[U-Boot] ARM: Incorrect ROM protection range?

[Albert ARIBAUD <albert.aribaud@free.fr>]

Le 24/02/2011 08:40, Po-Yu Chuang a ?crit :
> Hi Heiko,
>
> On Thu, Feb 24, 2011 at 3:33 PM, Heiko Schocher<hs@denx.de>  wrote:
>> Albert ARIBAUD wrote:
>> The bin length is calculated in arch/arm/lib/board.c, but it seems
>> to me not correct ... :-(
>>
>> in board_init_f():
>>
>> gd->mon_len = _bss_end_ofs;
>>
>> that seems correct to me, but later in board_init_r()
>>
>> monitor_flash_len = _bss_start_ofs;
>>
>> which is used for example in ./drivers/mtd/cfi_flash.c for protecting
>> the flash sectors ... so this should be fixed.

Correct.

>>> In the U-Boot image itself, knowing the image size could be achieved in
>>> ARM by using a general _end symbol that would be set after the last
>>> image output section, so _end-_start would equal the image size.
>>
>> we have such a "_end" in u-boot.lds files.

I *knew* this name did not pop up in my mind without a reason. :)

Apologies for not having checked.

> I guess we need a __dynsym_end in all u-boot.lds files.

I'd rather go for "_end", which does not tie the solution to dynsym 
being the last section in the image -- imagine for some reason we move 
dynsym inside the image rather than at the end, end the image size will 
be wrong again. With _end, and a suitable comment in the LD file... this 
will be averted where-ever dynsym ends up.

After all, the bug you uncovered was due to using the wrong symbol, a 
BSS related one rather than an image-related one, in the first place; so 
let's try and learn from past mistakes.

> best regards,
> Po-Yu Chuang

Amicalement,
-- 
Albert.