Date: Thu, 5 May 2005 12:04:39 +0200 From: Bernd Walter <ticso@cicely12.cicely.de> To: benash@eecs.umich.edu Cc: freebsd-alpha@freebsd.org Subject: Re: FreeBSD 5.3 on Digital Personal Workstation 500au Message-ID: <20050505100438.GL75629@cicely12.cicely.de> In-Reply-To: <1115247645.4279541da60e0@mail.eecs.umich.edu> References: <1115240233.42793729bff3f@mail.eecs.umich.edu> <20050504210854.GA82595@freebie.xs4all.nl> <1115247645.4279541da60e0@mail.eecs.umich.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, May 04, 2005 at 07:00:45PM -0400, benash@eecs.umich.edu wrote: > OK, it looks like I fixed the problem. I needed to have the SCSI card in a > 64-bit slot WHILE the other two cards were removed (so that nothing was plugged > into a 32-bit slot). The 5.3 installation seemed to complete without trouble, > and I am able to login as root. > > Any ideas as to why this fixed things? Alphas intialize IO ranges on chips that aren't configured on PCI-PCI bridges and FreeBSD doesn't check the path. Therefor a driver may access io addresses that aren't configured, or configured differently enough to trap. The whole story with FreeBSD PCI code and bridges on alpha systems is very unsatifiying. The main point is to get our PCI layer to do a path check and refused unavailable ranges, the other point is to get more drivers compatible with memory ranges. -- B.Walter BWCT http://www.bwct.de bernd@bwct.de info@bwct.de
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050505100438.GL75629>