[U-Boot] [PATCH v2] kwboot: boot Marvell Kirkwood SoCs over a serial link

[Daniel Stodden]

On Sun, 2012-05-06 at 12:53 -0500, David Purdy wrote:
> On Wed, May 2, 2012 at 2:16 PM, Luka Perkov <uboot at lukaperkov.net> wrote:
> > The kwboot program boots boards based on Marvell's Kirkwood platform
> > via Xmodem over their integrated UART.
> >
> > Signed-off-by: Daniel Stodden <daniel.stodden at googlemail.com>
> > Acked-by: Luka Perkov <uboot at lukaperkov.net>
> > ---
> >
> > Changes from version v1:
> >  * fix man page
> >  * minor cosmetic fixes in tools/kwboot.c
> >
> > I changed things pointed out on v1 of this patch which was originally
> > made by Daniel.
> >
> 
> @ Daniel, Luka, Prafulla, etal :
> 
> I've tried kwboot on two of my devices, and as I was expecting:
> 
> - it worked perfectly with the v1.21 BootROM device (a Seagate GoFlex
> Net) : Nicely done, well implemented, and the "timing" seems to be a
> non-issue w/ kwboot.  Other previous applications for UARTbooting have
> seemed to me to be less reliable - only booting 50% or 75% of the
> time.  I tried it with kwboot a dozen times or more and it was
> successful each time,  The -p patch-for-UART-booting option is a nice
> feature.
> 
> 
> - it did not seem to work w/  v1.11 BootROM devices (a Pogoplug
> V2/E02, Seagate Dockstar) {previous utilities/picocom-hacks showed the
> same behavior - the bootROM version 1.11 apparently has some
> difference/defect/shortcoming that does not allow this function to
> work as the docs suggest it should.  Too bad no one has pointed out
> exactly what the issue is, or found a patch/workaround it make this
> work on v1.11  BootROM devices.}.   NOTE: It did _not_ adversely
> affect the device otherwise.  I'd seen a post on a thread at another
> site that suggested that performing this procedure on a v1.11 device
> could corrupt the NAND - this was __not__ the case w/ my attempts
> (about a dozen trials, with UART uboot.kwb's and also using the -p
> patch option w/ NAND files.

Hey.

Unfortunately I only have a couple Buffalo boards and a Qnap, all
essentially the same chip.

Say, if you find the time, could you run it through strace and post the
full output? (strace -o /tmp/kwboot.log -s 256 -T ..). 

Just see where it bails. Maybe it's some issue worth trying to hack
around.

Thanks,
Daniel