Re: [avr-libc-dev] snprintf bug ?

[Bernard Fouché]

Russell Strong wrote:

> Hi,
>
> I may have found a bug, I'm not sure if it would be with avr-libc or
> gcc.  I'm using gcc-4.0.1, binutils-2.16.1 avr-libc-1.2.5.
>
> The problem is with the following function and snprintf, if heading_tmp
> is declared static it all works.  If it is not declared static, snprintf
> puts garbage into heading_tmp ( lot of 0xb1 ).
>
> Am I doing something illegal? Do snprintf pointers need to be static?  I
> couldn't find any thing like that in the Docs.
>
>  
The difference here between having 'static' or not for heading_tmp[] is, 
without 'static' to use the stack for memory storage, or, with 'static' 
to use the .bss segment (global variable). So that's RAM anyway and 
snprintf() does not care about what bunch of ram you use.

Usually the problem that you describe appears if:

- the memory used in the stack is referenced after that the function has 
returned. It is not the case here.
- the stack grew too much and something else scrapped it or the stack 
scrapped something else (some global variable just after the maximum 
allowable stack). Interrupt routines called while your function is being 
processed will grow the stack also. Obviously you have interrupt 
routines since you use cli()/sei().
- you ran in a yet unknown (and unlikely) compiler bug.

Can you show the result of 'avr-size your_program.elf' and tell us what 
target you are using?

 Bernard