Rockbox.org home
release
dev builds
extras
themes manual
wiki
device status forums
mailing lists
IRC bugs
patches
dev guide



Rockbox mail archive

Subject: Re: Rockbox won't start from flash, any ideas?
From: Rune Wiinberg (colourless_at_stud.ku.dk)
Date: 2003-06-25


Hi Jörg.

Did you try to reverse engineer a dump of the original flash content?
Wouldn't this enable you to figure out the necessary initialization code?

Just a thought :-)

/ Rune

----- Original Message -----
From: <idc-dragon_at_gmx.de>
To: <rockbox_at_cool.haxx.se>
Sent: Wednesday, June 25, 2003 2:21 PM
Subject: Rockbox won't start from flash, any ideas?

> Hello,
>
> frequent readers of this may know that I'm working on getting to Rockbox
> start from flash memory, kicking out the original firmware. The reward
will be a
> much faster "On"-to-music time. Irony has it that I have one of the few
> boxes with the wrong flash chip, not in-system programmable. But a new one
is on
> order, I'm desperately waiting for it.
>
> Meanwhile I have built my toolchain for UART boot (requires the serial
mod).
> With a little extra hardware mod (pulling the LCD lines low instead of
high)
> I can exploit a mode where the boot ROM expects code from the UART. A PC
> application now transfers a little monitor program to the internal RAM.
When
> that's done the two can talk, the current command set allows memory
read/write
> and jumping to an address.
>
> This is sufficient to program the flash (even if completely garbled,
because
> UART boot comes before flash boot) and other tricks. Currently I'm
> "practicing" to start Rockbox. To do that, I use the monitor to configure
the DRAM
> controller in the same way the flash boot does. When I have the DRAM
available,
> I load the rockbox binary into place and jump to it.
>
> Expectedly, this didn't really work. The big difference is that I "cold
> start" rockbox, with no firmware running ahead of it. In the normal case,
Rockbox
> is loaded into an environment where "the bed is already made", the
firmware
> has already initialized much of the hardware.
>
> All it currently does is switching on the backlight for a few seconds,
then
> shutting it off again. No splashscreen, no harddisk activity. A bit
> disappointing. Does this ring a bell to anybody, what initialization might
be missing?
>
> My next plan is to use my toolchain to not load rockbox, but the gdb stub.
> That can in turn load a debug version of rockbox and allows to step around
> there. However, my past experience with gdb (under Windows) is no good.
The
> target connection was very unreliable, the debug application was often
frozen and
> had to be killed. Is anybody debugging the box under Windows? Maybe I can
> persuade a Linux guy with gdb to do the UART boot mod?
>
> So long,
> Joerg
>
> --
> +++ GMX - Mail, Messaging & more http://www.gmx.net +++
> Bitte lächeln! Fotogalerie online mit GMX ohne eigene Homepage!



Page was last modified "Jan 10 2012" The Rockbox Crew
aaa