[E3-hacking] E3 Memory Map?

[Jonathan McDowell]

On Tue, May 16, 2006 at 11:27:26AM +0100, David Given wrote:
> Jasmine Strong wrote:
> [...]
> > The reason we have not successfully booted 2.6 from PBL is that
> > for some reason DEBUG_MUTEXES breaks 2.6 booting from PBL
> > NAND (but not from a serial load).  Therefore, to boot from PBL to
> > date we have needed to use 2.4, which is obsolete and therefore
> > not cool.
> 
> If the problem is that PBL has left some piece of hardware in a state
> that is confusing Linux, then simply booting Linux via littleLDR and
> uBoot won't help; unless something touches and resets that same piece of
> hardware, Linux will still fail.
> 
> Why not simply build a kernel without DEBUG_MUTEXES?

That's what I did. Longer term I'd like to figure out why it doesn't
work obviously.

With regards to using LDR and putting u-boot in a LINUX block I confess
I didn't think of that originally, but I'd rather be running code that
we understand rather than having to spend more time reverse engineering
the Amstrad block format and where things need to go in the NAND to make
it happy.

I have a solution that seems to work (except for u-boot being quite slow
to read from NAND; need to sort the ready bit support out) and as soon
as I get over the cold I've been dying of all weekend I'll sort out a
release for people.

J.

-- 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://www.earth.li/pipermail/e3-hacking/attachments/20060516/1fbece73/attachment.pgp