[U-Boot] [PATCH v2 08/22] omap: add spl support

Aneesh V aneesh at ti.com
Wed Jun 15 14:08:32 CEST 2011


Dear Wolfgang,

On Wednesday 15 June 2011 05:34 PM, Wolfgang Denk wrote:
> Dear Aneesh,
>
> In message<4DF88F45.9090905 at ti.com>  you wrote:
>>
>>> I don't get this.  Why don't we just pass the required make target
>>> from the top level Makefile?  If we want to build "onenand-ipl-2k.bin"
>>> then this would result in running "make onenand-ipl-2k.bin" in the
>>> respective directory - this is then also independent of whether this
>>> image contains support for one specific boot device or for any set of
>>> different devices.
>>
>> Then you will have to deal with many such images in the top-level
>> Makefile(u-boot/Makefile). Instead why not have just one rule like the
>> below in the top-level Makefile for all the different spl's:
>>
>>
>> SPL    :$(TIMESTAMP_FILE) $(VERSION_FILE) depend tools
>>       $(MAKE) -C spl/board/$(BOARDDIR) all
>>
>> And leave the rest to the board level Makefiles
>
> We just discussed that there are cases where there may be no board
> level Makefile be needed at all, because there is no board specific
> code to handle.
>
> Also, why would there be many such images? We might just want to use
> more reasonable names.  We have "u-boot.bin", and this works fine for
> all boards, so why cannot we make it that "u-boot-nand.bin" works for
> all boards booting from NAND, and in general "u-boot-<DEV>.bin" works
> for all boards booting from a<DEV>  boot device?
>
> Instead of "onenand-ipl-2k.bin", we would just have a generic
> "u-boot-onenand.bin"
>
>> Of course, this is assuming the existing Makefile structure. With the
>> new Makefile structure you are suggesting this may not hold good.
>
> Why not?

I was saying that my suggestion of delegating everything to board level
Makefile will not work with the new top-down approach you are
suggesting in the other mail-chain.

best regards,
Aneesh


More information about the U-Boot mailing list