On 2/6/07, Joe Ciccone <[EMAIL PROTECTED]> wrote:
> Colin Dean wrote:
> > I'm in the process of building CLFS with the intention of making a
> > bootable CD that proceeds to load the entire OS onto a RAMdisk, using
> > the remainder of the available RAM for regular...RAM purposes, hehe.
> >
> > Any of you know of good reading material that may point me towards that
> > which I seek?
> >
> > Some Googling yielded initrd tutorials, but I think that initrd is meant
> > more for a temporary part of the boot process than running the entire
> > system for an indeterminate amount of time.
> >
> > http://www-128.ibm.com/developerworks/linux/library/l-initrd.html
> >
> > That DeveloperWorks tutorial on initrd appeals to me. If ya'll think
> > that the initrd may work in my case, then I'll probably follow it.
> You may want to look into using device mapper to mount a root image (so
> that it's writable without massive memory requirements).
>
> Instead of an initrd I'd look into including an initramfs image that
> locates the CD, mounts it, does whatever it needs todo to mount the
> image, then does a pivot_root into the image's root. I think this is (or
> is atleast similar to) what the lfs-livecd does.
> >
> >
>
> _______________________________________________
> Clfs-support mailing list
> [email protected]
> http://lists.cross-lfs.org/cgi-bin/mailman/listinfo/clfs-support
>

Hmm, actually, look into the bootscripts for DSL, there's a "toram"
boot time option that does just that with their system ... but you may
learn more from an older version's bootscripts, as a lot of it's
harder to follow since they started using unionfs and such.

My idea, however: Load a small initramfs (containing mount and either
cp or tar+gzip, busybox could handle the job nicely) that mounts a
tmpfs based filesystem (make sure to set the max size large enough),
then copy a filesystem image into that and mount (or just untar a
system into the tmpfs), then pivotroot into the new system ... there's
no data persistency, but it should run very fast and relatively clean
(assuming no major memory leaks).

-- 
Poison [BLX]
Joshua M. Murphy
_______________________________________________
Clfs-support mailing list
[email protected]
http://lists.cross-lfs.org/cgi-bin/mailman/listinfo/clfs-support

Reply via email to