summaryrefslogtreecommitdiffstats
path: root/net/ipv4/tcp_input.c
AgeCommit message (Expand)AuthorFilesLines
2008-01-28[TCP]: Abstract tp->highest_sack accessing & point to next skbIlpo Järvinen1-12/+15
2008-01-28[TCP]: Cleanup local variables of clean_rtx_queueIlpo Järvinen1-14/+12
2008-01-28[TCP]: Add unlikely() to urgent handling in clean_rtx_queueIlpo Järvinen1-2/+2
2008-01-28[TCP]: Remove duplicated code block from clean_rtx_queueIlpo Järvinen1-29/+21
2008-01-28[TCP]: Cong.ctrl modules: remove unused good_ack from cong_avoidIlpo Järvinen1-5/+4
2008-01-28[TCP]: Unite identical code from two seqno split blocksIlpo Järvinen1-6/+1
2008-01-28[TCP]: Remove superflucious FLAG_DATA_SACKEDIlpo Järvinen1-7/+3
2008-01-28[TCP]: Move LOSTRETRANS MIB outside !(L|S) checkIlpo Järvinen1-1/+1
2008-01-28[TCP]: Two fixes to new sacktag codeIlpo Järvinen1-2/+2
2008-01-28[NET]: Name magic constants in sock_wake_async()Pavel Emelyanov1-6/+6
2008-01-28[TCP]: Correct DSACK check placingIlpo Järvinen1-10/+5
2008-01-28[TCP]: Rewrite SACK block processing & sack_recv_cache useIlpo Järvinen1-100/+171
2008-01-28[TCP]: Earlier SACK block verification & simplify access to themIlpo Järvinen1-34/+51
2008-01-28[TCP]: Create tcp_sacktag_one().Ilpo Järvinen1-96/+96
2008-01-28[TCP]: Prior_fackets can be replaced by highest_sack seqIlpo Järvinen1-3/+2
2008-01-28[TCP]: Make lost retrans detection more self-containedIlpo Järvinen1-9/+11
2008-01-28[TCP]: Convert highest_sack to sk_buff to allow direct accessIlpo Järvinen1-6/+6
2008-01-28[TCP]: non-FACK SACK follows conservative SACK loss recoveryIlpo Järvinen1-18/+62
2008-01-28[TCP]: Extend reordering detection to cover CA_Loss partiallyIlpo Järvinen1-1/+2
2008-01-28[TCP]: Move !in_sack test earlier in sacktag & reorganize if()sIlpo Järvinen1-17/+14
2007-12-29[TCP]: use non-delayed ACK for congestion control RTTGavin McCullagh1-8/+11
2007-12-16[IPV4]: Make tcp_input_metrics() get minimum RTO via tcp_rto_min()Satoru SATOH1-1/+1
2007-12-05[TCP]: Move prior_in_flight collect to more robust placeIlpo Järvinen1-2/+1
2007-12-05[TCP] FRTO: Use of existing funcs make code more obvious & robustIlpo Järvinen1-9/+5
2007-11-14[TCP] FRTO: Clear frto_highmark only after process_frto that uses itIlpo Järvinen1-2/+2
2007-11-14[TCP]: Make sure write_queue_from does not begin with NULL ptrIlpo Järvinen1-0/+5
2007-11-13[TCP] FRTO: Plug potential LOST-bit leakIlpo Järvinen1-1/+3
2007-11-13[TCP] FRTO: Limit snd_cwnd if TCP was application limitedIlpo Järvinen1-0/+3
2007-11-10[TCP]: Split SACK FRTO flag clearing (fixes FRTO corner case bug)Ilpo Järvinen1-18/+17
2007-11-10[TCP]: Add unlikely() to sacktag out-of-mem in fragment caseIlpo Järvinen1-1/+1
2007-11-10[TCP]: Fix reord detection due to snd_una covered holesIlpo Järvinen1-18/+32
2007-11-10[TCP]: Consider GSO while counting reord in sacktagIlpo Järvinen1-4/+8
2007-11-01[TCP]: Another TAGBITS -> SACKED_ACKED|LOST conversionIlpo Järvinen1-1/+1
2007-11-01[TCP]: Process DSACKs that reside within a SACK blockIlpo Järvinen1-3/+22
2007-10-26[TCP]: fix D-SACK cwnd handlingRyousei Takano1-7/+5
2007-10-26[TCP]: Make tcp_match_skb_to_sack() static.Adrian Bunk1-2/+2
2007-10-25[TCP]: Fix inconsistency of terms.Ryousei Takano1-8/+8
2007-10-23[NET]: Treat the sign of the result of skb_headroom() consistentlyChuck Lever1-1/+1
2007-10-18[TCP]: Remove lost_retrans zero seqno special casesIlpo Järvinen1-3/+3
2007-10-15[TCP]: high_seq parameter removed (all callers use tp->high_seq)Ilpo Järvinen1-6/+5
2007-10-11[TCP]: Limit processing lost_retrans loop to work-to-do casesIlpo Järvinen1-3/+11
2007-10-11[TCP]: Fix lost_retrans loop vs fastpath problemsIlpo Järvinen1-15/+22
2007-10-11[TCP]: No need to re-count fackets_out/sacked_out at RTOIlpo Järvinen1-10/+16
2007-10-11[TCP]: Extract tcp_match_queue_to_sack from sacktag codeIlpo Järvinen1-19/+35
2007-10-11[TCP]: Kill almost unused variable pcount from sacktagIlpo Järvinen1-6/+3
2007-10-11[TCP]: Fix mark_head_lost to ignore R-bit when trying to mark LIlpo Järvinen1-1/+1
2007-10-11[TCP]: Add bytes_acked (ABC) clearing to FRTO tooIlpo Järvinen1-0/+2
2007-10-10[TCP]: Separate lost_retrans loop into own functionIlpo Järvinen1-37/+43
2007-10-10[NET]: sparse warning fixesStephen Hemminger1-1/+1
2007-10-10[TCP]: "Annotate" another fackets_out state resetIlpo Järvinen1-1/+2