openssh static build - mission impossible?

Peter Stuge peter at stuge.se
Tue Mar 6 18:52:48 EST 2012


Mr Dash Four wrote:
> > I would expect armv6l-redhat-linux-gnueabi-gcc to be run here,
> > not gcc.
> 
> Nope! The compiler is quite capable of producing multi-arch
> binaries. This particular feature is called "multilib", not
> sure if you've heard about it, but it does exist.

I know multilib, but..


> >> I am cross-compiling from armv7l to armv6l as evident from the
> >> initial configure. The actual library - /usr/lib/libcrypt.a - is
> >> compiled and build for armv5tel (Fedora's own latest "stable").
> > 
> > Except if you are using the armv5tel gcc and libcrypt.a you are not
> > actually targeting armv6l anymore.
> 
> As I already mentioned above my gcc compiler is multilib-capable -

..multilib or no, the libcrypt.a being used is from /usr/lib and
since you say that this means armv7l then I don't think that is a
library you want to use on the armv6l target. The compiler used
having yet another tuple (armv5tel) doesn't reduce potential
problems and question marks.


> > That you get to study on your own. I will not learn about it in order
> > to teach you. You can find out easily by studying the OpenSSH code.
> 
> If you are going to act like a condescending prick, unwilling or
> incapable of offering anything of interest to this topic, then I'd
> suggest you move along - there's nothing to see here.

I've pointed to multiple issues worth investigating, and I've
explained one certain and one further possible high level solution
to your problem even if you choose to completely ignore the issues
worth investigating.

You can't expect anyone to provide you a tailor made solution that
you will just deploy. You have to do the work yourself. I'm sorry
if that seems condescending. I think it's just common sense.


//Peter


More information about the openssh-unix-dev mailing list