Date: Tue, 27 Jan 2004 17:03:39 +0100 From: Rudolf Cejka <cejkar@fit.vutbr.cz> To: Bill Paul <wpaul@FreeBSD.ORG> Cc: freebsd-current@FreeBSD.ORG Subject: Re: Project Evil: The Evil Continues Message-ID: <20040127160339.GA72958@fit.vutbr.cz> In-Reply-To: <20040127111802.GA96935@fit.vutbr.cz> References: <20040127111802.GA96935@fit.vutbr.cz>
next in thread | previous in thread | raw e-mail | index | archive | help
Rudolf Cejka wrote (2004/01/27): > > you have not provided a more precise means of duplicating the crash > > (I can't do cvsup from home: I only have a dialup). Cvsup is not > > a network diagnostic or traffic generation tool: it is not possible > > to generate consistent, reproducible results with it. Use ttcp or > > netperf instead, then show me _EXACTLY_ how you ran it to produce > > the crash so I can do it too. > > Ok, I do it. Now I have to go to lunch ;o) Do you have this or similar > card too? The driver is bcmwl5.sys and it seems that it is shared > among several cards. Hmm, I'm trying to reproduce the panic with multiple parallel nttcp/ttcp/netperf and dd, but still without any success. It seems that the panic is very dependent on the network, cpu and/or disk load. It occurs just in case, when cvsup is run with -P- argument, that is there are two sockets for four independent communication channels, two in and two out. If I run cvsup with -Pm, what means just one multiplexed socket for all four channels, the panics disappear. -- Rudolf Cejka <cejkar at fit.vutbr.cz> http://www.fit.vutbr.cz/~cejkar Brno University of Technology, Faculty of Information Technology Bozetechova 2, 612 66 Brno, Czech Republic
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040127160339.GA72958>