On Fri, 29 Apr 2011 06:43:41 -0700
Arjan van de Ven <[email protected]> wrote:

> On 4/29/2011 3:33 AM, Alan Cox wrote:
> > On Thu, 28 Apr 2011 15:07:00 -0700
> > Kristen Carlson Accardi<[email protected]>  wrote:
> >
> >> Despite the fact that these shim devices aren't being used,
> >> touching them causes bad things to happen unless you have
> >> a debug port connected.
> > Seems reasonable  - I take it other kinds of touching them don't
> > explode, eg if the kernel decides to play with them later or if
> > lspci touches them ?
> >
> > If not then probably the quirk should be revised to 'vanish' them
> > entirely.
> 
> touching the actual controller will make things go boom
> 
> merely looking at (or modifying) config space will not (since the
> config space for this guy is just a bit of dram)
> 

Re-read that code - its making a pci request to touch a config
register. The fact we do magic behinds its back on Moorestown is
neither here nor there. So your assertion isn't quite true it seems.

Maybe the PMU code should bulletproof this so we don't get surprises
elsewhere ?
_______________________________________________
MeeGo-kernel mailing list
[email protected]
http://lists.meego.com/listinfo/meego-kernel

Reply via email to