On Mon, Feb 21, 2000 at 11:32:11AM +1300, Chris Wedgwood wrote:
> On Sun, Feb 20, 2000 at 06:23:50AM +0100, Ralf Baechle wrote:
>
> > Doesn't make sense, most CPU's don't make a difference between read
> > and execute permissions for mappings or at least like the m68k don't
> > use their hardware capability to do so under Linux.
>
> Surely this is a m68k-linux bug -- not a linux bug?
It's not a bug but intentional. With the 68851 and 68030 builtin PMUs
you can only only distinguish between read and execute permission if
you either use an additional FCL lookup which adds an additional leve
to the page table tree or you use long descriptors which need twice as
much memory. The later 68k series PMUs don't even support that
feature in hardware apparently because nobody used it.
(Beware - I stopped hacking Linux/68k for the sake of the MIPS port in
'93 so my memory is somewhat rusty ...)
It's also not a bug because it is consistent with all UNIX specifications.
> I know with x86 you don't have the ability to allow read and prevent
> exec, so it could be a hang-over from there?
You could play games combining segments and MMU features to achive that.
See Solar's no-exec stack patch.
Ralf
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Feb 23 2000 - 21:00:26 EST