[U-Boot] [PATCH v7 09/11] arm: add customized boot command for Faraday Images

[Kuo-Jung Su]

2013/9/14 Albert ARIBAUD <albert.u.boot at aribaud.net>:
> Hi Kuo-Jung,
>
> On Mon, 29 Jul 2013 13:51:51 +0800, Kuo-Jung Su <dantesu at gmail.com>
> wrote:
>
>> + * At the time of writting, none of Faraday NAND & SPI controllers
>> + * supports XIP (eXecute In Place). So the Faraday A360/A369 SoC has
>> + * to implement a 1st level bootstrap code stored in the embedded ROM
>> + * inside the SoC.
>> + *
>> + * After power-on, the ROM code (1st level bootstrap code) would load
>> + * the 2nd bootstrap code into SRAM without any SDRAM initialization.
>> + *
>> + * The 2nd bootstrap code would then initialize SDRAM and load the
>> + * generic firmware (u-boot/linux) into SDRAM, and finally make
>> + * a long-jump to the firmware.
>> + *
>> + * Which means the SPL design of U-boot would never fit to A360/A369,
>> + * since it's usually not possible to alter a embedded ROM code.
>
> Sorry, but I don't see why SPL could not run in SRAM as the 2nd
> bootloader in your description; SPL certainly does not try to "alter a
> embedded ROM". So, please rewrite the paragraph with the correct reason
> why SPL cannot be the 2nd bootloader, e.g., is it
>
> - because the 2nd bootloader is actually in ROM?
> - because the SRAM is too small?
> - ...
>
> In any case:
>

Got it, thanks.
I'll try to study the SPL again, and rewrite the paragraph.

>> + * And because both the 1st & 2nd level bootstrap code use the private
>> + * Faraday Firmware Image Format, it would be better to drop U-boot
>> + * image support to simplify the design.
>
> "Drop"? Certainly not. "Introduce a new image format where U-Boot is
> prepended with a header defined as follow...", yes -- you can even
> make a case that, if SPL cannot be the 2nd bootloader, then SPL is
> actually "dropped" for this platform. Please rewrite.
>

Got it, thanks

> Also: could the whole description and rationale be moved to some
> README.* file either arch/arm/cpu/faraday or in doc/ so that readers
> oof the C file can see the start of the actual code without having to
> scroll through hundreds of comment lines?
>

Got it, thanks

> Amicalement,
> --
> Albert.



-- 
Best wishes,
Kuo-Jung Su