FW: Barebox on Karo TX25

[Jean-Christophe PLAGNIOL-VILLARD]

On 16:24 Thu 05 Dec     , Sascha Hauer wrote:
> On Tue, Dec 03, 2013 at 08:01:26PM +0100, Jean-Christophe PLAGNIOL-VILLARD wrote:
> > On 10:26 Tue 03 Dec     , Sascha Hauer wrote:
> > > BTW Please don't top-post.
> > > 
> > > On Mon, Dec 02, 2013 at 09:13:31AM +0000, John Parker wrote:
> > > > Hi again
> > > > 
> > > > So now my problem is that I really could do with the LED and UBIFS stuff that happened since February.
> > > > 
> > > > Should I try to fix mx25 in the latest code, or undo the problematic commit?
> > > 
> > > Well preferably you should fix the latest code and send a patch to make
> > > it work again ;) However, that could become tricky. For doing so the
> > > patch to enable the debug UART early may become useful. With this you
> > > can use putc_ll and puthex_ll in early code.
> > > 
> > > I did some investigation. What happens here is really strange. First I
> > > added a missing instruction cache invalidation. That alone doesn't fix
> > > the problem. Inserting a nop fixes booting for me, see the attached
> > > patch. If I move the nop up in the code it still works. If I move it
> > > down, then the board does not work. I have no idea what's happening here
> > > :(
> > 
> > could be that the sdram controler need times to stabelized and that you miss a
> > delay for this
> 
> I thought about this aswell, but I don't believe that a single nop could
> make such a reproducable difference. I tried all this with JTAG
> debugging aswell which should make all timing issues vanish.

on at91 we need to do a bus access to validate the sdram config so could be
an out of sync on the bus

Best Regards,
J.
> 
> Sascha
> 
> -- 
> Pengutronix e.K.                           |                             |
> Industrial Linux Solutions                 | http://www.pengutronix.de/  |
> Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
> Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |