Re: [jornada] de-brick

Javier,

I am glad that you have made so much progress!

It sounds like jflash is working just fine. I will locate a
bootldr image and send you a pointer.

The program does indeed take forever to run. There are
2 reasons for this: 1) it is toggling individual pins on the
parallel port by successive kernel ioctl calls (with lots
of overhead on each bit transmitted) and 2) there are built
in delays in the program (which are probably not necessary)
to make sure that the Jtag interface is not overrun with
data.

As I recall, it takes about 10 - 20 minutes (or maybe it just
feels like it) on my machine to download bootldr.

jca

Javier Marcet wrote:

> * John Ankcorn <ankcorn_at_hpl.hp.com> [020927 23:06]:
>
> >In case this is useful, I have put the source (and a very brief
> >Readme) for the jtag programming tool that I have been using
> >for Jornada56x devices at:
> > ftp://salsa.lcs.mit.edu/pub/jca/jflash.tgz
>
> >It requires a cable which is relatively easy to wire and some
> >way of connecting to the Jtag pads on the back of the 56x.
> >I have been using a custom milled block of nylon that uses pogo pins
> >(push pins) to connect to the test points.
>
> >I don't recommend using this tool, as you will definitely
> >erase your flash, perhaps making it unusable (until you
> >re-program it correctly!). I did want to make the source
> >and wiring information available for those among us who
> >are already the proud owners of expensive bricks in
> >hope that they can make them functional again.
>
> Hi John. Maybe you remember me. I had written to you looking for help to
> debrick my 567.
> Following the instructions on that archive I made the connection between
> the jornada and my desktop pc's parallel port.
>
> On this point I am not sure I understood the wiring correctly.
> I connected as follows:
>
> Jornada IR is here at the top, and the joranada is facing down:
> Jornada pin 1
> Jornada pin 2
> Jornada pin 3 -> 100 ohm resistor -> parallel port pin 2
> Jornada pin 4 -> 100 ohm resistor -> parallel port pin 4
> Jornada pin 5 -> 100 ohm resistor -> parallel port pin 14
> Jornada pin 6 -> 100 ohm resistor -> parallel port pin 11
> Jornada pin 7 -> 100 ohm resistor -> parallel port pin 3
>
> When the connection was done, I loaded the 'short' module, which gave no
> problem at all and then I run jflash.
> I tried both reading and writing what I believe is an old copy of the
> original wince bootloader which is 262144 bytes, and by looking at the
> contents it definately seems so.
> At first, though, it complained saying 'error reading flash attribute
> space', and that I should check my power connection and cables. The
> message was not constant. The PDA, however, was stable on the table,
> and the connections (wirings) where not changing, nor being touched at
> all.
>
> Since my PDA so far showed no sign of life I feared that I could need
> to somehow make it see it was being re-programmed, so I tried pushing
> the power button and also the pad and the central button this. This last
> one seems to have done the trick.
> I could re-program my old bootloader.
> It said one thing which I didn't like, though. It showed:
>
> JFLASH Version xxx
> current control 20
> SA-1110 revision B4
> Number of blocks in device = 128
> Block size = 65536 0x10000
> Device size = 16777216 0x1000000
> verify base 0 size = 8388608 0x800000
>
> which doesn't make too much sense since my unit has 32MBytes of flash,
> AFAIK. Yet I let it go to the end.
> During the process, which lasted like one minute -I'm not sure, I didn't
> time it, and being as ancxious as I was to see it come alive, I probably
> got a bad impression of the time-lapse -; the screen flashed a couple
> times, just a a white flash, together with a little noise.
> When the process was over, the unit was as before, not working at all.
>
> I am now tying to read it, which was a little bit harder, since the
> jflash tool refused to start giving the same message I quoted above.
>
> It's been a few minutes already - yes, a few ninutes for sure -, and it
> is still at 'Verifying flash at hex address 468da, 3.44% done', with
> the jflash tool taking +90% of my processor (1800 Athlon-XP).
> Since it is only reading I probable let it go a little bit more. After
> all I don't really know what's being read, if the boot sector, the whole
> flash ram, or what. And reading the contents of the entire flash was a
> lengthy process originally.
>
> Anyway, if you can point me to what I should flash on it thorugh the
> JTAG interface to see it alive.
> I am really interested in having Linux on it, although if I can get the
> wince back again quicker so that I am sure that everything is working
> fine, that'll be allright.
>
> --
> Javier Marcet <jmarcet_at_pobox.com>
> Saint Louis University, Madrid Campus
Received on Wed Oct 09 2002 - 01:24:44 EDT