[Server-devel] Booting F9 kernels on XO...
jvonau at shaw.ca
Wed Apr 15 19:06:50 EDT 2009
On Wed, 2009-04-15 at 12:14 -0400, Daniel Drake wrote:
> 2009/4/15 Martin Langhoff <martin.langhoff at gmail.com>:
> > - dsd has a nice "fixup F9 for XO" script described here
> > http://www.reactivated.net/weblog/archives/2008/08/regular-linux-desktops-on-the-xo/
> > The script builds a custom 2.6.25 which I think pulls in a few
> > additional patches (sound, etc), but most importantly, it builds the
> > mmc related drivers statically. It also sets up an initrd.
> > If we try to boot with the initd, it fails. But if we skip the initrd,
> > the system boots.
> What's the problem?
Think we need to be little clearer here.. Martin was building an ext3
image of an XS server for use on the XO. I added some bits
to /etc/sysconfig/mkinitrd and called mkinitrd in %post of the kickstart
file. This initrd doesn't play nice when booting F9's 2.6.27 kernel.
Remaking the initrd on the XO doesn't help, using the hints from:
> My custom initramfs definitely boots on my XO.
> Also the stock F9 one did too, but it landed a lot of errors during
> boot (which took several seconds to enumerate).
I have anaconda running on the XO, installs to usbkeys/booting from are
fine with F9. F11's anaconda runs straight away, but opps when installing
the rpms to the MMC card.
> > - Build a kernel exactly like Fedora 9 kernels, but with the mmc
> > drivers built statically. This is what I am am doing right now,
> > following the nice guide at
> > http://fedoraproject.org/wiki/Docs/CustomKernel
Martin: what kernel version?
> And boot without an initramfs?
> When I did this IIRC boot would fail approximately 1 in 5 attempts
> with sdhci errors. Never got a chance to figure out why, but putting
> in the initramfs hid the problem nicely.
I just d/l'ed http://dev.laptop.org/~cjb/rawhide-xo/f11-beta/20090403.bootable.gz
this kernel opps for me also when booting from the MMC card. Bad mmc card on my
More information about the Server-devel