When running a telnet session from or into the machine, it looks like there is a very noticable delay between typing CR and other characters and when these characters are displayed at the telnet session. It's as if there is a huge network delay (there is no traffic on the network). When logged into the machine directly via a console, I can not detect any of these "stalls". I have tried a different 3com 3c980c card and the second card had the same problem. Here are relevant sections of my dmesg's and ifconfig: Copyright (c) 1992-2001 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD 4.4-RELEASE #0: Tue Sep 18 11:57:08 PDT 2001 murray@builder.FreeBSD.org:/usr/src/sys/compile/GENERIC Timecounter "i8254" frequency 1193182 Hz CPU: AMD Athlon(tm) processor (1396.74-MHz 686-class CPU) Origin = "AuthenticAMD" Id = 0x644 Stepping = 4 Features=0x183f9ff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR,PGE,MCA,CMOV,PA T,PSE36,MMX,FXSR> AMD Features=0xc0440000<<b18>,AMIE,DSP,3DNow!> real memory = 268369920 (262080K bytes) avail memory = 256712704 (250696K bytes) ...snip... xl0: <3Com 3c980C Fast Etherlink XL> port 0xd000-0xd07f mem 0xe8000000-0xe800007 f irq 12 at device 8.0 on pci0 xl0: Ethernet address: 00:04:75:86:c4:8a miibus0: <MII bus> on xl0 xlphy0: <3c905C 10/100 internal PHY> on miibus0 xlphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto ... snip ... nsa# ifconfig xl0 xl0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500 inet 208.245.181.68 netmask 0xfffffff8 broadcast 208.245.181.71 inet6 fe80::204:75ff:fe86:c48a%xl0 prefixlen 64 scopeid 0x1 ether 00:04:75:86:c4:8a media: Ethernet autoselect (10baseT/UTP) status: active How-To-Repeat: This problem is easily repeatable.
State Changed From-To: open->feedback Can you try updating your FreeBSD system to a 4.5-RC? There have been various changes on the 4.x branch since 4.4-RELEASE, and in particular, there have been several very large TCP performance fixes in December.
Adding this to the audit trail.. G'luck, Peter -- What would this sentence be like if it weren't self-referential? ----- Forwarded message from Heberling <pjah@hicom.net> ----- From: "Heberling" <pjah@hicom.net> To: <roam@FreeBSD.org> Subject: Re: misc/34073: 3com 3c980c runs "bursty" / freezes-unfreezes Date: Thu, 24 Jan 2002 23:14:14 -0500 X-Priority: 3 X-MSMail-Priority: Normal X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700 I'm not familiar with "upgrading" to 4.5-RC from 4.4-stable. Will installing the Jan 27th release of 4.5 be a sufficient test? ----- End forwarded message -----
On Thu, Jan 24, 2002 at 11:14:14PM -0500, Heberling wrote: > I'm not familiar with "upgrading" to 4.5-RC from 4.4-stable. > Will installing the Jan 27th release of 4.5 be a sufficient test? Yes, testing against 4.5-RELEASE would be sufficient. The process of updating your FreeBSD system to a -STABLE snapshot is described in sections 19.2 and 19.3 of the FreeBSD Handbook at http://www.FreeBSD.org/handbook/synching.html; still, if you are not acquainted with it and have never done it before, it might be much simpler and easier to just upgrade to 4.5-RELEASE. G'luck, Peter -- When you are not looking at it, this sentence is in Spanish.
> State-Changed-From-To: open->feedback > State-Changed-By: roam > Can you try updating your FreeBSD system to a 4.5-RC? > There have been various changes on the 4.x branch since 4.4-RELEASE, > and in particular, there have been several very large TCP > performance fixes in December. Juergen, any news about this? Have you tried installing any newer releases?
Adding to the audit trail, Juergen's reply. | | On 2002-08-26 06:52 +0000, Juergen Heberling wrote: | > Giorgos, | > I know its hard to believe, but I have not had a chance to install a | > newer version of FreeBDS since this problem occured. | > Thank you for following up, however. | > | > Juergen |
This is a known problem with the 3Com driver. This bug is a duplicate of: http://www.freebsd.org/cgi/query-pr.cgi?pr=kern/18558 The problem is that receiving is dead slow, even though sending is running at full speed. The problem can be partialy fixed if you lower the receive buffers, but that won't eliminate it completely. (instead of getting 200kb/s you'll get 1mb/s etc). I'm running a 3Com 3c905-TX Fast Etherlink XL under fbsd 4.7-R, which is connected to a Netgear Switch, 100mbit full-duplex. -- Dimitris 'sehh' Michelinakis
State Changed From-To: feedback->closed Duplicate of kern/18558; this PR is referenced therein.