Date: Mon, 10 Feb 1997 09:29:04 +0100 From: j@uriah.heep.sax.de (J Wunsch) To: freebsd-hackers@freebsd.org (FreeBSD hackers) Subject: Re: Copious information on panic: ffs_valloc: dup alloc - IDEAS??? Message-ID: <Mutt.19970210092904.j@uriah.heep.sax.de> In-Reply-To: <32FE7015.2F1CF0FB@whistle.com>; from Julian Elischer on Feb 9, 1997 16:47:17 -0800 References: <199702091818.NAA23739@lakes.water.net> <Mutt.19970209220222.j@uriah.heep.sax.de> <32FE7015.2F1CF0FB@whistle.com>
next in thread | previous in thread | raw e-mail | index | archive | help
As Julian Elischer wrote: > if you make it have 2 heads, this code will be enabled > again and we'll lose the effect.. > you could HEAR the difference when we went to 1 head.. But unfortunately, you could also _see_ the panics afterwards. :-( Go and re-read Thomas's original article. It's not simply summarizable in three sentences, sorry. The only summary is this: ntracks == 1 causes cgmask = 0xffffffff. The latter seems to cause block miscalculations in the filesystem code. As for the exact chain of miscalculations, refer to Message-ID <199702090226.VAA22081@lakes.water.net>. If you need something to prove this, setup a small MFS, and beat upon it. I can send you my ``mount /dev/null for MFS'' patches, or maybe, i simply commit them. They are basically required anyway to create a MFS on a diskless station (where you don't have a swap partition to template the MFS parameters from). -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Mutt.19970210092904.j>