Date: Thu, 10 Dec 1998 09:40:48 -0800 (PST) From: David Kulp <dkulp@neomorphic.com> To: questions@FreeBSD.ORG Subject: How to kill a network card... Message-ID: <199812101740.JAA11495@board66.cruzers.com>
next in thread | raw e-mail | index | archive | help
I have two FreeBSD boxes on a LAN. Box A (a laptop) has a PCMCIA ethernet card with a dongle that has fallen and been beaten up quite a bit and a connector that is not particularly secure. Box B has an NE2000 card in it. Both are connected to a router off to the internet. When using either machine independently of the other and connecting to outside hosts (remote log-ins, ftp, web, etc.) I never have problems, but recently I've found that heavy traffic between A and B raises errors on B like: de0: receive: 00:e0:98:01:09:72: bad crc de0: receive: 00:e0:98:01:09:72: alignment error de0: receive: 00:e0:98:01:09:72: alignment error If B reports too many of these errors (5-10) it just stops working completely with no further errors reported. (Nothing will fix the situation except for power off/on. Just reboot won't do it. [NFS doesn't like this at all!]) On the other hand, A, which is the machine that I suspect is actually causing the trouble, since it has a suspect ethernet card, continues on its merry way. My question: Is this a driver problem, i.e. the driver is not being robust to bad packets? If so, this seems like a "bad thing" for FreeBSD "stability". Or is it a hardware tolerance problem? The fact that a bad ethernet packet here or there apparently hoses my little LAN is a rude annoyance. Shouldn't the system be able to deal with this better? thanks, David To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199812101740.JAA11495>