gta01/02 address space

Wolfgang Spraul wolfgang at
Thu May 1 11:35:58 CEST 2008

hmm, so maybe the addresses from 'nand bad' in u-boot are 0-based?
I have a GTA02 here (#114) with 4 bad blocks (0ff80000, 0ffa0000,  
0ffc0000, 0ffe0000).
That would mean that the last 4 blocks of NAND are bad. Before making  
that assumption I wanted to double-check whether this isn't memory  
mapped somewhere with a base other than 0, but you say it's not.
Werner asked a few times we should run tests over a number of units to  
find out  whether bad blocks are evenly distributed, or more prevalent  
in some places.
Until I hear otherwise, I will assume those numbers are 0-based. If I  
find the first >= 10000000 I know I'm wrong ;-)

On May 1, 2008, at 5:27 PM, Andy Green wrote:

> Hash: SHA1
> Somebody in the thread at some point said:
> | Andy,
> | great, good start!
> | Unfortunately I still don't know what the base address for  
> 0xFF800000
> | is. It is pointing into NAND, but where does NAND start?
> I think you are confusing NAND with some normal kind of decent  
> memory :-/
> NAND is a "peripheral", there are a bunch of peripheral registers at
> 0x4E000000 that you use to talk to this "peripheral".  It is not  
> memory
> mapped itself.  You send commands to the NAND "peripheral" from one
> register, write the address you want into another register and spam  
> the
> data out from another.
> So any addresses you see about NAND will be offsets inside the NAND
> peripheral and not having physical processor addresses.
> - -Andy
> BTW Pocky told me it is a holiday today
> Version: GnuPG v1.4.9 (GNU/Linux)
> Comment: Using GnuPG with Fedora -
> iEYEARECAAYFAkgZjPQACgkQOjLpvpq7dMry0QCeJjMF0tBRiCvrmGQnx9jwMf6C
> xuQAoI6bZu0GCs4u+m5a+i7/iYO3PT+b
> =vV+V

More information about the openmoko-kernel mailing list