[U-Boot] [PATCH v2] Make FIT support really optional

[Carlos Santos]

> From: "Carlos Santos" <casantos at datacom.ind.br>
> To: "Tom Rini" <trini at konsulko.com>
> Cc: u-boot at lists.denx.de
> Sent: Saturday, June 4, 2016 2:39:22 PM
> Subject: Re: [U-Boot] [PATCH v2] Make FIT support really optional

>> From: "Tom Rini" <trini at konsulko.com>
>> To: "Carlos Santos" <casantos at datacom.ind.br>
>> Cc: u-boot at lists.denx.de
>> Sent: Saturday, June 4, 2016 10:06:58 AM
>> Subject: Re: [U-Boot] [PATCH v2] Make FIT support really optional
> 
>> On Fri, Jun 03, 2016 at 04:16:26PM -0300, Carlos Santos wrote:
>> 
>>> Due to some mistakes in the source code, it was not possible to really
>>> turn FIT support off. This commit fixes the problem by means of the
>>> following changes:
>>> 
>>> - Enclose "bootm_host_load_image" and "bootm_host_load_images" between
>>>   checks for CONFIG_FIT_SIGNATURE, in common/bootm.c.
>>> 
>>> - Enclose the declaration of "bootm_host_load_images" between checks for
>>>   CONFIG_FIT_SIGNATURE, in common/bootm.h.
>>> 
>>> - Condition the compilation and linking of fit_common.o fit_image.o
>>>   image-host.o common/image-fit.o to CONFIG_FIT=y, in tools/Makefile.
>>> 
>>> Signed-off-by: Carlos Santos <casantos at datacom.ind.br>
>>> ---
>>> Changes v1 -> v2
>>>  Rebased to the top of master branch.
>>> 
>>>  common/bootm.c  | 2 ++
>>>  include/bootm.h | 2 ++
>>>  tools/Makefile  | 6 ++----
>>>  3 files changed, 6 insertions(+), 4 deletions(-)
>> 
>> So, why?  I don't like the idea of making FIT support in mkimage
>> conditional.
> 
> If FIT is not to be conditional then what's the purpose of the
> CONFIG_FIT_SIGNATURE configuration option? Looks like it exists exactly to make
> FIT support conditional, which seems to be a reasonable approach, since it
> helps to reduce the size of the boot loader.

Sorry, I meant "what is the purpose of the CONFIG_FIT option". 

>> This makes the life of distribution people harder, not
>> easier.  The functions in common/bootm.c should be being discarded in
>> U-Boot itself when we don't have CONFIG_FIT_SIGNATURE.  Thanks!
> 
> The patch exists because of "distribution people". I sent a patch to
> Buildroot[1] which was refused because it added dependencies on DTC and evolved
> to several follow-ups [2,3,4].
> 
> 1. http://patchwork.ozlabs.org/patch/618486/
> 2. http://patchwork.ozlabs.org/patch/619278/
> 3. http://patchwork.ozlabs.org/patch/619696/
> 4. http://patchwork.ozlabs.org/patch/629988/
> 
> Carlos Santos (Casantos)
> DATACOM, P&D