Hi, (regarding Debian on US III machines: please do not forget to read teh WARNING part at the end)
as the current installer doesn't support US III CPUs obviously (as they seem to need SMP support), I came up with a "hack" to install Debian. When I decided to build an installer with a very recent kernel I gave 2.6.23-rc4-SMP a try and it worked. Unfortunately the package which creates the kernel module udebs didn't work properly with the 2.6.23-rc4 kernel package, so I've kiced it into working by just disabling the failing stuff - of course, this resulted in a number of broken udebs. http://bzed.de/files/d-i-veryexperimental.tar.gz contains the following files/directories: * boot.img: a debian/etch netboot installer with a 2.6.23-rc4 SMP kernel. Boot it in expert mode, configure your network, install the ssh client (and probably the ssh netinstall stuff, so you can continue to install using ssh instead of the serial console). Enter the shell and use scp to copy the * 2.6.23-rc4-sparc64-smp directory into /lib/modules of your installer. The installer should be able to find your harddisks and controllers afterwards. If you need additional firmware like for the qla2xxx controllers (which are used in v880/v890 and other machines), download the firmware from ftp://ftp.qlogic.com/outgoing/linux/firmware/ and drop it in /lib/firmware. modprobe -r the module and modprobe it again. Same things should work for other firmware blobs. * linux-image-2.6.22-2-sparc64-smp_2.6.22-4+farm1_sparc.deb (together with linux-headers-2.6.22-2-sparc64-smp_2.6.22-4+farm1_sparc.deb) is a kernel which works well - at least on out US III machine. We've applied 179c85ea53bef807621f335767e41e23f86f01df to make sure that the system doesn't create unkillable processes anymore if you use the libc6 from _lenny_. Please read the following WARNING: using the libc6 from Etch on an US III machine results in a freeze (badly, as in not reacting to stop+a/break) of your system if you do things like using aptitude after becoming root by the use of su/sudo. This is not that bad with the libc6 from testing, but this is definitely NOT fixed. Take a look into http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=433187 for details, and provide any useful informations please. Hope that helps, Bernd -- Bernd Zeimetz <[EMAIL PROTECTED]> <http://bzed.de/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
