Date: Wed, 15 Nov 2000 09:30:14 -0500 (EST) From: Andrew Gallatin <gallatin@cs.duke.edu> To: wkb@freebie.demon.nl Cc: freebsd-alpha@FreeBSD.ORG Subject: Re: 4.2RC1 hosed on Tlaser Message-ID: <14866.39772.756614.268360@grasshopper.cs.duke.edu> In-Reply-To: <E13w2Kk-0007GP-00@post.mail.nl.demon.net> References: <E13w2Kk-0007GP-00@post.mail.nl.demon.net>
next in thread | previous in thread | raw e-mail | index | archive | help
wkb@freebie.demon.nl writes: > > [handwritten transcript from console] > .. boot succeeds ok... > dwlpx0 (DWLPA PCI at kft0 hose0) > base = 0xb80000000 > mask = 0xf80000000 > hacked = 0xb80000000 > panic: hose encoding hack would clobber base address > > I have no clue what this means, any takers? > > :-( :-( > > Wilko Yes. We have a grody hack in 4-stable to store off the hose in a memory or i/o address so that the low-level chipset code knows which hose a particular I/O request is for. In this case, the SRM console has assigned a memory address which overlaps the high bits we're using for the hose hack. This means that this particular box won't work until Doug's hose fixes are backported to -stable. (see, for example, the diff between rev 1.14 of sys/alpha/tlsb/dwlpx.c and 1.13.) This box would probably work if you were shuffle hardware around or remove it until it the SRM console happened to not use such a large memory address. BTW, this hack is responsible for the caveat about NCR/Sym controllers only working on the first hose of multi-hose boxes. Sigh. Drew To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?14866.39772.756614.268360>