On Thu, 2005-12-15 at 12:19 -0800, Steven Ringwald wrote: > Bob Chiodini wrote: > > >Okay, I was concentrating on the bio too big device loop0 message. > > > >Maybe a bug in busybox or the kernel: > > > >http://64.233.167.104/search?q=cache:1I7Sw6-T9EIJ:bugs.busybox.net/view.php%3Fid%3D498%26nbn%3D6+%22bio+too+big+device+loop0%22+busybox&hl=en&lr=&client=firefox-a&strip=1 > > > >The above is the cached, text-only, google link, busybox.org seems to be > >down or extremely slow. > > > >Are you running a 2.4 or 2.6 kernel? > > > > % uname -r > 2.6.14-1.1644_FC4 > % uname -p > x86_64 > > I went and downloaded a newer version of busybox, rebuilt it from > source, and installed it in my initrd. I also bit-the-bullet and got rid > of nash from that image. Seems to be behaving a little better. Now I get > as far as mountint the iso, and the pivot_root command seems to be > screwing up on me. My last line in my linuxrc file is msh, so that I can > have some interactivity. > > mount shows: > > /dev/hda2 on /dos type vfat (rw,noatime, ) > /dev/loop0 on /sysroot type iso9660 (ro) > /proc on /sysroot/proc type proc (rw,nodiratime) > > when I go: > busybox pivot_root /sysroot /sysroot/initrd > > I get: > > pivot_root: pivot_root: invalid argument > > but I can at least SEE the image > > > Steve, Do you ever sleep? Google shows a couple of hits with pivot_root and the 2.6.14rcX kernels, but they don't seem related to busybox. I assume your pivot_root is from busybox. You might try bringing in the binary from Fedora (/sbin/pivot_root). There is an example usage in the man (8) page. It recommends cd'ing to the new root before pivoting. There is an implication that something changed, or is changing. Also have a look at man 2 pivot_root. The man (2) pivot_root page indicates that -EINVAL will be returned if old root is not beneath new root. I assume that means the directory where old root will be mounted must exist. Can you post your new nash'less script? I'm only on my first cup of coffee... I could be way off-base. Bob...
