Messages in this thread | | | Date | Mon, 14 Apr 1997 17:34:30 +0200 (MET DST) | From | Ingo Molnar <> | Subject | Re: Linux kernel patch to remove stack exec permission |
| |
On Mon, 14 Apr 1997, Zefram wrote:
> Ingo Molnar wrote: > >would it be a good idea to strip off the highest bit in env[] and args[] > >when exec()-ing? > > Don't be silly. Some of us use full Latin-1 or UTF-8 for filenames, > and in grep patterns, and so on.
some of us like me ;) *blush*. And i was told that it's possible to build pure 7-bit ASCII based programs as well ...
Plus envp[] and argv[] are right above the stack, so the original patch already covers this case.
> >for the BSS/malloc() things we could theoretically get the kernel to put > >executable mmap()-ed areas into the 0-1G range, and the rest into the > >1G-2G range. > > This is the Right Thing to do w.r.t. trampolines. Fancy hacking that > into gcc?
actually i ment doing this on the kernel side, by sorting executable and nonexecutable mmap()-ed areas into segments ... not too clean, but it would prevent passing executable code via data channels >at all<.
80% of the more elusive security holes do rely on getting some code into a protected execution context somehow?
> Then again, it might be easier just to stop writing programs that smash > the stack. Amazingly enough, this confers other benefits too.
it's about decreasing certain risks. I would love having 10 security holes per 1 million lines of code, instead of hundreds.
-- mingo
| |