[U-Boot-Users] Why are some global vars part of the image, and some not?

[Wolfgang Denk]

In message <454F992D.7010402 at freescale.com> you wrote:
>
> Advantages:
> 
> 1) Eliminates the one (there may be more in the future) BSS-related bug

Let's keep this straight. The bug is not in U-Boot, but in newly
introduced code, which ignores the documented way to write code that
is running in the restrictred environment before relocation.

> 2) Eliminates the BSS initialization code

Yes, what a win. We save 50 bytes of code and add - wait a second:

-> size u-boot
   text    data     bss     dec     hex filename
 328632   30584  314076  673292   a460c u-boot

...and add 307 KiB of data. You lose by one to several thousands.

> 3) It's LESS complex.  I don't know why you think it's more complex.
> 4) Eliminates the obscure bug that could occur if you expect "int X = 0" to 
> actually be 0 instead of 0xFFFFFFFF.

You don't understand. No matter  what  your're  trying  to  do,  pre-
relocation  code  runs  in a severly restricted environment. All your
attapts to "fix" this are  void,  especially  since  the  whole  data
segment remains read-only. Solving the "int X=0;" case does not solve
the "X=1;" five lines later.

You have to be aware of this situation, and  trying  to  hide  it  by
providong  a  "working  solution" for a single special case is making
things more obscure.

I recommend just to keep in mind *all* the restrictions when  writing
pre-reloc  code,  and  to  minimize  the  amount of such code in your
systems.

Anything else is probably worse than what we have now.

> I agree that making the change globally would require regression testing.  For 
> instance, there could be buggy code out there that depends on the BSS section 
> being initialized to FF before the relocation.

You speculations are ... ummm... interesting. You find funny ways  to
defend a situatioin where the problem is all your's only.

> It could be implemented as a compile-time option that would be disabled by 
> default.

No. I will not accept any of such code. Period.

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Our business is run on trust.  We trust you will pay in advance.