[U-Boot-Users] [PATCH 3/5] Adds support for RedBoot

[Pantelis Antoniou]

On 29 Νοε 2006, at 11:21 ΜΜ, Wolfgang Denk wrote:

Hi Wolfgang,

> Dear Pantelis,
>
> in message <20061129172559.24638.5730.stgit at pantathon.hol.gr> you  
> wrote:
>>
>> +		- Print either all the RedBoot config variables or a specific one
>> +
>> +                  > redboot config printenv [config-var] [member]
>> +
>> +		  member is one of value, type, enable-sense, enable-key with  
>> value
>> +		  being the default one.
>> +
>> +		- Set an environment variable from a RedBoot config variable
>> +
>> +                  > redboot config setenv var config-var [member]
>> +
>> +		  For example when given a config variable of
>> +
>> +		  tsec1_esa=00:08:e5:11:32:33
>> +		          type=esa enable-sense=true
>> +
>> +		  > redboot config setenv ethaddr tsec1_esa
>> +
>> +		  would set the ethaddr environment variable to 00:08:e5:11:32:33
>
> Can you please explain why you  need  a  separate  implementation  of
> printenv  /  setenv  commands?  Why  do you make a difference between
> "RedBoot config variables" and normal (U-Boot) environment variables?
>

This is not the u-boot environment - it is the RedBoot environment.
I am using the RedBoot environment (which is Read-Only) for copying  
environment
variables. Some very important variables are stored there like the  
ethernet MAC
addresses & the ip address.

What this command does is copy the RedBoot config variable to the u- 
boot ethaddr.


>> +	        > redboot exec address	
>> +
>> +		The PPC bd_t differs for kernels compiled for redboot; this  
>> command
>> +		lets you boot such a kernel.
>
> I have to admit that I don't like this. I'd rathwer see you using the
> standard bootm command, and have the necessary  adaption  done  there
> based  on  the  image  type (which should indicate that thgis is a RB
> image). You have a lot of #ifdef's there anyway.
>
>

Tried it, didn't work. The amount of defines was getting obscene.
As an added bonus the registers used to pass control to the kernel  
are different,
plus you really need to disable the MMU & the caches the way that  
RedBot does it.

So in conclusion, yes, it is ugly, hacking bootm to make it work is even
uglier.

> Also, please cleanup your code (indentation by TABs, trailing white
> space, C++ comments, line length).
>

Some files were copied verbatim from RedBoot sources, so they are  
bound to be like this.
I've been told that Lattice modifies the FPGA loader regurally, so I  
tried to keep
the original formatting.

Anyway will do.

> Best regards,
>
> Wolfgang Denk
>
> -- 
> Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
> Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
> If all the Chinese simultaneously jumped into the Pacific  off  a  10
> foot platform erected 10 feet off their coast, it would cause a tidal
> wave that would destroy everything in this country west of Nebraska.

Regards

Pantelis

P.S. Two patches were discarded due to limit size - 40K is pretty  
small, when there's
a single file being 80K. Yes, the FPGA loader file.