aboutsummarylogtreecommitdiffstats
path: root/sys-kernel_arch-sources-g14_files-9007-squashed-net-tcp_bbr-bbr2-for-5.14.y.patch
blob: 4a242b28eda173719651551957a04deacc0605f7 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
From a0b13b7aa7bef3ed8ced3bcca60e3f7eac489f97 Mon Sep 17 00:00:00 2001
From: Scott B <arglebargle@arglebargle.dev>
Date: Sat, 4 Sep 2021 11:58:54 -0700
Subject: [PATCH] net-tcp_bbr: bbr2 for 5.14

Squashed commit of the following:

commit c44f6342dddc77fe7949fa655f38dba626089e8e
Author: Adithya Abraham Philip <abrahamphilip@google.com>
Date:   Fri Jun 11 21:56:10 2021 +0000

    net-tcp_bbr: v2: Fix missing ECT markings on retransmits for BBRv2

    Adds a new flag TCP_ECN_ECT_PERMANENT that is used by CCAs to
    indicate that retransmitted packets and pure ACKs must have the
    ECT bit set. This is a necessary fix for BBRv2, which when using
    ECN expects ECT to be set even on retransmitted packets and ACKs.
    Currently CCAs like BBRv2 which can use ECN but don't "need" it
    do not have a way to indicate that ECT should be set on
    retransmissions/ACKs.

    Signed-off-by: Adithya Abraham Philip <abrahamphilip@google.com>
    Signed-off-by: Neal Cardwell <ncardwell@google.com>

commit d7b290e75dfc3c6418efdd335a1426444f05a844
Author: Neal Cardwell <ncardwell@google.com>
Date:   Mon Dec 28 19:23:09 2020 -0500

    net-tcp_bbr: v2: don't assume prior_cwnd was set entering CA_Loss

    Fix WARN_ON_ONCE() warnings that were firing and pointing to a
    bbr->prior_cwnd of 0 when exiting CA_Loss and transitioning to
    CA_Open.

    The issue was that tcp_simple_retransmit() calls:

      tcp_set_ca_state(sk, TCP_CA_Loss);

    without first calling icsk_ca_ops->ssthresh(sk) (because
    tcp_simple_retransmit() is dealing with losses due to MTU issues and
    not congestion). The lack of this callback means that BBR did not get
    a chance to set bbr->prior_cwnd, and thus upon exiting CA_Loss in such
    cases the WARN_ON_ONCE() would fire due to a zero bbr->prior_cwnd.

    This commit removes that warning, since a bbr->prior_cwnd of 0 is a
    valid situation in this state transition.

    For setting inflight_lo upon entering CA_Loss, to avoid setting an
    inflight_lo of 0 in this case, this commit switches to taking the max
    of cwnd and prior_cwnd. We plan to remove that line of code when we
    switch to cautious (PRR-style) recovery, so that awkwardness will go
    away.

    Change-Id: I575dce871c2f20e91e3e9449e1706f42a07b8118

commit e9a0a3e20cd9ae66a7953f3acbd868888cb07109
Author: Neal Cardwell <ncardwell@google.com>
Date:   Mon Aug 17 19:10:21 2020 -0400

    net-tcp_bbr: v2: remove cycle_rand parameter that is unused in BBRv2

    Change-Id: Iee1df7e41e42de199068d7c89131ed3d228327c0

commit e452cc65ef66902908a8e20ad2ba1f01435ede08
Author: Neal Cardwell <ncardwell@google.com>
Date:   Mon Aug 17 19:08:41 2020 -0400

    net-tcp_bbr: v2: remove field bw_rtts that is unused in BBRv2

    Change-Id: I58e3346c707748a6f316f3ed060d2da84c32a79b

commit 097063f0b21fc7e8010b17517f16bf6872ca7ce5
Author: Neal Cardwell <ncardwell@google.com>
Date:   Thu Nov 21 15:28:01 2019 -0500

    net-tcp_bbr: v2: remove unnecessary rs.delivered_ce logic upon loss

    There is no reason to compute rs.delivered_ce upon loss.

    In fact, we specifically do not want to compute rs.delivered_ce upon loss.

    Two issues:

    (1) This would be the wrong thing to do, in behavior terms.  With
        RACK's dynamic reordering window, losses can be marked long after
        the sequence hole appears in the ACK/SACK stream. We want to to
        catch the ECN mark rate rising too high as quickly as possible,
        which means we want to check for high ECN mark rates at ACK time
        (as BBRv2 currently does) and not loss marking time.

    (2) This is dead code. The ECN mark rate cannot be detected as too
        high because the check needs rs->delivered to be > 0 as well:

           if (rs->delivered_ce > 0 && rs->delivered > 0 &&

        Since we are not setting rs->delivered upon loss, this check
        cannot succeed, so setting delivered_ce is pointless.

    This dead and wrong line was discovered by Randall Stewart at Netflix
    as he was reading the BBRv2 code.

    Change-Id: I37f83f418a259ec31d8f82de986db071b364b76a

commit 487953c510db2097dd095027a9c7f9f904ecdb17
Author: Neal Cardwell <ncardwell@google.com>
Date:   Tue Jun 11 12:54:22 2019 -0400

    net-tcp_bbr: v2: BBRv2 ("bbr2") congestion control for Linux TCP

    BBR v2 is an enhacement to the BBR v1 algorithm. It's designed to aim for lower
    queues, lower loss, and better Reno/CUBIC coexistence than BBR v1.

    BBR v2 maintains the core of BBR v1: an explicit model of the network
    path that is two-dimensional, adapting to estimate the (a) maximum
    available bandwidth and (b) maximum safe volume of data a flow can
    keep in-flight in the network. It maintains the estimated BDP as a
    core guide for estimating an appropriate level of in-flight data.

    BBR v2 makes several key enhancements:

    o Its bandwidth-probing time scale is adapted, within bounds, to allow improved
    coexistence with Reno and CUBIC. The bandwidth-probing time scale is (a)
    extended dynamically based on estimated BDP to improve coexistence with
    Reno/CUBIC; (b) bounded by an interactive wall-clock time-scale to be more
    scalable and responsive than Reno and CUBIC.

    o Rather than being largely agnostic to loss and ECN marks, it explicitly uses
    loss and (DCTCP-style) ECN signals to maintain its model.

    o It aims for lower losses than v1 by adjusting its model to attempt to stay
    within loss rate and ECN mark rate bounds (loss_thresh and ecn_thresh,
    respectively).

    o It adapts to loss/ECN signals even when the application is running out of
    data ("application-limited"), in case the "application-limited" flow is also
    "network-limited" (the bw and/or inflight available to this flow is lower than
    previously estimated when the flow ran out of data).

    o It has a three-part model: the model explicit three tracks operating points,
    where an operating point is a tuple: (bandwidth, inflight). The three operating
    points are:

      o latest:        the latest measurement from the current round trip
      o upper bound:   robust, optimistic, long-term upper bound
      o lower bound:   robust, conservative, short-term lower bound

    These are stored in the following state variables:

      o latest:  bw_latest, inflight_latest
      o lo:      bw_lo,     inflight_lo
      o hi:      bw_hi[2],  inflight_hi

    To gain intuition about the meaning of the three operating points, it
    may help to consider the analogs in CUBIC, which has a somewhat
    analogous three-part model used by its probing state machine:

      BBR param     CUBIC param
      -----------   -------------
      latest     ~  cwnd
      lo         ~  ssthresh
      hi         ~  last_max_cwnd

    The analogy is only a loose one, though, since the BBR operating
    points are calculated differently, and are 2-dimensional (bw,inflight)
    rather than CUBIC's one-dimensional notion of operating point
    (inflight).

    o It uses the three-part model to adapt the magnitude of its bandwidth
    to match the estimated space available in the buffer, rather than (as
    in BBR v1) assuming that it was always acceptable to place 0.25*BDP in
    the bottleneck buffer when probing (commodity datacenter switches
    commonly do not have that much buffer for WAN flows). When BBR v2
    estimates it hit a buffer limit during probing, its bandwidth probing
    then starts gently in case little space is still available in the
    buffer, and the accelerates, slowly at first and then rapidly if it
    can grow inflight without seeing congestion signals. In such cases,
    probing is bounded by inflight_hi + inflight_probe, where
    inflight_probe grows as: [0, 1, 2, 4, 8, 16,...]. This allows BBR to
    keep losses low and bounded if a bottleneck remains congested, while
    rapidly/scalably utilizing free bandwidth when it becomes available.

    o It has a slightly revised state machine, to achieve the goals above.
        BBR_BW_PROBE_UP:    pushes up inflight to probe for bw/vol
        BBR_BW_PROBE_DOWN:  drain excess inflight from the queue
        BBR_BW_PROBE_CRUISE: use pipe, w/ headroom in queue/pipe
        BBR_BW_PROBE_REFILL: try refill the pipe again to 100%, leaving queue empty

    o The estimated BDP: BBR v2 continues to maintain an estimate of the
    path's two-way propagation delay, by tracking a windowed min_rtt, and
    coordinating (on an as-ndeeded basis) to try to expose the two-way
    propagation delay by draining the bottleneck queue.

    BBR v2 continues to use its min_rtt and (currently-applicable) bandwidth
    estimate to estimate the current bandwidth-delay product. The estimated BDP
    still provides one important guideline for bounding inflight data. However,
    because any min-filtered RTT and max-filtered bw inherently tend to both
    overestimate, the estimated BDP is often too high; in this case loss or ECN
    marks can ensue, in which case BBR v2 adjusts inflight_hi and inflight_lo to
    adapt its sending rate and inflight down to match the available capacity of the
    path.

    o Space: Note that ICSK_CA_PRIV_SIZE increased. This is because BBR v2
    requires more space. Note that much of the space is due to support for
    per-socket parameterization and debugging in this release for research
    and debugging. With that state removed, the full "struct bbr" is 140
    bytes, or 144 with padding. This is an increase of 40 bytes over the
    existing ca_priv space.

    o Code: BBR v2 reuses many pieces from BBR v1. But it omits the following
      significant pieces:

      o "packet conservation" (bbr_set_cwnd_to_recover_or_restore(),
        bbr_can_grow_inflight())
      o long-term bandwidth estimator ("policer mode")

      The code layout tries to keep BBR v2 code near the bottom of the
      file, so that v1-applicable code in the top does not accidentally
      refer to v2 code.

    o Docs:
      See the following docs for more details and diagrams decsribing the BBR v2
      algorithm:
        https://datatracker.ietf.org/meeting/104/materials/slides-104-iccrg-an-update-on-bbr-00
        https://datatracker.ietf.org/meeting/102/materials/slides-102-iccrg-an-update-on-bbr-work-at-google-00

    o Internal notes:
      For this upstream rebase, Neal started from:
        git show fed518041ac6:net/ipv4/tcp_bbr.c > net/ipv4/tcp_bbr.c
      then removed dev instrumentation (dynamic get/set for parameters)
      and code that was only used by BBRv1

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: 2c84098e60bed6d67dde23cd7538c51dee273102
    Change-Id: I125cf26ba2a7a686f2fa5e87f4c2afceb65f7a05

commit 9fafce18324c72fb566a3c079b818d3cee757ef9
Author: Neal Cardwell <ncardwell@google.com>
Date:   Sat Nov 16 13:16:25 2019 -0500

    net-tcp: add fast_ack_mode=1: skip rwin check in tcp_fast_ack_mode__tcp_ack_snd_check()

    Add logic for an experimental TCP connection behavior, enabled with
    tp->fast_ack_mode = 1, which disables checking the receive window
    before sending an ack in __tcp_ack_snd_check(). If this behavior is
    enabled, the data receiver sends an ACK if the amount of data is >
    RCV.MSS.

    Change-Id: Iaa0a0fd7108221f883137a79d5bfa724f1b096d4

commit 65b716a609f5a6d4fd43d7185b2bd07d6525e904
Author: Neal Cardwell <ncardwell@google.com>
Date:   Fri Sep 27 17:10:26 2019 -0400

    net-tcp: re-generalize TSO sizing in TCP CC module API

    Reorganize the API for CC modules so that the CC module once again
    gets complete control of the TSO sizing decision. This is how the API
    was set up around 2016 and the initial BBRv1 upstreaming. Later Eric
    Dumazet simplified it. But with wider testing it now seems that to
    avoid CPU regressions BBR needs to have a different TSO sizing
    function.

    This is necessary to handle cases where there are many flows
    bottlenecked on the sender host's NIC, in which case BBR's pacing rate
    is much lower than CUBIC/Reno/DCTCP's. Why does this happen? Because
    BBR's pacing rate adapts to the low bandwidth share each flow sees. By
    contrast, CUBIC/Reno/DCTCP see no loss or ECN, so they grow a very
    large cwnd, and thus large pacing rate and large TSO burst size.

    Change-Id: Ic8ccfdbe4010ee8d4bf6a6334c48a2fceb2171ea

commit 857d19dbb3bbabb1d3da59405fc6ac57d17ebc14
Author: Yousuk Seung <ysseung@google.com>
Date:   Wed May 23 17:55:54 2018 -0700

    net-tcp: add new ca opts flag TCP_CONG_WANTS_CE_EVENTS

    Add a a new ca opts flag TCP_CONG_WANTS_CE_EVENTS that allows a
    congestion control module to receive CE events.

    Currently congestion control modules have to set the TCP_CONG_NEEDS_ECN
    bit in opts flag to receive CE events but this may incur changes in ECN
    behavior elsewhere. This patch adds a new bit TCP_CONG_WANTS_CE_EVENTS
    that allows congestion control modules to receive CE events
    independently of TCP_CONG_NEEDS_ECN.

    Effort: net-tcp
    Origin-9xx-SHA1: 9f7e14716cde760bc6c67ef8ef7e1ee48501d95b
    Change-Id: I2255506985242f376d910c6fd37daabaf4744f24

commit c1e3c81be5ef9832d7dc26dbe6c64dd7025f9503
Author: Neal Cardwell <ncardwell@google.com>
Date:   Tue May 7 22:37:19 2019 -0400

    net-tcp_bbr: v2: set tx.in_flight for skbs in repair write queue

    Syzkaller was able to use TCP_REPAIR to reproduce the new warning
    added in tcp_fragment():

      WARNING: CPU: 0 PID: 118174 at net/ipv4/tcp_output.c:1487
        tcp_fragment+0xdcc/0x10a0 net/ipv4/tcp_output.c:1487()
      inconsistent: tx.in_flight: 0 old_factor: 53

    The warning happens because skbs inserted into the tcp_rtx_queue
    during the repair process go through a sort of "fake send" process,
    and that process was seting pcount but not tx.in_flight, and thus the
    warnings (where old_factor is the old pcount).

    The fix of setting tx.in_flight in the TCP_REPAIR code path seems
    simple enough, and indeed makes the repro code from syzkaller stop
    producing warnings. Running through kokonut tests, and will send out
    for review when all tests pass.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: 330f825a08a6fe92cef74d799cc468864c479f63
    Change-Id: I0bc4a790f040fd4239620e1eedd5dc64666c6f05

commit 03c4ece7790dba1a88aebb68e960919ec9c587d2
Author: Neal Cardwell <ncardwell@google.com>
Date:   Wed May 1 20:16:25 2019 -0400

    net-tcp_bbr: v2: adjust skb tx.in_flight upon split in tcp_fragment()

    When we fragment an skb that has already been sent, we need to update
    the tx.in_flight for the first skb in the resulting pair ("buff").

    Because we were not updating the tx.in_flight, the tx.in_flight value
    was inconsistent with the pcount of the "buff" skb (tx.in_flight would
    be too high). That meant that if the "buff" skb was lost, then
    bbr2_inflight_hi_from_lost_skb() would calculate an inflight_hi value
    that is too high. This could result in longer queues and higher packet
    loss.

    Packetdrill testing verified that without this commit, when the second
    half of an skb is SACKed and then later the first half of that skb is
    marked lost, the calculated inflight_hi was incorrect.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: 385f1ddc610798fab2837f9f372857438b25f874
    Change-Id: I617f8cab4e9be7a0b8e8d30b047bf8645393354d

commit 1f57067a046edee861a5ab7daba10e9089ef7103
Author: Neal Cardwell <ncardwell@google.com>
Date:   Wed May 1 20:16:33 2019 -0400

    net-tcp_bbr: v2: adjust skb tx.in_flight upon merge in tcp_shifted_skb()

    When tcp_shifted_skb() updates state as adjacent SACKed skbs are
    coalesced, previously the tx.in_flight was not adjusted, so we could
    get contradictory state where the skb's recorded pcount was bigger
    than the tx.in_flight (the number of segments that were in_flight
    after sending the skb).

    Normally have a SACKed skb with contradictory pcount/tx.in_flight
    would not matter. However, with SACK reneging, the SACKed bit is
    removed, and an skb once again becomes eligible for retransmitting,
    fragmenting, SACKing, etc. Packetdrill testing verified the following
    sequence is possible in a kernel that does not have this commit:

     - skb N is SACKed
     - skb N+1 is SACKed and combined with skb N using tcp_shifted_skb()
       - tcp_shifted_skb() will increase the pcount of prev,
         but leave tx.in_flight as-is
       - so prev skb can have pcount > tx.in_flight
     - RTO, tcp_timeout_mark_lost(), detect reneg,
       remove "SACKed" bit, mark skb N as lost
       - find pcount of skb N is greater than its tx.in_flight

    I suspect this issue iw what caused the bbr2_inflight_hi_from_lost_skb():
      WARN_ON_ONCE(inflight_prev < 0)
    to fire in production machines using bbr2.

    Tested: See last commit in series for sponge link.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: 1a3e997e613d2dcf32b947992882854ebe873715
    Change-Id: I1b0b75c27519953430c7db51c6f358f104c7af55

commit dfaf2f3907da754b58996bdd6d22139d8ae7fbd7
Author: Neal Cardwell <ncardwell@google.com>
Date:   Tue May 7 22:36:36 2019 -0400

    net-tcp_bbr: v2: factor out tx.in_flight setting into tcp_set_tx_in_flight()

    Factor out the code to set an skb's tx.in_flight field into its own
    function, so that this code can be used for the TCP_REPAIR "fake send"
    code path that inserts skbs into the rtx queue without sending
    them. This is in preparation for the following patch, which fixes an
    issue with TCP_REPAIR and tx.in_flight.

    Tested: See last patch in series for sponge link.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: e880fc907d06ea7354333f60f712748ebce9497b
    Change-Id: I4fbd4a6e18a51ab06d50ab1c9ad820ce5bea89af

commit bf1996458d8348f65ec52b651ea386e47ba1db33
Author: Neal Cardwell <ncardwell@google.com>
Date:   Tue Aug 7 21:52:06 2018 -0400

    net-tcp_bbr: v2: introduce ca_ops->skb_marked_lost() CC module callback API

    For connections experiencing reordering, RACK can mark packets lost
    long after we receive the SACKs/ACKs hinting that the packets were
    actually lost.

    This means that CC modules cannot easily learn the volume of inflight
    data at which packet loss happens by looking at the current inflight
    or even the packets in flight when the most recently SACKed packet was
    sent. To learn this, CC modules need to know how many packets were in
    flight at the time lost packets were sent. This new callback, combined
    with TCP_SKB_CB(skb)->tx.in_flight, allows them to learn this.

    This also provides a consistent callback that is invoked whether
    packets are marked lost upon ACK processing, using the RACK reordering
    timer, or at RTO time.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: afcbebe3374e4632ac6714d39e4dc8a8455956f4
    Change-Id: I54826ab53df636be537e5d3c618a46145d12d51a

commit 8efc97365ed81bc15bb665fe036835ed511530e1
Author: Neal Cardwell <ncardwell@google.com>
Date:   Mon Nov 19 13:48:36 2018 -0500

    net-tcp_bbr: v2: export FLAG_ECE in rate_sample.is_ece

    For understanding the relationship between inflight and ECN signals,
    to try to find the highest inflight value that has acceptable levels
    ECN marking.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: 3eba998f2898541406c2666781182200934965a8
    Change-Id: I3a964e04cee83e11649a54507043d2dfe769a3b3

commit 81eada6998cb810e67a056b99175ef7ca82d180a
Author: Neal Cardwell <ncardwell@google.com>
Date:   Thu Oct 12 23:44:27 2017 -0400

    net-tcp_bbr: v2: count packets lost over TCP rate sampling interval

    For understanding the relationship between inflight and packet loss
    signals, to try to find the highest inflight value that has acceptable
    levels of packet losses.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: 4527e26b2bd7756a88b5b9ef1ada3da33dd609ab
    Change-Id: I594c2500868d9c530770e7ddd68ffc87c57f4fd5

commit ed57cac80e24315a6502fbff52218a95040ccba3
Author: Neal Cardwell <ncardwell@google.com>
Date:   Sat Aug 5 11:49:50 2017 -0400

    net-tcp_bbr: v2: snapshot packets in flight at transmit time and pass in rate_sample

    For understanding the relationship between inflight and losses or ECN
    signals, to try to find the highest inflight value that has acceptable
    levels of loss/ECN marking.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: b3eb4f2d20efab4ca001f32c9294739036c493ea
    Change-Id: I7314047d0ff14dd261a04b1969a46dc658c8836a

commit a33083242c9efc580e397b587d3c660e9d371098
Author: Neal Cardwell <ncardwell@google.com>
Date:   Sun Jun 24 21:55:59 2018 -0400

    net-tcp_bbr: v2: shrink delivered_mstamp, first_tx_mstamp to u32 to free up 8 bytes

    Free up some space for tracking inflight and losses for each
    bw sample, in upcoming commits.

    These timestamps are in microseconds, and are now stored in 32
    bits. So they can only hold time intervals up to roughly 2^12 = 4096
    seconds.  But Linux TCP RTT and RTO tracking has the same 32-bit
    microsecond implementation approach and resulting deployment
    limitations. So this is not introducing a new limit. And these should
    not be a limitation for the foreseeable future.

    Effort: net-tcp_bbr
    Origin-9xx-SHA1: 238a7e6b5d51625fef1ce7769826a7b21b02ae55
    Change-Id: I3b779603797263b52a61ad57c565eb91fe42680c

commit d3a59b8de6fc7493ed2db7d9d386e2c74f4e8938
Author: Yuchung Cheng <ycheng@google.com>
Date:   Tue Mar 27 18:01:46 2018 -0700

    net-tcp_rate: account for CE marks in rate sample

    This patch counts number of packets delivered have CE mark in the
    rate sample, using similar approach of delivery accounting.

    Effort: net-tcp_rate
    Origin-9xx-SHA1: 710644db434c3da335a7c8b72207a671ccbb5cf8
    Change-Id: I0968fb33fe19b5c774e8c3afd2685558a6ec8710

commit 21e06050c8ef2995a4c99bb9727452c6b0c005e3
Author: Yuchung Cheng <ycheng@google.com>
Date:   Tue Mar 27 18:33:29 2018 -0700

    net-tcp_rate: consolidate inflight tracking approaches in TCP

    In order to track CE marks per rate sample (one round trip), we'll
    need to snap the starting tcp delivered_ce acount in the packet
    meta header (tcp_skb_cb). But there's not enough space.

    Good news is that the "last_in_flight" in the header, used by
    NV congestion control, is almost equivalent as "delivered". In
    fact "delivered" is better by accounting out-of-order packets
    additionally.  Therefore we can remove it to make room for the
    CE tracking.

    This would make delayed ACK detection slightly less accurate but the
    impact is negligible since it's not used for any critical control.

    Effort: net-tcp_rate
    Origin-9xx-SHA1: ddcd46ec85d5f1c4454258af0c54b3254c0d64a7
    Change-Id: I1a184aad6d101c981ac7f2f275aa9417ff856910

commit a0f76e629209066ed4f0aa227ab61882952ba0d4
Author: Neal Cardwell <ncardwell@google.com>
Date:   Tue Jun 11 12:26:55 2019 -0400

    net-tcp_bbr: broaden app-limited rate sample detection

    This commit is a bug fix for the Linux TCP app-limited
    (application-limited) logic that is used for collecting rate
    (bandwidth) samples.

    Previously the app-limited logic only looked for "bubbles" of
    silence in between application writes, by checking at the start
    of each sendmsg. But "bubbles" of silence can also happen before
    retransmits: e.g. bubbles can happen between an application write
    and a retransmit, or between two retransmits.

    Retransmits are triggered by ACKs or timers. So this commit checks
    for bubbles of app-limited silence upon ACKs or timers.

    Why does this commit check for app-limited state at the start of
    ACKs and timer handling? Because at that point we know whether
    inflight was fully using the cwnd.  During processing the ACK or
    timer event we often change the cwnd; after changing the cwnd we
    can't know whether inflight was fully using the old cwnd.

    Origin-9xx-SHA1: 3fe9b53291e018407780fb8c356adb5666722cbc
    Change-Id: I37221506f5166877c2b110753d39bb0757985e68
---
 include/linux/tcp.h                |    3 +-
 include/net/inet_connection_sock.h |    5 +-
 include/net/tcp.h                  |   50 +-
 include/uapi/linux/inet_diag.h     |   33 +
 net/ipv4/Kconfig                   |   22 +
 net/ipv4/Makefile                  |    1 +
 net/ipv4/bpf_tcp_ca.c              |    2 +-
 net/ipv4/tcp.c                     |    1 +
 net/ipv4/tcp_bbr.c                 |   38 +-
 net/ipv4/tcp_bbr2.c                | 2674 ++++++++++++++++++++++++++++
 net/ipv4/tcp_cong.c                |    1 +
 net/ipv4/tcp_input.c               |   38 +-
 net/ipv4/tcp_output.c              |   28 +-
 net/ipv4/tcp_rate.c                |   36 +-
 net/ipv4/tcp_timer.c               |    1 +
 15 files changed, 2886 insertions(+), 47 deletions(-)
 create mode 100644 net/ipv4/tcp_bbr2.c

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 48d8a363319e..1bd559c69e83 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -225,7 +225,8 @@ struct tcp_sock {
 	u8	compressed_ack;
 	u8	dup_ack_counter:2,
 		tlp_retrans:1,	/* TLP is a retransmission */
-		unused:5;
+		fast_ack_mode:2, /* which fast ack mode ? */
+		unused:3;
 	u32	chrono_start;	/* Start time in jiffies of a TCP chrono */
 	u32	chrono_stat[3];	/* Time in jiffies for chrono_stat stats */
 	u8	chrono_type:2,	/* current chronograph type */
diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h
index b06c2d02ec84..2cdc5a0709fe 100644
--- a/include/net/inet_connection_sock.h
+++ b/include/net/inet_connection_sock.h
@@ -134,8 +134,9 @@ struct inet_connection_sock {
 	u32			  icsk_probes_tstamp;
 	u32			  icsk_user_timeout;
 
-	u64			  icsk_ca_priv[104 / sizeof(u64)];
-#define ICSK_CA_PRIV_SIZE	  sizeof_field(struct inet_connection_sock, icsk_ca_priv)
+/* XXX inflated by temporary internal debugging info */
+#define ICSK_CA_PRIV_SIZE      (216)
+	u64			  icsk_ca_priv[ICSK_CA_PRIV_SIZE / sizeof(u64)];
 };
 
 #define ICSK_TIME_RETRANS	1	/* Retransmit timer */
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 784d5c3ef1c5..daa878118bd3 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -371,6 +371,7 @@ static inline void tcp_dec_quickack_mode(struct sock *sk,
 #define	TCP_ECN_QUEUE_CWR	2
 #define	TCP_ECN_DEMAND_CWR	4
 #define	TCP_ECN_SEEN		8
+#define	TCP_ECN_ECT_PERMANENT	16
 
 enum tcp_tw_status {
 	TCP_TW_SUCCESS = 0,
@@ -807,6 +808,11 @@ static inline u32 tcp_stamp_us_delta(u64 t1, u64 t0)
 	return max_t(s64, t1 - t0, 0);
 }
 
+static inline u32 tcp_stamp32_us_delta(u32 t1, u32 t0)
+{
+	return max_t(s32, t1 - t0, 0);
+}
+
 static inline u32 tcp_skb_timestamp(const struct sk_buff *skb)
 {
 	return tcp_ns_to_ts(skb->skb_mstamp_ns);
@@ -874,16 +880,22 @@ struct tcp_skb_cb {
 	__u32		ack_seq;	/* Sequence number ACK'd	*/
 	union {
 		struct {
+#define TCPCB_DELIVERED_CE_MASK ((1U<<20) - 1)
 			/* There is space for up to 24 bytes */
-			__u32 in_flight:30,/* Bytes in flight at transmit */
-			      is_app_limited:1, /* cwnd not fully used? */
-			      unused:1;
+			__u32 is_app_limited:1, /* cwnd not fully used? */
+			      delivered_ce:20,
+			      unused:11;
 			/* pkts S/ACKed so far upon tx of skb, incl retrans: */
 			__u32 delivered;
 			/* start of send pipeline phase */
-			u64 first_tx_mstamp;
+			u32 first_tx_mstamp;
 			/* when we reached the "delivered" count */
-			u64 delivered_mstamp;
+			u32 delivered_mstamp;
+#define TCPCB_IN_FLIGHT_BITS 20
+#define TCPCB_IN_FLIGHT_MAX ((1U << TCPCB_IN_FLIGHT_BITS) - 1)
+			u32 in_flight:20,   /* packets in flight at transmit */
+			    unused2:12;
+			u32 lost;	/* packets lost so far upon tx of skb */
 		} tx;   /* only used for outgoing skbs */
 		union {
 			struct inet_skb_parm	h4;
@@ -1008,7 +1020,11 @@ enum tcp_ca_ack_event_flags {
 #define TCP_CONG_NON_RESTRICTED 0x1
 /* Requires ECN/ECT set on all packets */
 #define TCP_CONG_NEEDS_ECN	0x2
-#define TCP_CONG_MASK	(TCP_CONG_NON_RESTRICTED | TCP_CONG_NEEDS_ECN)
+/* Wants notification of CE events (CA_EVENT_ECN_IS_CE, CA_EVENT_ECN_NO_CE). */
+#define TCP_CONG_WANTS_CE_EVENTS	0x4
+#define TCP_CONG_MASK	(TCP_CONG_NON_RESTRICTED | \
+			 TCP_CONG_NEEDS_ECN | \
+			 TCP_CONG_WANTS_CE_EVENTS)
 
 union tcp_cc_info;
 
@@ -1028,8 +1044,13 @@ struct ack_sample {
  */
 struct rate_sample {
 	u64  prior_mstamp; /* starting timestamp for interval */
+	u32  prior_lost;	/* tp->lost at "prior_mstamp" */
 	u32  prior_delivered;	/* tp->delivered at "prior_mstamp" */
+	u32  prior_delivered_ce;/* tp->delivered_ce at "prior_mstamp" */
+	u32 tx_in_flight;	/* packets in flight at starting timestamp */
+	s32  lost;		/* number of packets lost over interval */
 	s32  delivered;		/* number of packets delivered over interval */
+	s32  delivered_ce;	/* packets delivered w/ CE mark over interval */
 	long interval_us;	/* time for tp->delivered to incr "delivered" */
 	u32 snd_interval_us;	/* snd interval for delivered packets */
 	u32 rcv_interval_us;	/* rcv interval for delivered packets */
@@ -1040,6 +1061,7 @@ struct rate_sample {
 	bool is_app_limited;	/* is sample from packet with bubble in pipe? */
 	bool is_retrans;	/* is sample from retransmission? */
 	bool is_ack_delayed;	/* is this (likely) a delayed ACK? */
+	bool is_ece;		/* did this ACK have ECN marked? */
 };
 
 struct tcp_congestion_ops {
@@ -1063,8 +1085,11 @@ struct tcp_congestion_ops {
 	/* hook for packet ack accounting (optional) */
 	void (*pkts_acked)(struct sock *sk, const struct ack_sample *sample);
 
-	/* override sysctl_tcp_min_tso_segs */
-	u32 (*min_tso_segs)(struct sock *sk);
+	/* pick target number of segments per TSO/GSO skb (optional): */
+	u32 (*tso_segs)(struct sock *sk, unsigned int mss_now);
+
+	/* react to a specific lost skb (optional) */
+	void (*skb_marked_lost)(struct sock *sk, const struct sk_buff *skb);
 
 	/* call when packets are delivered to update cwnd and pacing rate,
 	 * after all the ca_state processing. (optional)
@@ -1127,6 +1152,14 @@ static inline char *tcp_ca_get_name_by_key(u32 key, char *buffer)
 }
 #endif
 
+static inline bool tcp_ca_wants_ce_events(const struct sock *sk)
+{
+	const struct inet_connection_sock *icsk = inet_csk(sk);
+
+	return icsk->icsk_ca_ops->flags & (TCP_CONG_NEEDS_ECN |
+					   TCP_CONG_WANTS_CE_EVENTS);
+}
+
 static inline bool tcp_ca_needs_ecn(const struct sock *sk)
 {
 	const struct inet_connection_sock *icsk = inet_csk(sk);
@@ -1152,6 +1185,7 @@ static inline void tcp_ca_event(struct sock *sk, const enum tcp_ca_event event)
 }
 
 /* From tcp_rate.c */
+void tcp_set_tx_in_flight(struct sock *sk, struct sk_buff *skb);
 void tcp_rate_skb_sent(struct sock *sk, struct sk_buff *skb);
 void tcp_rate_skb_delivered(struct sock *sk, struct sk_buff *skb,
 			    struct rate_sample *rs);
diff --git a/include/uapi/linux/inet_diag.h b/include/uapi/linux/inet_diag.h
index 20ee93f0f876..96d52dd9c48a 100644
--- a/include/uapi/linux/inet_diag.h
+++ b/include/uapi/linux/inet_diag.h
@@ -231,9 +231,42 @@ struct tcp_bbr_info {
 	__u32	bbr_cwnd_gain;		/* cwnd gain shifted left 8 bits */
 };
 
+/* Phase as reported in netlink/ss stats. */
+enum tcp_bbr2_phase {
+	BBR2_PHASE_INVALID		= 0,
+	BBR2_PHASE_STARTUP		= 1,
+	BBR2_PHASE_DRAIN		= 2,
+	BBR2_PHASE_PROBE_RTT		= 3,
+	BBR2_PHASE_PROBE_BW_UP		= 4,
+	BBR2_PHASE_PROBE_BW_DOWN	= 5,
+	BBR2_PHASE_PROBE_BW_CRUISE	= 6,
+	BBR2_PHASE_PROBE_BW_REFILL	= 7
+};
+
+struct tcp_bbr2_info {
+	/* u64 bw: bandwidth (app throughput) estimate in Byte per sec: */
+	__u32	bbr_bw_lsb;		/* lower 32 bits of bw */
+	__u32	bbr_bw_msb;		/* upper 32 bits of bw */
+	__u32	bbr_min_rtt;		/* min-filtered RTT in uSec */
+	__u32	bbr_pacing_gain;	/* pacing gain shifted left 8 bits */
+	__u32	bbr_cwnd_gain;		/* cwnd gain shifted left 8 bits */
+	__u32	bbr_bw_hi_lsb;		/* lower 32 bits of bw_hi */
+	__u32	bbr_bw_hi_msb;		/* upper 32 bits of bw_hi */
+	__u32	bbr_bw_lo_lsb;		/* lower 32 bits of bw_lo */
+	__u32	bbr_bw_lo_msb;		/* upper 32 bits of bw_lo */
+	__u8	bbr_mode;		/* current bbr_mode in state machine */
+	__u8	bbr_phase;		/* current state machine phase */
+	__u8	unused1;		/* alignment padding; not used yet */
+	__u8	bbr_version;		/* MUST be at this offset in struct */
+	__u32	bbr_inflight_lo;	/* lower/short-term data volume bound */
+	__u32	bbr_inflight_hi;	/* higher/long-term data volume bound */
+	__u32	bbr_extra_acked;	/* max excess packets ACKed in epoch */
+};
+
 union tcp_cc_info {
 	struct tcpvegas_info	vegas;
 	struct tcp_dctcp_info	dctcp;
 	struct tcp_bbr_info	bbr;
+	struct tcp_bbr2_info	bbr2;
 };
 #endif /* _UAPI_INET_DIAG_H_ */
diff --git a/net/ipv4/Kconfig b/net/ipv4/Kconfig
index 87983e70f03f..a833a7a67ce7 100644
--- a/net/ipv4/Kconfig
+++ b/net/ipv4/Kconfig
@@ -669,6 +669,24 @@ config TCP_CONG_BBR
 	  AQM schemes that do not provide a delay signal. It requires the fq
 	  ("Fair Queue") pacing packet scheduler.
 
+config TCP_CONG_BBR2
+	tristate "BBR2 TCP"
+	default m
+	help
+
+	BBR2 TCP congestion control is a model-based congestion control
+	algorithm that aims to maximize network utilization, keep queues and
+	retransmit rates low, and to be able to coexist with Reno/CUBIC in
+	common scenarios. It builds an explicit model of the network path.  It
+	tolerates a targeted degree of random packet loss and delay that are
+	unrelated to congestion. It can operate over LAN, WAN, cellular, wifi,
+	or cable modem links, and can use DCTCP-L4S-style ECN signals.  It can
+	coexist with flows that use loss-based congestion control, and can
+	operate with shallow buffers, deep buffers, bufferbloat, policers, or
+	AQM schemes that do not provide a delay signal. It requires pacing,
+	using either TCP internal pacing or the fq ("Fair Queue") pacing packet
+	scheduler.
+
 choice
 	prompt "Default TCP congestion control"
 	default DEFAULT_CUBIC
@@ -706,6 +724,9 @@ choice
 	config DEFAULT_BBR
 		bool "BBR" if TCP_CONG_BBR=y
 
+	config DEFAULT_BBR2
+		bool "BBR2" if TCP_CONG_BBR2=y
+
 	config DEFAULT_RENO
 		bool "Reno"
 endchoice
@@ -730,6 +751,7 @@ config DEFAULT_TCP_CONG
 	default "dctcp" if DEFAULT_DCTCP
 	default "cdg" if DEFAULT_CDG
 	default "bbr" if DEFAULT_BBR
+	default "bbr2" if DEFAULT_BBR2
 	default "cubic"
 
 config TCP_MD5SIG
diff --git a/net/ipv4/Makefile b/net/ipv4/Makefile
index bbdd9c44f14e..8dee1547d820 100644
--- a/net/ipv4/Makefile
+++ b/net/ipv4/Makefile
@@ -46,6 +46,7 @@ obj-$(CONFIG_INET_TCP_DIAG) += tcp_diag.o
 obj-$(CONFIG_INET_UDP_DIAG) += udp_diag.o
 obj-$(CONFIG_INET_RAW_DIAG) += raw_diag.o
 obj-$(CONFIG_TCP_CONG_BBR) += tcp_bbr.o
+obj-$(CONFIG_TCP_CONG_BBR2) += tcp_bbr2.o
 obj-$(CONFIG_TCP_CONG_BIC) += tcp_bic.o
 obj-$(CONFIG_TCP_CONG_CDG) += tcp_cdg.o
 obj-$(CONFIG_TCP_CONG_CUBIC) += tcp_cubic.o
diff --git a/net/ipv4/bpf_tcp_ca.c b/net/ipv4/bpf_tcp_ca.c
index 9e41eff4a685..f26b0fcbb9f5 100644
--- a/net/ipv4/bpf_tcp_ca.c
+++ b/net/ipv4/bpf_tcp_ca.c
@@ -17,7 +17,7 @@ static u32 optional_ops[] = {
 	offsetof(struct tcp_congestion_ops, cwnd_event),
 	offsetof(struct tcp_congestion_ops, in_ack_event),
 	offsetof(struct tcp_congestion_ops, pkts_acked),
-	offsetof(struct tcp_congestion_ops, min_tso_segs),
+	offsetof(struct tcp_congestion_ops, tso_segs),
 	offsetof(struct tcp_congestion_ops, sndbuf_expand),
 	offsetof(struct tcp_congestion_ops, cong_control),
 };
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 8cb44040ec68..68c8eee7f831 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -3047,6 +3047,7 @@ int tcp_disconnect(struct sock *sk, int flags)
 	tp->rx_opt.dsack = 0;
 	tp->rx_opt.num_sacks = 0;
 	tp->rcv_ooopack = 0;
+	tp->fast_ack_mode = 0;
 
 
 	/* Clean up fastopen related fields */
diff --git a/net/ipv4/tcp_bbr.c b/net/ipv4/tcp_bbr.c
index 6274462b86b4..c0d5a4211fc1 100644
--- a/net/ipv4/tcp_bbr.c
+++ b/net/ipv4/tcp_bbr.c
@@ -292,26 +292,40 @@ static void bbr_set_pacing_rate(struct sock *sk, u32 bw, int gain)
 		sk->sk_pacing_rate = rate;
 }
 
-/* override sysctl_tcp_min_tso_segs */
 static u32 bbr_min_tso_segs(struct sock *sk)
 {
 	return sk->sk_pacing_rate < (bbr_min_tso_rate >> 3) ? 1 : 2;
 }
 
+/* Return the number of segments BBR would like in a TSO/GSO skb, given
+ * a particular max gso size as a constraint.
+ */
+static u32 bbr_tso_segs_generic(struct sock *sk, unsigned int mss_now,
+				u32 gso_max_size)
+{
+	u32 segs;
+	u64 bytes;
+
+	/* Budget a TSO/GSO burst size allowance based on bw (pacing_rate). */
+	bytes = sk->sk_pacing_rate >> sk->sk_pacing_shift;
+
+	bytes = min_t(u32, bytes, gso_max_size - 1 - MAX_TCP_HEADER);
+	segs = max_t(u32, bytes / mss_now, bbr_min_tso_segs(sk));
+	return segs;
+}
+
+/* Custom tcp_tso_autosize() for BBR, used at transmit time to cap skb size. */
+static u32  bbr_tso_segs(struct sock *sk, unsigned int mss_now)
+{
+	return bbr_tso_segs_generic(sk, mss_now, sk->sk_gso_max_size);
+}
+
+/* Like bbr_tso_segs(), using mss_cache, ignoring driver's sk_gso_max_size. */
 static u32 bbr_tso_segs_goal(struct sock *sk)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
-	u32 segs, bytes;
 
-	/* Sort of tcp_tso_autosize() but ignoring
-	 * driver provided sk_gso_max_size.
-	 */
-	bytes = min_t(unsigned long,
-		      sk->sk_pacing_rate >> READ_ONCE(sk->sk_pacing_shift),
-		      GSO_MAX_SIZE - 1 - MAX_TCP_HEADER);
-	segs = max_t(u32, bytes / tp->mss_cache, bbr_min_tso_segs(sk));
-
-	return min(segs, 0x7FU);
+	return  bbr_tso_segs_generic(sk, tp->mss_cache, GSO_MAX_SIZE);
 }
 
 /* Save "last known good" cwnd so we can restore it after losses or PROBE_RTT */
@@ -1147,7 +1161,7 @@ static struct tcp_congestion_ops tcp_bbr_cong_ops __read_mostly = {
 	.undo_cwnd	= bbr_undo_cwnd,
 	.cwnd_event	= bbr_cwnd_event,
 	.ssthresh	= bbr_ssthresh,
-	.min_tso_segs	= bbr_min_tso_segs,
+	.tso_segs	= bbr_tso_segs,
 	.get_info	= bbr_get_info,
 	.set_state	= bbr_set_state,
 };
diff --git a/net/ipv4/tcp_bbr2.c b/net/ipv4/tcp_bbr2.c
new file mode 100644
index 000000000000..fa49e17c47ca
--- /dev/null
+++ b/net/ipv4/tcp_bbr2.c
@@ -0,0 +1,2674 @@
+/* BBR (Bottleneck Bandwidth and RTT) congestion control, v2
+ *
+ * BBRv2 is a model-based congestion control algorithm that aims for low
+ * queues, low loss, and (bounded) Reno/CUBIC coexistence. To maintain a model
+ * of the network path, it uses measurements of bandwidth and RTT, as well as
+ * (if they occur) packet loss and/or DCTCP/L4S-style ECN signals.  Note that
+ * although it can use ECN or loss signals explicitly, it does not require
+ * either; it can bound its in-flight data based on its estimate of the BDP.
+ *
+ * The model has both higher and lower bounds for the operating range:
+ *   lo: bw_lo, inflight_lo: conservative short-term lower bound
+ *   hi: bw_hi, inflight_hi: robust long-term upper bound
+ * The bandwidth-probing time scale is (a) extended dynamically based on
+ * estimated BDP to improve coexistence with Reno/CUBIC; (b) bounded by
+ * an interactive wall-clock time-scale to be more scalable and responsive
+ * than Reno and CUBIC.
+ *
+ * Here is a state transition diagram for BBR:
+ *
+ *             |
+ *             V
+ *    +---> STARTUP  ----+
+ *    |        |         |
+ *    |        V         |
+ *    |      DRAIN   ----+
+ *    |        |         |
+ *    |        V         |
+ *    +---> PROBE_BW ----+
+ *    |      ^    |      |
+ *    |      |    |      |
+ *    |      +----+      |
+ *    |                  |
+ *    +---- PROBE_RTT <--+
+ *
+ * A BBR flow starts in STARTUP, and ramps up its sending rate quickly.
+ * When it estimates the pipe is full, it enters DRAIN to drain the queue.
+ * In steady state a BBR flow only uses PROBE_BW and PROBE_RTT.
+ * A long-lived BBR flow spends the vast majority of its time remaining
+ * (repeatedly) in PROBE_BW, fully probing and utilizing the pipe's bandwidth
+ * in a fair manner, with a small, bounded queue. *If* a flow has been
+ * continuously sending for the entire min_rtt window, and hasn't seen an RTT
+ * sample that matches or decreases its min_rtt estimate for 10 seconds, then
+ * it briefly enters PROBE_RTT to cut inflight to a minimum value to re-probe
+ * the path's two-way propagation delay (min_rtt). When exiting PROBE_RTT, if
+ * we estimated that we reached the full bw of the pipe then we enter PROBE_BW;
+ * otherwise we enter STARTUP to try to fill the pipe.
+ *
+ * BBR is described in detail in:
+ *   "BBR: Congestion-Based Congestion Control",
+ *   Neal Cardwell, Yuchung Cheng, C. Stephen Gunn, Soheil Hassas Yeganeh,
+ *   Van Jacobson. ACM Queue, Vol. 14 No. 5, September-October 2016.
+ *
+ * There is a public e-mail list for discussing BBR development and testing:
+ *   https://groups.google.com/forum/#!forum/bbr-dev
+ *
+ * NOTE: BBR might be used with the fq qdisc ("man tc-fq") with pacing enabled,
+ * otherwise TCP stack falls back to an internal pacing using one high
+ * resolution timer per TCP socket and may use more resources.
+ */
+#include <linux/module.h>
+#include <net/tcp.h>
+#include <linux/inet_diag.h>
+#include <linux/inet.h>
+#include <linux/random.h>
+
+#include "tcp_dctcp.h"
+
+/* Scale factor for rate in pkt/uSec unit to avoid truncation in bandwidth
+ * estimation. The rate unit ~= (1500 bytes / 1 usec / 2^24) ~= 715 bps.
+ * This handles bandwidths from 0.06pps (715bps) to 256Mpps (3Tbps) in a u32.
+ * Since the minimum window is >=4 packets, the lower bound isn't
+ * an issue. The upper bound isn't an issue with existing technologies.
+ */
+#define BW_SCALE 24
+#define BW_UNIT (1 << BW_SCALE)
+
+#define BBR_SCALE 8	/* scaling factor for fractions in BBR (e.g. gains) */
+#define BBR_UNIT (1 << BBR_SCALE)
+
+#define FLAG_DEBUG_VERBOSE	0x1	/* Verbose debugging messages */
+#define FLAG_DEBUG_LOOPBACK	0x2	/* Do NOT skip loopback addr */
+
+#define CYCLE_LEN		8	/* number of phases in a pacing gain cycle */
+
+/* BBR has the following modes for deciding how fast to send: */
+enum bbr_mode {
+	BBR_STARTUP,	/* ramp up sending rate rapidly to fill pipe */
+	BBR_DRAIN,	/* drain any queue created during startup */
+	BBR_PROBE_BW,	/* discover, share bw: pace around estimated bw */
+	BBR_PROBE_RTT,	/* cut inflight to min to probe min_rtt */
+};
+
+/* How does the incoming ACK stream relate to our bandwidth probing? */
+enum bbr_ack_phase {
+	BBR_ACKS_INIT,		  /* not probing; not getting probe feedback */
+	BBR_ACKS_REFILLING,	  /* sending at est. bw to fill pipe */
+	BBR_ACKS_PROBE_STARTING,  /* inflight rising to probe bw */
+	BBR_ACKS_PROBE_FEEDBACK,  /* getting feedback from bw probing */
+	BBR_ACKS_PROBE_STOPPING,  /* stopped probing; still getting feedback */
+};
+
+/* BBR congestion control block */
+struct bbr {
+	u32	min_rtt_us;	        /* min RTT in min_rtt_win_sec window */
+	u32	min_rtt_stamp;	        /* timestamp of min_rtt_us */
+	u32	probe_rtt_done_stamp;   /* end time for BBR_PROBE_RTT mode */
+	u32	probe_rtt_min_us;	/* min RTT in bbr_probe_rtt_win_ms window */
+	u32	probe_rtt_min_stamp;	/* timestamp of probe_rtt_min_us*/
+	u32     next_rtt_delivered; /* scb->tx.delivered at end of round */
+	u32	prior_rcv_nxt;	/* tp->rcv_nxt when CE state last changed */
+	u64	cycle_mstamp;	     /* time of this cycle phase start */
+	u32     mode:3,		     /* current bbr_mode in state machine */
+		prev_ca_state:3,     /* CA state on previous ACK */
+		packet_conservation:1,  /* use packet conservation? */
+		round_start:1,	     /* start of packet-timed tx->ack round? */
+		ce_state:1,          /* If most recent data has CE bit set */
+		bw_probe_up_rounds:5,   /* cwnd-limited rounds in PROBE_UP */
+		try_fast_path:1, 	/* can we take fast path? */
+		unused2:11,
+		idle_restart:1,	     /* restarting after idle? */
+		probe_rtt_round_done:1,  /* a BBR_PROBE_RTT round at 4 pkts? */
+		cycle_idx:3,	/* current index in pacing_gain cycle array */
+		has_seen_rtt:1;	     /* have we seen an RTT sample yet? */
+	u32	pacing_gain:11,	/* current gain for setting pacing rate */
+		cwnd_gain:11,	/* current gain for setting cwnd */
+		full_bw_reached:1,   /* reached full bw in Startup? */
+		full_bw_cnt:2,	/* number of rounds without large bw gains */
+		init_cwnd:7;	/* initial cwnd */
+	u32	prior_cwnd;	/* prior cwnd upon entering loss recovery */
+	u32	full_bw;	/* recent bw, to estimate if pipe is full */
+
+	/* For tracking ACK aggregation: */
+	u64	ack_epoch_mstamp;	/* start of ACK sampling epoch */
+	u16	extra_acked[2];		/* max excess data ACKed in epoch */
+	u32	ack_epoch_acked:20,	/* packets (S)ACKed in sampling epoch */
+		extra_acked_win_rtts:5,	/* age of extra_acked, in round trips */
+		extra_acked_win_idx:1,	/* current index in extra_acked array */
+	/* BBR v2 state: */
+		unused1:2,
+		startup_ecn_rounds:2,	/* consecutive hi ECN STARTUP rounds */
+		loss_in_cycle:1,	/* packet loss in this cycle? */
+		ecn_in_cycle:1;		/* ECN in this cycle? */
+	u32	loss_round_delivered; /* scb->tx.delivered ending loss round */
+	u32	undo_bw_lo;	     /* bw_lo before latest losses */
+	u32	undo_inflight_lo;    /* inflight_lo before latest losses */
+	u32	undo_inflight_hi;    /* inflight_hi before latest losses */
+	u32	bw_latest;	 /* max delivered bw in last round trip */
+	u32	bw_lo;		 /* lower bound on sending bandwidth */
+	u32	bw_hi[2];	 /* upper bound of sending bandwidth range*/
+	u32	inflight_latest; /* max delivered data in last round trip */
+	u32	inflight_lo;	 /* lower bound of inflight data range */
+	u32	inflight_hi;	 /* upper bound of inflight data range */
+	u32	bw_probe_up_cnt; /* packets delivered per inflight_hi incr */
+	u32	bw_probe_up_acks;  /* packets (S)ACKed since inflight_hi incr */
+	u32	probe_wait_us;	 /* PROBE_DOWN until next clock-driven probe */
+	u32	ecn_eligible:1,	/* sender can use ECN (RTT, handshake)? */
+		ecn_alpha:9,	/* EWMA delivered_ce/delivered; 0..256 */
+		bw_probe_samples:1,    /* rate samples reflect bw probing? */
+		prev_probe_too_high:1, /* did last PROBE_UP go too high? */
+		stopped_risky_probe:1, /* last PROBE_UP stopped due to risk? */
+		rounds_since_probe:8,  /* packet-timed rounds since probed bw */
+		loss_round_start:1,    /* loss_round_delivered round trip? */
+		loss_in_round:1,       /* loss marked in this round trip? */
+		ecn_in_round:1,	       /* ECN marked in this round trip? */
+		ack_phase:3,	       /* bbr_ack_phase: meaning of ACKs */
+		loss_events_in_round:4,/* losses in STARTUP round */
+		initialized:1;	       /* has bbr_init() been called? */
+	u32	alpha_last_delivered;	 /* tp->delivered    at alpha update */
+	u32	alpha_last_delivered_ce; /* tp->delivered_ce at alpha update */
+
+	/* Params configurable using setsockopt. Refer to correspoding
+	 * module param for detailed description of params.
+	 */
+	struct bbr_params {
+		u32	high_gain:11,		/* max allowed value: 2047 */
+			drain_gain:10,		/* max allowed value: 1023 */
+			cwnd_gain:11;		/* max allowed value: 2047 */
+		u32	cwnd_min_target:4,	/* max allowed value: 15 */
+			min_rtt_win_sec:5,	/* max allowed value: 31 */
+			probe_rtt_mode_ms:9,	/* max allowed value: 511 */
+			full_bw_cnt:3,		/* max allowed value: 7 */
+			cwnd_tso_budget:1,	/* allowed values: {0, 1} */
+			unused3:6,
+			drain_to_target:1,	/* boolean */
+			precise_ece_ack:1,	/* boolean */
+			extra_acked_in_startup:1, /* allowed values: {0, 1} */
+			fast_path:1;		/* boolean */
+		u32	full_bw_thresh:10,	/* max allowed value: 1023 */
+			startup_cwnd_gain:11,	/* max allowed value: 2047 */
+			bw_probe_pif_gain:9,	/* max allowed value: 511 */
+			usage_based_cwnd:1, 	/* boolean */
+			unused2:1;
+		u16	probe_rtt_win_ms:14,	/* max allowed value: 16383 */
+			refill_add_inc:2;	/* max allowed value: 3 */
+		u16	extra_acked_gain:11,	/* max allowed value: 2047 */
+			extra_acked_win_rtts:5; /* max allowed value: 31*/
+		u16	pacing_gain[CYCLE_LEN]; /* max allowed value: 1023 */
+		/* Mostly BBR v2 parameters below here: */
+		u32	ecn_alpha_gain:8,	/* max allowed value: 255 */
+			ecn_factor:8,		/* max allowed value: 255 */
+			ecn_thresh:8,		/* max allowed value: 255 */
+			beta:8;			/* max allowed value: 255 */
+		u32	ecn_max_rtt_us:19,	/* max allowed value: 524287 */
+			bw_probe_reno_gain:9,	/* max allowed value: 511 */
+			full_loss_cnt:4;	/* max allowed value: 15 */
+		u32	probe_rtt_cwnd_gain:8,	/* max allowed value: 255 */
+			inflight_headroom:8,	/* max allowed value: 255 */
+			loss_thresh:8,		/* max allowed value: 255 */
+			bw_probe_max_rounds:8;	/* max allowed value: 255 */
+		u32	bw_probe_rand_rounds:4, /* max allowed value: 15 */
+			bw_probe_base_us:26,	/* usecs: 0..2^26-1 (67 secs) */
+			full_ecn_cnt:2;		/* max allowed value: 3 */
+		u32	bw_probe_rand_us:26,	/* usecs: 0..2^26-1 (67 secs) */
+			undo:1,			/* boolean */
+			tso_rtt_shift:4,	/* max allowed value: 15 */
+			unused5:1;
+		u32	ecn_reprobe_gain:9,	/* max allowed value: 511 */
+			unused1:14,
+			ecn_alpha_init:9;	/* max allowed value: 256 */
+	} params;
+
+	struct {
+		u32	snd_isn; /* Initial sequence number */
+		u32	rs_bw; 	 /* last valid rate sample bw */
+		u32	target_cwnd; /* target cwnd, based on BDP */
+		u8	undo:1,  /* Undo even happened but not yet logged */
+			unused:7;
+		char	event;	 /* single-letter event debug codes */
+		u16	unused2;
+	} debug;
+};
+
+struct bbr_context {
+	u32 sample_bw;
+	u32 target_cwnd;
+	u32 log:1;
+};
+
+/* Window length of min_rtt filter (in sec). Max allowed value is 31 (0x1F) */
+static u32 bbr_min_rtt_win_sec = 10;
+/* Minimum time (in ms) spent at bbr_cwnd_min_target in BBR_PROBE_RTT mode.
+ * Max allowed value is 511 (0x1FF).
+ */
+static u32 bbr_probe_rtt_mode_ms = 200;
+/* Window length of probe_rtt_min_us filter (in ms), and consequently the
+ * typical interval between PROBE_RTT mode entries.
+ * Note that bbr_probe_rtt_win_ms must be <= bbr_min_rtt_win_sec * MSEC_PER_SEC
+ */
+static u32 bbr_probe_rtt_win_ms = 5000;
+/* Skip TSO below the following bandwidth (bits/sec): */
+static int bbr_min_tso_rate = 1200000;
+
+/* Use min_rtt to help adapt TSO burst size, with smaller min_rtt resulting
+ * in bigger TSO bursts. By default we cut the RTT-based allowance in half
+ * for every 2^9 usec (aka 512 us) of RTT, so that the RTT-based allowance
+ * is below 1500 bytes after 6 * ~500 usec = 3ms.
+ */
+static u32 bbr_tso_rtt_shift = 9;  /* halve allowance per 2^9 usecs, 512us */
+
+/* Select cwnd TSO budget approach:
+ *  0: padding
+ *  1: flooring
+ */
+static uint bbr_cwnd_tso_budget = 1;
+
+/* Pace at ~1% below estimated bw, on average, to reduce queue at bottleneck.
+ * In order to help drive the network toward lower queues and low latency while
+ * maintaining high utilization, the average pacing rate aims to be slightly
+ * lower than the estimated bandwidth. This is an important aspect of the
+ * design.
+ */
+static const int bbr_pacing_margin_percent = 1;
+
+/* We use a high_gain value of 2/ln(2) because it's the smallest pacing gain
+ * that will allow a smoothly increasing pacing rate that will double each RTT
+ * and send the same number of packets per RTT that an un-paced, slow-starting
+ * Reno or CUBIC flow would. Max allowed value is 2047 (0x7FF).
+ */
+static int bbr_high_gain  = BBR_UNIT * 2885 / 1000 + 1;
+/* The gain for deriving startup cwnd. Max allowed value is 2047 (0x7FF). */
+static int bbr_startup_cwnd_gain  = BBR_UNIT * 2885 / 1000 + 1;
+/* The pacing gain of 1/high_gain in BBR_DRAIN is calculated to typically drain
+ * the queue created in BBR_STARTUP in a single round. Max allowed value
+ * is 1023 (0x3FF).
+ */
+static int bbr_drain_gain = BBR_UNIT * 1000 / 2885;
+/* The gain for deriving steady-state cwnd tolerates delayed/stretched ACKs.
+ * Max allowed value is 2047 (0x7FF).
+ */
+static int bbr_cwnd_gain  = BBR_UNIT * 2;
+/* The pacing_gain values for the PROBE_BW gain cycle, to discover/share bw.
+ * Max allowed value for each element is 1023 (0x3FF).
+ */
+enum bbr_pacing_gain_phase {
+	BBR_BW_PROBE_UP		= 0,  /* push up inflight to probe for bw/vol */
+	BBR_BW_PROBE_DOWN	= 1,  /* drain excess inflight from the queue */
+	BBR_BW_PROBE_CRUISE	= 2,  /* use pipe, w/ headroom in queue/pipe */
+	BBR_BW_PROBE_REFILL	= 3,  /* v2: refill the pipe again to 100% */
+};
+static int bbr_pacing_gain[] = {
+	BBR_UNIT * 5 / 4,	/* probe for more available bw */
+	BBR_UNIT * 3 / 4,	/* drain queue and/or yield bw to other flows */
+	BBR_UNIT, BBR_UNIT, BBR_UNIT,	/* cruise at 1.0*bw to utilize pipe, */
+	BBR_UNIT, BBR_UNIT, BBR_UNIT	/* without creating excess queue... */
+};
+
+/* Try to keep at least this many packets in flight, if things go smoothly. For
+ * smooth functioning, a sliding window protocol ACKing every other packet
+ * needs at least 4 packets in flight. Max allowed value is 15 (0xF).
+ */
+static u32 bbr_cwnd_min_target = 4;
+
+/* Cwnd to BDP proportion in PROBE_RTT mode scaled by BBR_UNIT. Default: 50%.
+ * Use 0 to disable. Max allowed value is 255.
+ */
+static u32 bbr_probe_rtt_cwnd_gain = BBR_UNIT * 1 / 2;
+
+/* To estimate if BBR_STARTUP mode (i.e. high_gain) has filled pipe... */
+/* If bw has increased significantly (1.25x), there may be more bw available.
+ * Max allowed value is 1023 (0x3FF).
+ */
+static u32 bbr_full_bw_thresh = BBR_UNIT * 5 / 4;
+/* But after 3 rounds w/o significant bw growth, estimate pipe is full.
+ * Max allowed value is 7 (0x7).
+ */
+static u32 bbr_full_bw_cnt = 3;
+
+static u32 bbr_flags;		/* Debugging related stuff */
+
+/* Whether to debug using printk.
+ */
+static bool bbr_debug_with_printk;
+
+/* Whether to debug using ftrace event tcp:tcp_bbr_event.
+ * Ignored when bbr_debug_with_printk is set.
+ */
+static bool bbr_debug_ftrace;
+
+/* Experiment: each cycle, try to hold sub-unity gain until inflight <= BDP. */
+static bool bbr_drain_to_target = true;		/* default: enabled */
+
+/* Experiment: Flags to control BBR with ECN behavior.
+ */
+static bool bbr_precise_ece_ack = true;		/* default: enabled */
+
+/* The max rwin scaling shift factor is 14 (RFC 1323), so the max sane rwin is
+ * (2^(16+14) B)/(1024 B/packet) = 1M packets.
+ */
+static u32 bbr_cwnd_warn_val	= 1U << 20;
+
+static u16 bbr_debug_port_mask;
+
+/* BBR module parameters. These are module parameters only in Google prod.
+ * Upstream these are intentionally not module parameters.
+ */
+static int bbr_pacing_gain_size = CYCLE_LEN;
+
+/* Gain factor for adding extra_acked to target cwnd: */
+static int bbr_extra_acked_gain = 256;
+
+/* Window length of extra_acked window. Max allowed val is 31. */
+static u32 bbr_extra_acked_win_rtts = 5;
+
+/* Max allowed val for ack_epoch_acked, after which sampling epoch is reset */
+static u32 bbr_ack_epoch_acked_reset_thresh = 1U << 20;
+
+/* Time period for clamping cwnd increment due to ack aggregation */
+static u32 bbr_extra_acked_max_us = 100 * 1000;
+
+/* Use extra acked in startup ?
+ * 0: disabled
+ * 1: use latest extra_acked value from 1-2 rtt in startup
+ */
+static int bbr_extra_acked_in_startup = 1;		/* default: enabled */
+
+/* Experiment: don't grow cwnd beyond twice of what we just probed. */
+static bool bbr_usage_based_cwnd;		/* default: disabled */
+
+/* For lab testing, researchers can enable BBRv2 ECN support with this flag,
+ * when they know that any ECN marks that the connections experience will be
+ * DCTCP/L4S-style ECN marks, rather than RFC3168 ECN marks.
+ * TODO(ncardwell): Production use of the BBRv2 ECN functionality depends on
+ * negotiation or configuration that is outside the scope of the BBRv2
+ * alpha release.
+ */
+static bool bbr_ecn_enable = false;
+
+module_param_named(min_tso_rate,      bbr_min_tso_rate,      int,    0644);
+module_param_named(tso_rtt_shift,     bbr_tso_rtt_shift,     int,    0644);
+module_param_named(high_gain,         bbr_high_gain,         int,    0644);
+module_param_named(drain_gain,        bbr_drain_gain,        int,    0644);
+module_param_named(startup_cwnd_gain, bbr_startup_cwnd_gain, int,    0644);
+module_param_named(cwnd_gain,         bbr_cwnd_gain,         int,    0644);
+module_param_array_named(pacing_gain, bbr_pacing_gain,       int,
+			 &bbr_pacing_gain_size, 0644);
+module_param_named(cwnd_min_target,   bbr_cwnd_min_target,   uint,   0644);
+module_param_named(probe_rtt_cwnd_gain,
+		   bbr_probe_rtt_cwnd_gain,		     uint,   0664);
+module_param_named(cwnd_warn_val,     bbr_cwnd_warn_val,     uint,   0664);
+module_param_named(debug_port_mask,   bbr_debug_port_mask,   ushort, 0644);
+module_param_named(flags,             bbr_flags,             uint,   0644);
+module_param_named(debug_ftrace,      bbr_debug_ftrace, bool,   0644);
+module_param_named(debug_with_printk, bbr_debug_with_printk, bool,   0644);
+module_param_named(min_rtt_win_sec,   bbr_min_rtt_win_sec,   uint,   0644);
+module_param_named(probe_rtt_mode_ms, bbr_probe_rtt_mode_ms, uint,   0644);
+module_param_named(probe_rtt_win_ms,  bbr_probe_rtt_win_ms,  uint,   0644);
+module_param_named(full_bw_thresh,    bbr_full_bw_thresh,    uint,   0644);
+module_param_named(full_bw_cnt,       bbr_full_bw_cnt,       uint,   0644);
+module_param_named(cwnd_tso_bduget,   bbr_cwnd_tso_budget,   uint,   0664);
+module_param_named(extra_acked_gain,  bbr_extra_acked_gain,  int,    0664);
+module_param_named(extra_acked_win_rtts,
+		   bbr_extra_acked_win_rtts, uint,   0664);
+module_param_named(extra_acked_max_us,
+		   bbr_extra_acked_max_us, uint,   0664);
+module_param_named(ack_epoch_acked_reset_thresh,
+		   bbr_ack_epoch_acked_reset_thresh, uint,   0664);
+module_param_named(drain_to_target,   bbr_drain_to_target,   bool,   0664);
+module_param_named(precise_ece_ack,   bbr_precise_ece_ack,   bool,   0664);
+module_param_named(extra_acked_in_startup,
+		   bbr_extra_acked_in_startup, int, 0664);
+module_param_named(usage_based_cwnd, bbr_usage_based_cwnd, bool,   0664);
+module_param_named(ecn_enable,       bbr_ecn_enable,         bool,   0664);
+
+static void bbr2_exit_probe_rtt(struct sock *sk);
+static void bbr2_reset_congestion_signals(struct sock *sk);
+
+static void bbr_check_probe_rtt_done(struct sock *sk);
+
+/* Do we estimate that STARTUP filled the pipe? */
+static bool bbr_full_bw_reached(const struct sock *sk)
+{
+	const struct bbr *bbr = inet_csk_ca(sk);
+
+	return bbr->full_bw_reached;
+}
+
+/* Return the windowed max recent bandwidth sample, in pkts/uS << BW_SCALE. */
+static u32 bbr_max_bw(const struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	return max(bbr->bw_hi[0], bbr->bw_hi[1]);
+}
+
+/* Return the estimated bandwidth of the path, in pkts/uS << BW_SCALE. */
+static u32 bbr_bw(const struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	return min(bbr_max_bw(sk), bbr->bw_lo);
+}
+
+/* Return maximum extra acked in past k-2k round trips,
+ * where k = bbr_extra_acked_win_rtts.
+ */
+static u16 bbr_extra_acked(const struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	return max(bbr->extra_acked[0], bbr->extra_acked[1]);
+}
+
+/* Return rate in bytes per second, optionally with a gain.
+ * The order here is chosen carefully to avoid overflow of u64. This should
+ * work for input rates of up to 2.9Tbit/sec and gain of 2.89x.
+ */
+static u64 bbr_rate_bytes_per_sec(struct sock *sk, u64 rate, int gain,
+				  int margin)
+{
+	unsigned int mss = tcp_sk(sk)->mss_cache;
+
+	rate *= mss;
+	rate *= gain;
+	rate >>= BBR_SCALE;
+	rate *= USEC_PER_SEC / 100 * (100 - margin);
+	rate >>= BW_SCALE;
+	rate = max(rate, 1ULL);
+	return rate;
+}
+
+static u64 bbr_bw_bytes_per_sec(struct sock *sk, u64 rate)
+{
+	return bbr_rate_bytes_per_sec(sk, rate, BBR_UNIT, 0);
+}
+
+static u64 bbr_rate_kbps(struct sock *sk, u64 rate)
+{
+	rate = bbr_bw_bytes_per_sec(sk, rate);
+	rate *= 8;
+	do_div(rate, 1000);
+	return rate;
+}
+
+static u32 bbr_tso_segs_goal(struct sock *sk);
+static void bbr_debug(struct sock *sk, u32 acked,
+		      const struct rate_sample *rs, struct bbr_context *ctx)
+{
+	static const char ca_states[] = {
+		[TCP_CA_Open]		= 'O',
+		[TCP_CA_Disorder]	= 'D',
+		[TCP_CA_CWR]		= 'C',
+		[TCP_CA_Recovery]	= 'R',
+		[TCP_CA_Loss]		= 'L',
+	};
+	static const char mode[] = {
+		'G',  /* Growing   - BBR_STARTUP */
+		'D',  /* Drain     - BBR_DRAIN */
+		'W',  /* Window    - BBR_PROBE_BW */
+		'M',  /* Min RTT   - BBR_PROBE_RTT */
+	};
+	static const char ack_phase[] = { /* bbr_ack_phase strings */
+		'I',	/* BBR_ACKS_INIT	   - 'Init' */
+		'R',	/* BBR_ACKS_REFILLING	   - 'Refilling' */
+		'B',	/* BBR_ACKS_PROBE_STARTING - 'Before' */
+		'F',	/* BBR_ACKS_PROBE_FEEDBACK - 'Feedback' */
+		'A',	/* BBR_ACKS_PROBE_STOPPING - 'After' */
+	};
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	const u32 una = tp->snd_una - bbr->debug.snd_isn;
+	const u32 fack = tcp_highest_sack_seq(tp);
+	const u16 dport = ntohs(inet_sk(sk)->inet_dport);
+	bool is_port_match = (bbr_debug_port_mask &&
+			      ((dport & bbr_debug_port_mask) == 0));
+	char debugmsg[320];
+
+	if (sk->sk_state == TCP_SYN_SENT)
+		return;  /* no bbr_init() yet if SYN retransmit -> CA_Loss */
+
+	if (!tp->snd_cwnd || tp->snd_cwnd > bbr_cwnd_warn_val) {
+		char addr[INET6_ADDRSTRLEN + 10] = { 0 };
+
+		if (sk->sk_family == AF_INET)
+			snprintf(addr, sizeof(addr), "%pI4:%u",
+				 &inet_sk(sk)->inet_daddr, dport);
+		else if (sk->sk_family == AF_INET6)
+			snprintf(addr, sizeof(addr), "%pI6:%u",
+				 &sk->sk_v6_daddr, dport);
+
+		WARN_ONCE(1,
+			"BBR %s cwnd alert: %u "
+			"snd_una: %u ca: %d pacing_gain: %u cwnd_gain: %u "
+			"bw: %u rtt: %u min_rtt: %u "
+			"acked: %u tso_segs: %u "
+			"bw: %d %ld %d pif: %u\n",
+			addr, tp->snd_cwnd,
+			una, inet_csk(sk)->icsk_ca_state,
+			bbr->pacing_gain, bbr->cwnd_gain,
+			bbr_max_bw(sk), (tp->srtt_us >> 3), bbr->min_rtt_us,
+			acked, bbr_tso_segs_goal(sk),
+			rs->delivered, rs->interval_us, rs->is_retrans,
+			tcp_packets_in_flight(tp));
+	}
+
+	if (likely(!bbr_debug_with_printk && !bbr_debug_ftrace))
+		return;
+
+	if (!sock_flag(sk, SOCK_DBG) && !is_port_match)
+		return;
+
+	if (!ctx->log && !tp->app_limited && !(bbr_flags & FLAG_DEBUG_VERBOSE))
+		return;
+
+	if (ipv4_is_loopback(inet_sk(sk)->inet_daddr) &&
+	    !(bbr_flags & FLAG_DEBUG_LOOPBACK))
+		return;
+
+	snprintf(debugmsg, sizeof(debugmsg) - 1,
+		 "BBR %pI4:%-5u %5u,%03u:%-7u %c "
+		 "%c %2u br %2u cr %2d rtt %5ld d %2d i %5ld mrtt %d %cbw %llu "
+		 "bw %llu lb %llu ib %llu qb %llu "
+		 "a %u if %2u %c %c dl %u l %u al %u # %u t %u %c %c "
+		 "lr %d er %d ea %d bwl %lld il %d ih %d c %d "
+		 "v %d %c %u %c %s\n",
+		 &inet_sk(sk)->inet_daddr, dport,
+		 una / 1000, una % 1000, fack - tp->snd_una,
+		 ca_states[inet_csk(sk)->icsk_ca_state],
+		 bbr->debug.undo ? '@' : mode[bbr->mode],
+		 tp->snd_cwnd,
+		 bbr_extra_acked(sk),	/* br (legacy): extra_acked */
+		 rs->tx_in_flight,	/* cr (legacy): tx_inflight */
+		 rs->rtt_us,
+		 rs->delivered,
+		 rs->interval_us,
+		 bbr->min_rtt_us,
+		 rs->is_app_limited ? '_' : 'l',
+		 bbr_rate_kbps(sk, ctx->sample_bw), /* lbw: latest sample bw */
+		 bbr_rate_kbps(sk, bbr_max_bw(sk)), /* bw: max bw */
+		 0ULL,				    /* lb: [obsolete] */
+		 0ULL,				    /* ib: [obsolete] */
+		 (u64)sk->sk_pacing_rate * 8 / 1000,
+		 acked,
+		 tcp_packets_in_flight(tp),
+		 rs->is_ack_delayed ? 'd' : '.',
+		 bbr->round_start ? '*' : '.',
+		 tp->delivered, tp->lost,
+		 tp->app_limited,
+		 0,			    	    /* #: [obsolete] */
+		 ctx->target_cwnd,
+		 tp->reord_seen ? 'r' : '.',  /* r: reordering seen? */
+		 ca_states[bbr->prev_ca_state],
+		 (rs->lost + rs->delivered) > 0 ?
+		 (1000 * rs->lost /
+		  (rs->lost + rs->delivered)) : 0,    /* lr: loss rate x1000 */
+		 (rs->delivered) > 0 ?
+		 (1000 * rs->delivered_ce /
+		  (rs->delivered)) : 0,		      /* er: ECN rate x1000 */
+		 1000 * bbr->ecn_alpha >> BBR_SCALE,  /* ea: ECN alpha x1000 */
+		 bbr->bw_lo == ~0U ?
+		   -1 : (s64)bbr_rate_kbps(sk, bbr->bw_lo), /* bwl */
+		 bbr->inflight_lo,	/* il */
+		 bbr->inflight_hi,	/* ih */
+		 bbr->bw_probe_up_cnt,	/* c */
+		 2,			/* v: version */
+		 bbr->debug.event,
+		 bbr->cycle_idx,
+		 ack_phase[bbr->ack_phase],
+		 bbr->bw_probe_samples ? "Y" : "N");
+	debugmsg[sizeof(debugmsg) - 1] = 0;
+
+	/* printk takes a higher precedence. */
+	if (bbr_debug_with_printk)
+		printk(KERN_DEBUG "%s", debugmsg);
+
+	if (unlikely(bbr->debug.undo))
+		bbr->debug.undo = 0;
+}
+
+/* Convert a BBR bw and gain factor to a pacing rate in bytes per second. */
+static unsigned long bbr_bw_to_pacing_rate(struct sock *sk, u32 bw, int gain)
+{
+	u64 rate = bw;
+
+	rate = bbr_rate_bytes_per_sec(sk, rate, gain,
+				      bbr_pacing_margin_percent);
+	rate = min_t(u64, rate, sk->sk_max_pacing_rate);
+	return rate;
+}
+
+/* Initialize pacing rate to: high_gain * init_cwnd / RTT. */
+static void bbr_init_pacing_rate_from_rtt(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	u64 bw;
+	u32 rtt_us;
+
+	if (tp->srtt_us) {		/* any RTT sample yet? */
+		rtt_us = max(tp->srtt_us >> 3, 1U);
+		bbr->has_seen_rtt = 1;
+	} else {			 /* no RTT sample yet */
+		rtt_us = USEC_PER_MSEC;	 /* use nominal default RTT */
+	}
+	bw = (u64)tp->snd_cwnd * BW_UNIT;
+	do_div(bw, rtt_us);
+	sk->sk_pacing_rate = bbr_bw_to_pacing_rate(sk, bw, bbr->params.high_gain);
+}
+
+/* Pace using current bw estimate and a gain factor. */
+static void bbr_set_pacing_rate(struct sock *sk, u32 bw, int gain)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	unsigned long rate = bbr_bw_to_pacing_rate(sk, bw, gain);
+
+	if (unlikely(!bbr->has_seen_rtt && tp->srtt_us))
+		bbr_init_pacing_rate_from_rtt(sk);
+	if (bbr_full_bw_reached(sk) || rate > sk->sk_pacing_rate)
+		sk->sk_pacing_rate = rate;
+}
+
+static u32 bbr_min_tso_segs(struct sock *sk)
+{
+	return sk->sk_pacing_rate < (bbr_min_tso_rate >> 3) ? 1 : 2;
+}
+
+/* Return the number of segments BBR would like in a TSO/GSO skb, given
+ * a particular max gso size as a constraint.
+ */
+static u32 bbr_tso_segs_generic(struct sock *sk, unsigned int mss_now,
+				u32 gso_max_size)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 segs, r;
+	u64 bytes;
+
+	/* Budget a TSO/GSO burst size allowance based on bw (pacing_rate). */
+	bytes = sk->sk_pacing_rate >> sk->sk_pacing_shift;
+
+	/* Budget a TSO/GSO burst size allowance based on min_rtt. For every
+	 * K = 2^tso_rtt_shift microseconds of min_rtt, halve the burst.
+	 * The min_rtt-based burst allowance is: 64 KBytes / 2^(min_rtt/K)
+	 */
+	if (bbr->params.tso_rtt_shift) {
+		r = bbr->min_rtt_us >> bbr->params.tso_rtt_shift;
+		if (r < BITS_PER_TYPE(u32))   /* prevent undefined behavior */
+			bytes += GSO_MAX_SIZE >> r;
+	}
+
+	bytes = min_t(u32, bytes, gso_max_size - 1 - MAX_TCP_HEADER);
+	segs = max_t(u32, bytes / mss_now, bbr_min_tso_segs(sk));
+	return segs;
+}
+
+/* Custom tcp_tso_autosize() for BBR, used at transmit time to cap skb size. */
+static u32  bbr_tso_segs(struct sock *sk, unsigned int mss_now)
+{
+	return bbr_tso_segs_generic(sk, mss_now, sk->sk_gso_max_size);
+}
+
+/* Like bbr_tso_segs(), using mss_cache, ignoring driver's sk_gso_max_size. */
+static u32 bbr_tso_segs_goal(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+
+	return  bbr_tso_segs_generic(sk, tp->mss_cache, GSO_MAX_SIZE);
+}
+
+/* Save "last known good" cwnd so we can restore it after losses or PROBE_RTT */
+static void bbr_save_cwnd(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (bbr->prev_ca_state < TCP_CA_Recovery && bbr->mode != BBR_PROBE_RTT)
+		bbr->prior_cwnd = tp->snd_cwnd;  /* this cwnd is good enough */
+	else  /* loss recovery or BBR_PROBE_RTT have temporarily cut cwnd */
+		bbr->prior_cwnd = max(bbr->prior_cwnd, tp->snd_cwnd);
+}
+
+static void bbr_cwnd_event(struct sock *sk, enum tcp_ca_event event)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (event == CA_EVENT_TX_START && tp->app_limited) {
+		bbr->idle_restart = 1;
+		bbr->ack_epoch_mstamp = tp->tcp_mstamp;
+		bbr->ack_epoch_acked = 0;
+		/* Avoid pointless buffer overflows: pace at est. bw if we don't
+		 * need more speed (we're restarting from idle and app-limited).
+		 */
+		if (bbr->mode == BBR_PROBE_BW)
+			bbr_set_pacing_rate(sk, bbr_bw(sk), BBR_UNIT);
+		else if (bbr->mode == BBR_PROBE_RTT)
+			bbr_check_probe_rtt_done(sk);
+	} else if ((event == CA_EVENT_ECN_IS_CE ||
+		    event == CA_EVENT_ECN_NO_CE) &&
+		    bbr_ecn_enable &&
+		    bbr->params.precise_ece_ack) {
+		u32 state = bbr->ce_state;
+		dctcp_ece_ack_update(sk, event, &bbr->prior_rcv_nxt, &state);
+		bbr->ce_state = state;
+		if (tp->fast_ack_mode == 2 && event == CA_EVENT_ECN_IS_CE)
+			tcp_enter_quickack_mode(sk, TCP_MAX_QUICKACKS);
+	}
+}
+
+/* Calculate bdp based on min RTT and the estimated bottleneck bandwidth:
+ *
+ * bdp = ceil(bw * min_rtt * gain)
+ *
+ * The key factor, gain, controls the amount of queue. While a small gain
+ * builds a smaller queue, it becomes more vulnerable to noise in RTT
+ * measurements (e.g., delayed ACKs or other ACK compression effects). This
+ * noise may cause BBR to under-estimate the rate.
+ */
+static u32 bbr_bdp(struct sock *sk, u32 bw, int gain)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 bdp;
+	u64 w;
+
+	/* If we've never had a valid RTT sample, cap cwnd at the initial
+	 * default. This should only happen when the connection is not using TCP
+	 * timestamps and has retransmitted all of the SYN/SYNACK/data packets
+	 * ACKed so far. In this case, an RTO can cut cwnd to 1, in which
+	 * case we need to slow-start up toward something safe: initial cwnd.
+	 */
+	if (unlikely(bbr->min_rtt_us == ~0U))	 /* no valid RTT samples yet? */
+		return bbr->init_cwnd;  /* be safe: cap at initial cwnd */
+
+	w = (u64)bw * bbr->min_rtt_us;
+
+	/* Apply a gain to the given value, remove the BW_SCALE shift, and
+	 * round the value up to avoid a negative feedback loop.
+	 */
+	bdp = (((w * gain) >> BBR_SCALE) + BW_UNIT - 1) / BW_UNIT;
+
+	return bdp;
+}
+
+/* To achieve full performance in high-speed paths, we budget enough cwnd to
+ * fit full-sized skbs in-flight on both end hosts to fully utilize the path:
+ *   - one skb in sending host Qdisc,
+ *   - one skb in sending host TSO/GSO engine
+ *   - one skb being received by receiver host LRO/GRO/delayed-ACK engine
+ * Don't worry, at low rates (bbr_min_tso_rate) this won't bloat cwnd because
+ * in such cases tso_segs_goal is 1. The minimum cwnd is 4 packets,
+ * which allows 2 outstanding 2-packet sequences, to try to keep pipe
+ * full even with ACK-every-other-packet delayed ACKs.
+ */
+static u32 bbr_quantization_budget(struct sock *sk, u32 cwnd)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 tso_segs_goal;
+
+	tso_segs_goal = 3 * bbr_tso_segs_goal(sk);
+
+	/* Allow enough full-sized skbs in flight to utilize end systems. */
+	if (bbr->params.cwnd_tso_budget == 1) {
+		cwnd = max_t(u32, cwnd, tso_segs_goal);
+		cwnd = max_t(u32, cwnd, bbr->params.cwnd_min_target);
+	} else {
+		cwnd += tso_segs_goal;
+		cwnd = (cwnd + 1) & ~1U;
+	}
+	/* Ensure gain cycling gets inflight above BDP even for small BDPs. */
+	if (bbr->mode == BBR_PROBE_BW && bbr->cycle_idx == BBR_BW_PROBE_UP)
+		cwnd += 2;
+
+	return cwnd;
+}
+
+/* Find inflight based on min RTT and the estimated bottleneck bandwidth. */
+static u32 bbr_inflight(struct sock *sk, u32 bw, int gain)
+{
+	u32 inflight;
+
+	inflight = bbr_bdp(sk, bw, gain);
+	inflight = bbr_quantization_budget(sk, inflight);
+
+	return inflight;
+}
+
+/* With pacing at lower layers, there's often less data "in the network" than
+ * "in flight". With TSQ and departure time pacing at lower layers (e.g. fq),
+ * we often have several skbs queued in the pacing layer with a pre-scheduled
+ * earliest departure time (EDT). BBR adapts its pacing rate based on the
+ * inflight level that it estimates has already been "baked in" by previous
+ * departure time decisions. We calculate a rough estimate of the number of our
+ * packets that might be in the network at the earliest departure time for the
+ * next skb scheduled:
+ *   in_network_at_edt = inflight_at_edt - (EDT - now) * bw
+ * If we're increasing inflight, then we want to know if the transmit of the
+ * EDT skb will push inflight above the target, so inflight_at_edt includes
+ * bbr_tso_segs_goal() from the skb departing at EDT. If decreasing inflight,
+ * then estimate if inflight will sink too low just before the EDT transmit.
+ */
+static u32 bbr_packets_in_net_at_edt(struct sock *sk, u32 inflight_now)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	u64 now_ns, edt_ns, interval_us;
+	u32 interval_delivered, inflight_at_edt;
+
+	now_ns = tp->tcp_clock_cache;
+	edt_ns = max(tp->tcp_wstamp_ns, now_ns);
+	interval_us = div_u64(edt_ns - now_ns, NSEC_PER_USEC);
+	interval_delivered = (u64)bbr_bw(sk) * interval_us >> BW_SCALE;
+	inflight_at_edt = inflight_now;
+	if (bbr->pacing_gain > BBR_UNIT)              /* increasing inflight */
+		inflight_at_edt += bbr_tso_segs_goal(sk);  /* include EDT skb */
+	if (interval_delivered >= inflight_at_edt)
+		return 0;
+	return inflight_at_edt - interval_delivered;
+}
+
+/* Find the cwnd increment based on estimate of ack aggregation */
+static u32 bbr_ack_aggregation_cwnd(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 max_aggr_cwnd, aggr_cwnd = 0;
+
+	if (bbr->params.extra_acked_gain &&
+	    (bbr_full_bw_reached(sk) || bbr->params.extra_acked_in_startup)) {
+		max_aggr_cwnd = ((u64)bbr_bw(sk) * bbr_extra_acked_max_us)
+				/ BW_UNIT;
+		aggr_cwnd = (bbr->params.extra_acked_gain * bbr_extra_acked(sk))
+			     >> BBR_SCALE;
+		aggr_cwnd = min(aggr_cwnd, max_aggr_cwnd);
+	}
+
+	return aggr_cwnd;
+}
+
+/* Returns the cwnd for PROBE_RTT mode. */
+static u32 bbr_probe_rtt_cwnd(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (bbr->params.probe_rtt_cwnd_gain == 0)
+		return bbr->params.cwnd_min_target;
+	return max_t(u32, bbr->params.cwnd_min_target,
+		     bbr_bdp(sk, bbr_bw(sk), bbr->params.probe_rtt_cwnd_gain));
+}
+
+/* Slow-start up toward target cwnd (if bw estimate is growing, or packet loss
+ * has drawn us down below target), or snap down to target if we're above it.
+ */
+static void bbr_set_cwnd(struct sock *sk, const struct rate_sample *rs,
+			 u32 acked, u32 bw, int gain, u32 cwnd,
+			 struct bbr_context *ctx)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 target_cwnd = 0, prev_cwnd = tp->snd_cwnd, max_probe;
+
+	if (!acked)
+		goto done;  /* no packet fully ACKed; just apply caps */
+
+	target_cwnd = bbr_bdp(sk, bw, gain);
+
+	/* Increment the cwnd to account for excess ACKed data that seems
+	 * due to aggregation (of data and/or ACKs) visible in the ACK stream.
+	 */
+	target_cwnd += bbr_ack_aggregation_cwnd(sk);
+	target_cwnd = bbr_quantization_budget(sk, target_cwnd);
+
+	/* If we're below target cwnd, slow start cwnd toward target cwnd. */
+	bbr->debug.target_cwnd = target_cwnd;
+
+	/* Update cwnd and enable fast path if cwnd reaches target_cwnd. */
+	bbr->try_fast_path = 0;
+	if (bbr_full_bw_reached(sk)) { /* only cut cwnd if we filled the pipe */
+		cwnd += acked;
+		if (cwnd >= target_cwnd) {
+			cwnd = target_cwnd;
+			bbr->try_fast_path = 1;
+		}
+	} else if (cwnd < target_cwnd || cwnd  < 2 * bbr->init_cwnd) {
+		cwnd += acked;
+	} else {
+		bbr->try_fast_path = 1;
+	}
+
+	/* When growing cwnd, don't grow beyond twice what we just probed. */
+	if (bbr->params.usage_based_cwnd) {
+		max_probe = max(2 * tp->max_packets_out, tp->snd_cwnd);
+		cwnd = min(cwnd, max_probe);
+	}
+
+	cwnd = max_t(u32, cwnd, bbr->params.cwnd_min_target);
+done:
+	tp->snd_cwnd = min(cwnd, tp->snd_cwnd_clamp);	/* apply global cap */
+	if (bbr->mode == BBR_PROBE_RTT)  /* drain queue, refresh min_rtt */
+		tp->snd_cwnd = min_t(u32, tp->snd_cwnd, bbr_probe_rtt_cwnd(sk));
+
+	ctx->target_cwnd = target_cwnd;
+	ctx->log = (tp->snd_cwnd != prev_cwnd);
+}
+
+/* See if we have reached next round trip */
+static void bbr_update_round_start(struct sock *sk,
+		const struct rate_sample *rs, struct bbr_context *ctx)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr->round_start = 0;
+
+	/* See if we've reached the next RTT */
+	if (rs->interval_us > 0 &&
+	    !before(rs->prior_delivered, bbr->next_rtt_delivered)) {
+		bbr->next_rtt_delivered = tp->delivered;
+		bbr->round_start = 1;
+	}
+}
+
+/* Calculate the bandwidth based on how fast packets are delivered */
+static void bbr_calculate_bw_sample(struct sock *sk,
+			const struct rate_sample *rs, struct bbr_context *ctx)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u64 bw = 0;
+
+	/* Divide delivered by the interval to find a (lower bound) bottleneck
+	 * bandwidth sample. Delivered is in packets and interval_us in uS and
+	 * ratio will be <<1 for most connections. So delivered is first scaled.
+	 * Round up to allow growth at low rates, even with integer division.
+	 */
+	if (rs->interval_us > 0) {
+		if (WARN_ONCE(rs->delivered < 0,
+			      "negative delivered: %d interval_us: %ld\n",
+			      rs->delivered, rs->interval_us))
+			return;
+
+		bw = DIV_ROUND_UP_ULL((u64)rs->delivered * BW_UNIT, rs->interval_us);
+	}
+
+	ctx->sample_bw = bw;
+	bbr->debug.rs_bw = bw;
+}
+
+/* Estimates the windowed max degree of ack aggregation.
+ * This is used to provision extra in-flight data to keep sending during
+ * inter-ACK silences.
+ *
+ * Degree of ack aggregation is estimated as extra data acked beyond expected.
+ *
+ * max_extra_acked = "maximum recent excess data ACKed beyond max_bw * interval"
+ * cwnd += max_extra_acked
+ *
+ * Max extra_acked is clamped by cwnd and bw * bbr_extra_acked_max_us (100 ms).
+ * Max filter is an approximate sliding window of 5-10 (packet timed) round
+ * trips for non-startup phase, and 1-2 round trips for startup.
+ */
+static void bbr_update_ack_aggregation(struct sock *sk,
+				       const struct rate_sample *rs)
+{
+	u32 epoch_us, expected_acked, extra_acked;
+	struct bbr *bbr = inet_csk_ca(sk);
+	struct tcp_sock *tp = tcp_sk(sk);
+	u32 extra_acked_win_rtts_thresh = bbr->params.extra_acked_win_rtts;
+
+	if (!bbr->params.extra_acked_gain || rs->acked_sacked <= 0 ||
+	    rs->delivered < 0 || rs->interval_us <= 0)
+		return;
+
+	if (bbr->round_start) {
+		bbr->extra_acked_win_rtts = min(0x1F,
+						bbr->extra_acked_win_rtts + 1);
+		if (bbr->params.extra_acked_in_startup &&
+		    !bbr_full_bw_reached(sk))
+			extra_acked_win_rtts_thresh = 1;
+		if (bbr->extra_acked_win_rtts >=
+		    extra_acked_win_rtts_thresh) {
+			bbr->extra_acked_win_rtts = 0;
+			bbr->extra_acked_win_idx = bbr->extra_acked_win_idx ?
+						   0 : 1;
+			bbr->extra_acked[bbr->extra_acked_win_idx] = 0;
+		}
+	}
+
+	/* Compute how many packets we expected to be delivered over epoch. */
+	epoch_us = tcp_stamp_us_delta(tp->delivered_mstamp,
+				      bbr->ack_epoch_mstamp);
+	expected_acked = ((u64)bbr_bw(sk) * epoch_us) / BW_UNIT;
+
+	/* Reset the aggregation epoch if ACK rate is below expected rate or
+	 * significantly large no. of ack received since epoch (potentially
+	 * quite old epoch).
+	 */
+	if (bbr->ack_epoch_acked <= expected_acked ||
+	    (bbr->ack_epoch_acked + rs->acked_sacked >=
+	     bbr_ack_epoch_acked_reset_thresh)) {
+		bbr->ack_epoch_acked = 0;
+		bbr->ack_epoch_mstamp = tp->delivered_mstamp;
+		expected_acked = 0;
+	}
+
+	/* Compute excess data delivered, beyond what was expected. */
+	bbr->ack_epoch_acked = min_t(u32, 0xFFFFF,
+				   bbr->ack_epoch_acked + rs->acked_sacked);
+	extra_acked = bbr->ack_epoch_acked - expected_acked;
+	extra_acked = min(extra_acked, tp->snd_cwnd);
+	if (extra_acked > bbr->extra_acked[bbr->extra_acked_win_idx])
+		bbr->extra_acked[bbr->extra_acked_win_idx] = extra_acked;
+}
+
+/* Estimate when the pipe is full, using the change in delivery rate: BBR
+ * estimates that STARTUP filled the pipe if the estimated bw hasn't changed by
+ * at least bbr_full_bw_thresh (25%) after bbr_full_bw_cnt (3) non-app-limited
+ * rounds. Why 3 rounds: 1: rwin autotuning grows the rwin, 2: we fill the
+ * higher rwin, 3: we get higher delivery rate samples. Or transient
+ * cross-traffic or radio noise can go away. CUBIC Hystart shares a similar
+ * design goal, but uses delay and inter-ACK spacing instead of bandwidth.
+ */
+static void bbr_check_full_bw_reached(struct sock *sk,
+				      const struct rate_sample *rs)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 bw_thresh;
+
+	if (bbr_full_bw_reached(sk) || !bbr->round_start || rs->is_app_limited)
+		return;
+
+	bw_thresh = (u64)bbr->full_bw * bbr->params.full_bw_thresh >> BBR_SCALE;
+	if (bbr_max_bw(sk) >= bw_thresh) {
+		bbr->full_bw = bbr_max_bw(sk);
+		bbr->full_bw_cnt = 0;
+		return;
+	}
+	++bbr->full_bw_cnt;
+	bbr->full_bw_reached = bbr->full_bw_cnt >= bbr->params.full_bw_cnt;
+}
+
+/* If pipe is probably full, drain the queue and then enter steady-state. */
+static bool bbr_check_drain(struct sock *sk, const struct rate_sample *rs,
+			    struct bbr_context *ctx)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (bbr->mode == BBR_STARTUP && bbr_full_bw_reached(sk)) {
+		bbr->mode = BBR_DRAIN;	/* drain queue we created */
+		tcp_sk(sk)->snd_ssthresh =
+				bbr_inflight(sk, bbr_max_bw(sk), BBR_UNIT);
+		bbr2_reset_congestion_signals(sk);
+	}	/* fall through to check if in-flight is already small: */
+	if (bbr->mode == BBR_DRAIN &&
+	    bbr_packets_in_net_at_edt(sk, tcp_packets_in_flight(tcp_sk(sk))) <=
+	    bbr_inflight(sk, bbr_max_bw(sk), BBR_UNIT))
+		return true;  /* exiting DRAIN now */
+	return false;
+}
+
+static void bbr_check_probe_rtt_done(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (!(bbr->probe_rtt_done_stamp &&
+	      after(tcp_jiffies32, bbr->probe_rtt_done_stamp)))
+		return;
+
+	bbr->probe_rtt_min_stamp = tcp_jiffies32; /* schedule next PROBE_RTT */
+	tp->snd_cwnd = max(tp->snd_cwnd, bbr->prior_cwnd);
+	bbr2_exit_probe_rtt(sk);
+}
+
+/* The goal of PROBE_RTT mode is to have BBR flows cooperatively and
+ * periodically drain the bottleneck queue, to converge to measure the true
+ * min_rtt (unloaded propagation delay). This allows the flows to keep queues
+ * small (reducing queuing delay and packet loss) and achieve fairness among
+ * BBR flows.
+ *
+ * The min_rtt filter window is 10 seconds. When the min_rtt estimate expires,
+ * we enter PROBE_RTT mode and cap the cwnd at bbr_cwnd_min_target=4 packets.
+ * After at least bbr_probe_rtt_mode_ms=200ms and at least one packet-timed
+ * round trip elapsed with that flight size <= 4, we leave PROBE_RTT mode and
+ * re-enter the previous mode. BBR uses 200ms to approximately bound the
+ * performance penalty of PROBE_RTT's cwnd capping to roughly 2% (200ms/10s).
+ *
+ * Note that flows need only pay 2% if they are busy sending over the last 10
+ * seconds. Interactive applications (e.g., Web, RPCs, video chunks) often have
+ * natural silences or low-rate periods within 10 seconds where the rate is low
+ * enough for long enough to drain its queue in the bottleneck. We pick up
+ * these min RTT measurements opportunistically with our min_rtt filter. :-)
+ */
+static void bbr_update_min_rtt(struct sock *sk, const struct rate_sample *rs)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	bool probe_rtt_expired, min_rtt_expired;
+	u32 expire;
+
+	/* Track min RTT in probe_rtt_win_ms to time next PROBE_RTT state. */
+	expire = bbr->probe_rtt_min_stamp +
+		 msecs_to_jiffies(bbr->params.probe_rtt_win_ms);
+	probe_rtt_expired = after(tcp_jiffies32, expire);
+	if (rs->rtt_us >= 0 &&
+	    (rs->rtt_us <= bbr->probe_rtt_min_us ||
+	     (probe_rtt_expired && !rs->is_ack_delayed))) {
+		bbr->probe_rtt_min_us = rs->rtt_us;
+		bbr->probe_rtt_min_stamp = tcp_jiffies32;
+	}
+	/* Track min RTT seen in the min_rtt_win_sec filter window: */
+	expire = bbr->min_rtt_stamp + bbr->params.min_rtt_win_sec * HZ;
+	min_rtt_expired = after(tcp_jiffies32, expire);
+	if (bbr->probe_rtt_min_us <= bbr->min_rtt_us ||
+	    min_rtt_expired) {
+		bbr->min_rtt_us = bbr->probe_rtt_min_us;
+		bbr->min_rtt_stamp = bbr->probe_rtt_min_stamp;
+	}
+
+	if (bbr->params.probe_rtt_mode_ms > 0 && probe_rtt_expired &&
+	    !bbr->idle_restart && bbr->mode != BBR_PROBE_RTT) {
+		bbr->mode = BBR_PROBE_RTT;  /* dip, drain queue */
+		bbr_save_cwnd(sk);  /* note cwnd so we can restore it */
+		bbr->probe_rtt_done_stamp = 0;
+		bbr->ack_phase = BBR_ACKS_PROBE_STOPPING;
+		bbr->next_rtt_delivered = tp->delivered;
+	}
+
+	if (bbr->mode == BBR_PROBE_RTT) {
+		/* Ignore low rate samples during this mode. */
+		tp->app_limited =
+			(tp->delivered + tcp_packets_in_flight(tp)) ? : 1;
+		/* Maintain min packets in flight for max(200 ms, 1 round). */
+		if (!bbr->probe_rtt_done_stamp &&
+		    tcp_packets_in_flight(tp) <= bbr_probe_rtt_cwnd(sk)) {
+			bbr->probe_rtt_done_stamp = tcp_jiffies32 +
+				msecs_to_jiffies(bbr->params.probe_rtt_mode_ms);
+			bbr->probe_rtt_round_done = 0;
+			bbr->next_rtt_delivered = tp->delivered;
+		} else if (bbr->probe_rtt_done_stamp) {
+			if (bbr->round_start)
+				bbr->probe_rtt_round_done = 1;
+			if (bbr->probe_rtt_round_done)
+				bbr_check_probe_rtt_done(sk);
+		}
+	}
+	/* Restart after idle ends only once we process a new S/ACK for data */
+	if (rs->delivered > 0)
+		bbr->idle_restart = 0;
+}
+
+static void bbr_update_gains(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	switch (bbr->mode) {
+	case BBR_STARTUP:
+		bbr->pacing_gain = bbr->params.high_gain;
+		bbr->cwnd_gain	 = bbr->params.startup_cwnd_gain;
+		break;
+	case BBR_DRAIN:
+		bbr->pacing_gain = bbr->params.drain_gain;  /* slow, to drain */
+		bbr->cwnd_gain = bbr->params.startup_cwnd_gain;  /* keep cwnd */
+		break;
+	case BBR_PROBE_BW:
+		bbr->pacing_gain = bbr->params.pacing_gain[bbr->cycle_idx];
+		bbr->cwnd_gain = bbr->params.cwnd_gain;
+		break;
+	case BBR_PROBE_RTT:
+		bbr->pacing_gain = BBR_UNIT;
+		bbr->cwnd_gain = BBR_UNIT;
+		break;
+	default:
+		WARN_ONCE(1, "BBR bad mode: %u\n", bbr->mode);
+		break;
+	}
+}
+
+static void bbr_init(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	int i;
+
+	WARN_ON_ONCE(tp->snd_cwnd >= bbr_cwnd_warn_val);
+
+	bbr->initialized = 1;
+	bbr->params.high_gain = min(0x7FF, bbr_high_gain);
+	bbr->params.drain_gain = min(0x3FF, bbr_drain_gain);
+	bbr->params.startup_cwnd_gain = min(0x7FF, bbr_startup_cwnd_gain);
+	bbr->params.cwnd_gain = min(0x7FF, bbr_cwnd_gain);
+	bbr->params.cwnd_tso_budget = min(0x1U, bbr_cwnd_tso_budget);
+	bbr->params.cwnd_min_target = min(0xFU, bbr_cwnd_min_target);
+	bbr->params.min_rtt_win_sec = min(0x1FU, bbr_min_rtt_win_sec);
+	bbr->params.probe_rtt_mode_ms = min(0x1FFU, bbr_probe_rtt_mode_ms);
+	bbr->params.full_bw_cnt = min(0x7U, bbr_full_bw_cnt);
+	bbr->params.full_bw_thresh = min(0x3FFU, bbr_full_bw_thresh);
+	bbr->params.extra_acked_gain = min(0x7FF, bbr_extra_acked_gain);
+	bbr->params.extra_acked_win_rtts = min(0x1FU, bbr_extra_acked_win_rtts);
+	bbr->params.drain_to_target = bbr_drain_to_target ? 1 : 0;
+	bbr->params.precise_ece_ack = bbr_precise_ece_ack ? 1 : 0;
+	bbr->params.extra_acked_in_startup = bbr_extra_acked_in_startup ? 1 : 0;
+	bbr->params.probe_rtt_cwnd_gain = min(0xFFU, bbr_probe_rtt_cwnd_gain);
+	bbr->params.probe_rtt_win_ms =
+		min(0x3FFFU,
+		    min_t(u32, bbr_probe_rtt_win_ms,
+			  bbr->params.min_rtt_win_sec * MSEC_PER_SEC));
+	for (i = 0; i < CYCLE_LEN; i++)
+		bbr->params.pacing_gain[i] = min(0x3FF, bbr_pacing_gain[i]);
+	bbr->params.usage_based_cwnd = bbr_usage_based_cwnd ? 1 : 0;
+	bbr->params.tso_rtt_shift =  min(0xFU, bbr_tso_rtt_shift);
+
+	bbr->debug.snd_isn = tp->snd_una;
+	bbr->debug.target_cwnd = 0;
+	bbr->debug.undo = 0;
+
+	bbr->init_cwnd = min(0x7FU, tp->snd_cwnd);
+	bbr->prior_cwnd = tp->prior_cwnd;
+	tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
+	bbr->next_rtt_delivered = 0;
+	bbr->prev_ca_state = TCP_CA_Open;
+	bbr->packet_conservation = 0;
+
+	bbr->probe_rtt_done_stamp = 0;
+	bbr->probe_rtt_round_done = 0;
+	bbr->probe_rtt_min_us = tcp_min_rtt(tp);
+	bbr->probe_rtt_min_stamp = tcp_jiffies32;
+	bbr->min_rtt_us = tcp_min_rtt(tp);
+	bbr->min_rtt_stamp = tcp_jiffies32;
+
+	bbr->has_seen_rtt = 0;
+	bbr_init_pacing_rate_from_rtt(sk);
+
+	bbr->round_start = 0;
+	bbr->idle_restart = 0;
+	bbr->full_bw_reached = 0;
+	bbr->full_bw = 0;
+	bbr->full_bw_cnt = 0;
+	bbr->cycle_mstamp = 0;
+	bbr->cycle_idx = 0;
+	bbr->mode = BBR_STARTUP;
+	bbr->debug.rs_bw = 0;
+
+	bbr->ack_epoch_mstamp = tp->tcp_mstamp;
+	bbr->ack_epoch_acked = 0;
+	bbr->extra_acked_win_rtts = 0;
+	bbr->extra_acked_win_idx = 0;
+	bbr->extra_acked[0] = 0;
+	bbr->extra_acked[1] = 0;
+
+	bbr->ce_state = 0;
+	bbr->prior_rcv_nxt = tp->rcv_nxt;
+	bbr->try_fast_path = 0;
+
+	cmpxchg(&sk->sk_pacing_status, SK_PACING_NONE, SK_PACING_NEEDED);
+}
+
+static u32 bbr_sndbuf_expand(struct sock *sk)
+{
+	/* Provision 3 * cwnd since BBR may slow-start even during recovery. */
+	return 3;
+}
+
+/* __________________________________________________________________________
+ *
+ * Functions new to BBR v2 ("bbr") congestion control are below here.
+ * __________________________________________________________________________
+ */
+
+/* Incorporate a new bw sample into the current window of our max filter. */
+static void bbr2_take_bw_hi_sample(struct sock *sk, u32 bw)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr->bw_hi[1] = max(bw, bbr->bw_hi[1]);
+}
+
+/* Keep max of last 1-2 cycles. Each PROBE_BW cycle, flip filter window. */
+static void bbr2_advance_bw_hi_filter(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (!bbr->bw_hi[1])
+		return;  /* no samples in this window; remember old window */
+	bbr->bw_hi[0] = bbr->bw_hi[1];
+	bbr->bw_hi[1] = 0;
+}
+
+/* How much do we want in flight? Our BDP, unless congestion cut cwnd. */
+static u32 bbr2_target_inflight(struct sock *sk)
+{
+	u32 bdp = bbr_inflight(sk, bbr_bw(sk), BBR_UNIT);
+
+	return min(bdp, tcp_sk(sk)->snd_cwnd);
+}
+
+static bool bbr2_is_probing_bandwidth(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	return (bbr->mode == BBR_STARTUP) ||
+		(bbr->mode == BBR_PROBE_BW &&
+		 (bbr->cycle_idx == BBR_BW_PROBE_REFILL ||
+		  bbr->cycle_idx == BBR_BW_PROBE_UP));
+}
+
+/* Has the given amount of time elapsed since we marked the phase start? */
+static bool bbr2_has_elapsed_in_phase(const struct sock *sk, u32 interval_us)
+{
+	const struct tcp_sock *tp = tcp_sk(sk);
+	const struct bbr *bbr = inet_csk_ca(sk);
+
+	return tcp_stamp_us_delta(tp->tcp_mstamp,
+				  bbr->cycle_mstamp + interval_us) > 0;
+}
+
+static void bbr2_handle_queue_too_high_in_startup(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr->full_bw_reached = 1;
+	bbr->inflight_hi = bbr_inflight(sk, bbr_max_bw(sk), BBR_UNIT);
+}
+
+/* Exit STARTUP upon N consecutive rounds with ECN mark rate > ecn_thresh. */
+static void bbr2_check_ecn_too_high_in_startup(struct sock *sk, u32 ce_ratio)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (bbr_full_bw_reached(sk) || !bbr->ecn_eligible ||
+	    !bbr->params.full_ecn_cnt || !bbr->params.ecn_thresh)
+		return;
+
+	if (ce_ratio >= bbr->params.ecn_thresh)
+		bbr->startup_ecn_rounds++;
+	else
+		bbr->startup_ecn_rounds = 0;
+
+	if (bbr->startup_ecn_rounds >= bbr->params.full_ecn_cnt) {
+		bbr->debug.event = 'E';  /* ECN caused STARTUP exit */
+		bbr2_handle_queue_too_high_in_startup(sk);
+		return;
+	}
+}
+
+static void bbr2_update_ecn_alpha(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	s32 delivered, delivered_ce;
+	u64 alpha, ce_ratio;
+	u32 gain;
+
+	if (bbr->params.ecn_factor == 0)
+		return;
+
+	delivered = tp->delivered - bbr->alpha_last_delivered;
+	delivered_ce = tp->delivered_ce - bbr->alpha_last_delivered_ce;
+
+	if (delivered == 0 ||		/* avoid divide by zero */
+	    WARN_ON_ONCE(delivered < 0 || delivered_ce < 0))  /* backwards? */
+		return;
+
+	/* See if we should use ECN sender logic for this connection. */
+	if (!bbr->ecn_eligible && bbr_ecn_enable &&
+	    (bbr->min_rtt_us <= bbr->params.ecn_max_rtt_us ||
+	     !bbr->params.ecn_max_rtt_us))
+		bbr->ecn_eligible = 1;
+
+	ce_ratio = (u64)delivered_ce << BBR_SCALE;
+	do_div(ce_ratio, delivered);
+	gain = bbr->params.ecn_alpha_gain;
+	alpha = ((BBR_UNIT - gain) * bbr->ecn_alpha) >> BBR_SCALE;
+	alpha += (gain * ce_ratio) >> BBR_SCALE;
+	bbr->ecn_alpha = min_t(u32, alpha, BBR_UNIT);
+
+	bbr->alpha_last_delivered = tp->delivered;
+	bbr->alpha_last_delivered_ce = tp->delivered_ce;
+
+	bbr2_check_ecn_too_high_in_startup(sk, ce_ratio);
+}
+
+/* Each round trip of BBR_BW_PROBE_UP, double volume of probing data. */
+static void bbr2_raise_inflight_hi_slope(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 growth_this_round, cnt;
+
+	/* Calculate "slope": packets S/Acked per inflight_hi increment. */
+	growth_this_round = 1 << bbr->bw_probe_up_rounds;
+	bbr->bw_probe_up_rounds = min(bbr->bw_probe_up_rounds + 1, 30);
+	cnt = tp->snd_cwnd / growth_this_round;
+	cnt = max(cnt, 1U);
+	bbr->bw_probe_up_cnt = cnt;
+	bbr->debug.event = 'G';  /* Grow inflight_hi slope */
+}
+
+/* In BBR_BW_PROBE_UP, not seeing high loss/ECN/queue, so raise inflight_hi. */
+static void bbr2_probe_inflight_hi_upward(struct sock *sk,
+					  const struct rate_sample *rs)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 delta;
+
+	if (!tp->is_cwnd_limited || tp->snd_cwnd < bbr->inflight_hi) {
+		bbr->bw_probe_up_acks = 0;  /* don't accmulate unused credits */
+		return;  /* not fully using inflight_hi, so don't grow it */
+	}
+
+	/* For each bw_probe_up_cnt packets ACKed, increase inflight_hi by 1. */
+	bbr->bw_probe_up_acks += rs->acked_sacked;
+	if (bbr->bw_probe_up_acks >=  bbr->bw_probe_up_cnt) {
+		delta = bbr->bw_probe_up_acks / bbr->bw_probe_up_cnt;
+		bbr->bw_probe_up_acks -= delta * bbr->bw_probe_up_cnt;
+		bbr->inflight_hi += delta;
+		bbr->debug.event = 'I';  /* Increment inflight_hi */
+	}
+
+	if (bbr->round_start)
+		bbr2_raise_inflight_hi_slope(sk);
+}
+
+/* Does loss/ECN rate for this sample say inflight is "too high"?
+ * This is used by both the bbr_check_loss_too_high_in_startup() function,
+ * which can be used in either v1 or v2, and the PROBE_UP phase of v2, which
+ * uses it to notice when loss/ECN rates suggest inflight is too high.
+ */
+static bool bbr2_is_inflight_too_high(const struct sock *sk,
+				     const struct rate_sample *rs)
+{
+	const struct bbr *bbr = inet_csk_ca(sk);
+	u32 loss_thresh, ecn_thresh;
+
+	if (rs->lost > 0 && rs->tx_in_flight) {
+		loss_thresh = (u64)rs->tx_in_flight * bbr->params.loss_thresh >>
+				BBR_SCALE;
+		if (rs->lost > loss_thresh)
+			return true;
+	}
+
+	if (rs->delivered_ce > 0 && rs->delivered > 0 &&
+	    bbr->ecn_eligible && bbr->params.ecn_thresh) {
+		ecn_thresh = (u64)rs->delivered * bbr->params.ecn_thresh >>
+				BBR_SCALE;
+		if (rs->delivered_ce >= ecn_thresh)
+			return true;
+	}
+
+	return false;
+}
+
+/* Calculate the tx_in_flight level that corresponded to excessive loss.
+ * We find "lost_prefix" segs of the skb where loss rate went too high,
+ * by solving for "lost_prefix" in the following equation:
+ *   lost                     /  inflight                     >= loss_thresh
+ *  (lost_prev + lost_prefix) / (inflight_prev + lost_prefix) >= loss_thresh
+ * Then we take that equation, convert it to fixed point, and
+ * round up to the nearest packet.
+ */
+static u32 bbr2_inflight_hi_from_lost_skb(const struct sock *sk,
+					  const struct rate_sample *rs,
+					  const struct sk_buff *skb)
+{
+	const struct bbr *bbr = inet_csk_ca(sk);
+	u32 loss_thresh  = bbr->params.loss_thresh;
+	u32 pcount, divisor, inflight_hi;
+	s32 inflight_prev, lost_prev;
+	u64 loss_budget, lost_prefix;
+
+	pcount = tcp_skb_pcount(skb);
+
+	/* How much data was in flight before this skb? */
+	inflight_prev = rs->tx_in_flight - pcount;
+	if (WARN_ONCE(inflight_prev < 0,
+		      "tx_in_flight: %u pcount: %u reneg: %u",
+		      rs->tx_in_flight, pcount, tcp_sk(sk)->is_sack_reneg))
+		return ~0U;
+
+	/* How much inflight data was marked lost before this skb? */
+	lost_prev = rs->lost - pcount;
+	if (WARN_ON_ONCE(lost_prev < 0))
+		return ~0U;
+
+	/* At what prefix of this lost skb did losss rate exceed loss_thresh? */
+	loss_budget = (u64)inflight_prev * loss_thresh + BBR_UNIT - 1;
+	loss_budget >>= BBR_SCALE;
+	if (lost_prev >= loss_budget) {
+		lost_prefix = 0;   /* previous losses crossed loss_thresh */
+	} else {
+		lost_prefix = loss_budget - lost_prev;
+		lost_prefix <<= BBR_SCALE;
+		divisor = BBR_UNIT - loss_thresh;
+		if (WARN_ON_ONCE(!divisor))  /* loss_thresh is 8 bits */
+			return ~0U;
+		do_div(lost_prefix, divisor);
+	}
+
+	inflight_hi = inflight_prev + lost_prefix;
+	return inflight_hi;
+}
+
+/* If loss/ECN rates during probing indicated we may have overfilled a
+ * buffer, return an operating point that tries to leave unutilized headroom in
+ * the path for other flows, for fairness convergence and lower RTTs and loss.
+ */
+static u32 bbr2_inflight_with_headroom(const struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 headroom, headroom_fraction;
+
+	if (bbr->inflight_hi == ~0U)
+		return ~0U;
+
+	headroom_fraction = bbr->params.inflight_headroom;
+	headroom = ((u64)bbr->inflight_hi * headroom_fraction) >> BBR_SCALE;
+	headroom = max(headroom, 1U);
+	return max_t(s32, bbr->inflight_hi - headroom,
+		     bbr->params.cwnd_min_target);
+}
+
+/* Bound cwnd to a sensible level, based on our current probing state
+ * machine phase and model of a good inflight level (inflight_lo, inflight_hi).
+ */
+static void bbr2_bound_cwnd_for_inflight_model(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 cap;
+
+	/* tcp_rcv_synsent_state_process() currently calls tcp_ack()
+	 * and thus cong_control() without first initializing us(!).
+	 */
+	if (!bbr->initialized)
+		return;
+
+	cap = ~0U;
+	if (bbr->mode == BBR_PROBE_BW &&
+	    bbr->cycle_idx != BBR_BW_PROBE_CRUISE) {
+		/* Probe to see if more packets fit in the path. */
+		cap = bbr->inflight_hi;
+	} else {
+		if (bbr->mode == BBR_PROBE_RTT ||
+		    (bbr->mode == BBR_PROBE_BW &&
+		     bbr->cycle_idx == BBR_BW_PROBE_CRUISE))
+			cap = bbr2_inflight_with_headroom(sk);
+	}
+	/* Adapt to any loss/ECN since our last bw probe. */
+	cap = min(cap, bbr->inflight_lo);
+
+	cap = max_t(u32, cap, bbr->params.cwnd_min_target);
+	tp->snd_cwnd = min(cap, tp->snd_cwnd);
+}
+
+/* Estimate a short-term lower bound on the capacity available now, based
+ * on measurements of the current delivery process and recent history. When we
+ * are seeing loss/ECN at times when we are not probing bw, then conservatively
+ * move toward flow balance by multiplicatively cutting our short-term
+ * estimated safe rate and volume of data (bw_lo and inflight_lo). We use a
+ * multiplicative decrease in order to converge to a lower capacity in time
+ * logarithmic in the magnitude of the decrease.
+ *
+ * However, we do not cut our short-term estimates lower than the current rate
+ * and volume of delivered data from this round trip, since from the current
+ * delivery process we can estimate the measured capacity available now.
+ *
+ * Anything faster than that approach would knowingly risk high loss, which can
+ * cause low bw for Reno/CUBIC and high loss recovery latency for
+ * request/response flows using any congestion control.
+ */
+static void bbr2_adapt_lower_bounds(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 ecn_cut, ecn_inflight_lo, beta;
+
+	/* We only use lower-bound estimates when not probing bw.
+	 * When probing we need to push inflight higher to probe bw.
+	 */
+	if (bbr2_is_probing_bandwidth(sk))
+		return;
+
+	/* ECN response. */
+	if (bbr->ecn_in_round && bbr->ecn_eligible && bbr->params.ecn_factor) {
+		/* Reduce inflight to (1 - alpha*ecn_factor). */
+		ecn_cut = (BBR_UNIT -
+			   ((bbr->ecn_alpha * bbr->params.ecn_factor) >>
+			    BBR_SCALE));
+		if (bbr->inflight_lo == ~0U)
+			bbr->inflight_lo = tp->snd_cwnd;
+		ecn_inflight_lo = (u64)bbr->inflight_lo * ecn_cut >> BBR_SCALE;
+	} else {
+		ecn_inflight_lo = ~0U;
+	}
+
+	/* Loss response. */
+	if (bbr->loss_in_round) {
+		/* Reduce bw and inflight to (1 - beta). */
+		if (bbr->bw_lo == ~0U)
+			bbr->bw_lo = bbr_max_bw(sk);
+		if (bbr->inflight_lo == ~0U)
+			bbr->inflight_lo = tp->snd_cwnd;
+		beta = bbr->params.beta;
+		bbr->bw_lo =
+			max_t(u32, bbr->bw_latest,
+			      (u64)bbr->bw_lo *
+			      (BBR_UNIT - beta) >> BBR_SCALE);
+		bbr->inflight_lo =
+			max_t(u32, bbr->inflight_latest,
+			      (u64)bbr->inflight_lo *
+			      (BBR_UNIT - beta) >> BBR_SCALE);
+	}
+
+	/* Adjust to the lower of the levels implied by loss or ECN. */
+	bbr->inflight_lo = min(bbr->inflight_lo, ecn_inflight_lo);
+}
+
+/* Reset any short-term lower-bound adaptation to congestion, so that we can
+ * push our inflight up.
+ */
+static void bbr2_reset_lower_bounds(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr->bw_lo = ~0U;
+	bbr->inflight_lo = ~0U;
+}
+
+/* After bw probing (STARTUP/PROBE_UP), reset signals before entering a state
+ * machine phase where we adapt our lower bound based on congestion signals.
+ */
+static void bbr2_reset_congestion_signals(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr->loss_in_round = 0;
+	bbr->ecn_in_round = 0;
+	bbr->loss_in_cycle = 0;
+	bbr->ecn_in_cycle = 0;
+	bbr->bw_latest = 0;
+	bbr->inflight_latest = 0;
+}
+
+/* Update (most of) our congestion signals: track the recent rate and volume of
+ * delivered data, presence of loss, and EWMA degree of ECN marking.
+ */
+static void bbr2_update_congestion_signals(
+	struct sock *sk, const struct rate_sample *rs, struct bbr_context *ctx)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	u64 bw;
+
+	bbr->loss_round_start = 0;
+	if (rs->interval_us <= 0 || !rs->acked_sacked)
+		return; /* Not a valid observation */
+	bw = ctx->sample_bw;
+
+	if (!rs->is_app_limited || bw >= bbr_max_bw(sk))
+		bbr2_take_bw_hi_sample(sk, bw);
+
+	bbr->loss_in_round |= (rs->losses > 0);
+
+	/* Update rate and volume of delivered data from latest round trip: */
+	bbr->bw_latest       = max_t(u32, bbr->bw_latest,       ctx->sample_bw);
+	bbr->inflight_latest = max_t(u32, bbr->inflight_latest, rs->delivered);
+
+	if (before(rs->prior_delivered, bbr->loss_round_delivered))
+		return;		/* skip the per-round-trip updates */
+	/* Now do per-round-trip updates. */
+	bbr->loss_round_delivered = tp->delivered;  /* mark round trip */
+	bbr->loss_round_start = 1;
+	bbr2_adapt_lower_bounds(sk);
+
+	/* Update windowed "latest" (single-round-trip) filters. */
+	bbr->loss_in_round = 0;
+	bbr->ecn_in_round  = 0;
+	bbr->bw_latest = ctx->sample_bw;
+	bbr->inflight_latest = rs->delivered;
+}
+
+/* Bandwidth probing can cause loss. To help coexistence with loss-based
+ * congestion control we spread out our probing in a Reno-conscious way. Due to
+ * the shape of the Reno sawtooth, the time required between loss epochs for an
+ * idealized Reno flow is a number of round trips that is the BDP of that
+ * flow. We count packet-timed round trips directly, since measured RTT can
+ * vary widely, and Reno is driven by packet-timed round trips.
+ */
+static bool bbr2_is_reno_coexistence_probe_time(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 inflight, rounds, reno_gain, reno_rounds;
+
+	/* Random loss can shave some small percentage off of our inflight
+	 * in each round. To survive this, flows need robust periodic probes.
+	 */
+	rounds = bbr->params.bw_probe_max_rounds;
+
+	reno_gain = bbr->params.bw_probe_reno_gain;
+	if (reno_gain) {
+		inflight = bbr2_target_inflight(sk);
+		reno_rounds = ((u64)inflight * reno_gain) >> BBR_SCALE;
+		rounds = min(rounds, reno_rounds);
+	}
+	return bbr->rounds_since_probe >= rounds;
+}
+
+/* How long do we want to wait before probing for bandwidth (and risking
+ * loss)? We randomize the wait, for better mixing and fairness convergence.
+ *
+ * We bound the Reno-coexistence inter-bw-probe time to be 62-63 round trips.
+ * This is calculated to allow fairness with a 25Mbps, 30ms Reno flow,
+ * (eg 4K video to a broadband user):
+ *   BDP = 25Mbps * .030sec /(1514bytes) = 61.9 packets
+ *
+ * We bound the BBR-native inter-bw-probe wall clock time to be:
+ *  (a) higher than 2 sec: to try to avoid causing loss for a long enough time
+ *      to allow Reno at 30ms to get 4K video bw, the inter-bw-probe time must
+ *      be at least: 25Mbps * .030sec / (1514bytes) * 0.030sec = 1.9secs
+ *  (b) lower than 3 sec: to ensure flows can start probing in a reasonable
+ *      amount of time to discover unutilized bw on human-scale interactive
+ *      time-scales (e.g. perhaps traffic from a web page download that we
+ *      were competing with is now complete).
+ */
+static void bbr2_pick_probe_wait(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	/* Decide the random round-trip bound for wait until probe: */
+	bbr->rounds_since_probe =
+		prandom_u32_max(bbr->params.bw_probe_rand_rounds);
+	/* Decide the random wall clock bound for wait until probe: */
+	bbr->probe_wait_us = bbr->params.bw_probe_base_us +
+			     prandom_u32_max(bbr->params.bw_probe_rand_us);
+}
+
+static void bbr2_set_cycle_idx(struct sock *sk, int cycle_idx)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr->cycle_idx = cycle_idx;
+	/* New phase, so need to update cwnd and pacing rate. */
+	bbr->try_fast_path = 0;
+}
+
+/* Send at estimated bw to fill the pipe, but not queue. We need this phase
+ * before PROBE_UP, because as soon as we send faster than the available bw
+ * we will start building a queue, and if the buffer is shallow we can cause
+ * loss. If we do not fill the pipe before we cause this loss, our bw_hi and
+ * inflight_hi estimates will underestimate.
+ */
+static void bbr2_start_bw_probe_refill(struct sock *sk, u32 bw_probe_up_rounds)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr2_reset_lower_bounds(sk);
+	if (bbr->inflight_hi != ~0U)
+		bbr->inflight_hi += bbr->params.refill_add_inc;
+	bbr->bw_probe_up_rounds = bw_probe_up_rounds;
+	bbr->bw_probe_up_acks = 0;
+	bbr->stopped_risky_probe = 0;
+	bbr->ack_phase = BBR_ACKS_REFILLING;
+	bbr->next_rtt_delivered = tp->delivered;
+	bbr2_set_cycle_idx(sk, BBR_BW_PROBE_REFILL);
+}
+
+/* Now probe max deliverable data rate and volume. */
+static void bbr2_start_bw_probe_up(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr->ack_phase = BBR_ACKS_PROBE_STARTING;
+	bbr->next_rtt_delivered = tp->delivered;
+	bbr->cycle_mstamp = tp->tcp_mstamp;
+	bbr2_set_cycle_idx(sk, BBR_BW_PROBE_UP);
+	bbr2_raise_inflight_hi_slope(sk);
+}
+
+/* Start a new PROBE_BW probing cycle of some wall clock length. Pick a wall
+ * clock time at which to probe beyond an inflight that we think to be
+ * safe. This will knowingly risk packet loss, so we want to do this rarely, to
+ * keep packet loss rates low. Also start a round-trip counter, to probe faster
+ * if we estimate a Reno flow at our BDP would probe faster.
+ */
+static void bbr2_start_bw_probe_down(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr2_reset_congestion_signals(sk);
+	bbr->bw_probe_up_cnt = ~0U;     /* not growing inflight_hi any more */
+	bbr2_pick_probe_wait(sk);
+	bbr->cycle_mstamp = tp->tcp_mstamp;		/* start wall clock */
+	bbr->ack_phase = BBR_ACKS_PROBE_STOPPING;
+	bbr->next_rtt_delivered = tp->delivered;
+	bbr2_set_cycle_idx(sk, BBR_BW_PROBE_DOWN);
+}
+
+/* Cruise: maintain what we estimate to be a neutral, conservative
+ * operating point, without attempting to probe up for bandwidth or down for
+ * RTT, and only reducing inflight in response to loss/ECN signals.
+ */
+static void bbr2_start_bw_probe_cruise(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (bbr->inflight_lo != ~0U)
+		bbr->inflight_lo = min(bbr->inflight_lo, bbr->inflight_hi);
+
+	bbr2_set_cycle_idx(sk, BBR_BW_PROBE_CRUISE);
+}
+
+/* Loss and/or ECN rate is too high while probing.
+ * Adapt (once per bw probe) by cutting inflight_hi and then restarting cycle.
+ */
+static void bbr2_handle_inflight_too_high(struct sock *sk,
+					  const struct rate_sample *rs)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	const u32 beta = bbr->params.beta;
+
+	bbr->prev_probe_too_high = 1;
+	bbr->bw_probe_samples = 0;  /* only react once per probe */
+	bbr->debug.event = 'L';     /* Loss/ECN too high */
+	/* If we are app-limited then we are not robustly
+	 * probing the max volume of inflight data we think
+	 * might be safe (analogous to how app-limited bw
+	 * samples are not known to be robustly probing bw).
+	 */
+	if (!rs->is_app_limited)
+		bbr->inflight_hi = max_t(u32, rs->tx_in_flight,
+					 (u64)bbr2_target_inflight(sk) *
+					 (BBR_UNIT - beta) >> BBR_SCALE);
+	if (bbr->mode == BBR_PROBE_BW && bbr->cycle_idx == BBR_BW_PROBE_UP)
+		bbr2_start_bw_probe_down(sk);
+}
+
+/* If we're seeing bw and loss samples reflecting our bw probing, adapt
+ * using the signals we see. If loss or ECN mark rate gets too high, then adapt
+ * inflight_hi downward. If we're able to push inflight higher without such
+ * signals, push higher: adapt inflight_hi upward.
+ */
+static bool bbr2_adapt_upper_bounds(struct sock *sk,
+				   const struct rate_sample *rs)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	/* Track when we'll see bw/loss samples resulting from our bw probes. */
+	if (bbr->ack_phase == BBR_ACKS_PROBE_STARTING && bbr->round_start)
+		bbr->ack_phase = BBR_ACKS_PROBE_FEEDBACK;
+	if (bbr->ack_phase == BBR_ACKS_PROBE_STOPPING && bbr->round_start) {
+		/* End of samples from bw probing phase. */
+		bbr->bw_probe_samples = 0;
+		bbr->ack_phase = BBR_ACKS_INIT;
+		/* At this point in the cycle, our current bw sample is also
+		 * our best recent chance at finding the highest available bw
+		 * for this flow. So now is the best time to forget the bw
+		 * samples from the previous cycle, by advancing the window.
+		 */
+		if (bbr->mode == BBR_PROBE_BW && !rs->is_app_limited)
+			bbr2_advance_bw_hi_filter(sk);
+		/* If we had an inflight_hi, then probed and pushed inflight all
+		 * the way up to hit that inflight_hi without seeing any
+		 * high loss/ECN in all the resulting ACKs from that probing,
+		 * then probe up again, this time letting inflight persist at
+		 * inflight_hi for a round trip, then accelerating beyond.
+		 */
+		if (bbr->mode == BBR_PROBE_BW &&
+		    bbr->stopped_risky_probe && !bbr->prev_probe_too_high) {
+			bbr->debug.event = 'R';  /* reprobe */
+			bbr2_start_bw_probe_refill(sk, 0);
+			return true;  /* yes, decided state transition */
+		}
+	}
+
+	if (bbr2_is_inflight_too_high(sk, rs)) {
+		if (bbr->bw_probe_samples)  /*  sample is from bw probing? */
+			bbr2_handle_inflight_too_high(sk, rs);
+	} else {
+		/* Loss/ECN rate is declared safe. Adjust upper bound upward. */
+		if (bbr->inflight_hi == ~0U)  /* no excess queue signals yet? */
+			return false;
+
+		/* To be resilient to random loss, we must raise inflight_hi
+		 * if we observe in any phase that a higher level is safe.
+		 */
+		if (rs->tx_in_flight > bbr->inflight_hi) {
+			bbr->inflight_hi = rs->tx_in_flight;
+			bbr->debug.event = 'U';  /* raise up inflight_hi */
+		}
+
+		if (bbr->mode == BBR_PROBE_BW &&
+		    bbr->cycle_idx == BBR_BW_PROBE_UP)
+			bbr2_probe_inflight_hi_upward(sk, rs);
+	}
+
+	return false;
+}
+
+/* Check if it's time to probe for bandwidth now, and if so, kick it off. */
+static bool bbr2_check_time_to_probe_bw(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 n;
+
+	/* If we seem to be at an operating point where we are not seeing loss
+	 * but we are seeing ECN marks, then when the ECN marks cease we reprobe
+	 * quickly (in case a burst of cross-traffic has ceased and freed up bw,
+	 * or in case we are sharing with multiplicatively probing traffic).
+	 */
+	if (bbr->params.ecn_reprobe_gain && bbr->ecn_eligible &&
+	    bbr->ecn_in_cycle && !bbr->loss_in_cycle &&
+	    inet_csk(sk)->icsk_ca_state == TCP_CA_Open) {
+		bbr->debug.event = 'A';  /* *A*ll clear to probe *A*gain */
+		/* Calculate n so that when bbr2_raise_inflight_hi_slope()
+		 * computes growth_this_round as 2^n it will be roughly the
+		 * desired volume of data (inflight_hi*ecn_reprobe_gain).
+		 */
+		n = ilog2((((u64)bbr->inflight_hi *
+			    bbr->params.ecn_reprobe_gain) >> BBR_SCALE));
+		bbr2_start_bw_probe_refill(sk, n);
+		return true;
+	}
+
+	if (bbr2_has_elapsed_in_phase(sk, bbr->probe_wait_us) ||
+	    bbr2_is_reno_coexistence_probe_time(sk)) {
+		bbr2_start_bw_probe_refill(sk, 0);
+		return true;
+	}
+	return false;
+}
+
+/* Is it time to transition from PROBE_DOWN to PROBE_CRUISE? */
+static bool bbr2_check_time_to_cruise(struct sock *sk, u32 inflight, u32 bw)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	bool is_under_bdp, is_long_enough;
+
+	/* Always need to pull inflight down to leave headroom in queue. */
+	if (inflight > bbr2_inflight_with_headroom(sk))
+		return false;
+
+	is_under_bdp = inflight <= bbr_inflight(sk, bw, BBR_UNIT);
+	if (bbr->params.drain_to_target)
+		return is_under_bdp;
+
+	is_long_enough = bbr2_has_elapsed_in_phase(sk, bbr->min_rtt_us);
+	return is_under_bdp || is_long_enough;
+}
+
+/* PROBE_BW state machine: cruise, refill, probe for bw, or drain? */
+static void bbr2_update_cycle_phase(struct sock *sk,
+				    const struct rate_sample *rs)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	bool is_risky = false, is_queuing = false;
+	u32 inflight, bw;
+
+	if (!bbr_full_bw_reached(sk))
+		return;
+
+	/* In DRAIN, PROBE_BW, or PROBE_RTT, adjust upper bounds. */
+	if (bbr2_adapt_upper_bounds(sk, rs))
+		return;		/* already decided state transition */
+
+	if (bbr->mode != BBR_PROBE_BW)
+		return;
+
+	inflight = bbr_packets_in_net_at_edt(sk, rs->prior_in_flight);
+	bw = bbr_max_bw(sk);
+
+	switch (bbr->cycle_idx) {
+	/* First we spend most of our time cruising with a pacing_gain of 1.0,
+	 * which paces at the estimated bw, to try to fully use the pipe
+	 * without building queue. If we encounter loss/ECN marks, we adapt
+	 * by slowing down.
+	 */
+	case BBR_BW_PROBE_CRUISE:
+		if (bbr2_check_time_to_probe_bw(sk))
+			return;		/* already decided state transition */
+		break;
+
+	/* After cruising, when it's time to probe, we first "refill": we send
+	 * at the estimated bw to fill the pipe, before probing higher and
+	 * knowingly risking overflowing the bottleneck buffer (causing loss).
+	 */
+	case BBR_BW_PROBE_REFILL:
+		if (bbr->round_start) {
+			/* After one full round trip of sending in REFILL, we
+			 * start to see bw samples reflecting our REFILL, which
+			 * may be putting too much data in flight.
+			 */
+			bbr->bw_probe_samples = 1;
+			bbr2_start_bw_probe_up(sk);
+		}
+		break;
+
+	/* After we refill the pipe, we probe by using a pacing_gain > 1.0, to
+	 * probe for bw. If we have not seen loss/ECN, we try to raise inflight
+	 * to at least pacing_gain*BDP; note that this may take more than
+	 * min_rtt if min_rtt is small (e.g. on a LAN).
+	 *
+	 * We terminate PROBE_UP bandwidth probing upon any of the following:
+	 *
+	 * (1) We've pushed inflight up to hit the inflight_hi target set in the
+	 *     most recent previous bw probe phase. Thus we want to start
+	 *     draining the queue immediately because it's very likely the most
+	 *     recently sent packets will fill the queue and cause drops.
+	 *     (checked here)
+	 * (2) We have probed for at least 1*min_rtt_us, and the
+	 *     estimated queue is high enough (inflight > 1.25 * estimated_bdp).
+	 *     (checked here)
+	 * (3) Loss filter says loss rate is "too high".
+	 *     (checked in bbr_is_inflight_too_high())
+	 * (4) ECN filter says ECN mark rate is "too high".
+	 *     (checked in bbr_is_inflight_too_high())
+	 */
+	case BBR_BW_PROBE_UP:
+		if (bbr->prev_probe_too_high &&
+		    inflight >= bbr->inflight_hi) {
+			bbr->stopped_risky_probe = 1;
+			is_risky = true;
+			bbr->debug.event = 'D';   /* D for danger */
+		} else if (bbr2_has_elapsed_in_phase(sk, bbr->min_rtt_us) &&
+			   inflight >=
+			   bbr_inflight(sk, bw,
+					bbr->params.bw_probe_pif_gain)) {
+			is_queuing = true;
+			bbr->debug.event = 'Q'; /* building Queue */
+		}
+		if (is_risky || is_queuing) {
+			bbr->prev_probe_too_high = 0;  /* no loss/ECN (yet) */
+			bbr2_start_bw_probe_down(sk);  /* restart w/ down */
+		}
+		break;
+
+	/* After probing in PROBE_UP, we have usually accumulated some data in
+	 * the bottleneck buffer (if bw probing didn't find more bw). We next
+	 * enter PROBE_DOWN to try to drain any excess data from the queue. To
+	 * do this, we use a pacing_gain < 1.0. We hold this pacing gain until
+	 * our inflight is less then that target cruising point, which is the
+	 * minimum of (a) the amount needed to leave headroom, and (b) the
+	 * estimated BDP. Once inflight falls to match the target, we estimate
+	 * the queue is drained; persisting would underutilize the pipe.
+	 */
+	case BBR_BW_PROBE_DOWN:
+		if (bbr2_check_time_to_probe_bw(sk))
+			return;		/* already decided state transition */
+		if (bbr2_check_time_to_cruise(sk, inflight, bw))
+			bbr2_start_bw_probe_cruise(sk);
+		break;
+
+	default:
+		WARN_ONCE(1, "BBR invalid cycle index %u\n", bbr->cycle_idx);
+	}
+}
+
+/* Exiting PROBE_RTT, so return to bandwidth probing in STARTUP or PROBE_BW. */
+static void bbr2_exit_probe_rtt(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr2_reset_lower_bounds(sk);
+	if (bbr_full_bw_reached(sk)) {
+		bbr->mode = BBR_PROBE_BW;
+		/* Raising inflight after PROBE_RTT may cause loss, so reset
+		 * the PROBE_BW clock and schedule the next bandwidth probe for
+		 * a friendly and randomized future point in time.
+		 */
+		bbr2_start_bw_probe_down(sk);
+		/* Since we are exiting PROBE_RTT, we know inflight is
+		 * below our estimated BDP, so it is reasonable to cruise.
+		 */
+		bbr2_start_bw_probe_cruise(sk);
+	} else {
+		bbr->mode = BBR_STARTUP;
+	}
+}
+
+/* Exit STARTUP based on loss rate > 1% and loss gaps in round >= N. Wait until
+ * the end of the round in recovery to get a good estimate of how many packets
+ * have been lost, and how many we need to drain with a low pacing rate.
+ */
+static void bbr2_check_loss_too_high_in_startup(struct sock *sk,
+					       const struct rate_sample *rs)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (bbr_full_bw_reached(sk))
+		return;
+
+	/* For STARTUP exit, check the loss rate at the end of each round trip
+	 * of Recovery episodes in STARTUP. We check the loss rate at the end
+	 * of the round trip to filter out noisy/low loss and have a better
+	 * sense of inflight (extent of loss), so we can drain more accurately.
+	 */
+	if (rs->losses && bbr->loss_events_in_round < 0xf)
+		bbr->loss_events_in_round++;  /* update saturating counter */
+	if (bbr->params.full_loss_cnt && bbr->loss_round_start &&
+	    inet_csk(sk)->icsk_ca_state == TCP_CA_Recovery &&
+	    bbr->loss_events_in_round >= bbr->params.full_loss_cnt &&
+	    bbr2_is_inflight_too_high(sk, rs)) {
+		bbr->debug.event = 'P';  /* Packet loss caused STARTUP exit */
+		bbr2_handle_queue_too_high_in_startup(sk);
+		return;
+	}
+	if (bbr->loss_round_start)
+		bbr->loss_events_in_round = 0;
+}
+
+/* If we are done draining, advance into steady state operation in PROBE_BW. */
+static void bbr2_check_drain(struct sock *sk, const struct rate_sample *rs,
+			     struct bbr_context *ctx)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (bbr_check_drain(sk, rs, ctx)) {
+		bbr->mode = BBR_PROBE_BW;
+		bbr2_start_bw_probe_down(sk);
+	}
+}
+
+static void bbr2_update_model(struct sock *sk, const struct rate_sample *rs,
+			      struct bbr_context *ctx)
+{
+	bbr2_update_congestion_signals(sk, rs, ctx);
+	bbr_update_ack_aggregation(sk, rs);
+	bbr2_check_loss_too_high_in_startup(sk, rs);
+	bbr_check_full_bw_reached(sk, rs);
+	bbr2_check_drain(sk, rs, ctx);
+	bbr2_update_cycle_phase(sk, rs);
+	bbr_update_min_rtt(sk, rs);
+}
+
+/* Fast path for app-limited case.
+ *
+ * On each ack, we execute bbr state machine, which primarily consists of:
+ * 1) update model based on new rate sample, and
+ * 2) update control based on updated model or state change.
+ *
+ * There are certain workload/scenarios, e.g. app-limited case, where
+ * either we can skip updating model or we can skip update of both model
+ * as well as control. This provides signifcant softirq cpu savings for
+ * processing incoming acks.
+ *
+ * In case of app-limited, if there is no congestion (loss/ecn) and
+ * if observed bw sample is less than current estimated bw, then we can
+ * skip some of the computation in bbr state processing:
+ *
+ * - if there is no rtt/mode/phase change: In this case, since all the
+ *   parameters of the network model are constant, we can skip model
+ *   as well control update.
+ *
+ * - else we can skip rest of the model update. But we still need to
+ *   update the control to account for the new rtt/mode/phase.
+ *
+ * Returns whether we can take fast path or not.
+ */
+static bool bbr2_fast_path(struct sock *sk, bool *update_model,
+		const struct rate_sample *rs, struct bbr_context *ctx)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+	u32 prev_min_rtt_us, prev_mode;
+
+	if (bbr->params.fast_path && bbr->try_fast_path &&
+	    rs->is_app_limited && ctx->sample_bw < bbr_max_bw(sk) &&
+	    !bbr->loss_in_round && !bbr->ecn_in_round) {
+		prev_mode = bbr->mode;
+		prev_min_rtt_us = bbr->min_rtt_us;
+		bbr2_check_drain(sk, rs, ctx);
+		bbr2_update_cycle_phase(sk, rs);
+		bbr_update_min_rtt(sk, rs);
+
+		if (bbr->mode == prev_mode &&
+		    bbr->min_rtt_us == prev_min_rtt_us &&
+		    bbr->try_fast_path)
+			return true;
+
+		/* Skip model update, but control still needs to be updated */
+		*update_model = false;
+	}
+	return false;
+}
+
+static void bbr2_main(struct sock *sk, const struct rate_sample *rs)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	struct bbr_context ctx = { 0 };
+	bool update_model = true;
+	u32 bw;
+
+	bbr->debug.event = '.';  /* init to default NOP (no event yet) */
+
+	bbr_update_round_start(sk, rs, &ctx);
+	if (bbr->round_start) {
+		bbr->rounds_since_probe =
+			min_t(s32, bbr->rounds_since_probe + 1, 0xFF);
+		bbr2_update_ecn_alpha(sk);
+	}
+
+	bbr->ecn_in_round  |= rs->is_ece;
+	bbr_calculate_bw_sample(sk, rs, &ctx);
+
+	if (bbr2_fast_path(sk, &update_model, rs, &ctx))
+		goto out;
+
+	if (update_model)
+		bbr2_update_model(sk, rs, &ctx);
+
+	bbr_update_gains(sk);
+	bw = bbr_bw(sk);
+	bbr_set_pacing_rate(sk, bw, bbr->pacing_gain);
+	bbr_set_cwnd(sk, rs, rs->acked_sacked, bw, bbr->cwnd_gain,
+		     tp->snd_cwnd, &ctx);
+	bbr2_bound_cwnd_for_inflight_model(sk);
+
+out:
+	bbr->prev_ca_state = inet_csk(sk)->icsk_ca_state;
+	bbr->loss_in_cycle |= rs->lost > 0;
+	bbr->ecn_in_cycle  |= rs->delivered_ce > 0;
+
+	bbr_debug(sk, rs->acked_sacked, rs, &ctx);
+}
+
+/* Module parameters that are settable by TCP_CONGESTION_PARAMS are declared
+ * down here, so that the algorithm functions that use the parameters must use
+ * the per-socket parameters; if they accidentally use the global version
+ * then there will be a compile error.
+ * TODO(ncardwell): move all per-socket parameters down to this section.
+ */
+
+/* On losses, scale down inflight and pacing rate by beta scaled by BBR_SCALE.
+ * No loss response when 0. Max allwed value is 255.
+ */
+static u32 bbr_beta = BBR_UNIT * 30 / 100;
+
+/* Gain factor for ECN mark ratio samples, scaled by BBR_SCALE.
+ * Max allowed value is 255.
+ */
+static u32 bbr_ecn_alpha_gain = BBR_UNIT * 1 / 16;  /* 1/16 = 6.25% */
+
+/* The initial value for the ecn_alpha state variable. Default and max
+ * BBR_UNIT (256), representing 1.0. This allows a flow to respond quickly
+ * to congestion if the bottleneck is congested when the flow starts up.
+ */
+static u32 bbr_ecn_alpha_init = BBR_UNIT;	/* 1.0, to respond quickly */
+
+/* On ECN, cut inflight_lo to (1 - ecn_factor * ecn_alpha) scaled by BBR_SCALE.
+ * No ECN based bounding when 0. Max allwed value is 255.
+ */
+static u32 bbr_ecn_factor = BBR_UNIT * 1 / 3;	    /* 1/3 = 33% */
+
+/* Estimate bw probing has gone too far if CE ratio exceeds this threshold.
+ * Scaled by BBR_SCALE. Disabled when 0. Max allowed is 255.
+ */
+static u32 bbr_ecn_thresh = BBR_UNIT * 1 / 2;  /* 1/2 = 50% */
+
+/* Max RTT (in usec) at which to use sender-side ECN logic.
+ * Disabled when 0 (ECN allowed at any RTT).
+ * Max allowed for the parameter is 524287 (0x7ffff) us, ~524 ms.
+ */
+static u32 bbr_ecn_max_rtt_us = 5000;
+
+/* If non-zero, if in a cycle with no losses but some ECN marks, after ECN
+ * clears then use a multiplicative increase to quickly reprobe bw by
+ * starting inflight probing at the given multiple of inflight_hi.
+ * Default for this experimental knob is 0 (disabled).
+ * Planned value for experiments: BBR_UNIT * 1 / 2 = 128, representing 0.5.
+ */
+static u32 bbr_ecn_reprobe_gain;
+
+/* Estimate bw probing has gone too far if loss rate exceeds this level. */
+static u32 bbr_loss_thresh = BBR_UNIT * 2 / 100;  /* 2% loss */
+
+/* Exit STARTUP if number of loss marking events in a Recovery round is >= N,
+ * and loss rate is higher than bbr_loss_thresh.
+ * Disabled if 0. Max allowed value is 15 (0xF).
+ */
+static u32 bbr_full_loss_cnt = 8;
+
+/* Exit STARTUP if number of round trips with ECN mark rate above ecn_thresh
+ * meets this count. Max allowed value is 3.
+ */
+static u32 bbr_full_ecn_cnt = 2;
+
+/* Fraction of unutilized headroom to try to leave in path upon high loss. */
+static u32 bbr_inflight_headroom = BBR_UNIT * 15 / 100;
+
+/* Multiplier to get target inflight (as multiple of BDP) for PROBE_UP phase.
+ * Default is 1.25x, as in BBR v1. Max allowed is 511.
+ */
+static u32 bbr_bw_probe_pif_gain = BBR_UNIT * 5 / 4;
+
+/* Multiplier to get Reno-style probe epoch duration as: k * BDP round trips.
+ * If zero, disables this BBR v2 Reno-style BDP-scaled coexistence mechanism.
+ * Max allowed is 511.
+ */
+static u32 bbr_bw_probe_reno_gain = BBR_UNIT;
+
+/* Max number of packet-timed rounds to wait before probing for bandwidth.  If
+ * we want to tolerate 1% random loss per round, and not have this cut our
+ * inflight too much, we must probe for bw periodically on roughly this scale.
+ * If low, limits Reno/CUBIC coexistence; if high, limits loss tolerance.
+ * We aim to be fair with Reno/CUBIC up to a BDP of at least:
+ *  BDP = 25Mbps * .030sec /(1514bytes) = 61.9 packets
+ */
+static u32 bbr_bw_probe_max_rounds = 63;
+
+/* Max amount of randomness to inject in round counting for Reno-coexistence.
+ * Max value is 15.
+ */
+static u32 bbr_bw_probe_rand_rounds = 2;
+
+/* Use BBR-native probe time scale starting at this many usec.
+ * We aim to be fair with Reno/CUBIC up to an inter-loss time epoch of at least:
+ *  BDP*RTT = 25Mbps * .030sec /(1514bytes) * 0.030sec = 1.9 secs
+ */
+static u32 bbr_bw_probe_base_us = 2 * USEC_PER_SEC;  /* 2 secs */
+
+/* Use BBR-native probes spread over this many usec: */
+static u32 bbr_bw_probe_rand_us = 1 * USEC_PER_SEC;  /* 1 secs */
+
+/* Undo the model changes made in loss recovery if recovery was spurious? */
+static bool bbr_undo = true;
+
+/* Use fast path if app-limited, no loss/ECN, and target cwnd was reached? */
+static bool bbr_fast_path = true;	/* default: enabled */
+
+/* Use fast ack mode ? */
+static int bbr_fast_ack_mode = 1;	/* default: rwnd check off */
+
+/* How much to additively increase inflight_hi when entering REFILL? */
+static u32 bbr_refill_add_inc;		/* default: disabled */
+
+module_param_named(beta,                 bbr_beta,                 uint, 0644);
+module_param_named(ecn_alpha_gain,       bbr_ecn_alpha_gain,       uint, 0644);
+module_param_named(ecn_alpha_init,       bbr_ecn_alpha_init,       uint, 0644);
+module_param_named(ecn_factor,           bbr_ecn_factor,           uint, 0644);
+module_param_named(ecn_thresh,           bbr_ecn_thresh,           uint, 0644);
+module_param_named(ecn_max_rtt_us,       bbr_ecn_max_rtt_us,       uint, 0644);
+module_param_named(ecn_reprobe_gain,     bbr_ecn_reprobe_gain,     uint, 0644);
+module_param_named(loss_thresh,          bbr_loss_thresh,          uint, 0664);
+module_param_named(full_loss_cnt,        bbr_full_loss_cnt,        uint, 0664);
+module_param_named(full_ecn_cnt,         bbr_full_ecn_cnt,         uint, 0664);
+module_param_named(inflight_headroom,    bbr_inflight_headroom,    uint, 0664);
+module_param_named(bw_probe_pif_gain,    bbr_bw_probe_pif_gain,    uint, 0664);
+module_param_named(bw_probe_reno_gain,   bbr_bw_probe_reno_gain,   uint, 0664);
+module_param_named(bw_probe_max_rounds,  bbr_bw_probe_max_rounds,  uint, 0664);
+module_param_named(bw_probe_rand_rounds, bbr_bw_probe_rand_rounds, uint, 0664);
+module_param_named(bw_probe_base_us,     bbr_bw_probe_base_us,     uint, 0664);
+module_param_named(bw_probe_rand_us,     bbr_bw_probe_rand_us,     uint, 0664);
+module_param_named(undo,                 bbr_undo,                 bool, 0664);
+module_param_named(fast_path,		 bbr_fast_path,		   bool, 0664);
+module_param_named(fast_ack_mode,	 bbr_fast_ack_mode,	   uint, 0664);
+module_param_named(refill_add_inc,       bbr_refill_add_inc,       uint, 0664);
+
+static void bbr2_init(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr_init(sk);	/* run shared init code for v1 and v2 */
+
+	/* BBR v2 parameters: */
+	bbr->params.beta = min_t(u32, 0xFFU, bbr_beta);
+	bbr->params.ecn_alpha_gain = min_t(u32, 0xFFU, bbr_ecn_alpha_gain);
+	bbr->params.ecn_alpha_init = min_t(u32, BBR_UNIT, bbr_ecn_alpha_init);
+	bbr->params.ecn_factor = min_t(u32, 0xFFU, bbr_ecn_factor);
+	bbr->params.ecn_thresh = min_t(u32, 0xFFU, bbr_ecn_thresh);
+	bbr->params.ecn_max_rtt_us = min_t(u32, 0x7ffffU, bbr_ecn_max_rtt_us);
+	bbr->params.ecn_reprobe_gain = min_t(u32, 0x1FF, bbr_ecn_reprobe_gain);
+	bbr->params.loss_thresh = min_t(u32, 0xFFU, bbr_loss_thresh);
+	bbr->params.full_loss_cnt = min_t(u32, 0xFU, bbr_full_loss_cnt);
+	bbr->params.full_ecn_cnt = min_t(u32, 0x3U, bbr_full_ecn_cnt);
+	bbr->params.inflight_headroom =
+		min_t(u32, 0xFFU, bbr_inflight_headroom);
+	bbr->params.bw_probe_pif_gain =
+		min_t(u32, 0x1FFU, bbr_bw_probe_pif_gain);
+	bbr->params.bw_probe_reno_gain =
+		min_t(u32, 0x1FFU, bbr_bw_probe_reno_gain);
+	bbr->params.bw_probe_max_rounds =
+		min_t(u32, 0xFFU, bbr_bw_probe_max_rounds);
+	bbr->params.bw_probe_rand_rounds =
+		min_t(u32, 0xFU, bbr_bw_probe_rand_rounds);
+	bbr->params.bw_probe_base_us =
+		min_t(u32, (1 << 26) - 1, bbr_bw_probe_base_us);
+	bbr->params.bw_probe_rand_us =
+		min_t(u32, (1 << 26) - 1, bbr_bw_probe_rand_us);
+	bbr->params.undo = bbr_undo;
+	bbr->params.fast_path = bbr_fast_path ? 1 : 0;
+	bbr->params.refill_add_inc = min_t(u32, 0x3U, bbr_refill_add_inc);
+
+	/* BBR v2 state: */
+	bbr->initialized = 1;
+	/* Start sampling ECN mark rate after first full flight is ACKed: */
+	bbr->loss_round_delivered = tp->delivered + 1;
+	bbr->loss_round_start = 0;
+	bbr->undo_bw_lo = 0;
+	bbr->undo_inflight_lo = 0;
+	bbr->undo_inflight_hi = 0;
+	bbr->loss_events_in_round = 0;
+	bbr->startup_ecn_rounds = 0;
+	bbr2_reset_congestion_signals(sk);
+	bbr->bw_lo = ~0U;
+	bbr->bw_hi[0] = 0;
+	bbr->bw_hi[1] = 0;
+	bbr->inflight_lo = ~0U;
+	bbr->inflight_hi = ~0U;
+	bbr->bw_probe_up_cnt = ~0U;
+	bbr->bw_probe_up_acks = 0;
+	bbr->bw_probe_up_rounds = 0;
+	bbr->probe_wait_us = 0;
+	bbr->stopped_risky_probe = 0;
+	bbr->ack_phase = BBR_ACKS_INIT;
+	bbr->rounds_since_probe = 0;
+	bbr->bw_probe_samples = 0;
+	bbr->prev_probe_too_high = 0;
+	bbr->ecn_eligible = 0;
+	bbr->ecn_alpha = bbr->params.ecn_alpha_init;
+	bbr->alpha_last_delivered = 0;
+	bbr->alpha_last_delivered_ce = 0;
+
+	tp->fast_ack_mode = min_t(u32, 0x2U, bbr_fast_ack_mode);
+
+	if ((tp->ecn_flags & TCP_ECN_OK) && bbr_ecn_enable)
+		tp->ecn_flags |= TCP_ECN_ECT_PERMANENT;
+}
+
+/* Core TCP stack informs us that the given skb was just marked lost. */
+static void bbr2_skb_marked_lost(struct sock *sk, const struct sk_buff *skb)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+	struct tcp_skb_cb *scb = TCP_SKB_CB(skb);
+	struct rate_sample rs;
+
+	/* Capture "current" data over the full round trip of loss,
+	 * to have a better chance to see the full capacity of the path.
+	*/
+	if (!bbr->loss_in_round)  /* first loss in this round trip? */
+		bbr->loss_round_delivered = tp->delivered;  /* set round trip */
+	bbr->loss_in_round = 1;
+	bbr->loss_in_cycle = 1;
+
+	if (!bbr->bw_probe_samples)
+		return;  /* not an skb sent while probing for bandwidth */
+	if (unlikely(!scb->tx.delivered_mstamp))
+		return;  /* skb was SACKed, reneged, marked lost; ignore it */
+	/* We are probing for bandwidth. Construct a rate sample that
+	 * estimates what happened in the flight leading up to this lost skb,
+	 * then see if the loss rate went too high, and if so at which packet.
+	 */
+	memset(&rs, 0, sizeof(rs));
+	rs.tx_in_flight = scb->tx.in_flight;
+	rs.lost = tp->lost - scb->tx.lost;
+	rs.is_app_limited = scb->tx.is_app_limited;
+	if (bbr2_is_inflight_too_high(sk, &rs)) {
+		rs.tx_in_flight = bbr2_inflight_hi_from_lost_skb(sk, &rs, skb);
+		bbr2_handle_inflight_too_high(sk, &rs);
+	}
+}
+
+/* Revert short-term model if current loss recovery event was spurious. */
+static u32 bbr2_undo_cwnd(struct sock *sk)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr->debug.undo = 1;
+	bbr->full_bw = 0;   /* spurious slow-down; reset full pipe detection */
+	bbr->full_bw_cnt = 0;
+	bbr->loss_in_round = 0;
+
+	if (!bbr->params.undo)
+		return tp->snd_cwnd;
+
+	/* Revert to cwnd and other state saved before loss episode. */
+	bbr->bw_lo = max(bbr->bw_lo, bbr->undo_bw_lo);
+	bbr->inflight_lo = max(bbr->inflight_lo, bbr->undo_inflight_lo);
+	bbr->inflight_hi = max(bbr->inflight_hi, bbr->undo_inflight_hi);
+	return bbr->prior_cwnd;
+}
+
+/* Entering loss recovery, so save state for when we undo recovery. */
+static u32 bbr2_ssthresh(struct sock *sk)
+{
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	bbr_save_cwnd(sk);
+	/* For undo, save state that adapts based on loss signal. */
+	bbr->undo_bw_lo		= bbr->bw_lo;
+	bbr->undo_inflight_lo	= bbr->inflight_lo;
+	bbr->undo_inflight_hi	= bbr->inflight_hi;
+	return tcp_sk(sk)->snd_ssthresh;
+}
+
+static enum tcp_bbr2_phase bbr2_get_phase(struct bbr *bbr)
+{
+	switch (bbr->mode) {
+	case BBR_STARTUP:
+		return BBR2_PHASE_STARTUP;
+	case BBR_DRAIN:
+		return BBR2_PHASE_DRAIN;
+	case BBR_PROBE_BW:
+		break;
+	case BBR_PROBE_RTT:
+		return BBR2_PHASE_PROBE_RTT;
+	default:
+		return BBR2_PHASE_INVALID;
+	}
+	switch (bbr->cycle_idx) {
+	case BBR_BW_PROBE_UP:
+		return BBR2_PHASE_PROBE_BW_UP;
+	case BBR_BW_PROBE_DOWN:
+		return BBR2_PHASE_PROBE_BW_DOWN;
+	case BBR_BW_PROBE_CRUISE:
+		return BBR2_PHASE_PROBE_BW_CRUISE;
+	case BBR_BW_PROBE_REFILL:
+		return BBR2_PHASE_PROBE_BW_REFILL;
+	default:
+		return BBR2_PHASE_INVALID;
+	}
+}
+
+static size_t bbr2_get_info(struct sock *sk, u32 ext, int *attr,
+			    union tcp_cc_info *info)
+{
+	if (ext & (1 << (INET_DIAG_BBRINFO - 1)) ||
+	    ext & (1 << (INET_DIAG_VEGASINFO - 1))) {
+		struct bbr *bbr = inet_csk_ca(sk);
+		u64 bw = bbr_bw_bytes_per_sec(sk, bbr_bw(sk));
+		u64 bw_hi = bbr_bw_bytes_per_sec(sk, bbr_max_bw(sk));
+		u64 bw_lo = bbr->bw_lo == ~0U ?
+			~0ULL : bbr_bw_bytes_per_sec(sk, bbr->bw_lo);
+
+		memset(&info->bbr2, 0, sizeof(info->bbr2));
+		info->bbr2.bbr_bw_lsb		= (u32)bw;
+		info->bbr2.bbr_bw_msb		= (u32)(bw >> 32);
+		info->bbr2.bbr_min_rtt		= bbr->min_rtt_us;
+		info->bbr2.bbr_pacing_gain	= bbr->pacing_gain;
+		info->bbr2.bbr_cwnd_gain	= bbr->cwnd_gain;
+		info->bbr2.bbr_bw_hi_lsb	= (u32)bw_hi;
+		info->bbr2.bbr_bw_hi_msb	= (u32)(bw_hi >> 32);
+		info->bbr2.bbr_bw_lo_lsb	= (u32)bw_lo;
+		info->bbr2.bbr_bw_lo_msb	= (u32)(bw_lo >> 32);
+		info->bbr2.bbr_mode		= bbr->mode;
+		info->bbr2.bbr_phase		= (__u8)bbr2_get_phase(bbr);
+		info->bbr2.bbr_version		= (__u8)2;
+		info->bbr2.bbr_inflight_lo	= bbr->inflight_lo;
+		info->bbr2.bbr_inflight_hi	= bbr->inflight_hi;
+		info->bbr2.bbr_extra_acked	= bbr_extra_acked(sk);
+		*attr = INET_DIAG_BBRINFO;
+		return sizeof(info->bbr2);
+	}
+	return 0;
+}
+
+static void bbr2_set_state(struct sock *sk, u8 new_state)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	struct bbr *bbr = inet_csk_ca(sk);
+
+	if (new_state == TCP_CA_Loss) {
+		struct rate_sample rs = { .losses = 1 };
+		struct bbr_context ctx = { 0 };
+
+		bbr->prev_ca_state = TCP_CA_Loss;
+		bbr->full_bw = 0;
+		if (!bbr2_is_probing_bandwidth(sk) && bbr->inflight_lo == ~0U) {
+			/* bbr_adapt_lower_bounds() needs cwnd before
+			 * we suffered an RTO, to update inflight_lo:
+			 */
+			bbr->inflight_lo =
+				max(tp->snd_cwnd, bbr->prior_cwnd);
+		}
+		bbr_debug(sk, 0, &rs, &ctx);
+	} else if (bbr->prev_ca_state == TCP_CA_Loss &&
+		   new_state != TCP_CA_Loss) {
+		tp->snd_cwnd = max(tp->snd_cwnd, bbr->prior_cwnd);
+		bbr->try_fast_path = 0; /* bound cwnd using latest model */
+	}
+}
+
+static struct tcp_congestion_ops tcp_bbr2_cong_ops __read_mostly = {
+	.flags		= TCP_CONG_NON_RESTRICTED | TCP_CONG_WANTS_CE_EVENTS,
+	.name		= "bbr2",
+	.owner		= THIS_MODULE,
+	.init		= bbr2_init,
+	.cong_control	= bbr2_main,
+	.sndbuf_expand	= bbr_sndbuf_expand,
+	.skb_marked_lost = bbr2_skb_marked_lost,
+	.undo_cwnd	= bbr2_undo_cwnd,
+	.cwnd_event	= bbr_cwnd_event,
+	.ssthresh	= bbr2_ssthresh,
+	.tso_segs	= bbr_tso_segs,
+	.get_info	= bbr2_get_info,
+	.set_state	= bbr2_set_state,
+};
+
+static int __init bbr_register(void)
+{
+	BUILD_BUG_ON(sizeof(struct bbr) > ICSK_CA_PRIV_SIZE);
+	return tcp_register_congestion_control(&tcp_bbr2_cong_ops);
+}
+
+static void __exit bbr_unregister(void)
+{
+	tcp_unregister_congestion_control(&tcp_bbr2_cong_ops);
+}
+
+module_init(bbr_register);
+module_exit(bbr_unregister);
+
+MODULE_AUTHOR("Van Jacobson <vanj@google.com>");
+MODULE_AUTHOR("Neal Cardwell <ncardwell@google.com>");
+MODULE_AUTHOR("Yuchung Cheng <ycheng@google.com>");
+MODULE_AUTHOR("Soheil Hassas Yeganeh <soheil@google.com>");
+MODULE_AUTHOR("Priyaranjan Jha <priyarjha@google.com>");
+MODULE_AUTHOR("Yousuk Seung <ysseung@google.com>");
+MODULE_AUTHOR("Kevin Yang <yyd@google.com>");
+MODULE_AUTHOR("Arjun Roy <arjunroy@google.com>");
+
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_DESCRIPTION("TCP BBR (Bottleneck Bandwidth and RTT)");
diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c
index db5831e6c136..153ed9010c0c 100644
--- a/net/ipv4/tcp_cong.c
+++ b/net/ipv4/tcp_cong.c
@@ -179,6 +179,7 @@ void tcp_init_congestion_control(struct sock *sk)
 	struct inet_connection_sock *icsk = inet_csk(sk);
 
 	tcp_sk(sk)->prior_ssthresh = 0;
+	tcp_sk(sk)->fast_ack_mode = 0;
 	if (icsk->icsk_ca_ops->init)
 		icsk->icsk_ca_ops->init(sk);
 	if (tcp_ca_needs_ecn(sk))
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 149ceb5c94ff..af1f5742b4f8 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -348,7 +348,7 @@ static void __tcp_ecn_check_ce(struct sock *sk, const struct sk_buff *skb)
 			tcp_enter_quickack_mode(sk, 2);
 		break;
 	case INET_ECN_CE:
-		if (tcp_ca_needs_ecn(sk))
+		if (tcp_ca_wants_ce_events(sk))
 			tcp_ca_event(sk, CA_EVENT_ECN_IS_CE);
 
 		if (!(tp->ecn_flags & TCP_ECN_DEMAND_CWR)) {
@@ -359,7 +359,7 @@ static void __tcp_ecn_check_ce(struct sock *sk, const struct sk_buff *skb)
 		tp->ecn_flags |= TCP_ECN_SEEN;
 		break;
 	default:
-		if (tcp_ca_needs_ecn(sk))
+		if (tcp_ca_wants_ce_events(sk))
 			tcp_ca_event(sk, CA_EVENT_ECN_NO_CE);
 		tp->ecn_flags |= TCP_ECN_SEEN;
 		break;
@@ -1039,7 +1039,12 @@ static void tcp_verify_retransmit_hint(struct tcp_sock *tp, struct sk_buff *skb)
  */
 static void tcp_notify_skb_loss_event(struct tcp_sock *tp, const struct sk_buff *skb)
 {
+	struct sock *sk = (struct sock *)tp;
+	const struct tcp_congestion_ops *ca_ops = inet_csk(sk)->icsk_ca_ops;
+
 	tp->lost += tcp_skb_pcount(skb);
+	if (ca_ops->skb_marked_lost)
+		ca_ops->skb_marked_lost(sk, skb);
 }
 
 void tcp_mark_skb_lost(struct sock *sk, struct sk_buff *skb)
@@ -1420,6 +1425,17 @@ static bool tcp_shifted_skb(struct sock *sk, struct sk_buff *prev,
 	WARN_ON_ONCE(tcp_skb_pcount(skb) < pcount);
 	tcp_skb_pcount_add(skb, -pcount);
 
+	/* Adjust tx.in_flight as pcount is shifted from skb to prev. */
+	if (WARN_ONCE(TCP_SKB_CB(skb)->tx.in_flight < pcount,
+		      "prev in_flight: %u skb in_flight: %u pcount: %u",
+		      TCP_SKB_CB(prev)->tx.in_flight,
+		      TCP_SKB_CB(skb)->tx.in_flight,
+		      pcount))
+		TCP_SKB_CB(skb)->tx.in_flight = 0;
+	else
+		TCP_SKB_CB(skb)->tx.in_flight -= pcount;
+	TCP_SKB_CB(prev)->tx.in_flight += pcount;
+
 	/* When we're adding to gso_segs == 1, gso_size will be zero,
 	 * in theory this shouldn't be necessary but as long as DSACK
 	 * code can come after this skb later on it's better to keep
@@ -3189,7 +3205,6 @@ static int tcp_clean_rtx_queue(struct sock *sk, const struct sk_buff *ack_skb,
 	long seq_rtt_us = -1L;
 	long ca_rtt_us = -1L;
 	u32 pkts_acked = 0;
-	u32 last_in_flight = 0;
 	bool rtt_update;
 	int flag = 0;
 
@@ -3225,7 +3240,6 @@ static int tcp_clean_rtx_queue(struct sock *sk, const struct sk_buff *ack_skb,
 			if (!first_ackt)
 				first_ackt = last_ackt;
 
-			last_in_flight = TCP_SKB_CB(skb)->tx.in_flight;
 			if (before(start_seq, reord))
 				reord = start_seq;
 			if (!after(scb->end_seq, tp->high_seq))
@@ -3291,8 +3305,8 @@ static int tcp_clean_rtx_queue(struct sock *sk, const struct sk_buff *ack_skb,
 		seq_rtt_us = tcp_stamp_us_delta(tp->tcp_mstamp, first_ackt);
 		ca_rtt_us = tcp_stamp_us_delta(tp->tcp_mstamp, last_ackt);
 
-		if (pkts_acked == 1 && last_in_flight < tp->mss_cache &&
-		    last_in_flight && !prior_sacked && fully_acked &&
+		if (pkts_acked == 1 && fully_acked && !prior_sacked &&
+		    (tp->snd_una - prior_snd_una) < tp->mss_cache &&
 		    sack->rate->prior_delivered + 1 == tp->delivered &&
 		    !(flag & (FLAG_CA_ALERT | FLAG_SYN_ACKED))) {
 			/* Conservatively mark a delayed ACK. It's typically
@@ -3349,9 +3363,10 @@ static int tcp_clean_rtx_queue(struct sock *sk, const struct sk_buff *ack_skb,
 
 	if (icsk->icsk_ca_ops->pkts_acked) {
 		struct ack_sample sample = { .pkts_acked = pkts_acked,
-					     .rtt_us = sack->rate->rtt_us,
-					     .in_flight = last_in_flight };
+					     .rtt_us = sack->rate->rtt_us };
 
+		sample.in_flight = tp->mss_cache *
+			(tp->delivered - sack->rate->prior_delivered);
 		icsk->icsk_ca_ops->pkts_acked(sk, &sample);
 	}
 
@@ -3749,6 +3764,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag)
 
 	prior_fack = tcp_is_sack(tp) ? tcp_highest_sack_seq(tp) : tp->snd_una;
 	rs.prior_in_flight = tcp_packets_in_flight(tp);
+	tcp_rate_check_app_limited(sk);
 
 	/* ts_recent update must be made after we are sure that the packet
 	 * is in window.
@@ -3846,6 +3862,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag)
 	delivered = tcp_newly_delivered(sk, delivered, flag);
 	lost = tp->lost - lost;			/* freshly marked lost */
 	rs.is_ack_delayed = !!(flag & FLAG_ACK_MAYBE_DELAYED);
+	rs.is_ece = !!(flag & FLAG_ECE);
 	tcp_rate_gen(sk, delivered, lost, is_sack_reneg, sack_state.rate);
 	tcp_cong_control(sk, ack, delivered, flag, sack_state.rate);
 	tcp_xmit_recovery(sk, rexmit);
@@ -5414,13 +5431,14 @@ static void __tcp_ack_snd_check(struct sock *sk, int ofo_possible)
 
 	    /* More than one full frame received... */
 	if (((tp->rcv_nxt - tp->rcv_wup) > inet_csk(sk)->icsk_ack.rcv_mss &&
+	     (tp->fast_ack_mode == 1 ||
 	     /* ... and right edge of window advances far enough.
 	      * (tcp_recvmsg() will send ACK otherwise).
 	      * If application uses SO_RCVLOWAT, we want send ack now if
 	      * we have not received enough bytes to satisfy the condition.
 	      */
-	    (tp->rcv_nxt - tp->copied_seq < sk->sk_rcvlowat ||
-	     __tcp_select_window(sk) >= tp->rcv_wnd)) ||
+	      (tp->rcv_nxt - tp->copied_seq < sk->sk_rcvlowat ||
+	       __tcp_select_window(sk) >= tp->rcv_wnd))) ||
 	    /* We ACK each frame or... */
 	    tcp_in_quickack_mode(sk) ||
 	    /* Protocol state mandates a one-time immediate ACK */
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 29553fce8502..0eb462008718 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -377,7 +377,8 @@ static void tcp_ecn_send(struct sock *sk, struct sk_buff *skb,
 				th->cwr = 1;
 				skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;
 			}
-		} else if (!tcp_ca_needs_ecn(sk)) {
+		} else if (!(tp->ecn_flags & TCP_ECN_ECT_PERMANENT) &&
+			!tcp_ca_needs_ecn(sk)) {
 			/* ACK or retransmitted segment: clear ECT|CE */
 			INET_ECN_dontxmit(sk);
 		}
@@ -1256,8 +1257,6 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb,
 	tp->tcp_wstamp_ns = max(tp->tcp_wstamp_ns, tp->tcp_clock_cache);
 	skb->skb_mstamp_ns = tp->tcp_wstamp_ns;
 	if (clone_it) {
-		TCP_SKB_CB(skb)->tx.in_flight = TCP_SKB_CB(skb)->end_seq
-			- tp->snd_una;
 		oskb = skb;
 
 		tcp_skb_tsorted_save(oskb) {
@@ -1536,7 +1535,7 @@ int tcp_fragment(struct sock *sk, enum tcp_queue tcp_queue,
 {
 	struct tcp_sock *tp = tcp_sk(sk);
 	struct sk_buff *buff;
-	int nsize, old_factor;
+	int nsize, old_factor, inflight_prev;
 	long limit;
 	int nlen;
 	u8 flags;
@@ -1615,6 +1614,15 @@ int tcp_fragment(struct sock *sk, enum tcp_queue tcp_queue,
 
 		if (diff)
 			tcp_adjust_pcount(sk, skb, diff);
+
+		/* Set buff tx.in_flight as if buff were sent by itself. */
+		inflight_prev = TCP_SKB_CB(skb)->tx.in_flight - old_factor;
+		if (WARN_ONCE(inflight_prev < 0,
+			      "inconsistent: tx.in_flight: %u old_factor: %d",
+			      TCP_SKB_CB(skb)->tx.in_flight, old_factor))
+			inflight_prev = 0;
+		TCP_SKB_CB(buff)->tx.in_flight = inflight_prev +
+						 tcp_skb_pcount(buff);
 	}
 
 	/* Link BUFF into the send queue. */
@@ -1983,13 +1991,12 @@ static u32 tcp_tso_autosize(const struct sock *sk, unsigned int mss_now,
 static u32 tcp_tso_segs(struct sock *sk, unsigned int mss_now)
 {
 	const struct tcp_congestion_ops *ca_ops = inet_csk(sk)->icsk_ca_ops;
-	u32 min_tso, tso_segs;
+	u32 tso_segs;
 
-	min_tso = ca_ops->min_tso_segs ?
-			ca_ops->min_tso_segs(sk) :
-			sock_net(sk)->ipv4.sysctl_tcp_min_tso_segs;
-
-	tso_segs = tcp_tso_autosize(sk, mss_now, min_tso);
+	tso_segs = ca_ops->tso_segs ?
+		ca_ops->tso_segs(sk, mss_now) :
+		tcp_tso_autosize(sk, mss_now,
+				 sock_net(sk)->ipv4.sysctl_tcp_min_tso_segs);
 	return min_t(u32, tso_segs, sk->sk_gso_max_segs);
 }
 
@@ -2629,6 +2636,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
 			skb->skb_mstamp_ns = tp->tcp_wstamp_ns = tp->tcp_clock_cache;
 			list_move_tail(&skb->tcp_tsorted_anchor, &tp->tsorted_sent_queue);
 			tcp_init_tso_segs(skb, mss_now);
+			tcp_set_tx_in_flight(sk, skb);
 			goto repair; /* Skip network transmission */
 		}
 
diff --git a/net/ipv4/tcp_rate.c b/net/ipv4/tcp_rate.c
index 0de693565963..796fa6e5310c 100644
--- a/net/ipv4/tcp_rate.c
+++ b/net/ipv4/tcp_rate.c
@@ -34,6 +34,24 @@
  * ready to send in the write queue.
  */
 
+void tcp_set_tx_in_flight(struct sock *sk, struct sk_buff *skb)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	u32 in_flight;
+
+	/* Check, sanitize, and record packets in flight after skb was sent. */
+	in_flight = tcp_packets_in_flight(tp) + tcp_skb_pcount(skb);
+	if (WARN_ONCE(in_flight > TCPCB_IN_FLIGHT_MAX,
+		      "insane in_flight %u cc %s mss %u "
+		      "cwnd %u pif %u %u %u %u\n",
+		      in_flight, inet_csk(sk)->icsk_ca_ops->name,
+		      tp->mss_cache, tp->snd_cwnd,
+		      tp->packets_out, tp->retrans_out,
+		      tp->sacked_out, tp->lost_out))
+		in_flight = TCPCB_IN_FLIGHT_MAX;
+	TCP_SKB_CB(skb)->tx.in_flight = in_flight;
+}
+
 /* Snapshot the current delivery information in the skb, to generate
  * a rate sample later when the skb is (s)acked in tcp_rate_skb_delivered().
  */
@@ -65,7 +83,10 @@ void tcp_rate_skb_sent(struct sock *sk, struct sk_buff *skb)
 	TCP_SKB_CB(skb)->tx.first_tx_mstamp	= tp->first_tx_mstamp;
 	TCP_SKB_CB(skb)->tx.delivered_mstamp	= tp->delivered_mstamp;
 	TCP_SKB_CB(skb)->tx.delivered		= tp->delivered;
+	TCP_SKB_CB(skb)->tx.delivered_ce	= tp->delivered_ce;
+	TCP_SKB_CB(skb)->tx.lost		= tp->lost;
 	TCP_SKB_CB(skb)->tx.is_app_limited	= tp->app_limited ? 1 : 0;
+	tcp_set_tx_in_flight(sk, skb);
 }
 
 /* When an skb is sacked or acked, we fill in the rate sample with the (prior)
@@ -86,16 +107,20 @@ void tcp_rate_skb_delivered(struct sock *sk, struct sk_buff *skb,
 
 	if (!rs->prior_delivered ||
 	    after(scb->tx.delivered, rs->prior_delivered)) {
+		rs->prior_lost	     = scb->tx.lost;
+		rs->prior_delivered_ce  = scb->tx.delivered_ce;
 		rs->prior_delivered  = scb->tx.delivered;
 		rs->prior_mstamp     = scb->tx.delivered_mstamp;
 		rs->is_app_limited   = scb->tx.is_app_limited;
 		rs->is_retrans	     = scb->sacked & TCPCB_RETRANS;
+		rs->tx_in_flight     = scb->tx.in_flight;
 
 		/* Record send time of most recently ACKed packet: */
 		tp->first_tx_mstamp  = tcp_skb_timestamp_us(skb);
 		/* Find the duration of the "send phase" of this window: */
-		rs->interval_us = tcp_stamp_us_delta(tp->first_tx_mstamp,
-						     scb->tx.first_tx_mstamp);
+		rs->interval_us      = tcp_stamp32_us_delta(
+						tp->first_tx_mstamp,
+						scb->tx.first_tx_mstamp);
 
 	}
 	/* Mark off the skb delivered once it's sacked to avoid being
@@ -137,6 +162,11 @@ void tcp_rate_gen(struct sock *sk, u32 delivered, u32 lost,
 		return;
 	}
 	rs->delivered   = tp->delivered - rs->prior_delivered;
+	rs->lost        = tp->lost - rs->prior_lost;
+
+	rs->delivered_ce = tp->delivered_ce - rs->prior_delivered_ce;
+	/* delivered_ce occupies less than 32 bits in the skb control block */
+	rs->delivered_ce &= TCPCB_DELIVERED_CE_MASK;
 
 	/* Model sending data and receiving ACKs as separate pipeline phases
 	 * for a window. Usually the ACK phase is longer, but with ACK
@@ -144,7 +174,7 @@ void tcp_rate_gen(struct sock *sk, u32 delivered, u32 lost,
 	 * longer phase.
 	 */
 	snd_us = rs->interval_us;				/* send phase */
-	ack_us = tcp_stamp_us_delta(tp->tcp_mstamp,
+	ack_us = tcp_stamp32_us_delta(tp->tcp_mstamp,
 				    rs->prior_mstamp); /* ack phase */
 	rs->interval_us = max(snd_us, ack_us);
 
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 20cf4a98c69d..b5f7e49a003a 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -607,6 +607,7 @@ void tcp_write_timer_handler(struct sock *sk)
 		goto out;
 	}
 
+	tcp_rate_check_app_limited(sk);
 	tcp_mstamp_refresh(tcp_sk(sk));
 	event = icsk->icsk_pending;
 
-- 
2.33.0