Bob Proulx wrote: > > Daniel B. wrote: > > Can anyone tell me what has changed in the kernel regarding initrd > > images between woody and unstable? > > Isn't the initrd built on the fly at install time by mkinitrd? > Therefore you would need to look at what changed in that program or > its config files.
I unpacked sid's version of initrd-tools and tried manually stuffing its /sbin/init script into the initrd image, but it didn't make much difference. (I couldn't actually install sid's initrd-tools because it depends on sid's module-init-tools, which conflicts with woody's module utilities, on which my current kernel depends. Actually, I think the question is what changed in the kernel regarding module lookup. Module configuration in the initrd file seems to be insufficient for 2.4.22 even though it was sufficient for 2.4.18. >From sid's module-init-tools, it looks like the kernel needs alias declarations like "alias block-major-22 ide-detect". However, since mkinitrd seems to copy module.conf from the current system, and I don't want to risk breaking my current system by adding references to "ide-detect" (which isn't in woody), I can't proceed along that line either. Agh! Should it really be this hard to upgrade a kernel? At a higher level, should it really be this hard to avoid disk corruption (without dropping back to PIO speeds and without missing part of my 200GB disk)? I'm trying to get a new-enough kernel that I can: - use DMA (preferably UDMA) - see all 200GB of my disk - not have occasional disk corruption - not have frequent several-second hangs - use both my processors Daniel -- Daniel Barclay [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]