Re: don't let mmap allocate down to zero

[Mikael Pettersson]

William Lee Irwin III writes:
 > William Lee Irwin III writes:
 > >> There's a long discussion here, in which no one appears to have noticed
 > >> that SHLIB_BASE does not exist in mainline. Is anyone else awake here?
 > 
 > On Thu, Jan 27, 2005 at 10:29:12AM +0100, Mikael Pettersson wrote:
 > > About the only kernel-level enforcement I would feel comfortable with is
 > > to have non-fixed mmap()s refuse to grab the _page_ at address 0. Any range
 > > larger than this is policy, and hence needs a user-space override mechanism.
 > > Also, if user-space wants to catch accesses in a larger region above 0 then
 > > it can do that itself, by checking the result of mmap(), or by doing a fixed
 > > mmap() at address 0 with suitable size and rwx protection disabled.
 > 
 > FIRST_USER_PGD_NR is a matter of killing the entire box dead where it
 > exists, not any kind of process' preference. Userspace should be
 > prevented from setting up vmas below FIRST_USER_PGD_NR. It has zero to
 > do with what userspace's own concerns, but rather the kernel trying to
 > avoid system-critical data from being stomped on by userspace. It would
 > be like accidentally allowing userspace to use the IDT for malloc() on
 > x86, destroying one's ability to handle interrupts, page faults, etc.,
 > to allow userspace to go below FIRST_USER_PGD_NR on ARM.

My argument only applied to the "protect the user" discussion.

_If_ the kernel needs to reserve parts of the user's address
space for IDTs etc, then certainly mmap() mustn't map those.
But that's orthogonal (in spirit, if not in implementation)
from the "protect the user" discussion.

/Mkael
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/