Re: [RFC][PATCH] linux-2.6.2-rc2_vsyscall-gtod_B1.patch

[Andrea Arcangeli <andrea@suse.de>]

On Mon, Feb 02, 2004 at 08:35:40PM -0800, Ulrich Drepper wrote:
> Andrea Arcangeli wrote:
> > I definitely call it a great success,
> 
> You got to be kidding.  Some object fixed in the address space which can
> perform system calls.  Nothing is more welcome to somebody trying to
> exploit some bugs.
> 
> The vdso must be randomized.  This is completely impossible with this
> stupid fixed address scheme and it must be changed as soon as possible.

sorry, no idea what you're talking about. I can't see any valid single
reason to randomize the addresses. (the only effect is that it will hurt
performance)

Whatever problem you found, feel free to post an exploit so I will
certainly be able to understand your problem, if you can't to me it
means there's no problem.

the closer thing that your statements reminds me, is the discussion
about the reentrancy of the gettimeofday, basically to allow
virtualization, if that's what you meant that can be addressed just fine
with a modification to the ptes with a syscall, no valid reason to
slowdown the production fast path with an inefficient API just for the
re-virtualization of the vsyscalls.