Re: [patch, rfc] Semantics of attribute (aligned), misscompilation of crtstuff

[Zdenek Dvorak <rakdver at kam dot mff dot cuni dot cz>]

Hi,

> > as discussed in http://gcc.gnu.org/ml/gcc-patches/2007-10/msg00771.html,
> > aligned(sizeof(func_ptr)) on variables specifies only that the alignment
> > must be at least that of sizeof(func_ptr), so crtstuff should not
> > rely on that the alignment of such a variable is not increased above
> > this bound.
> > 
> > This patch makes crtstuff work even if the alignment is increased, which
> > makes gcc bootstrap with vectorization enabled on ppc64-linux.  I have
> > also bootstrapped & regtested the patch on i686.
> 
> How about __attribute__((aligned(sizeof(func_ptr)),packed)) instead of
> the runtime test?

do we support attribute packed on variables? I get

../../gcc/gcc/crtstuff.c:196: warning: âpackedâ attribute ignored
../../gcc/gcc/crtstuff.c:209: warning: âpackedâ attribute ignored

when I try to use it.

Zdenek