[PATCH RFC v2] image: apply FDTOs on FDT image node without a load property

[Marek Vasut]

On 1/21/25 9:25 AM, Quentin Schulz wrote:

[...]

>>> --- a/boot/image-fit.c
>>> +++ b/boot/image-fit.c
>>> @@ -2348,10 +2348,10 @@ int boot_get_fdt_fit(struct bootm_headers 
>>> *images, ulong addr,
>>>       char *next_config = NULL;
>>>       ulong load, len;
>>>   #ifdef CONFIG_OF_LIBFDT_OVERLAY
>>> -    ulong image_start, image_end;
>>>       ulong ovload, ovlen, ovcopylen;
>>>       const char *uconfig;
>>>       const char *uname;
>>> +    char *of_flat_tree;
>>
>> Can this be void * instead ?
>>
> 
> The compiler seems happy with that. FWIW, map_sysmem() returns a void*, 
> and boot_relocate_fdt() expects a char** and we anyway cast it to a 
> ulong later.

Indeed, that's what struck me.

> The question I have is why? What does this bring? Maybe there's 
> something I can learn here :)
Nothing much, except for type opaqueness. The ideal fix would be to 
align the types, but that would be a huge undertaking and I cannot ask 
that of you as a prerequisite for this one fix.