> > > 
> > 
> >   I believe that the primary reason for privilege is that it extracts the
> > tables via /dev/mem.
> 
> Seriously ?

  The lesswatts version does (which was the last version I used), but 
there is an indication that the one we bring in will use /dev/xsvc.

> 
> Is there no other way to get this ?

  Not with this case (at least as far as I know).

> 
> If that is the case then I have some very serious concerns that this case
> raises.  Some of the security related but more about the stablity of the
> interfaces this case is using to get the data if it is reading memory
> directly.  Surely we have a better way to get this data, and ideally given it
> is monitoring data we don't want such massive amount of privilege to read it.

  Certainly, an acpi driver could be written to allow mapping, 
accessing, or managing this space, but I don't see this occuring any 
time soon.

> 
> Then again maybe this is in the same category as iasl which (ab)uses /dev/xsvc
> for a similar purpose.

  Pretty much.  These tools will primarily be used for debugging, so I 
see less of an issue.  But the primary maintainer is on the interest 
list, it is possible that he has some alternate ideas.


        ---- Randy


Reply via email to