[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Partial success with bootstrapping mipsel

Martin Schulze wrote:
> this is a partial success report about bootstrapping a mipsel
> machine.

This is a success report now. :)

> We've booted the machine using BOOTP/TFTP.

And entered the installer, yeah!!!

> The trick is to add the initial ramdisk directly to the kernel which
> is fetched via tftp.  A "regular" mipsel box should be able to boot a
> 16MB kernel.  Our initrd was 10MB large uncompressed, 2.9MB
> compressed.

The second trick is to kick mklibs.sh, reduced libraries didn't work,
non-reduced ones do work and size is no issue as long as we stay below
16MB for tftp.

> After booting, the kernel recognizes the compressed ramdisk image and
> boots it.  Unfortunately afterwards busybox was hanging after it
> displayed its initial message "Busybox 0.xx booting ...".  I guess
> there is some realistic hope to get the remaining integration done in
> time for woody.

This is fixed now.

Next step: fiddle with the keymap, do test installations to find out
if something else fails.

> For Anthony: Woglinde has walked a different path with an NFS-Root
> install which also works until a certain step, which also looks
> promising.
> That means, that binary-mips has hope (Woglinde), and binary-mipsel
> has hope as well (Karsten).

Anthony, it would be nice if this would classify as sufficient for
inclusion of both architectures mipsel and mips in woody - unless
something else fails at a later stage, but that's an issue for the



Linux - the choice of a GNU generation.

Please always Cc to me when replying to me on the lists.

Reply to: