Re: [PATCH 3/4] add ksm kernel shared memory driver.

[Chris Wright <chrisw@redhat.com>]

* Alan Cox (alan@lxorguk.ukuu.org.uk) wrote:
> On Tue, 2 Dec 2008 10:07:24 -0800
> Chris Wright <chrisw@redhat.com> wrote:
> > * Alan Cox (alan@lxorguk.ukuu.org.uk) wrote:
> > > > +	r = !memcmp(old_digest, sha1_item->sha1val, SHA1_DIGEST_SIZE);
> > > > +	mutex_unlock(&sha1_lock);
> > > > +	if (r) {
> > > > +		char *old_addr, *new_addr;
> > > > +		old_addr = kmap_atomic(oldpage, KM_USER0);
> > > > +		new_addr = kmap_atomic(newpage, KM_USER1);
> > > > +		r = !memcmp(old_addr+PAGEHASH_LEN, new_addr+PAGEHASH_LEN,
> > > > +			    PAGE_SIZE-PAGEHASH_LEN);
> > > 
> > > NAK - this isn't guaranteed to be robust so you could end up merging
> > > different pages one provided by a malicious attacker.
> > 
> > I presume you're referring to the digest comparison.  While there's
> > theoretical concern of hash collision, it's mitigated by hmac(sha1)
> > so the attacker can't brute force for known collisions.
> 
> Using current known techniques. A random collision is just as bad news.

And, just to clarify, your concern would extend to any digest based
comparison?  Or are you specifically concerned about sha1?