http://www.research.ibm.com/trl/projects/security/ssp/

On Mon, 2003-11-17 at 13:37, Peter S. Mazinger wrote:
> On 17 Nov 2003, Ned Ludd wrote:
> 
> > On Sun, 2003-11-16 at 19:46, Peter S. Mazinger wrote:
> > > On 12 Nov 2003, Ned Ludd wrote:
> > > 
> > > > On Wed, 2003-11-12 at 16:31, Peter S. Mazinger wrote:
> > > > > There are problems building some of the binaries with propolice enabled 
> > > > > gcc, mainly the .hidden support in binutils has to be "hidden" from gcc, 
> > > > > but as I can see (read), the glibc version does not work flawlessly 
> > > > > either.
> > > > > 
> > > > Have you successfully used ssp with uclibc?
> > > It worked (somehow, not quite correct, and some packages where not 
> > > rebuildable) at the time where uClibc had support for LIBGCC_FUNCTIONS 
> > > (adding functions from libgcc to libc) This was the only way I got a 
> > > system running it (__guard and __smash... where added to libc). Since then 
> > > I cannot modify buildroot so that I get a system running, segfaults on all 
> > > the line (tested only with gcc-3.3.x and protector 3.3-5). I think the 
> > > solution would be to add the needed functions to uClibc (and remove them 
> > > from libgcc!!!).
> > 
> > One of the hardened-gcc version introduced guard symbols to libgcc.a.
> > I'm pretty sure this no longer should be the case if your running ~arch
> > as they are in fact getting introduced into glibc. 
> > If you care to hack the functionaly into uClibc I'll point you at the
> > glibc version from etoh http://dev.gentoo.org/~solar/ssp/guard.c
> Where do I find the modified protector patch for gcc?
> 
> Peter
-- 
Ned Ludd <solar@g.o>
Gentoo Linux Developer