i would think so, but i ask since, if you check this:
http://elinux.org/BeagleBoard#Binaries
that page links you to the angstrom distro page to grab copies of MLO
and u-boot.bin, perhaps leading a new reader to think that different
versions of those binaries are tied to specific distros.
if they *are* distro-independent, it might be worth mentioning that
on that page so that there's no confusion.
rday
p.s. part of the presentation might involve discussing why you might
want to build your own MLO and u-boot.bin images from scratch, but i
think i'll leave that until later. i'm assuming that build process is
pretty straightforward.
--
========================================================================
Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca
Linked In: http://www.linkedin.com/in/rpjday
Twitter: http://twitter.com/rpjday
========================================================================
No, they are not distro dependent. It's just that angstrom most
regularly build the binaries from recent sources. Yes, it would be
nice to clarify this in above wiki. It's a wiki, so do you like to
clarify this?
> p.s. part of the presentation might involve discussing why you might
> want to build your own MLO and u-boot.bin images from scratch, but i
> think i'll leave that until later. i'm assuming that build process is
> pretty straightforward.
Yes, it is.
http://elinux.org/BeagleBoard#U-Boot
Many thanks and best regards
Dirk
Actually the proper answer is: MLO and u-boot.bin are not distro-agnostic.
However. MLO and u-boot.bin *are* kernel agnostic.
There are intricate dependencies. E.g. the latest u-boot won't
properly boot the angstrom 2.6.27 kernel, and the 2.6.27 u-boot will
not boot the 2.6.29 angstrom kernel. One of the causes seem to be that
there are (were?) some dependencies between u-boot and kernel where
the kernel relied on certain initialisations being done by u-boot.
Especially this seems to be an issue wrt USB.
I'm not an MLO wizard, but there the incompatibilities seem to be much less.
BTW: a while ago (in april I think) I build the 2.6.27 kernel, but I
did not manage to find a u-boot which would work with this kernel
(although I am sure I had this early jan). Rationale is that I wanted
to compare SD handling between .27 and .29 (as my SDHC card worked
with .26 and if I recall correctly also .27, but won't work properly
any more with .28 and .29. If anyone has a solution, please let me
know.
Frans
I just checked the MLO source and it's really simple basically only loading
u-boot from MMC in case present. In case not present, it will try to load it
from the NAND flash instead.
As such it's kernel/distro independent, but some kernels/distros might rely
on the fact, that MLO sets up the MMC interface in order to search for a
u-boot binary.
In case this is one day changed due to whatever reason some stuff might
eventually go wrong...
As such, I think it's pretty safe to assume that the kernel and u-boot
should be independent off the MLO. In case they aren't I at least would
consider it a problem that should be solved in u-boot and/or kernel...
Best regards
Søren
that's the impression i got -- even if u-boot is somehow tied to the
kernel in a distro-dependent way, MLO is so basic that it shouldn't
have any issues working with *any* properly-built u-boot.bin.
however, just to play it safe, it's always a good idea to get
everything from the same place, so that's how i'll word it on my wiki.
thanks.
rday
i still plan on asking trivially dumb questions on a regular basis
and documenting the answers. i long ago stopped being embarrassed by
making a fool of myself that way. :-)
coming soon: what's up with toolchains?
rday