Re: [PATCH 3/6] ksm: change the KSM_REMOVE_MEMORY_REGION ioctl.

[Hugh Dickins <hugh@veritas.com>]

On Wed, 6 May 2009, Chris Wright wrote:
> * Hugh Dickins (hugh@veritas.com) wrote:
> > 
> > Is the phrase "covert channel" going to come up somehow?
> 
> There's two (still hand wavy) conerns I see there.  First is the security
> implication: timing writes to see cow and guess the shared data for
> another apps VM_LOCKED region,

Mmm, yes, there's fun to be had there; though I don't see it as having
anything to do with VM_LOCKED, beyond that the paranoid have reason to
place their most anxious data in VM_LOCKED areas.

I'm thinking of an app which prepares pages full of scurrilous rumour,
then waits around looking at its /proc/self/smaps to see if anyone else
is writing stories like that!

> second is just plain old complaints of
> those rt latency sensitive apps that somehow have VM_LOCKED|VM_MERGE
> and complain of COW fault time, probably just "don't do that".

Right.  There are sensitive sites which ought not to configure such
merging on; but I don't think we should disallow merging locked.

Hugh

--
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>