On Sun, Nov 26, 2006 at 09:10:19PM -0200, Otavio Salvador wrote: > Sven Luther <[EMAIL PROTECTED]> writes: > > >> Well, it's right and wrong. Those modules don't need to be available > >> while booting but be loaded just after init is being run (I might be > >> wrong here) but if I got right there's a problem on 2.6.18 kernel that > >> doesn't instruct udev to load them and including them on > >> initramfs-tools/modules or /etc/modules is a nice workaround for the > >> problem. > >> > >> Am I wrong? > > > > Indeed, but somewhere in initramfs, there is a place where additional > > modules > > are loaded, and this loads i2c-powermac and therm_pm72 right now, but not > > the > > windfarm modules. > > > > We could do an initramfs hook package which would provide suitable stuff > > there, or include it in initramfs proper. > > Yes. We could but it would add another package on archive while the > fix is a simple change on initramfs itself. I don't think it's a good > trade of.
Well, the initramfs-tools is a hook based technology. We could add that hook to a package dependeded upon by the kernel or always installed in powerpc installs. Like mkvmlinuz for example. But yes, fixing it in initramfs-tools is nicer. Now that the d-i problem is solved, i will ping again maks, and maybe provide the d-i patch to him again, or send him a patch directly. That said, the problem is more subtle in initramfs, because in d-i, loading extra modules on powerpc64 hardware is not really a problem, but we may not wish to do so in the installed system, to not clutter lsmod output if nothing else. Friendly, Sven Luther -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
