Re: [ECOS] Stack access violations in eCos

[Larice Robert <larice at vidisys dot de>]

> Larice Robert wrote:
> >>Renaming CYGBLD_ANNOTATE_ALIGNED is just a search and replace. And I've 
> >>done what I can to help with the ChangeLog!
> > 
> > 
> > i've done so, will send it a bit later.
> > 
> > 
> >>CYGBLD_ATTRIB_ALIGN should stay as it is, accepting an argument. We're 
> >>adding a new CYGBLD_ATTRIB_ALIGNED macro that takes the maximum alignment 
> >>required for a thread stack. Perhaps the best thing to do is in fact to 
> >>rename your CYGBLD_ANNOTATE_ALIGNED to CYGBLD_ATTRIB_ALIGNED_MAX to make 
> >>this intention clearer.
> > 
> > 
> > meanwhile i've learned a possibility how to change the definition of
> > CYGBLD_ATTRIB_ALIGN in such a way as to allow no argument.
> > 
> > -#define CYGBLD_ATTRIB_ALIGN(__align__)    __attribute__((aligned(__align__)))
> > +#define CYGBLD_ATTRIB_ALIGN(__align__...) __attribute__((aligned(__align__)))
> 
> I don't want to introduce a dependency on vararg macros. That completely 
> ties us to GCC removing a lot of the point of cyg_type.h in the first place!
> 
> > or would you prefer something like
> > char stack[200] CYGBLD_ATTRIB_ALIGN(CYGARC_ALIGNMENT);
> > not so beautiful i think
> 
> I think I must have written unclearly, sorry. I would prefer:
> 
> char stack[200] CYGBLD_ATTRIB_ALIGNED_MAX;
> 
> i.e. a separate macro.

Hello Jifl,

 i'm sending modified versions of my patch
 
 it comes as a set of files, so you can modify it yourself easier.
 the patch is seperated into 4 files, somehow sorted, see %readme.txt
 the changelogs comes as a perl script, which will generate them
 _doit.sh shows how to use this stuff

 furthermore i've send _diff, which is the grand result here with
 my pretty old ecos tree.

 i've done no test whatsoever

Robert Larice

Attachment: alignkit.tgz
Description: alignkit.tgz

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss