Re: Improving Ada's image - Was: 7E7 Flight Controls Electronics

[Hyman Rosen <hyrosen@mail.com>]

Warren W. Gay VE3WWG wrote:
> Today's slashdot points out yet another exploit in Linux.

I believe the link is actually
<http://slashdot.org/article.pl?sid=04/06/14/118209&mode=flat&tid=106&tid=126&tid=128&tid=185&tid=190>.

> We like to deride Windows, but Linux, *NIX in general all suffer from
> the same problem. Some of these are no doubt due to design issues
> (which Ada may or may not help with). But surely, one has to wonder/hope
> if/that some other fundamental change (language) can offer an
> improvement. I think many are simply saying that they believe it
> to be so.

Which illustrates that you did not actually bother to read about
what the problem is. The exploit code (which crashes the kernel,
so it's not precisely a security violation) involves user code
which executes a floating-point processor instruction in a signal
handler that should cause a floating-point exception, immediately
followed by a system call. I think that the hardware exception
occurs a little bit later, such that kernel mode has taken over,
and when the kernel goes to restore the floating-point state, the
exception finally kicks in and brings it down.

I suspect the ultimnate fix involves the kernel checking for
deferred user-mode floating point exceptions so that it can reraise
them in user mode, but in any case, this has nothing to do with the
programming language in which the kernel was written - at the point
of death, it is executing an assembly code insertion.