Hi! I have a fresh CURRENT server with TSO (TCP Segmentation Offload) on a em(4) network interface and Squid-2.6.6. And I have Windows 98 users which use the Squid via PPP connection. They are have MTU 576 bytes. And they can't use the Squid because slow connection to the Squid or timeouts. This problem disappeared if users increase their MTU to 1500 or if I disable TSO on the server. tcpdump of one of failed requests: 10:35:03.592164 IP 192.168.0.100.1035 > 192.168.0.2.3128: S 1423332:1423332(0) win 8192 <mss 536,nop,nop,sackOK> 10:35:03.592274 IP 192.168.0.2.3128 > 192.168.0.100.1035: S 2428819522:2428819522(0) ack 1423333 win 0 <mss 536> 10:35:03.799735 IP 192.168.0.100.1035 > 192.168.0.2.3128: . ack 1 win 8576 10:35:03.800244 IP 192.168.0.2.3128 > 192.168.0.100.1035: . ack 1 win 65535 10:35:03.958630 IP 192.168.0.100.1035 > 192.168.0.2.3128: F 1:1(0) ack 1 win 8576 10:35:03.958950 IP 192.168.0.2.3128 > 192.168.0.100.1035: . ack 2 win 65535 10:35:03.958990 IP 192.168.0.2.3128 > 192.168.0.100.1035: F 1:1(0) ack 2 win 65535 10:35:04.104539 IP 192.168.0.100.1035 > 192.168.0.2.3128: . ack 2 win 8576 Does anybody know about this problem? Thanks! -- DSS5-RIPE DSS-RIPN 2:550/5068_at_fidonet 2:550/5069_at_fidonet xmpp:dsh_at_vlink.ru mailto:dsh_at_vlink.ru http://neva.vlink.ru/~dsh/Received on Thu Jan 11 2007 - 11:19:42 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:04 UTC