[U-Boot] [PATCH v2] sbc8548: document how to write main flash from alternate flash

[Paul Gortmaker]

On 15-03-16 04:01 PM, York Sun wrote:
> 
> 
> On 03/16/2015 12:53 PM, Paul Gortmaker wrote:
>> If you are running on the alternate flash in order to fix a
>> corrupted main flash image, it might be good to have the steps
>> for that documented as well.
>>
>> Also delete any protect commands from non-active flash banks;
>> they are only required for the instances where you are writing
>> to the currently in use flash image.
>>
>> Signed-off-by: Paul Gortmaker <paul.gortmaker at windriver.com>
>> ---
>>
>> [v2: drop protect commands from instances where we flash the non
>>      active flash bank ; it wasn't protected anyway.]
>>
>>  board/sbc8548/README | 12 ++++++++++--
>>  1 file changed, 10 insertions(+), 2 deletions(-)
>>
>> diff --git a/board/sbc8548/README b/board/sbc8548/README
>> index feac5e3e63e5..92731645f266 100644
>> --- a/board/sbc8548/README
>> +++ b/board/sbc8548/README
>> @@ -132,11 +132,19 @@ image in the SODIMM that is built with CONFIG_SYS_ALT_BOOT enabled,
>>  
>>  	tftp u-boot.bin
>>  	md 200000 10
>> -	protect off all
>>  	era eff00000 efffffff
>>  	cp.b 200000 eff00000 100000
>>  	md eff00000 10
>> -	protect on all
>> +
>> +If you are running the alternate 64MB /CS0 settings and want to update
>> +the "normal" default 8MB u-boot image, then (built with CONFIG_SYS_ALT_BOOT
>> +disabled) the steps will become:
>> +
>> +	tftp u-boot.bin
>> +	md 200000 10
>> +	erase effa0000 efffffff
>> +	cp.b 200000 effa0000 60000
>> +	md effa0000 10
>>  
>>  Finally, if you are running the alternate 64MB /CS0 settings and want
>>  to update the in-use u-boot image, then (again with CONFIG_SYS_ALT_BOOT
>>
> 
> Paul,
> 
> Sorry I didn't catch this in previous version. In your instruction, you presume
> the tftp location is 0x200000, which may not be the case. I suggest you add
> specific location in tftp command so it is clear.

Sure, I can do that.

> 
> I also noticed the file size seems different in your README, comparing with the
> section above. It may be worth to double check.

I think the reason for that was that the alternate flash has a large
sector size, and we can't get the exact number of 0x60000 alignment
on 0x80000 sized sectors.   From "fli" output:

  FF980000 E      FFA00000 E      FFA80000 E      FFB00000 E      FFB80000 E    
  FFC00000 E      FFC80000 E      FFD00000 E      FFD80000 E      FFE00000 E    
  FFE80000 E      FFF00000   RO   FFF80000   RO 

I guess it could be 0x80000 for write instead of 0x100000, but for
some reason I recall it being useful to invalidate the environment
sector at the same time on the alt bank ; I can't recall why anymore...

The eTSEC aren't detected when I build v2015.04-rc3 ; if I solve that
then I'll try a 80000 write for alt flash and see if it comes up OK.
(if you have a hint about the eTSEC, let me know ; will save me a bisect)

P.
--


> 
> York
>