RE: [Kernel-discuss] hx2750 kernel work

From: Torbjörn Jansson <torbjorn.jansson_at_mbox200.swipnet.se>
Date: Mon, 14 Mar 2005 22:23:58 +0100

kernel-discuss-bounces_at_handhelds.org <> scribbled on :
> My biggest problem at the minute is I'm missing some key
> gpios and can't
> find the DoC device in memory (I'm fairly sure it's somewhere
> 0x04000000-0x040fffff).

On the h4150 the disc on chip is located at address zero, and a few kb in
size (can't remeber the exact size)
And it also enters reset mode whenever you try to read it, so that means you
will always read parts of it as zeros.
My guess is that your doc chip is also located at address zero, if so you
shoud have a region of zeros at 0x1000 (if i'm not mistaken)

I tried to get the DiscOnChip working for the h4150 but i didn't get very
far.
There is currently two drivers, one that's obsolete and have bugs and a new
one.
The problem is that the new driver don't support the milleniup plus version
of doc but the old one does, kind of.

To get the old driver atleast partialy working i had to modify it so it
probes for the doc chip at address zero and also fix a few macros it uses to
read memory, i think it was trying 32bit access instead of 8bit or something
like that.
With these 2 modifications i got the module to load but the data that comes
from the /dev/mtd device looks as if you read 32bits and throw away 16, so
every 2 bytes is missing.
I compared it against a known rom dump and that seems to be the case.

Another wierd thing with the old driver is that it sometimes takes a few
"modprobe docprobe" to get docprobe to detect it and get the modules loaded,
my guess is that it doesn't take it out of reset mode properly.
Received on Mon Mar 14 2005 - 16:26:36 EST

This archive was generated by hypermail 2.2.0 : Mon Jul 25 2005 - 17:21:03 EDT