Re: [rfc] increase struct page size?!

[Matthew Wilcox <matthew@wil.cx>]

On Sat, May 19, 2007 at 10:53:20AM -0700, William Lee Irwin III wrote:
> On Fri, May 18, 2007 at 04:42:10PM +0100, Hugh Dickins wrote:
> > Sooner rather than later, don't we need those 8 bytes to expand from
> > atomic_t to atomic64_t _count and _mapcount?  Not that we really need
> > all 64 bits of both, but I don't know how to work atomically with less.
> > (Why do I have this sneaking feeling that you're actually wanting
> > to stick something into the lower bits of page->virtual?)
> 
> I wonder how close we get to overflow on ->_mapcount and ->_count.
> (untested/uncompiled).

I think the problem is that an attacker can deliberately overflow
->_count, not that it can happen innocuously.  By mmaping, say, the page
of libc that contains memcpy() several million times, and forking
enough, can't you make ->_mapcount hit 0?  I'm not a VM guy, I just
vaguely remember people talking about this before.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>