On Thursday 24 April 2008, [EMAIL PROTECTED] wrote: > Alan McKinnon <[EMAIL PROTECTED]> writes:
> >> kernel /kernel-2.6.25-r1 root=(hd0)/sda3 > >> > >> Fails > > > > Won't work. (hd0) is a grub thing. You need a /dev/sda3 or similar > > in there > > I think you are wrong about that. But just a fine point and not > central to the problem. > For example I know for sure you can use the grub notation at the > kernel address like: > > kernel (hd0,0)/kernel-XXXXXX > > At least I know for sure it was possible at one time.. I haven't > actully used that notation in grub for quite a while. I do have that > notation as the address for the splash image in several working > grub.confs. (like splashimage=(hd0,0)/grub/splash.xpm.gz) To clarify, your kernel line will work just fine with that notation. It's an instruction to grub in the form of a path that it understands. Essentially you are telling grub which partition to look on for the kernel and it knows what (hd0,0) is. /dev/sda1 won't work there, as grub does not understand Linux kernel names (There might be a hack to get around this but the code would be ugly as hell). You have a "root" directive to grub, so the notation is redundant. "root" is there so you don;t have to keep typing (hd0,0) everywhere <snip> > In the screen shot provided note that it appears grub is expecting an > intramfs and only lists those types of devices, rejecting both > (hd0,0) and /dev/sda3. > > http://www.jtan.com/~reader/vu/disp.cgi Aaaaaaaaaah, now I see. It's one of two things, and neither is your grub.conf. That's the kernel spitting that garbage at you, so your grub.conf is just fine. You have either: 1. Compiled in the need for an initrd and have not supplied one, or 2. (more likely) you do not have support for your chipset, ata and/or root filesystem compiled into the kernel (NOT as modules). "VFS: Unable to mount root fs" is almost invariable due to this Do you want to use an initrd, or due the highly customized thing and dispense with it? -- Alan McKinnon alan dot mckinnon at gmail dot com -- [email protected] mailing list
