> > We should not fix the problem in the case where it _can_ be solved,
> > just because there are cases where it cannot?
>
> Exactly. Of course, everybody is free to use that patch, if she things
> that there is an advantage in using it. However, it would be very bad
> for Linux if the patch was included and it became public knowledge that
> Linux is protected against executable-stack attacks. If later somebody
> finds the work-around (as somebody certainly will), it will be a PR
> desaster.
Finding bugs ever few days is far less encouraging!
> There are much better ways of fighting these kinds of attacks. About
> the best way is to fix the applications that risk buffer overflows.
> glibc 2 does a very good job in this: it produces linker warnings for
> unsafe functions. So every time somebody installs such a package from
> the sources, she will get a linker warning, until she finally does
> something about it. Works much better in the long run, IMHO.
This is like saying protected memory isn't needed, you better debug the
apps to not overwrite innocent memory!
If the patch is a CONFIG option then everybody can decide for themselves.
Karl.
// Electronic Mail - SMTP: kvo@seagha.com
\X/ - X400: c=BE; a=RTT; p=SEAGHA; s=VOGEL; g=KARL