From: <idc-dragon_at_gmx.de>
Date: Mon, 4 Aug 2003 15:51:33 +0200 (MEST)

Hi,

This is to find charging issues, with the people for whom flashing caused a
general charging problem with _both_(!) F1-started original software and
Rockbox. Maybe there is an undiscovered bit controlling the charging of the FM.

I've temporarily added some diagnostic code to Rockbox. On start, it
captures the whole I/O space of the processor into a buffer, before Rockbox changes
any bit of it. This buffer can then later be saved via debug screen entry,
produced a file "startup_io.bin" in the root dir. So we know the start scenario
as far as accessible to the CPU (not I2C yet).

I'd like to compare 2 situations:
1. Rockbox starting the old fashioned way: not flashed, original firmware
2. Rockbox starting from flash

If anybody would like to help with this, I'll need those 2 dumps. Generating
it is uncomfortable because you have to flash the original firmware back in,
which only works with the rockbox version of my flash package (else
incompatible plugin), etc.

The next step which I haven't coded yet is the reverse: to load such a file
into the I/O before rolo-ing a software. If we have a dump of the known-good
case, it would be interesting if this fixes the charging when manually
rolo-ing Rockbox.

Thanks for field-testing,
Jörg

-- 
COMPUTERBILD 15/03: Premium-e-mail-Dienste im Test
--------------------------------------------------
1. GMX TopMail - Platz 1 und Testsieger!
2. GMX ProMail - Platz 2 und Preis-Qualitätssieger!
3. Arcor - 4. web.de - 5. T-Online - 6. freenet.de - 7. daybyday - 8. e-Post

Received on 2003-08-04