_DISSECTOR_KEY_IP))) {
		NL_SET_ERR_MSG_FMT_MOD(extack, "Unsupported flower keys %#llx",
				       dissector->used_keys);
		return -EOPNOTSUPP;
	}

	MAP_KEY_AND_MASK(BASIC, basic, n_proto, eth_proto);
	/* Make sure we're IP if any L3/L4 keys used. */
	if (!IS_ALL_ONES(match->mask.eth_proto) ||
	    !(match->value.eth_proto == htons(ETH_P_IP) ||
	      match->value.eth_proto == htons(ETH_P_IPV6)))
		if (dissector->used_keys &
		    (BIT_ULL(FLOW_DISSECTOR_KEY_IPV4_ADDRS) |
		     BIT_ULL(FLOW_DISSECTOR_KEY_IPV6_ADDRS) |
		     BIT_ULL(FLOW_DISSECTOR_KEY_PORTS) |
		     BIT_ULL(FLOW_DISSECTOR_KEY_IP) |
		     BIT_ULL(FLOW_DISSECTOR_KEY_TCP))) {
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "L3/L4 flower keys %#llx require protocol ipv[46]",
					       dissector->used_keys);
			return -EINVAL;
		}

	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_VLAN)) {
		struct flow_match_vlan fm;

		flow_rule_match_vlan(rule, &fm);
		if (fm.mask->vlan_id || fm.mask->vlan_priority || fm.mask->vlan_tpid) {
			match->value.vlan_proto[0] = fm.key->vlan_tpid;
			match->mask.vlan_proto[0] = fm.mask->vlan_tpid;
			match->value.vlan_tci[0] = cpu_to_be16(fm.key->vlan_priority << 13 |
							       fm.key->vlan_id);
			match->mask.vlan_tci[0] = cpu_to_be16(fm.mask->vlan_priority << 13 |
							      fm.mask->vlan_id);
		}
	}

	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_CVLAN)) {
		struct flow_match_vlan fm;

		flow_rule_match_cvlan(rule, &fm);
		if (fm.mask->vlan_id || fm.mask->vlan_priority || fm.mask->vlan_tpid) {
			match->value.vlan_proto[1] = fm.key->vlan_tpid;
			match->mask.vlan_proto[1] = fm.mask->vlan_tpid;
			match->value.vlan_tci[1] = cpu_to_be16(fm.key->vlan_priority << 13 |
							       fm.key->vlan_id);
			match->mask.vlan_tci[1] = cpu_to_be16(fm.mask->vlan_priority << 13 |
							      fm.mask->vlan_id);
		}
	}

	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ETH_ADDRS)) {
		struct flow_match_eth_addrs fm;

		flow_rule_match_eth_addrs(rule, &fm);
		ether_addr_copy(match->value.eth_saddr, fm.key->src);
		ether_addr_copy(match->value.eth_daddr, fm.key->dst);
		ether_addr_copy(match->mask.eth_saddr, fm.mask->src);
		ether_addr_copy(match->mask.eth_daddr, fm.mask->dst);
	}

	MAP_KEY_AND_MASK(BASIC, basic, ip_proto, ip_proto);
	/* Make sure we're TCP/UDP if any L4 keys used. */
	if ((match->value.ip_proto != IPPROTO_UDP &&
	     match->value.ip_proto != IPPROTO_TCP) || !IS_ALL_ONES(match->mask.ip_proto))
		if (dissector->used_keys &
		    (BIT_ULL(FLOW_DISSECTOR_KEY_PORTS) |
		     BIT_ULL(FLOW_DISSECTOR_KEY_TCP))) {
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "L4 flower keys %#llx require ipproto udp or tcp",
					       dissector->used_keys);
			return -EINVAL;
		}
	MAP_KEY_AND_MASK(IP, ip, tos, ip_tos);
	MAP_KEY_AND_MASK(IP, ip, ttl, ip_ttl);
	if (ipv == 4) {
		MAP_KEY_AND_MASK(IPV4_ADDRS, ipv4_addrs, src, src_ip);
		MAP_KEY_AND_MASK(IPV4_ADDRS, ipv4_addrs, dst, dst_ip);
	}
#ifdef CONFIG_IPV6
	else if (ipv == 6) {
		MAP_KEY_AND_MASK(IPV6_ADDRS, ipv6_addrs, src, src_ip6);
		MAP_KEY_AND_MASK(IPV6_ADDRS, ipv6_addrs, dst, dst_ip6);
	}
#endif
	MAP_KEY_AND_MASK(PORTS, ports, src, l4_sport);
	MAP_KEY_AND_MASK(PORTS, ports, dst, l4_dport);
	MAP_KEY_AND_MASK(TCP, tcp, flags, tcp_flags);
	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL)) {
		struct flow_match_control fm;

		flow_rule_match_enc_control(rule, &fm);
		if (fm.mask->flags) {
			NL_SET_ERR_MSG_FMT_MOD(extack, "Unsupported match on enc_control.flags %#x",
					       fm.mask->flags);
			return -EOPNOTSUPP;
		}
		if (!IS_ALL_ONES(fm.mask->addr_type)) {
			NL_SET_ERR_MSG_FMT_MOD(extack, "Unsupported enc addr_type mask %u (key %u)",
					       fm.mask->addr_type,
					       fm.key->addr_type);
			return -EOPNOTSUPP;
		}
		switch (fm.key->addr_type) {
		case FLOW_DISSECTOR_KEY_IPV4_ADDRS:
			MAP_ENC_KEY_AND_MASK(IPV4_ADDRS, ipv4_addrs, enc_ipv4_addrs,
					     src, enc_src_ip);
			MAP_ENC_KEY_AND_MASK(IPV4_ADDRS, ipv4_addrs, enc_ipv4_addrs,
					     dst, enc_dst_ip);
			break;
#ifdef CONFIG_IPV6
		case FLOW_DISSECTOR_KEY_IPV6_ADDRS:
			MAP_ENC_KEY_AND_MASK(IPV6_ADDRS, ipv6_addrs, enc_ipv6_addrs,
					     src, enc_src_ip6);
			MAP_ENC_KEY_AND_MASK(IPV6_ADDRS, ipv6_addrs, enc_ipv6_addrs,
					     dst, enc_dst_ip6);
			break;
#endif
		default:
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "Unsupported enc addr_type %u (supported are IPv4, IPv6)",
					       fm.key->addr_type);
			return -EOPNOTSUPP;
		}
		MAP_ENC_KEY_AND_MASK(IP, ip, enc_ip, tos, enc_ip_tos);
		MAP_ENC_KEY_AND_MASK(IP, ip, enc_ip, ttl, enc_ip_ttl);
		MAP_ENC_KEY_AND_MASK(PORTS, ports, enc_ports, src, enc_sport);
		MAP_ENC_KEY_AND_MASK(PORTS, ports, enc_ports, dst, enc_dport);
		MAP_ENC_KEY_AND_MASK(KEYID, enc_keyid, enc_keyid, keyid, enc_keyid);
	} else if (dissector->used_keys &
		   (BIT_ULL(FLOW_DISSECTOR_KEY_ENC_KEYID) |
		    BIT_ULL(FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS) |
		    BIT_ULL(FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS) |
		    BIT_ULL(FLOW_DISSECTOR_KEY_ENC_IP) |
		    BIT_ULL(FLOW_DISSECTOR_KEY_ENC_PORTS))) {
		NL_SET_ERR_MSG_FMT_MOD(extack,
				       "Flower enc keys require enc_control (keys: %#llx)",
				       dissector->used_keys);
		return -EOPNOTSUPP;
	}
	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_CT)) {
		struct flow_match_ct fm;

		flow_rule_match_ct(rule, &fm);
		match->value.ct_state_trk = !!(fm.key->ct_state & TCA_FLOWER_KEY_CT_FLAGS_TRACKED);
		match->mask.ct_state_trk = !!(fm.mask->ct_state & TCA_FLOWER_KEY_CT_FLAGS_TRACKED);
		match->value.ct_state_est = !!(fm.key->ct_state & TCA_FLOWER_KEY_CT_FLAGS_ESTABLISHED);
		match->mask.ct_state_est = !!(fm.mask->ct_state & TCA_FLOWER_KEY_CT_FLAGS_ESTABLISHED);
		if (fm.mask->ct_state & ~(TCA_FLOWER_KEY_CT_FLAGS_TRACKED |
					  TCA_FLOWER_KEY_CT_FLAGS_ESTABLISHED)) {
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "Unsupported ct_state match %#x",
					       fm.mask->ct_state);
			return -EOPNOTSUPP;
		}
		match->value.ct_mark = fm.key->ct_mark;
		match->mask.ct_mark = fm.mask->ct_mark;
		match->value.ct_zone = fm.key->ct_zone;
		match->mask.ct_zone = fm.mask->ct_zone;

		if (memchr_inv(fm.mask->ct_labels, 0, sizeof(fm.mask->ct_labels))) {
			NL_SET_ERR_MSG_MOD(extack, "Matching on ct_label not supported");
			return -EOPNOTSUPP;
		}
	}

	return 0;
}

static void efx_tc_flower_release_encap_match(struct efx_nic *efx,
					      struct efx_tc_encap_match *encap)
{
	int rc;

	if (!refcount_dec_and_test(&encap->ref))
		return; /* still in use */

	if (encap->type == EFX_TC_EM_DIRECT) {
		rc = efx_mae_unregister_encap_match(efx, encap);
		if (rc)
			/* Display message but carry on and remove entry from our
			 * SW tables, because there's not much we can do about it.
			 */
			netif_err(efx, drv, efx->net_dev,
				  "Failed to release encap match %#x, rc %d\n",
				  encap->fw_id, rc);
	}
	rhashtable_remove_fast(&efx->tc->encap_match_ht, &encap->linkage,
			       efx_tc_encap_match_ht_params);
	if (encap->pseudo)
		efx_tc_flower_release_encap_match(efx, encap->pseudo);
	kfree(encap);
}

static int efx_tc_flower_record_encap_match(struct efx_nic *efx,
					    struct efx_tc_match *match,
					    enum efx_encap_type type,
					    enum efx_tc_em_pseudo_type em_type,
					    u8 child_ip_tos_mask,
					    __be16 child_udp_sport_mask,
					    struct netlink_ext_ack *extack)
{
	struct efx_tc_encap_match *encap, *old, *pseudo = NULL;
	bool ipv6 = false;
	int rc;

	/* We require that the socket-defining fields (IP addrs and UDP dest
	 * port) are present and exact-match.  Other fields may only be used
	 * if the field-set (and any masks) are the same for all encap
	 * matches on the same <sip,dip,dport> tuple; this is enforced by
	 * pseudo encap matches.
	 */
	if (match->mask.enc_dst_ip | match->mask.enc_src_ip) {
		if (!IS_ALL_ONES(match->mask.enc_dst_ip)) {
			NL_SET_ERR_MSG_MOD(extack,
					   "Egress encap match is not exact on dst IP address");
			return -EOPNOTSUPP;
		}
		if (!IS_ALL_ONES(match->mask.enc_src_ip)) {
			NL_SET_ERR_MSG_MOD(extack,
					   "Egress encap match is not exact on src IP address");
			return -EOPNOTSUPP;
		}
#ifdef CONFIG_IPV6
		if (!ipv6_addr_any(&match->mask.enc_dst_ip6) ||
		    !ipv6_addr_any(&match->mask.enc_src_ip6)) {
			NL_SET_ERR_MSG_MOD(extack,
					   "Egress encap match on both IPv4 and IPv6, don't understand");
			return -EOPNOTSUPP;
		}
	} else {
		ipv6 = true;
		if (!efx_ipv6_addr_all_ones(&match->mask.enc_dst_ip6)) {
			NL_SET_ERR_MSG_MOD(extack,
					   "Egress encap match is not exact on dst IP address");
			return -EOPNOTSUPP;
		}
		if (!efx_ipv6_addr_all_ones(&match->mask.enc_src_ip6)) {
			NL_SET_ERR_MSG_MOD(extack,
					   "Egress encap match is not exact on src IP address");
			return -EOPNOTSUPP;
		}
#endif
	}
	if (!IS_ALL_ONES(match->mask.enc_dport)) {
		NL_SET_ERR_MSG_MOD(extack, "Egress encap match is not exact on dst UDP port");
		return -EOPNOTSUPP;
	}
	if (match->mask.enc_sport || match->mask.enc_ip_tos) {
		struct efx_tc_match pmatch = *match;

		if (em_type == EFX_TC_EM_PSEUDO_MASK) { /* can't happen */
			NL_SET_ERR_MSG_MOD(extack, "Bad recursion in egress encap match handler");
			return -EOPNOTSUPP;
		}
		pmatch.value.enc_ip_tos = 0;
		pmatch.mask.enc_ip_tos = 0;
		pmatch.value.enc_sport = 0;
		pmatch.mask.enc_sport = 0;
		rc = efx_tc_flower_record_encap_match(efx, &pmatch, type,
						      EFX_TC_EM_PSEUDO_MASK,
						      match->mask.enc_ip_tos,
						      match->mask.enc_sport,
						      extack);
		if (rc)
			return rc;
		pseudo = pmatch.encap;
	}
	if (match->mask.enc_ip_ttl) {
		NL_SET_ERR_MSG_MOD(extack, "Egress encap match on IP TTL not supported");
		rc = -EOPNOTSUPP;
		goto fail_pseudo;
	}

	rc = efx_mae_check_encap_match_caps(efx, ipv6, match->mask.enc_ip_tos,
					    match->mask.enc_sport, extack);
	if (rc)
		goto fail_pseudo;

	encap = kzalloc(sizeof(*encap), GFP_USER);
	if (!encap) {
		rc = -ENOMEM;
		goto fail_pseudo;
	}
	encap->src_ip = match->value.enc_src_ip;
	encap->dst_ip = match->value.enc_dst_ip;
#ifdef CONFIG_IPV6
	encap->src_ip6 = match->value.enc_src_ip6;
	encap->dst_ip6 = match->value.enc_dst_ip6;
#endif
	encap->udp_dport = match->value.enc_dport;
	encap->tun_type = type;
	encap->ip_tos = match->value.enc_ip_tos;
	encap->ip_tos_mask = match->mask.enc_ip_tos;
	encap->child_ip_tos_mask = child_ip_tos_mask;
	encap->udp_sport = match->value.enc_sport;
	encap->udp_sport_mask = match->mask.enc_sport;
	encap->child_udp_sport_mask = child_udp_sport_mask;
	encap->type = em_type;
	encap->pseudo = pseudo;
	old = rhashtable_lookup_get_insert_fast(&efx->tc->encap_match_ht,
						&encap->linkage,
						efx_tc_encap_match_ht_params);
	if (old) {
		/* don't need our new entry */
		kfree(encap);
		if (pseudo) /* don't need our new pseudo either */
			efx_tc_flower_release_encap_match(efx, pseudo);
		if (IS_ERR(old)) /* oh dear, it's actually an error */
			return PTR_ERR(old);
		/* check old and new em_types are compatible */
		switch (old->type) {
		case EFX_TC_EM_DIRECT:
			/* old EM is in hardware, so mustn't overlap with a
			 * pseudo, but may be shared with another direct EM
			 */
			if (em_type == EFX_TC_EM_DIRECT)
				break;
			NL_SET_ERR_MSG_MOD(extack, "Pseudo encap match conflicts with existing direct entry");
			return -EEXIST;
		case EFX_TC_EM_PSEUDO_MASK:
			/* old EM is protecting a ToS- or src port-qualified
			 * filter, so may only be shared with another pseudo
			 * for the same ToS and src port masks.
			 */
			if (em_type != EFX_TC_EM_PSEUDO_MASK) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "%s encap match conflicts with existing pseudo(MASK) entry",
						       em_type ? "Pseudo" : "Direct");
				return -EEXIST;
			}
			if (child_ip_tos_mask != old->child_ip_tos_mask) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Pseudo encap match for TOS mask %#04x conflicts with existing pseudo(MASK) entry for TOS mask %#04x",
						       child_ip_tos_mask,
						       old->child_ip_tos_mask);
				return -EEXIST;
			}
			if (child_udp_sport_mask != old->child_udp_sport_mask) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Pseudo encap match for UDP src port mask %#x conflicts with existing pseudo(MASK) entry for mask %#x",
						       child_udp_sport_mask,
						       old->child_udp_sport_mask);
				return -EEXIST;
			}
			break;
		default: /* Unrecognised pseudo-type.  Just say no */
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "%s encap match conflicts with existing pseudo(%d) entry",
					       em_type ? "Pseudo" : "Direct",
					       old->type);
			return -EEXIST;
		}
		/* check old and new tun_types are compatible */
		if (old->tun_type != type) {
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "Egress encap match with conflicting tun_type %u != %u",
					       old->tun_type, type);
			return -EEXIST;
		}
		if (!refcount_inc_not_zero(&old->ref))
			return -EAGAIN;
		/* existing entry found */
		encap = old;
	} else {
		if (em_type == EFX_TC_EM_DIRECT) {
			rc = efx_mae_register_encap_match(efx, encap);
			if (rc) {
				NL_SET_ERR_MSG_MOD(extack, "Failed to record egress encap match in HW");
				goto fail;
			}
		}
		refcount_set(&encap->ref, 1);
	}
	match->encap = encap;
	return 0;
fail:
	rhashtable_remove_fast(&efx->tc->encap_match_ht, &encap->linkage,
			       efx_tc_encap_match_ht_params);
	kfree(encap);
fail_pseudo:
	if (pseudo)
		efx_tc_flower_release_encap_match(efx, pseudo);
	return rc;
}

static struct efx_tc_recirc_id *efx_tc_get_recirc_id(struct efx_nic *efx,
						     u32 chain_index,
						     struct net_device *net_dev)
{
	struct efx_tc_recirc_id *rid, *old;
	int rc;

	rid = kzalloc(sizeof(*rid), GFP_USER);
	if (!rid)
		return ERR_PTR(-ENOMEM);
	rid->chain_index = chain_index;
	/* We don't take a reference here, because it's implied - if there's
	 * a rule on the net_dev that's been offloaded to us, then the net_dev
	 * can't go away until the rule has been deoffloaded.
	 */
	rid->net_dev = net_dev;
	old = rhashtable_lookup_get_insert_fast(&efx->tc->recirc_ht,
						&rid->linkage,
						efx_tc_recirc_ht_params);
	if (old) {
		/* don't need our new entry */
		kfree(rid);
		if (IS_ERR(old)) /* oh dear, it's actually an error */
			return ERR_CAST(old);
		if (!refcount_inc_not_zero(&old->ref))
			return ERR_PTR(-EAGAIN);
		/* existing entry found */
		rid = old;
	} else {
		rc = ida_alloc_range(&efx->tc->recirc_ida, 1, U8_MAX, GFP_USER);
		if (rc < 0) {
			rhashtable_remove_fast(&efx->tc->recirc_ht,
					       &rid->linkage,
					       efx_tc_recirc_ht_params);
			kfree(rid);
			return ERR_PTR(rc);
		}
		rid->fw_id = rc;
		refcount_set(&rid->ref, 1);
	}
	return rid;
}

static void efx_tc_put_recirc_id(struct efx_nic *efx, struct efx_tc_recirc_id *rid)
{
	if (!refcount_dec_and_test(&rid->ref))
		return; /* still in use */
	rhashtable_remove_fast(&efx->tc->recirc_ht, &rid->linkage,
			       efx_tc_recirc_ht_params);
	ida_free(&efx->tc->recirc_ida, rid->fw_id);
	kfree(rid);
}

static void efx_tc_delete_rule(struct efx_nic *efx, struct efx_tc_flow_rule *rule)
{
	efx_mae_delete_rule(efx, rule->fw_id);

	/* Release entries in subsidiary tables */
	efx_tc_free_action_set_list(efx, &rule->acts, true);
	if (rule->match.rid)
		efx_tc_put_recirc_id(efx, rule->match.rid);
	if (rule->match.encap)
		efx_tc_flower_release_encap_match(efx, rule->match.encap);
	rule->fw_id = MC_CMD_MAE_ACTION_RULE_INSERT_OUT_ACTION_RULE_ID_NULL;
}

static const char *efx_tc_encap_type_name(enum efx_encap_type typ)
{
	switch (typ) {
	case EFX_ENCAP_TYPE_NONE:
		return "none";
	case EFX_ENCAP_TYPE_VXLAN:
		return "vxlan";
	case EFX_ENCAP_TYPE_GENEVE:
		return "geneve";
	default:
		pr_warn_once("Unknown efx_encap_type %d encountered\n", typ);
		return "unknown";
	}
}

/* For details of action order constraints refer to SF-123102-TC-1§12.6.1 */
enum efx_tc_action_order {
	EFX_TC_AO_DECAP,
	EFX_TC_AO_DEC_TTL,
	EFX_TC_AO_PEDIT_MAC_ADDRS,
	EFX_TC_AO_VLAN_POP,
	EFX_TC_AO_VLAN_PUSH,
	EFX_TC_AO_COUNT,
	EFX_TC_AO_ENCAP,
	EFX_TC_AO_DELIVER
};
/* Determine whether we can add @new action without violating order */
static bool efx_tc_flower_action_order_ok(const struct efx_tc_action_set *act,
					  enum efx_tc_action_order new)
{
	switch (new) {
	case EFX_TC_AO_DECAP:
		if (act->decap)
			return false;
		/* PEDIT_MAC_ADDRS must not happen before DECAP, though it
		 * can wait until much later
		 */
		if (act->dst_mac || act->src_mac)
			return false;

		/* Decrementing ttl must not happen before DECAP */
		if (act->do_ttl_dec)
			return false;
		fallthrough;
	case EFX_TC_AO_VLAN_POP:
		if (act->vlan_pop >= 2)
			return false;
		/* If we've already pushed a VLAN, we can't then pop it;
		 * the hardware would instead try to pop an existing VLAN
		 * before pushing the new one.
		 */
		if (act->vlan_push)
			return false;
		fallthrough;
	case EFX_TC_AO_VLAN_PUSH:
		if (act->vlan_push >= 2)
			return false;
		fallthrough;
	case EFX_TC_AO_COUNT:
		if (act->count)
			return false;
		fallthrough;
	case EFX_TC_AO_PEDIT_MAC_ADDRS:
	case EFX_TC_AO_ENCAP:
		if (act->encap_md)
			return false;
		fallthrough;
	case EFX_TC_AO_DELIVER:
		return !act->deliver;
	case EFX_TC_AO_DEC_TTL:
		if (act->encap_md)
			return false;
		return !act->do_ttl_dec;
	default:
		/* Bad caller.  Whatever they wanted to do, say they can't. */
		WARN_ON_ONCE(1);
		return false;
	}
}

/**
 * DOC: TC conntrack sequences
 *
 * The MAE hardware can handle at most two rounds of action rule matching,
 * consequently we support conntrack through the notion of a "left-hand side
 * rule".  This is a rule which typically contains only the actions "ct" and
 * "goto chain N", and corresponds to one or more "right-hand side rules" in
 * chain N, which typically match on +trk+est, and may perform ct(nat) actions.
 * RHS rules go in the Action Rule table as normal but with a nonzero recirc_id
 * (the hardware equivalent of chain_index), while LHS rules may go in either
 * the Action Rule or the Outer Rule table, the latter being preferred for
 * performance reasons, and set both DO_CT and a recirc_id in their response.
 *
 * Besides the RHS rules, there are often also similar rules matching on
 * +trk+new which perform the ct(commit) action.  These are not offloaded.
 */

static bool efx_tc_rule_is_lhs_rule(struct flow_rule *fr,
				    struct efx_tc_match *match)
{
	const struct flow_action_entry *fa;
	int i;

	flow_action_for_each(i, fa, &fr->action) {
		switch (fa->id) {
		case FLOW_ACTION_GOTO:
			return true;
		case FLOW_ACTION_CT:
			/* If rule is -trk, or doesn't mention trk at all, then
			 * a CT action implies a conntrack lookup (hence it's an
			 * LHS rule).  If rule is +trk, then a CT action could
			 * just be ct(nat) or even ct(commit) (though the latter
			 * can't be offloaded).
			 */
			if (!match->mask.ct_state_trk || !match->value.ct_state_trk)
				return true;
			break;
		default:
			break;
		}
	}
	return false;
}

static int efx_tc_flower_handle_lhs_actions(struct efx_nic *efx,
					    struct flow_cls_offload *tc,
					    struct flow_rule *fr,
					    struct net_device *net_dev,
					    struct efx_tc_lhs_rule *rule)

{
	struct netlink_ext_ack *extack = tc->common.extack;
	struct efx_tc_lhs_action *act = &rule->lhs_act;
	const struct flow_action_entry *fa;
	bool pipe = true;
	int i;

	flow_action_for_each(i, fa, &fr->action) {
		struct efx_tc_ct_zone *ct_zone;
		struct efx_tc_recirc_id *rid;

		if (!pipe) {
			/* more actions after a non-pipe action */
			NL_SET_ERR_MSG_MOD(extack, "Action follows non-pipe action");
			return -EINVAL;
		}
		switch (fa->id) {
		case FLOW_ACTION_GOTO:
			if (!fa->chain_index) {
				NL_SET_ERR_MSG_MOD(extack, "Can't goto chain 0, no looping in hw");
				return -EOPNOTSUPP;
			}
			rid = efx_tc_get_recirc_id(efx, fa->chain_index,
						   net_dev);
			if (IS_ERR(rid)) {
				NL_SET_ERR_MSG_MOD(extack, "Failed to allocate a hardware recirculation ID for this chain_index");
				return PTR_ERR(rid);
			}
			act->rid = rid;
			if (fa->hw_stats) {
				struct efx_tc_counter_index *cnt;

				if (!(fa->hw_stats & FLOW_ACTION_HW_STATS_DELAYED)) {
					NL_SET_ERR_MSG_FMT_MOD(extack,
							       "hw_stats_type %u not supported (only 'delayed')",
							       fa->hw_stats);
					return -EOPNOTSUPP;
				}
				cnt = efx_tc_flower_get_counter_index(efx, tc->cookie,
								      EFX_TC_COUNTER_TYPE_OR);
				if (IS_ERR(cnt)) {
					NL_SET_ERR_MSG_MOD(extack, "Failed to obtain a counter");
					return PTR_ERR(cnt);
				}
				WARN_ON(act->count); /* can't happen */
				act->count = cnt;
			}
			pipe = false;
			break;
		case FLOW_ACTION_CT:
			if (act->zone) {
				NL_SET_ERR_MSG_MOD(extack, "Can't offload multiple ct actions");
				return -EOPNOTSUPP;
			}
			if (fa->ct.action & (TCA_CT_ACT_COMMIT |
					     TCA_CT_ACT_FORCE)) {
				NL_SET_ERR_MSG_MOD(extack, "Can't offload ct commit/force");
				return -EOPNOTSUPP;
			}
			if (fa->ct.action & TCA_CT_ACT_CLEAR) {
				NL_SET_ERR_MSG_MOD(extack, "Can't clear ct in LHS rule");
				return -EOPNOTSUPP;
			}
			if (fa->ct.action & (TCA_CT_ACT_NAT |
					     TCA_CT_ACT_NAT_SRC |
					     TCA_CT_ACT_NAT_DST)) {
				NL_SET_ERR_MSG_MOD(extack, "Can't perform NAT in LHS rule - packet isn't conntracked yet");
				return -EOPNOTSUPP;
			}
			if (fa->ct.action) {
				NL_SET_ERR_MSG_FMT_MOD(extack, "Unhandled ct.action %u for LHS rule\n",
						       fa->ct.action);
				return -EOPNOTSUPP;
			}
			ct_zone = efx_tc_ct_register_zone(efx, fa->ct.zone,
							  fa->ct.flow_table);
			if (IS_ERR(ct_zone)) {
				NL_SET_ERR_MSG_MOD(extack, "Failed to register for CT updates");
				return PTR_ERR(ct_zone);
			}
			act->zone = ct_zone;
			break;
		default:
			NL_SET_ERR_MSG_FMT_MOD(extack, "Unhandled action %u for LHS rule\n",
					       fa->id);
			return -EOPNOTSUPP;
		}
	}

	if (pipe) {
		NL_SET_ERR_MSG_MOD(extack, "Missing goto chain in LHS rule");
		return -EOPNOTSUPP;
	}
	return 0;
}

static void efx_tc_flower_release_lhs_actions(struct efx_nic *efx,
					      struct efx_tc_lhs_action *act)
{
	if (act->rid)
		efx_tc_put_recirc_id(efx, act->rid);
	if (act->zone)
		efx_tc_ct_unregister_zone(efx, act->zone);
	if (act->count)
		efx_tc_flower_put_counter_index(efx, act->count);
}

/**
 * struct efx_tc_mangler_state - accumulates 32-bit pedits into fields
 *
 * @dst_mac_32:	dst_mac[0:3] has been populated
 * @dst_mac_16:	dst_mac[4:5] has been populated
 * @src_mac_16:	src_mac[0:1] has been populated
 * @src_mac_32:	src_mac[2:5] has been populated
 * @dst_mac:	h_dest field of ethhdr
 * @src_mac:	h_source field of ethhdr
 *
 * Since FLOW_ACTION_MANGLE comes in 32-bit chunks that do not
 * necessarily equate to whole fields of the packet header, this
 * structure is used to hold the cumulative effect of the partial
 * field pedits that have been processed so far.
 */
struct efx_tc_mangler_state {
	u8 dst_mac_32:1; /* eth->h_dest[0:3] */
	u8 dst_mac_16:1; /* eth->h_dest[4:5] */
	u8 src_mac_16:1; /* eth->h_source[0:1] */
	u8 src_mac_32:1; /* eth->h_source[2:5] */
	unsigned char dst_mac[ETH_ALEN];
	unsigned char src_mac[ETH_ALEN];
};

/** efx_tc_complete_mac_mangle() - pull complete field pedits out of @mung
 * @efx:	NIC we're installing a flow rule on
 * @act:	action set (cursor) to update
 * @mung:	accumulated partial mangles
 * @extack:	netlink extended ack for reporting errors
 *
 * Check @mung to find any combinations of partial mangles that can be
 * combined into a complete packet field edit, add that edit to @act,
 * and consume the partial mangles from @mung.
 */

static int efx_tc_complete_mac_mangle(struct efx_nic *efx,
				      struct efx_tc_action_set *act,
				      struct efx_tc_mangler_state *mung,
				      struct netlink_ext_ack *extack)
{
	struct efx_tc_mac_pedit_action *ped;

	if (mung->dst_mac_32 && mung->dst_mac_16) {
		ped = efx_tc_flower_get_mac(efx, mung->dst_mac, extack);
		if (IS_ERR(ped))
			return PTR_ERR(ped);

		/* Check that we have not already populated dst_mac */
		if (act->dst_mac)
			efx_tc_flower_put_mac(efx, act->dst_mac);

		act->dst_mac = ped;

		/* consume the incomplete state */
		mung->dst_mac_32 = 0;
		mung->dst_mac_16 = 0;
	}
	if (mung->src_mac_16 && mung->src_mac_32) {
		ped = efx_tc_flower_get_mac(efx, mung->src_mac, extack);
		if (IS_ERR(ped))
			return PTR_ERR(ped);

		/* Check that we have not already populated src_mac */
		if (act->src_mac)
			efx_tc_flower_put_mac(efx, act->src_mac);

		act->src_mac = ped;

		/* consume the incomplete state */
		mung->src_mac_32 = 0;
		mung->src_mac_16 = 0;
	}
	return 0;
}

static int efx_tc_pedit_add(struct efx_nic *efx, struct efx_tc_action_set *act,
			    const struct flow_action_entry *fa,
			    struct netlink_ext_ack *extack)
{
	switch (fa->mangle.htype) {
	case FLOW_ACT_MANGLE_HDR_TYPE_IP4:
		switch (fa->mangle.offset) {
		case offsetof(struct iphdr, ttl):
			/* check that pedit applies to ttl only */
			if (fa->mangle.mask != ~EFX_TC_HDR_TYPE_TTL_MASK)
				break;

			/* Adding 0xff is equivalent to decrementing the ttl.
			 * Other added values are not supported.
			 */
			if ((fa->mangle.val & EFX_TC_HDR_TYPE_TTL_MASK) != U8_MAX)
				break;

			/* check that we do not decrement ttl twice */
			if (!efx_tc_flower_action_order_ok(act,
							   EFX_TC_AO_DEC_TTL)) {
				NL_SET_ERR_MSG_MOD(extack, "Unsupported: multiple dec ttl");
				return -EOPNOTSUPP;
			}
			act->do_ttl_dec = 1;
			return 0;
		default:
			break;
		}
		break;
	case FLOW_ACT_MANGLE_HDR_TYPE_IP6:
		switch (fa->mangle.offset) {
		case round_down(offsetof(struct ipv6hdr, hop_limit), 4):
			/* check that pedit applies to hoplimit only */
			if (fa->mangle.mask != EFX_TC_HDR_TYPE_HLIMIT_MASK)
				break;

			/* Adding 0xff is equivalent to decrementing the hoplimit.
			 * Other added values are not supported.
			 */
			if ((fa->mangle.val >> 24) != U8_MAX)
				break;

			/* check that we do not decrement hoplimit twice */
			if (!efx_tc_flower_action_order_ok(act,
							   EFX_TC_AO_DEC_TTL)) {
				NL_SET_ERR_MSG_MOD(extack, "Unsupported: multiple dec ttl");
				return -EOPNOTSUPP;
			}
			act->do_ttl_dec = 1;
			return 0;
		default:
			break;
		}
		break;
	default:
		break;
	}

	NL_SET_ERR_MSG_FMT_MOD(extack,
			       "Unsupported: ttl add action type %x %x %x/%x",
			       fa->mangle.htype, fa->mangle.offset,
			       fa->mangle.val, fa->mangle.mask);
	return -EOPNOTSUPP;
}

/**
 * efx_tc_mangle() - handle a single 32-bit (or less) pedit
 * @efx:	NIC we're installing a flow rule on
 * @act:	action set (cursor) to update
 * @fa:		FLOW_ACTION_MANGLE action metadata
 * @mung:	accumulator for partial mangles
 * @extack:	netlink extended ack for reporting errors
 * @match:	original match used along with the mangle action
 *
 * Identify the fields written by a FLOW_ACTION_MANGLE, and record
 * the partial mangle state in @mung.  If this mangle completes an
 * earlier partial mangle, consume and apply to @act by calling
 * efx_tc_complete_mac_mangle().
 */

static int efx_tc_mangle(struct efx_nic *efx, struct efx_tc_action_set *act,
			 const struct flow_action_entry *fa,
			 struct efx_tc_mangler_state *mung,
			 struct netlink_ext_ack *extack,
			 struct efx_tc_match *match)
{
	__le32 mac32;
	__le16 mac16;
	u8 tr_ttl;

	switch (fa->mangle.htype) {
	case FLOW_ACT_MANGLE_HDR_TYPE_ETH:
		BUILD_BUG_ON(offsetof(struct ethhdr, h_dest) != 0);
		BUILD_BUG_ON(offsetof(struct ethhdr, h_source) != 6);
		if (!efx_tc_flower_action_order_ok(act, EFX_TC_AO_PEDIT_MAC_ADDRS)) {
			NL_SET_ERR_MSG_MOD(extack,
					   "Pedit mangle mac action violates action order");
			return -EOPNOTSUPP;
		}
		switch (fa->mangle.offset) {
		case 0:
			if (fa->mangle.mask) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Unsupported: mask (%#x) of eth.dst32 mangle",
						       fa->mangle.mask);
				return -EOPNOTSUPP;
			}
			/* Ethernet address is little-endian */
			mac32 = cpu_to_le32(fa->mangle.val);
			memcpy(mung->dst_mac, &mac32, sizeof(mac32));
			mung->dst_mac_32 = 1;
			return efx_tc_complete_mac_mangle(efx, act, mung, extack);
		case 4:
			if (fa->mangle.mask == 0xffff) {
				mac16 = cpu_to_le16(fa->mangle.val >> 16);
				memcpy(mung->src_mac, &mac16, sizeof(mac16));
				mung->src_mac_16 = 1;
			} else if (fa->mangle.mask == 0xffff0000) {
				mac16 = cpu_to_le16((u16)fa->mangle.val);
				memcpy(mung->dst_mac + 4, &mac16, sizeof(mac16));
				mung->dst_mac_16 = 1;
			} else {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Unsupported: mask (%#x) of eth+4 mangle is not high or low 16b",
						       fa->mangle.mask);
				return -EOPNOTSUPP;
			}
			return efx_tc_complete_mac_mangle(efx, act, mung, extack);
		case 8:
			if (fa->mangle.mask) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Unsupported: mask (%#x) of eth.src32 mangle",
						       fa->mangle.mask);
				return -EOPNOTSUPP;
			}
			mac32 = cpu_to_le32(fa->mangle.val);
			memcpy(mung->src_mac + 2, &mac32, sizeof(mac32));
			mung->src_mac_32 = 1;
			return efx_tc_complete_mac_mangle(efx, act, mung, extack);
		default:
			NL_SET_ERR_MSG_FMT_MOD(extack, "Unsupported: mangle eth+%u %x/%x",
					       fa->mangle.offset, fa->mangle.val, fa->mangle.mask);
			return -EOPNOTSUPP;
		}
		break;
	case FLOW_ACT_MANGLE_HDR_TYPE_IP4:
		switch (fa->mangle.offset) {
		case offsetof(struct iphdr, ttl):
			/* we currently only support pedit IP4 when it applies
			 * to TTL and then only when it can be achieved with a
			 * decrement ttl action
			 */

			/* check that pedit applies to ttl only */
			if (fa->mangle.mask != ~EFX_TC_HDR_TYPE_TTL_MASK) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Unsupported: mask (%#x) out of range, only support mangle action on ipv4.ttl",
						       fa->mangle.mask);
				return -EOPNOTSUPP;
			}

			/* we can only convert to a dec ttl when we have an
			 * exact match on the ttl field
			 */
			if (match->mask.ip_ttl != U8_MAX) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Unsupported: only support mangle ipv4.ttl when we have an exact match on ttl, mask used for match (%#x)",
						       match->mask.ip_ttl);
				return -EOPNOTSUPP;
			}

			/* check that we don't try to decrement 0, which equates
			 * to setting the ttl to 0xff
			 */
			if (match->value.ip_ttl == 0) {
				NL_SET_ERR_MSG_MOD(extack,
						   "Unsupported: we cannot decrement ttl past 0");
				return -EOPNOTSUPP;
			}

			/* check that we do not decrement ttl twice */
			if (!efx_tc_flower_action_order_ok(act,
							   EFX_TC_AO_DEC_TTL)) {
				NL_SET_ERR_MSG_MOD(extack,
						   "Unsupported: multiple dec ttl");
				return -EOPNOTSUPP;
			}

			/* check pedit can be achieved with decrement action */
			tr_ttl = match->value.ip_ttl - 1;
			if ((fa->mangle.val & EFX_TC_HDR_TYPE_TTL_MASK) == tr_ttl) {
				act->do_ttl_dec = 1;
				return 0;
			}

			fallthrough;
		default:
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "Unsupported: only support mangle on the ttl field (offset is %u)",
					       fa->mangle.offset);
			return -EOPNOTSUPP;
		}
		break;
	case FLOW_ACT_MANGLE_HDR_TYPE_IP6:
		switch (fa->mangle.offset) {
		case round_down(offsetof(struct ipv6hdr, hop_limit), 4):
			/* we currently only support pedit IP6 when it applies
			 * to the hoplimit and then only when it can be achieved
			 * with a decrement hoplimit action
			 */

			/* check that pedit applies to ttl only */
			if (fa->mangle.mask != EFX_TC_HDR_TYPE_HLIMIT_MASK) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Unsupported: mask (%#x) out of range, only support mangle action on ipv6.hop_limit",
						       fa->mangle.mask);

				return -EOPNOTSUPP;
			}

			/* we can only convert to a dec ttl when we have an
			 * exact match on the ttl field
			 */
			if (match->mask.ip_ttl != U8_MAX) {
				NL_SET_ERR_MSG_FMT_MOD(extack,
						       "Unsupported: only support mangle ipv6.hop_limit when we have an exact match on ttl, mask used for match (%#x)",
						       match->mask.ip_ttl);
				return -EOPNOTSUPP;
			}

			/* check that we don't try to decrement 0, which equates
			 * to setting the ttl to 0xff
			 */
			if (match->value.ip_ttl == 0) {
				NL_SET_ERR_MSG_MOD(extack,
						   "Unsupported: we cannot decrement hop_limit past 0");
				return -EOPNOTSUPP;
			}

			/* check that we do not decrement hoplimit twice */
			if (!efx_tc_flower_action_order_ok(act,
							   EFX_TC_AO_DEC_TTL)) {
				NL_SET_ERR_MSG_MOD(extack,
						   "Unsupported: multiple dec ttl");
				return -EOPNOTSUPP;
			}

			/* check pedit can be achieved with decrement action */
			tr_ttl = match->value.ip_ttl - 1;
			if ((fa->mangle.val >> 24) == tr_ttl) {
				act->do_ttl_dec = 1;
				return 0;
			}

			fallthrough;
		default:
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "Unsupported: only support mangle on the hop_limit field");
			return -EOPNOTSUPP;
		}
	default:
		NL_SET_ERR_MSG_FMT_MOD(extack, "Unhandled mangle htype %u for action rule",
				       fa->mangle.htype);
		return -EOPNOTSUPP;
	}
	return 0;
}

/**
 * efx_tc_incomplete_mangle() - check for leftover partial pedits
 * @mung:	accumulator for partial mangles
 * @extack:	netlink extended ack for reporting errors
 *
 * Since the MAE can only overwrite whole fields, any partial
 * field mangle left over on reaching packet delivery (mirred or
 * end of TC actions) cannot be offloaded.  Check for any such
 * and reject them with -%EOPNOTSUPP.
 */

static int efx_tc_incomplete_mangle(struct efx_tc_mangler_state *mung,
				    struct netlink_ext_ack *extack)
{
	if (mung->dst_mac_32 || mung->dst_mac_16) {
		NL_SET_ERR_MSG_MOD(extack, "Incomplete pedit of destination MAC address");
		return -EOPNOTSUPP;
	}
	if (mung->src_mac_16 || mung->src_mac_32) {
		NL_SET_ERR_MSG_MOD(extack, "Incomplete pedit of source MAC address");
		return -EOPNOTSUPP;
	}
	return 0;
}

static int efx_tc_flower_replace_foreign(struct efx_nic *efx,
					 struct net_device *net_dev,
					 struct flow_cls_offload *tc)
{
	struct flow_rule *fr = flow_cls_offload_flow_rule(tc);
	struct netlink_ext_ack *extack = tc->common.extack;
	struct efx_tc_flow_rule *rule = NULL, *old = NULL;
	struct efx_tc_action_set *act = NULL;
	bool found = false, uplinked = false;
	const struct flow_action_entry *fa;
	struct efx_tc_match match;
	struct efx_rep *to_efv;
	s64 rc;
	int i;

	/* Parse match */
	memset(&match, 0, sizeof(match));
	rc = efx_tc_flower_parse_match(efx, fr, &match, NULL);
	if (rc)
		return rc;
	/* The rule as given to us doesn't specify a source netdevice.
	 * But, determining whether packets from a VF should match it is
	 * complicated, so leave those to the software slowpath: qualify
	 * the filter with source m-port == wire.
	 */
	rc = efx_tc_flower_external_mport(efx, EFX_EFV_PF);
	if (rc < 0) {
		NL_SET_ERR_MSG_MOD(extack, "Failed to identify ingress m-port for foreign filter");
		return rc;
	}
	match.value.ingress_port = rc;
	match.mask.ingress_port = ~0;

	if (tc->common.chain_index) {
		struct efx_tc_recirc_id *rid;

		rid = efx_tc_get_recirc_id(efx, tc->common.chain_index, net_dev);
		if (IS_ERR(rid)) {
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "Failed to allocate a hardware recirculation ID for chain_index %u",
					       tc->common.chain_index);
			return PTR_ERR(rid);
		}
		match.rid = rid;
		match.value.recirc_id = rid->fw_id;
	}
	match.mask.recirc_id = 0xff;

	/* AR table can't match on DO_CT (+trk).  But a commonly used pattern is
	 * +trk+est, which is strictly implied by +est, so rewrite it to that.
	 */
	if (match.mask.ct_state_trk && match.value.ct_state_trk &&
	    match.mask.ct_state_est && match.value.ct_state_est)
		match.mask.ct_state_trk = 0;
	/* Thanks to CT_TCP_FLAGS_INHIBIT, packets with interesting flags could
	 * match +trk-est (CT_HIT=0) despite being on an established connection.
	 * So make -est imply -tcp_syn_fin_rst match to ensure these packets
	 * still hit the software path.
	 */
	if (match.mask.ct_state_est && !match.value.ct_state_est) {
		if (match.value.tcp_syn_fin_rst) {
			/* Can't offload this combination */
			rc = -EOPNOTSUPP;
			goto release;
		}
		match.mask.tcp_syn_fin_rst = true;
	}

	flow_action_for_each(i, fa, &fr->action) {
		switch (fa->id) {
		case FLOW_ACTION_REDIRECT:
		case FLOW_ACTION_MIRRED: /* mirred means mirror here */
			to_efv = efx_tc_flower_lookup_efv(efx, fa->dev);
			if (IS_ERR(to_efv))
				continue;
			found = true;
			break;
		default:
			break;
		}
	}
	if (!found) { /* We don't care. */
		netif_dbg(efx, drv, efx->net_dev,
			  "Ignoring foreign filter that doesn't egdev us\n");
		rc = -EOPNOTSUPP;
		goto release;
	}

	rc = efx_mae_match_check_caps(efx, &match.mask, NULL);
	if (rc)
		goto release;

	if (efx_tc_match_is_encap(&match.mask)) {
		enum efx_encap_type type;

		type = efx_tc_indr_netdev_type(net_dev);
		if (type == EFX_ENCAP_TYPE_NONE) {
			NL_SET_ERR_MSG_MOD(extack,
					   "Egress encap match on unsupported tunnel device");
			rc = -EOPNOTSUPP;
			goto release;
		}

		rc = efx_mae_check_encap_type_supported(efx, type);
		if (rc) {
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "Firmware reports no support for %s encap match",
					       efx_tc_encap_type_name(type));
			goto release;
		}

		rc = efx_tc_flower_record_encap_match(efx, &match, type,
						      EFX_TC_EM_DIRECT, 0, 0,
						      extack);
		if (rc)
			goto release;
	} else {
		/* This is not a tunnel decap rule, ignore it */
		netif_dbg(efx, drv, efx->net_dev,
			  "Ignoring foreign filter without encap match\n");
		rc = -EOPNOTSUPP;
		goto release;
	}

	rule = kzalloc(sizeof(*rule), GFP_USER);
	if (!rule) {
		rc = -ENOMEM;
		goto release;
	}
	INIT_LIST_HEAD(&rule->acts.list);
	rule->cookie = tc->cookie;
	old = rhashtable_lookup_get_insert_fast(&efx->tc->match_action_ht,
						&rule->linkage,
						efx_tc_match_action_ht_params);
	if (IS_ERR(old)) {
		rc = PTR_ERR(old);
		goto release;
	} else if (old) {
		netif_dbg(efx, drv, efx->net_dev,
			  "Ignoring already-offloaded rule (cookie %lx)\n",
			  tc->cookie);
		rc = -EEXIST;
		goto release;
	}

	act = kzalloc(sizeof(*act), GFP_USER);
	if (!act) {
		rc = -ENOMEM;
		goto release;
	}

	/* Parse actions.  For foreign rules we only support decap & redirect.
	 * See corresponding code in efx_tc_flower_replace() for theory of
	 * operation & how 'act' cursor is used.
	 */
	flow_action_for_each(i, fa, &fr->action) {
		struct efx_tc_action_set save;

		switch (fa->id) {
		case FLOW_ACTION_REDIRECT:
		case FLOW_ACTION_MIRRED:
			/* See corresponding code in efx_tc_flower_replace() for
			 * long explanations of what's going on here.
			 */
			save = *act;
			if (fa->hw_stats) {
				struct efx_tc_counter_index *ctr;

				if (!(fa->hw_stats & FLOW_ACTION_HW_STATS_DELAYED)) {
					NL_SET_ERR_MSG_FMT_MOD(extack,
							       "hw_stats_type %u not supported (only 'delayed')",
							       fa->hw_stats);
					rc = -EOPNOTSUPP;
					goto release;
				}
				if (!efx_tc_flower_action_order_ok(act, EFX_TC_AO_COUNT)) {
					rc = -EOPNOTSUPP;
					goto release;
				}

				ctr = efx_tc_flower_get_counter_index(efx,
								      tc->cookie,
								      EFX_TC_COUNTER_TYPE_AR);
				if (IS_ERR(ctr)) {
					rc = PTR_ERR(ctr);
					NL_SET_ERR_MSG_MOD(extack, "Failed to obtain a counter");
					goto release;
				}
				act->count = ctr;
				INIT_LIST_HEAD(&act->count_user);
			}

			if (!efx_tc_flower_action_order_ok(act, EFX_TC_AO_DELIVER)) {
				/* can't happen */
				rc = -EOPNOTSUPP;
				NL_SET_ERR_MSG_MOD(extack,
						   "Deliver action violates action order (can't happen)");
				goto release;
			}
			to_efv = efx_tc_flower_lookup_efv(efx, fa->dev);
			/* PF implies egdev is us, in which case we really
			 * want to deliver to the uplink (because this is an
			 * ingress filter).  If we don't recognise the egdev
			 * at all, then we'd better trap so SW can handle it.
			 */
			if (IS_ERR(to_efv))
				to_efv = EFX_EFV_PF;
			if (to_efv == EFX_EFV_PF) {
				if (uplinked)
					break;
				uplinked = true;
			}
			rc = efx_tc_flower_internal_mport(efx, to_efv);
			if (rc < 0) {
				NL_SET_ERR_MSG_MOD(extack, "Failed to identify egress m-port");
				goto release;
			}
			act->dest_mport = rc;
			act->deliver = 1;
			rc = efx_mae_alloc_action_set(efx, act);
			if (rc) {
				NL_SET_ERR_MSG_MOD(extack,
						   "Failed to write action set to hw (mirred)");
				goto release;
			}
			list_add_tail(&act->list, &rule->acts.list);
			act = NULL;
			if (fa->id == FLOW_ACTION_REDIRECT)
				break; /* end of the line */
			/* Mirror, so continue on with saved act */
			act = kzalloc(sizeof(*act), GFP_USER);
			if (!act) {
				rc = -ENOMEM;
				goto release;
			}
			*act = save;
			break;
		case FLOW_ACTION_TUNNEL_DECAP:
			if (!efx_tc_flower_action_order_ok(act, EFX_TC_AO_DECAP)) {
				rc = -EINVAL;
				NL_SET_ERR_MSG_MOD(extack, "Decap action violates action order");
				goto release;
			}
			act->decap = 1;
			/* If we previously delivered/trapped to uplink, now
			 * that we've decapped we'll want another copy if we
			 * try to deliver/trap to uplink again.
			 */
			uplinked = false;
			break;
		default:
			NL_SET_ERR_MSG_FMT_MOD(extack, "Unhandled action %u",
					       fa->id);
			rc = -EOPNOTSUPP;
			goto release;
		}
	}

	if (act) {
		if (!uplinked) {
			/* Not shot/redirected, so deliver to default dest (which is
			 * the uplink, as this is an ingress filter)
			 */
			efx_mae_mport_uplink(efx, &act->dest_mport);
			act->deliver = 1;
		}
		rc = efx_mae_alloc_action_set(efx, act);
		if (rc) {
			NL_SET_ERR_MSG_MOD(extack, "Failed to write action set to hw (deliver)");
			goto release;
		}
		list_add_tail(&act->list, &rule->acts.list);
		act = NULL; /* Prevent double-free in error path */
	}

	rule->match = match;

	netif_dbg(efx, drv, efx->net_dev,
		  "Successfully parsed foreign filter (cookie %lx)\n",
		  tc->cookie);

	rc = efx_mae_alloc_action_set_list(efx, &rule->acts);
	if (rc) {
		NL_SET_ERR_MSG_MOD(extack, "Failed to write action set list to hw");
		goto release;
	}
	rc = efx_mae_insert_rule(efx, &rule->match, EFX_TC_PRIO_TC,
				 rule->acts.fw_id, &rule->fw_id);
	if (rc) {
		NL_SET_ERR_MSG_MOD(extack, "Failed to insert rule in hw");
		goto release_acts;
	}
	return 0;

release_acts:
	efx_mae_free_action_set_list(efx, &rule->acts);
release:
	/* We failed to insert the rule, so free up any entries we created in
	 * subsidiary tables.
	 */
	if (match.rid)
		efx_tc_put_recirc_id(efx, match.rid);
	if (act)
		efx_tc_free_action_set(efx, act, false);
	if (rule) {
		if (!old)
			rhashtable_remove_fast(&efx->tc->match_action_ht,
					       &rule->linkage,
					       efx_tc_match_action_ht_params);
		efx_tc_free_action_set_list(efx, &rule->acts, false);
	}
	kfree(rule);
	if (match.encap)
		efx_tc_flower_release_encap_match(efx, match.encap);
	return rc;
}

static int efx_tc_flower_replace_lhs(struct efx_nic *efx,
				     struct flow_cls_offload *tc,
				     struct flow_rule *fr,
				     struct efx_tc_match *match,
				     struct efx_rep *efv,
				     struct net_device *net_dev)
{
	struct netlink_ext_ack *extack = tc->common.extack;
	struct efx_tc_lhs_rule *rule, *old;
	int rc;

	if (tc->common.chain_index) {
		NL_SET_ERR_MSG_MOD(extack, "LHS rule only allowed in chain 0");
		return -EOPNOTSUPP;
	}

	if (match->mask.ct_state_trk && match->value.ct_state_trk) {
		NL_SET_ERR_MSG_MOD(extack, "LHS rule can never match +trk");
		return -EOPNOTSUPP;
	}
	/* LHS rules are always -trk, so we don't need to match on that */
	match->mask.ct_state_trk = 0;
	match->value.ct_state_trk = 0;

	rc = efx_mae_match_check_caps_lhs(efx, &match->mask, extack);
	if (rc)
		return rc;

	rule = kzalloc(sizeof(*rule), GFP_USER);
	if (!rule)
		return -ENOMEM;
	rule->cookie = tc->cookie;
	old = rhashtable_lookup_get_insert_fast(&efx->tc->lhs_rule_ht,
						&rule->linkage,
						efx_tc_lhs_rule_ht_params);
	if (IS_ERR(old)) {
		rc = PTR_ERR(old);
		goto release;
	} else if (old) {
		netif_dbg(efx, drv, efx->net_dev,
			  "Already offloaded rule (cookie %lx)\n", tc->cookie);
		rc = -EEXIST;
		NL_SET_ERR_MSG_MOD(extack, "Rule already offloaded");
		goto release;
	}

	/* Parse actions */
	/* See note in efx_tc_flower_replace() regarding passed net_dev
	 * (used for efx_tc_get_recirc_id()).
	 */
	rc = efx_tc_flower_handle_lhs_actions(efx, tc, fr, efx->net_dev, rule);
	if (rc)
		goto release;

	rule->match = *match;

	rc = efx_mae_insert_lhs_rule(efx, rule, EFX_TC_PRIO_TC);
	if (rc) {
		NL_SET_ERR_MSG_MOD(extack, "Failed to insert rule in hw");
		goto release;
	}
	netif_dbg(efx, drv, efx->net_dev,
		  "Successfully parsed lhs rule (cookie %lx)\n",
		  tc->cookie);
	return 0;

release:
	efx_tc_flower_release_lhs_actions(efx, &rule->lhs_act);
	if (!old)
		rhashtable_remove_fast(&efx->tc->lhs_rule_ht, &rule->linkage,
				       efx_tc_lhs_rule_ht_params);
	kfree(rule);
	return rc;
}

static int efx_tc_flower_replace(struct efx_nic *efx,
				 struct net_device *net_dev,
				 struct flow_cls_offload *tc,
				 struct efx_rep *efv)
{
	struct flow_rule *fr = flow_cls_offload_flow_rule(tc);
	struct netlink_ext_ack *extack = tc->common.extack;
	const struct ip_tunnel_info *encap_info = NULL;
	struct efx_tc_flow_rule *rule = NULL, *old;
	struct efx_tc_mangler_state mung = {};
	struct efx_tc_action_set *act = NULL;
	const struct flow_action_entry *fa;
	struct efx_rep *from_efv, *to_efv;
	struct efx_tc_match match;
	u32 acts_id;
	s64 rc;
	int i;

	if (!tc_can_offload_extack(efx->net_dev, extack))
		return -EOPNOTSUPP;
	if (WARN_ON(!efx->tc))
		return -ENETDOWN;
	if (WARN_ON(!efx->tc->up))
		return -ENETDOWN;

	from_efv = efx_tc_flower_lookup_efv(efx, net_dev);
	if (IS_ERR(from_efv)) {
		/* Not from our PF or representors, so probably a tunnel dev */
		return efx_tc_flower_replace_foreign(efx, net_dev, tc);
	}

	if (efv != from_efv) {
		/* can't happen */
		NL_SET_ERR_MSG_FMT_MOD(extack, "for %s efv is %snull but from_efv is %snull (can't happen)",
				       netdev_name(net_dev), efv ? "non-" : "",
				       from_efv ? "non-" : "");
		return -EINVAL;
	}

	/* Parse match */
	memset(&match, 0, sizeof(match));
	rc = efx_tc_flower_external_mport(efx, from_efv);
	if (rc < 0) {
		NL_SET_ERR_MSG_MOD(extack, "Failed to identify ingress m-port");
		return rc;
	}
	match.value.ingress_port = rc;
	match.mask.ingress_port = ~0;
	rc = efx_tc_flower_parse_match(efx, fr, &match, extack);
	if (rc)
		return rc;
	if (efx_tc_match_is_encap(&match.mask)) {
		NL_SET_ERR_MSG_MOD(extack, "Ingress enc_key matches not supported");
		return -EOPNOTSUPP;
	}

	if (efx_tc_rule_is_lhs_rule(fr, &match))
		return efx_tc_flower_replace_lhs(efx, tc, fr, &match, efv,
						 net_dev);

	/* chain_index 0 is always recirc_id 0 (and does not appear in recirc_ht).
	 * Conveniently, match.rid == NULL and match.value.recirc_id == 0 owing
	 * to the initial memset(), so we don't need to do anything in that case.
	 */
	if (tc->common.chain_index) {
		struct efx_tc_recirc_id *rid;

		/* Note regarding passed net_dev:
		 * VFreps and PF can share chain namespace, as they have
		 * distinct ingress_mports.  So we don't need to burn an
		 * extra recirc_id if both use the same chain_index.
		 * (Strictly speaking, we could give each VFrep its own
		 * recirc_id namespace that doesn't take IDs away from the
		 * PF, but that would require a bunch of additional IDAs -
		 * one for each representor - and that's not likely to be
		 * the main cause of recirc_id exhaustion anyway.)
		 */
		rid = efx_tc_get_recirc_id(efx, tc->common.chain_index,
					   efx->net_dev);
		if (IS_ERR(rid)) {
			NL_SET_ERR_MSG_FMT_MOD(extack,
					       "Failed to allocate a hardware recirculation ID for chain_index %u",
					       tc->common.chain_index);
			return PTR_ERR(rid);
		}
		match.rid = rid;
		match.value.recirc_id = rid->fw_id;
	}
	match.mask.recirc_id = 0xff;

	/* AR table can't match on DO_CT (+trk).  But a commonly used pattern is
	 * +trk+est, which is strictly implied by +est, so rewrite it to that.
	 */
	if (match.mask.ct_state_trk && match.value.ct_state_trk &&
	    match.mask.ct_state_est && match.value.ct_state_est)
		match.mask.ct_state_trk = 0;
	/* Thanks to CT_TCP_FLAGS_INHIBIT, packets with interesting flags could
	 * match +trk-est (CT_HIT=0) despite being on an established connection.
	 * So make -est imply -tcp_syn_fin_rst match to ensure these packets
	 * still hit the software path.
	 */
	if (match.mask.ct_state_est && !match.value.ct_state_est) {
		if (match.value.tcp_syn_fin_rst) {
			/* Can't offload this combination */
			rc = -EOPNOTSUPP;
			goto release;
		}
		match.mask.tcp_syn_fin_rst = true;
	}

	rc = efx_mae_match_check_caps(efx, &match.mask, extack);
	if (rc)
		goto release;

	rule = kzalloc(sizeof(*rule), GFP_USER);
	if (!rule) {
		rc = -ENOMEM;
		goto release;
	}
	INIT_LIST_HEAD(&rule->acts.list);
	rule->cookie = tc->cookie;
	old = rhashtable_lookup_get_insert_fast(&efx->tc->match_action_ht,
						&rule->linkage,
						efx_tc_match_action_ht_params);
	if (IS_ERR(old)) {
		rc = PTR_ERR(old);
		goto release;
	} else if (old) {
		netif_dbg(efx, drv, efx->net_dev,
			  "Already offloaded rule (cookie %lx)\n", tc->cookie);
		NL_SET_ERR_MSG_MOD(extack, "Rule already offloaded");
		rc = -EEXIST;
		goto release;
	}

	/* Parse actions */
	act = kzalloc(sizeof(*act), GFP_USER);
	if (!act) {
		rc = -ENOMEM;
		goto release;
	}

	/**
	 * DOC: TC action translation
	 *
	 * Actions in TC are sequential and cumulative, with delivery actions
	 * potentially anywhere in the order.  The EF100 MAE, however, takes
	 * an 'action set list' consisting of 'action sets', each of which is
	 * applied to the _original_ packet, and consists of a set of optional
	 * actions in a fixed order with delivery at the end.
	 * To translate between these two models, we maintain a 'cursor', @act,
	 * which describes the cumulative effect of all the packet-mutating
	 * actions encountered so far; on handling a delivery (mirred or drop)
	 * action, once the action-set has been inserted into hardware, we
	 * append @act to the action-set list (@rule->acts); if this is a pipe
	 * action (mirred mirror) we then allocate a new @act with a copy of
	 * the cursor state _before_ the delivery action, otherwise we set @act
	 * to %NULL.
	 * This ensures that every allocated action-set is either attached to
	 * @rule->acts or pointed to by @act (and never both), and that only
	 * those action-sets in @rule->acts exist in hardware.  Consequently,
	 * in the failure path, @act only needs to be freed in memory, whereas
	 * for @rule->acts we remove each action-set from hardware before
	 * freeing it (efx_tc_free_action_set_list()), even if the action-set
	 * list itself is not in hardware.
	 */
	flow_action_for_each(i, fa, &fr->action) {
		struct efx_tc_action_set save;
		u16 tci;

		if (!act) {
			/* more actions after a non-pipe action */
			NL_SET_ERR_MSG_MOD(extack, "Action follows non-pipe action");
			rc = -EINVAL;
			goto release;
		}

		if ((fa->id == FLOW_ACTION_REDIRECT ||
		     fa->id == FLOW_ACTION_MIRRED ||
		     fa->id == FLOW_ACTION_DROP) && fa->hw_stats) {
			struct efx_tc_counter_index *ctr;

			/* Currently the only actions that want stats are
			 * mirred and gact (ok, shot, trap, goto-chain), which
			 * means we want stats just before delivery.  Also,
			 * note that tunnel_key set shouldn't change the length
			 * — it's only the subsequent mirred that does that,
			 * and the stats are taken _before_ the mirred action
			 * happens.
			 */
			if (!efx_tc_flower_action_order_ok(act, EFX_TC_AO_COUNT)) {
				/* All supported actions that count either steal
				 * (gact shot, mirred redirect) or clone act
				 * (mirred mirror), so we should never get two
				 * count actions on one action_set.
				 */
				NL_SET_ERR_MSG_MOD(extack, "Count-action conflict (can't happen)");
				rc = -EOPNOTSUPP;
				goto release;
			}

			if (!(fa->hw_stats & FLOW_ACTION_HW_STATS_DELAYED)) {
				NL_SET_ERR_MSG_FMT_MOD(extack, "hw_stats_type %u not supported (only 'delayed')",
						       fa->hw_stats);
				rc = -EOPNOTSUPP;
				goto release;
			}

			ctr = efx_tc_flower_get_counter_index(efx, tc->cookie,
							      EFX_TC_COUNTER_TYPE_AR);
			if (IS_ERR(ctr)) {
				rc = PTR_ERR(ctr);
				NL_SET_ERR_MSG_MOD(extack, "Failed to obtain a counter");
				goto release;
			}
			act->count = ctr;
			INIT_LIST_HEAD(&act->count_user);
		}

		switch (fa->id) {
		case FLOW_ACTION_DROP:
			rc = efx_mae_alloc_action_set(efx, act);
			if (rc) {
				NL_SET_ERR_MSG_MOD(extack, "Failed to write action set to hw (drop)");
				goto release;
			}
			list_add_tail(&act->list, &rule->acts.list);
			act = NULL; /* end of the line */
			break;
		case FLOW_ACTION_REDIRECT:
		case FLOW_ACTION_MIRRED:
			save = *act;

			if (encap_info) {
				struct efx_tc_encap_action *encap;

				if (!efx_tc_flower_action_order_ok(act,
								   EFX_TC_AO_ENCAP)) {
					rc = -EOPNOTSUPP;
					NL_SET_ERR_MSG_MOD(extack, "Encap action violates action order");
					goto release;
				}
				encap = efx_tc_flower_create_encap_md(
						efx, encap_info, fa->dev, extack);
				if (IS_ERR_OR_NULL(encap)) {
					rc = PTR_ERR(encap);
					if (!rc)
						rc = -EIO; /* arbitrary */
					goto release;
				}
				act->encap_md = encap;
				list_add_tail(&act->encap_user, &encap->users);
				act->dest_mport = encap->dest_mport;
				act->deliver = 1;
				if (act->count && !WARN_ON(!act->count->cnt)) {
					/* This counter is used by an encap
					 * action, which needs a reference back
					 * so it can prod neighbouring whenever
					 * traffic is seen.
					 */
					spin_lock_bh(&act->count->cnt->lock);
					list_add_tail(&act->count_user,
						      &act->count->cnt->users);
					spin_unlock_bh(&act->count->cnt->lock);
				}
				rc = efx_mae_alloc_action_set(efx, act);
				if (rc) {
					NL_SET_ERR_MSG_MOD(extack, "Failed to write action set to hw (encap)");
					goto release;
				}
				list_add_tail(&act->list, &rule->acts.list);
				act->user = &rule->acts;
				act = NULL;
				if (fa->id == FLOW_ACTION_REDIRECT)
					break; /* end of the line */
				/* Mirror, so continue on with saved act */
				save.count = NULL;
				act = kzalloc(sizeof(*act), GFP_USER);
				if (!act) {
					rc = -ENOMEM;
					goto release;
				}
				*act = save;
				break;
			}

			if (!efx_tc_flower_action_order_ok(act, EFX_TC_AO_DELIVER)) {
				/* can't happen */
				rc = -EOPNOTSUPP;
				NL_SET_ERR_MSG_MOD(extack, "Deliver action violates action order (can't happen)");
				goto release;
			}

			to_efv = efx_tc_flower_lookup_efv(efx, fa->dev);
			if (IS_ERR(to_efv)) {
				NL_SET_ERR_MSG_MOD(extack, "Mirred egress device not on switch");
				rc = PTR_ERR(to_efv);
				goto release;
			}
			rc = efx_tc_flower_external_mport(efx, to_efv);
			if (rc < 0) {
				NL_SET_ERR_MSG_MOD(extack, "Failed to identify egress m-port");
				goto release;
			}
			act->dest_mport = rc;
			act->deliver = 1;
			rc = efx_mae_alloc_action_set(efx, act);
			if (rc) {
				NL_SET_ERR_MSG_MOD(extack, "Failed to write action set to hw (mirred)");
				goto release;
			}
			list_add_tail(&act->list, &rule->acts.list);
			act = NULL;
			if (fa->id == FLOW_ACTION_REDIRECT)
				break; /* end of the line */
			/* Mirror, so continue on with saved act */
			save.count = NULL;
			act = kzalloc(sizeof(*act), GFP_USER);
			if (!act) {
				rc = -ENOMEM;
				goto release;
			}
			*act = save;
			break;
		case FLOW_ACTION_VLAN_POP:
			if (act->vlan_push) {
				act->vlan_push--;
			} else if (efx_tc_flower_action_order_ok(act, EFX_TC_AO_VLAN_POP)) {
				act->vlan_pop++;
			} else {
				NL_SET_ERR_MSG_MOD(extack,
						   "More than two VLAN pops, or action order violated");
				rc = -EINVAL;
				goto release;
			}
			break;
		case FLOW_ACTION_VLAN_PUSH:
			if (!efx_tc_flower_action_order_ok(act, EFX_TC_AO_VLAN_PUSH)) {
				rc = -EINVAL;
				NL_SET_ERR_MSG_MOD(extack,
						   "More than two VLAN pushes, or action order violated");
				goto release;
			}
			tci = fa->vlan.vid & VLAN_VID_MASK;
			tci |= fa->vlan.prio << VLAN_PRIO_SHIFT;
			act->vlan_tci[act->vlan_push] = cpu_to_be16(tci);
			act->vlan_proto[act->vlan_push] = fa->vlan.proto;
			act->vlan_push++;
			break;
		case FLOW_ACTION_ADD:
			rc = efx_tc_pedit_add(efx, act, fa, extack);
			if (rc < 0)
				goto release;
			break;
		case FLOW_ACTION_MANGLE:
			rc = efx_tc_mangle(efx, act, fa, &mung, extack, &match);
			if (rc < 0)
				goto release;
			break;
		case FLOW_ACTION_TUNNEL_ENCAP:
			if (encap_info) {
				/* Can't specify encap multiple times.
				 * If you want to overwrite an existing
				 * encap_info, use an intervening
				 * FLOW_ACTION_TUNNEL_DECAP to clear it.
				 */
				NL_SET_ERR_MSG_MOD(extack, "Tunnel key set when already set");
				rc = -EINVAL;
				goto release;
			}
			if (!fa->tunnel) {
				NL_SET_ERR_MSG_MOD(extack, "Tunnel key set is missing key");
				rc = -EOPNOTSUPP;
				goto release;
			}
			encap_info = fa->tunnel;
			break;
		case FLOW_ACTION_TUNNEL_DECAP:
			if (encap_info) {
				encap_info = NULL;
				break;
			}
			/* Since we don't support enc_key matches on ingress
			 * (and if we did there'd be no tunnel-device to give
			 * us a type), we can't offload a decap that's not
			 * just undoing a previous encap action.
			 */
			NL_SET_ERR_MSG_MOD(extack, "Cannot offload tunnel decap action without tunnel device");
			rc = -EOPNOTSUPP;
			goto release;
		default:
			NL_SET_ERR_MSG_FMT_MOD(extack, "Unhandled action %u",
					       fa->id);
			rc = -EOPNOTSUPP;
			goto release;
		}
	}

	rc = efx_tc_incomplete_mangle(&mung, extack);
	if (rc < 0)
		goto release;
	if (act) {
		/* Not shot/redirected, so deliver to default dest */
		if (from_efv == EFX_EFV_PF)
			/* Rule applies to traffic from the wire,
			 * and default dest is thus the PF
			 */
			efx_mae_mport_uplink(efx, &act->dest_mport);
		else
			/* Representor, so rule applies to traffic from
			 * representee, and default dest is thus the rep.
			 * All reps use the same mport for delivery
			 */
			efx_mae_mport_mport(efx, efx->tc->reps_mport_id,
					    &act->dest_mport);
		act->deliver = 1;
		rc = efx_mae_alloc_action_set(efx, act);
		if (rc) {
			NL_SET_ERR_MSG_MOD(extack, "Failed to write action set to hw (deliver)");
			goto release;
		}
		list_add_tail(&act->list, &rule->acts.list);
		act = NULL; /* Prevent double-free in error path */
	}

	netif_dbg(efx, drv, efx->net_dev,
		  "Successfully parsed filter (cookie %lx)\n",
		  tc->cookie);

	rule->match = match;

	rc = efx_mae_alloc_action_set_list(efx, &rule->acts);
	if (rc) {
		NL_SET_ERR_MSG_MOD(extack, "Failed to write action set list to hw");
		goto release;
	}
	if (from_efv == EFX_EFV_PF)
		/* PF netdev, so rule applies to traffic from wire */
		rule->fallback = &efx->tc->facts.pf;
	else
		/* repdev, so rule applies to traffic from representee */
		rule->fallback = &efx->tc->facts.reps;
	if (!efx_tc_check_ready(efx, rule)) {
		netif_dbg(efx, drv, efx->net_dev, "action not ready for hw\n");
		acts_id = rule->fallback->fw_id;
	} else {
		netif_dbg(efx, drv, efx->net_dev, "ready for hw\n");
		acts_id = rule->acts.fw_id;
	}
	rc = efx_mae_insert_rule(efx, &rule->match, EFX_TC_PRIO_TC,
				 acts_id, &rule->fw_id);
	if (rc) {
		NL_SET_ERR_MSG_MOD(extack, "Failed to insert rule in hw");
		goto release_acts;
	}
	return 0;

release_acts:
	efx_mae_free_action_set_list(efx, &rule->acts);
release:
	/* We failed to insert the rule, so free up any entries we created in
	 * subsidiary tables.
	 */
	if (match.rid)
		efx_tc_put_recirc_id(efx, match.rid);
	if (act)
		efx_tc_free_action_set(efx, act, false);
	if (rule) {
		if (!old)
			rhashtable_remove_fast(&efx->tc->match_action_ht,
					       &rule->linkage,
					       efx_tc_match_action_ht_params);
		efx_tc_free_action_set_list(efx, &rule->acts, false);
	}
	kfree(rule);
	return rc;
}

static int efx_tc_flower_destroy(struct efx_nic *efx,
				 struct net_device *net_dev,
				 struct flow_cls_offload *tc)
{
	struct netlink_ext_ack *extack = tc->common.extack;
	struct efx_tc_lhs_rule *lhs_rule;
	struct efx_tc_flow_rule *rule;

	lhs_rule = rhashtable_lookup_fast(&efx->tc->lhs_rule_ht, &tc->cookie,
					  efx_tc_lhs_rule_ht_params);
	if (lhs_rule) {
		/* Remove it from HW */
		efx_mae_remove_lhs_rule(efx, lhs_rule);
		/* Delete it from SW */
		efx_tc_flower_release_lhs_actions(efx, &lhs_rule->lhs_act);
		rhashtable_remove_fast(&efx->tc->lhs_rule_ht, &lhs_rule->linkage,
				       efx_tc_lhs_rule_ht_params);
		if (lhs_rule->match.encap)
			efx_tc_flower_release_encap_match(efx, lhs_rule->match.encap);
		netif_dbg(efx, drv, efx->net_dev, "Removed (lhs) filter %lx\n",
			  lhs_rule->cookie);
		kfree(lhs_rule);
		return 0;
	}

	rule = rhashtable_lookup_fast(&efx->tc->match_action_ht, &tc->cookie,
				      efx_tc_match_action_ht_params);
	if (!rule) {
		/* Only log a message if we're the ingress device.  Otherwise
		 * it's a foreign filter and we might just not have been
		 * interested (e.g. we might not have been the egress device
		 * either).
		 */
		if (!IS_ERR(efx_tc_flower_lookup_efv(efx, net_dev)))
			netif_warn(efx, drv, efx->net_dev,
				   "Filter %lx not found to remove\n", tc->cookie);
		NL_SET_ERR_MSG_MOD(extack, "Flow cookie not found in offloaded rules");
		return -ENOENT;
	}

	/* Remove it from HW */
	efx_tc_delete_rule(efx, rule);
	/* Delete it from SW */
	rhashtable_remove_fast(&efx->tc->match_action_ht, &rule->linkage,
			       efx_tc_match_action_ht_params);
	netif_dbg(efx, drv, efx->net_dev, "Removed filter %lx\n", rule->cookie);
	kfree(rule);
	return 0;
}

static int efx_tc_flower_stats(struct efx_nic *efx, struct net_device *net_dev,
			       struct flow_cls_offload *tc)
{
	struct netlink_ext_ack *extack = tc->common.extack;
	struct efx_tc_counter_index *ctr;
	struct efx_tc_counter *cnt;
	u64 packets, bytes;

	ctr = efx_tc_flower_find_counter_index(efx, tc->cookie);
	if (!ctr) {
		/* See comment in efx_tc_flower_destroy() */
		if (!IS_ERR(efx_tc_flower_lookup_efv(efx, net_dev)))
			if (net_ratelimit())
				netif_warn(efx, drv, efx->net_dev,
					   "Filter %lx not found for stats\n",
					   tc->cookie);
		NL_SET_ERR_MSG_MOD(extack, "Flow cookie not found in offloaded rules");
		return -ENOENT;
	}
	if (WARN_ON(!ctr->cnt)) /* can't happen */
		return -EIO;
	cnt = ctr->cnt;

	spin_lock_bh(&cnt->lock);
	/* Report only new pkts/bytes since last time TC asked */
	packets = cnt->packets;
	bytes = cnt->bytes;
	flow_stats_update(&tc->stats, bytes - cnt->old_bytes,
			  packets - cnt->old_packets, 0, cnt->touched,
			  FLOW_ACTION_HW_STATS_DELAYED);
	cnt->old_packets = packets;
	cnt->old_bytes = bytes;
	spin_unlock_bh(&cnt->lock);
	return 0;
}

int efx_tc_flower(struct efx_nic *efx, struct net_device *net_dev,
		  struct flow_cls_offload *tc, struct efx_rep *efv)
{
	int rc;

	if (!efx->tc)
		return -EOPNOTSUPP;

	mutex_lock(&efx->tc->mutex);
	switch (tc->command) {
	case FLOW_CLS_REPLACE:
		rc = efx_tc_flower_replace(efx, net_dev, tc, efv);
		break;
	case FLOW_CLS_DESTROY:
		rc = efx_tc_flower_destroy(efx, net_dev, tc);
		break;
	case FLOW_CLS_STATS:
		rc = efx_tc_flower_stats(efx, net_dev, tc);
		break;
	default:
		rc = -EOPNOTSUPP;
		break;
	}
	mutex_unlock(&efx->tc->mutex);
	return rc;
}

static int efx_tc_configure_default_rule(struct efx_nic *efx, u32 ing_port,
					 u32 eg_port, struct efx_tc_flow_rule *rule)
{
	struct efx_tc_action_set_list *acts = &rule->acts;
	struct efx_tc_match *match = &rule->match;
	struct efx_tc_action_set *act;
	int rc;

	match->value.ingress_port = ing_port;
	match->mask.ingress_port = ~0;
	act = kzalloc(sizeof(*act), GFP_KERNEL);
	if (!act)
		return -ENOMEM;
	act->deliver = 1;
	act->dest_mport = eg_port;
	rc = efx_mae_alloc_action_set(efx, act);
	if (rc)
		goto fail1;
	EFX_WARN_ON_PARANOID(!list_empty(&acts->list));
	list_add_tail(&act->list, &acts->list);
	rc = efx_mae_alloc_action_set_list(efx, acts);
	if (rc)
		goto fail2;
	rc = efx_mae_insert_rule(efx, match, EFX_TC_PRIO_DFLT,
				 acts->fw_id, &rule->fw_id);
	if (rc)
		goto fail3;
	return 0;
fail3:
	efx_mae_free_action_set_list(efx, acts);
fail2:
	list_del(&act->list);
	efx_mae_free_action_set(efx, act->fw_id);
fail1:
	kfree(act);
	return rc;
}

static int efx_tc_configure_default_rule_pf(struct efx_nic *efx)
{
	struct efx_tc_flow_rule *rule = &efx->tc->dflt.pf;
	u32 ing_port, eg_port;

	efx_mae_mport_uplink(efx, &ing_port);
	efx_mae_mport_wire(efx, &eg_port);
	return efx_tc_configure_default_rule(efx, ing_port, eg_port, rule);
}

static int efx_tc_configure_default_rule_wire(struct efx_nic *efx)
{
	struct efx_tc_flow_rule *rule = &efx->tc->dflt.wire;
	u32 ing_port, eg_port;

	efx_mae_mport_wire(efx, &ing_port);
	efx_mae_mport_uplink(efx, &eg_port);
	return efx_tc_configure_default_rule(efx, ing_port, eg_port, rule);
}

int efx_tc_configure_default_rule_rep(struct efx_rep *efv)
{
	struct efx_tc_flow_rule *rule = &efv->dflt;
	struct efx_nic *efx = efv->parent;
	u32 ing_port, eg_port;

	efx_mae_mport_mport(efx, efv->mport, &ing_port);
	efx_mae_mport_mport(efx, efx->tc->reps_mport_id, &eg_port);
	return efx_tc_configure_default_rule(efx, ing_port, eg_port, rule);
}

void efx_tc_deconfigure_default_rule(struct efx_nic *efx,
				     struct efx_tc_flow_rule *rule)
{
	if (rule->fw_id != MC_CMD_MAE_ACTION_RULE_INSERT_OUT_ACTION_RULE_ID_NULL)
		efx_tc_delete_rule(efx, rule);
	rule->fw_id = MC_CMD_MAE_ACTION_RULE_INSERT_OUT_ACTION_RULE_ID_NULL;
}

static int efx_tc_configure_fallback_acts(struct efx_nic *efx, u32 eg_port,
					  struct efx_tc_action_set_list *acts)
{
	struct efx_tc_action_set *act;
	int rc;

	act = kzalloc(sizeof(*act), GFP_KERNEL);
	if (!act)
		return -ENOMEM;
	act->deliver = 1;
	act->dest_mport = eg_port;
	rc = efx_mae_alloc_action_set(efx, act);
	if (rc)
		goto fail1;
	EFX_WARN_ON_PARANOID(!list_empty(&acts->list));
	list_add_tail(&act->list, &acts->list);
	rc = efx_mae_alloc_action_set_list(efx, acts);
	if (rc)
		goto fail2;
	return 0;
fail2:
	list_del(&act->list);
	efx_mae_free_action_set(efx, act->fw_id);
fail1:
	kfree(act);
	return rc;
}

static int efx_tc_configure_fallback_acts_pf(struct efx_nic *efx)
{
	struct efx_tc_action_set_list *acts = &efx->tc->facts.pf;
	u32 eg_port;

	efx_mae_mport_uplink(efx, &eg_port);
	return efx_tc_configure_fallback_acts(efx, eg_port, acts);
}

static int efx_tc_configure_fallback_acts_reps(struct efx_nic *efx)
{
	struct efx_tc_action_set_list *acts = &efx->tc->facts.reps;
	u32 eg_port;

	efx_mae_mport_mport(efx, efx->tc->reps_mport_id, &eg_port);
	return efx_tc_configure_fallback_acts(efx, eg_port, acts);
}

static void efx_tc_deconfigure_fallback_acts(struct efx_nic *efx,
					     struct efx_tc_action_set_list *acts)
{
	efx_tc_free_action_set_list(efx, acts, true);
}

static int efx_tc_configure_rep_mport(struct efx_nic *efx)
{
	u32 rep_mport_label;
	int rc;

	rc = efx_mae_allocate_mport(efx, &efx->tc->reps_mport_id, &rep_mport_label);
	if (rc)
		return rc;
	pci_dbg(efx->pci_dev, "created rep mport 0x%08x (0x%04x)\n",
		efx->tc->reps_mport_id, rep_mport_label);
	/* Use mport *selector* as vport ID */
	efx_mae_mport_mport(efx, efx->tc->reps_mport_id,
			    &efx->tc->reps_mport_vport_id);
	return 0;
}

static void efx_tc_deconfigure_rep_mport(struct efx_nic *efx)
{
	efx_mae_free_mport(efx, efx->tc->reps_mport_id);
	efx->tc->reps_mport_id = MAE_MPORT_SELECTOR_NULL;
}

int efx_tc_insert_rep_filters(struct efx_nic *efx)
{
	struct efx_filter_spec promisc, allmulti;
	int rc;

	if (efx->type->is_vf)
		return 0;
	if (!efx->tc)
		return 0;
	efx_filter_init_rx(&promisc, EFX_FILTER_PRI_REQUIRED, 0, 0);
	efx_filter_set_uc_def(&promisc);
	efx_filter_set_vport_id(&promisc, efx->tc->reps_mport_vport_id);
	rc = efx_filter_insert_filter(efx, &promisc, false);
	if (rc < 0)
		return rc;
	efx->tc->reps_filter_uc = rc;
	efx_filter_init_rx(&allmulti, EFX_FILTER_PRI_REQUIRED, 0, 0);
	efx_filter_set_mc_def(&allmulti);
	efx_filter_set_vport_id(&allmulti, efx->tc->reps_mport_vport_id);
	rc = efx_filter_insert_filter(efx, &allmulti, false);
	if (rc < 0)
		return rc;
	efx->tc->reps_filter_mc = rc;
	return 0;
}

void efx_tc_remove_rep_filters(struct efx_nic *efx)
{
	if (efx->type->is_vf)
		return;
	if (!efx->tc)
		return;
	if (efx->tc->reps_filter_mc >= 0)
		efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED, efx->tc->reps_filter_mc);
	efx->tc->reps_filter_mc = -1;
	if (efx->tc->reps_filter_uc >= 0)
		efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED, efx->tc->reps_filter_uc);
	efx->tc->reps_filter_uc = -1;
}

int efx_init_tc(struct efx_nic *efx)
{
	int rc;

	rc = efx_mae_get_caps(efx, efx->tc->caps);
	if (rc)
		return rc;
	if (efx->tc->caps->match_field_count > MAE_NUM_FIELDS)
		/* Firmware supports some match fields the driver doesn't know
		 * about.  Not fatal, unless any of those fields are required
		 * (MAE_FIELD_SUPPORTED_MATCH_ALWAYS) but if so we don't know.
		 */
		netif_warn(efx, probe, efx->net_dev,
			   "FW reports additional match fields %u\n",
			   efx->tc->caps->match_field_count);
	if (efx->tc->caps->action_prios < EFX_TC_PRIO__NUM) {
		netif_err(efx, probe, efx->net_dev,
			  "Too few action prios supported (have %u, need %u)\n",
			  efx->tc->caps->action_prios, EFX_TC_PRIO__NUM);
		return -EIO;
	}
	rc = efx_tc_configure_default_rule_pf(efx);
	if (rc)
		return rc;
	rc = efx_tc_configure_default_rule_wire(efx);
	if (rc)
		return rc;
	rc = efx_tc_configure_rep_mport(efx);
	if (rc)
		return rc;
	rc = efx_tc_configure_fallback_acts_pf(efx);
	if (rc)
		return rc;
	rc = efx_tc_configure_fallback_acts_reps(efx);
	if (rc)
		return rc;
	rc = efx_mae_get_tables(efx);
	if (rc)
		return rc;
	rc = flow_indr_dev_register(efx_tc_indr_setup_cb, efx);
	if (rc)
		goto out_free;
	efx->tc->up = true;
	return 0;
out_free:
	efx_mae_free_tables(efx);
	return rc;
}

void efx_fini_tc(struct efx_nic *efx)
{
	/* We can get called even if efx_init_struct_tc() failed */
	if (!efx->tc)
		return;
	if (efx->tc->up)
		flow_indr_dev_unregister(efx_tc_indr_setup_cb, efx, efx_tc_block_unbind);
	efx_tc_deconfigure_rep_mport(efx);
	efx_tc_deconfigure_default_rule(efx, &efx->tc->dflt.pf);
	efx_tc_deconfigure_default_rule(efx, &efx->tc->dflt.wire);
	efx_tc_deconfigure_fallback_acts(efx, &efx->tc->facts.pf);
	efx_tc_deconfigure_fallback_acts(efx, &efx->tc->facts.reps);
	efx->tc->up = false;
	efx_mae_free_tables(efx);
}

/* At teardown time, all TC filter rules (and thus all resources they created)
 * should already have been removed.  If we find any in our hashtables, make a
 * cursory attempt to clean up the software side.
 */
static void efx_tc_encap_match_free(void *ptr, void *__unused)
{
	struct efx_tc_encap_match *encap = ptr;

	WARN_ON(refcount_read(&encap->ref));
	kfree(encap);
}

static void efx_tc_recirc_free(void *ptr, void *arg)
{
	struct efx_tc_recirc_id *rid = ptr;
	struct efx_nic *efx = arg;

	WARN_ON(refcount_read(&rid->ref));
	ida_free(&efx->tc->recirc_ida, rid->fw_id);
	kfree(rid);
}

static void efx_tc_lhs_free(void *ptr, void *arg)
{
	struct efx_tc_lhs_rule *rule = ptr;
	struct efx_nic *efx = arg;

	netif_err(efx, drv, efx->net_dev,
		  "tc lhs_rule %lx still present at teardown, removing\n",
		  rule->cookie);

	if (rule->lhs_act.zone)
		efx_tc_ct_unregister_zone(efx, rule->lhs_act.zone);
	if (rule->lhs_act.count)
		efx_tc_flower_put_counter_index(efx, rule->lhs_act.count);
	efx_mae_remove_lhs_rule(efx, rule);

	kfree(rule);
}

static void efx_tc_mac_free(void *ptr, void *__unused)
{
	struct efx_tc_mac_pedit_action *ped = ptr;

	WARN_ON(refcount_read(&ped->ref));
	kfree(ped);
}

static void efx_tc_flow_free(void *ptr, void *arg)
{
	struct efx_tc_flow_rule *rule = ptr;
	struct efx_nic *efx = arg;

	netif_err(efx, drv, efx->net_dev,
		  "tc rule %lx still present at teardown, removing\n",
		  rule->cookie);

	/* Also releases entries in subsidiary tables */
	efx_tc_delete_rule(efx, rule);

	kfree(rule);
}

int efx_init_struct_tc(struct efx_nic *efx)
{
	int rc;

	if (efx->type->is_vf)
		return 0;

	efx->tc = kzalloc(sizeof(*efx->tc), GFP_KERNEL);
	if (!efx->tc)
		return -ENOMEM;
	efx->tc->caps = kzalloc(sizeof(struct mae_caps), GFP_KERNEL);
	if (!efx->tc->caps) {
		rc = -ENOMEM;
		goto fail_alloc_caps;
	}
	INIT_LIST_HEAD(&efx->tc->block_list);

	mutex_init(&efx->tc->mutex);
	init_waitqueue_head(&efx->tc->flush_wq);
	rc = efx_tc_init_encap_actions(efx);
	if (rc < 0)
		goto fail_encap_actions;
	rc = efx_tc_init_counters(efx);
	if (rc < 0)
		goto fail_counters;
	rc = rhashtable_init(&efx->tc->mac_ht, &efx_tc_mac_ht_params);
	if (rc < 0)
		goto fail_mac_ht;
	rc = rhashtable_init(&efx->tc->encap_match_ht, &efx_tc_encap_match_ht_params);
	if (rc < 0)
		goto fail_encap_match_ht;
	rc = rhashtable_init(&efx->tc->match_action_ht, &efx_tc_match_action_ht_params);
	if (rc < 0)
		goto fail_match_action_ht;
	rc = rhashtable_init(&efx->tc->lhs_rule_ht, &efx_tc_lhs_rule_ht_params);
	if (rc < 0)
		goto fail_lhs_rule_ht;
	rc = efx_tc_init_conntrack(efx);
	if (rc < 0)
		goto fail_conntrack;
	rc = rhashtable_init(&efx->tc->recirc_ht, &efx_tc_recirc_ht_params);
	if (rc < 0)
		goto fail_recirc_ht;
	ida_init(&efx->tc->recirc_ida);
	efx->tc->reps_filter_uc = -1;
	efx->tc->reps_filter_mc = -1;
	INIT_LIST_HEAD(&efx->tc->dflt.pf.acts.list);
	efx->tc->dflt.pf.fw_id = MC_CMD_MAE_ACTION_RULE_INSERT_OUT_ACTION_RULE_ID_NULL;
	INIT_LIST_HEAD(&efx->tc->dflt.wire.acts.list);
	efx->tc->dflt.wire.fw_id = MC_CMD_MAE_ACTION_RULE_INSERT_OUT_ACTION_RULE_ID_NULL;
	INIT_LIST_HEAD(&efx->tc->facts.pf.list);
	efx->tc->facts.pf.fw_id = MC_CMD_MAE_ACTION_SET_ALLOC_OUT_ACTION_SET_ID_NULL;
	INIT_LIST_HEAD(&efx->tc->facts.reps.list);
	efx->tc->facts.reps.fw_id = MC_CMD_MAE_ACTION_SET_ALLOC_OUT_ACTION_SET_ID_NULL;
	efx->extra_channel_type[EFX_EXTRA_CHANNEL_TC] = &efx_tc_channel_type;
	return 0;
fail_recirc_ht:
	efx_tc_destroy_conntrack(efx);
fail_conntrack:
	rhashtable_destroy(&efx->tc->lhs_rule_ht);
fail_lhs_rule_ht:
	rhashtable_destroy(&efx->tc->match_action_ht);
fail_match_action_ht:
	rhashtable_destroy(&efx->tc->encap_match_ht);
fail_encap_match_ht:
	rhashtable_destroy(&efx->tc->mac_ht);
fail_mac_ht:
	efx_tc_destroy_counters(efx);
fail_counters:
	efx_tc_destroy_encap_actions(efx);
fail_encap_actions:
	mutex_destroy(&efx->tc->mutex);
	kfree(efx->tc->caps);
fail_alloc_caps:
	kfree(efx->tc);
	efx->tc = NULL;
	return rc;
}

void efx_fini_struct_tc(struct efx_nic *efx)
{
	if (!efx->tc)
		return;

	mutex_lock(&efx->tc->mutex);
	EFX_WARN_ON_PARANOID(efx->tc->dflt.pf.fw_id !=
			     MC_CMD_MAE_ACTION_RULE_INSERT_OUT_ACTION_RULE_ID_NULL);
	EFX_WARN_ON_PARANOID(efx->tc->dflt.wire.fw_id !=
			     MC_CMD_MAE_ACTION_RULE_INSERT_OUT_ACTION_RULE_ID_NULL);
	EFX_WARN_ON_PARANOID(efx->tc->facts.pf.fw_id !=
			     MC_CMD_MAE_ACTION_SET_LIST_ALLOC_OUT_ACTION_SET_LIST_ID_NULL);
	EFX_WARN_ON_PARANOID(efx->tc->facts.reps.fw_id !=
			     MC_CMD_MAE_ACTION_SET_LIST_ALLOC_OUT_ACTION_SET_LIST_ID_NULL);
	rhashtable_free_and_destroy(&efx->tc->lhs_rule_ht, efx_tc_lhs_free, efx);
	rhashtable_free_and_destroy(&efx->tc->match_action_ht, efx_tc_flow_free,
				    efx);
	rhashtable_free_and_destroy(&efx->tc->encap_match_ht,
				    efx_tc_encap_match_free, NULL);
	efx_tc_fini_conntrack(efx);
	rhashtable_free_and_destroy(&efx->tc->recirc_ht, efx_tc_recirc_free, efx);
	WARN_ON(!ida_is_empty(&efx->tc->recirc_ida));
	ida_destroy(&efx->tc->recirc_ida);
	rhashtable_free_and_destroy(&efx->tc->mac_ht, efx_tc_mac_free, NULL);
	efx_tc_fini_counters(efx);
	efx_tc_fini_encap_actions(efx);
	mutex_unlock(&efx->tc->mutex);
	mutex_destroy(&efx->tc->mutex);
	kfree(efx->tc->caps);
	kfree(efx->tc);
	efx->tc = NULL;
}