[PATCH] PR optimization/12324

[Bernardo Innocenti]

Jan Hubicka wrote:

>>I've submitted this PR. The attached test case is a simplified case of
>>a real case from linux/drivers/net/fec.c:
> 
> I do understand the problem you have.  What I don't follow however is
> why the function needs to be deferred.  It seems to me that it will be
> output anyway as long as it is referenced but I may be wrong (this is
> all bit nasty), so I would like to understand it better.

Those functions get stripped away from the output as long as you either
remove the static qualifier, reference them directly or compile without
-Os/-O2.

> I see we got suck here somewhat,
> I will try to search for archives first on why that line got in at first
> place.

cvs annotate -> ChangeLog -> gcc-patches

This path might lead to the truth ;-)

-- 
  // Bernardo Innocenti - Develer S.r.l., R&D dept.
\X/  http://www.develer.com/

Please don't send Word attachments - http://www.gnu.org/philosophy/no-word-attachments.html