Re: [jornada] de-brick

* Jamey Hicks <jamey.hicks_at_hp.com> [021009 18:28]:

>>>The intermittent nature of the problem with JTAG makes me think it is a
>>>signal integrity problem. I have run into the same thing here on
>>>occasion. If you soldered connections, make sure none are cold solder

[...]

>There are two things that are important here. First, that the wires are
>not shorted together. Second, that there is a good solid connection for
>each wire.

[...]

>The value of the resistors is not important. The resistors are there to
>prevent ringing on the jtag signals. However, such long wires can cause
>noise on the signals, so if you still have problems I recommend
>shortening it to a half meter or so.

I've been thinking about the problem I have, and also tried everything I
could think off. What I saw is that without the Jornada connected to the
parallel port, jflash gives one of the messages I sometimes get with it
connected.
Also, since I compile my own kernel and have not been using the parallel
port for a long time, it might something to do with my kernel config. I
have the parport compiled as module, and also have the lp module.
Loading 'lp', as to use a printer, I cannot load short.o, since it says
the port is busy.
Could you tell me what is exactly needed to use jflash? Besides, in case
I insmod a short.o compiled on my system, and thus not needing to use
short_load, would that be a problem?

I made an 'alias /dev/shortint short' on my modules.conf, for devfsd to
be able to locate it.
BTW, I am using kernel 2.4.20-pre8-ac3

-- 
Javier Marcet <jmarcet_at_pobox.com>
Saint Louis University, Madrid Campus