[U-Boot-Users] [PATCH] CFI driver AMD Command Set Top boot geometry reversal, etc. [Updated]

[Tolunay Orkun <listmember@orkun.us>]

Stefan Roese wrote:

>>> -               flash_write_cmd(info, 0, 0, FLASH_CMD_READ_ID);
>>> +               flash_write_cmd(info, 0, AMD_ADDR_START,
>>> FLASH_CMD_READ_ID);
>> I think this change is what was missing which I overlooked. Intel does
>> not need a specific address to write the command nor unlock sequence.
> 
> And that's the reason I don't like those Intel devices. I got some devices 
> accidentally screwed in a project once.

I hear you. I was not talking about merits.

> So FLASH ID's are now read back correctly.

Great. Your flash device should validate CFI = 1.0 code path for geometry 
reversal detection as well.

>> I will update the patch tonight and resend it yet again.
> 
> Not necessary. I'll forward this patch with the fix to Wolfgang. Thanks again.

I did it anyway [DNX#2006111242000019]. Plus a few more minor changes: I've 
added my copyright line. Added a few more comments and removed some other 
change history comments included with previous copyright lines. Copyrights 
are still there as it should be.

I prefer if you can use this patch if it is not too late.

> And when it's included we can continue to extend the code with additional 
> features, like fixing this ST FLASH geometry reversal (or others) by 
> identifying it's device ID.

It looks the sector addresses you included (below) is now that of top boot. 
Is it not working for you? When I enabled DEBUG the flash driver would have 
trouble with writes even for Intel  (possibily clearing the status register 
prematurely). Make sure you test with DEBUG disabled for cp command.

> Sector Start Addresses:
> FFE00000        FFE10000        FFE20000 E      FFE30000 E      FFE40000 E
> FFE50000 E      FFE60000 E      FFE70000 E      FFE80000 E      FFE90000 E
> FFEA0000 E      FFEB0000 E      FFEC0000 E      FFED0000 E      FFEE0000 E
> FFEF0000 E      FFF00000 E      FFF10000 E      FFF20000 E      FFF30000 E
> FFF40000 E      FFF50000 E      FFF60000 E      FFF70000 E      FFF80000 E
> FFF90000 E      FFFA0000   RO   FFFB0000   RO   FFFC0000   RO   FFFD0000   RO
> FFFE0000   RO   FFFF0000   RO   FFFF8000 E      FFFFA000 E      FFFFC000

I think should probably re-org the code sometime to be more table driven. We 
can avoid testing for vendor everywhere and make code much easier to read 
and extend. If a vendor is sticking to AMD style chips we could also have 
option to exclude the Intel support (or vice versa) completely to save some 
space that way.

Best regards,
Tolunay