More on the armv4l segfault.

[Joakim Tjernlund]

> -----Original Message-----
> From: Rob Landley [mailto:rob at landley.net] 
> Sent: den 30 december 2006 05:44
> To: Joakim Tjernlund
> Cc: 'Paul Brook'; uclibc at uclibc.org
> Subject: Re: More on the armv4l segfault.
> 
> On Thursday 28 December 2006 6:30 pm, Joakim Tjernlund wrote:
> > Rob, found this link that might explain your problem:
> >  http://sourceware.org/ml/binutils/2005-10/msg00107.html
> 
> Ah, cool.
> 
> So my problem is that if there are _no_ non-weak references 
> to a symbol, the 
> linker won't pull that symbol in at all?

I think so, but it is only a guess. readelf -a on the binary
with and without your patch should tell us.
Does removing the weak_function on any other prototype work?

> 
> That kind of sucks, but also explains why my patch worked.  
> Any suggestions of 
> a better way to fix this in uClibc?

Sadly no, maybe some #ifdef STATIC trick to force a non weak reference.

 Jocke

> 
> >  Jocke
> 
> Rob