On (01/03/11 21:26), Arjan van de Ven wrote:
> On 1/3/2011 8:07 PM, Andreas Mohr wrote:
> >Hi,
> >
> >On Mon, Jan 03, 2011 at 05:37:45PM +0200, Sergey Senozhatsky wrote:
> >>On (01/02/11 20:35), Arjan van de Ven wrote:
> >>>On 1/2/2011 1:21 PM, Andreas Mohr wrote:
> >>>>Nice stuff, except perhaps for some network weirdness:
> >>>>
> >>>>    Good          Wake-on-lan status for device eth0
> >>>>
> >>>>Well.... I don't even _have_ eth0. Even ifconfig -a lists eth1 only.
> >>>>Unless it's that darn device renaming (it probably is).
> >>>>In which case powertop had better adopt device renaming support, too
> >>>>(you wouldn't want to display names of interfaces which don't actually
> >>>>exist on the user side, right?).
> >>>ok will check how to resolve this.. does make sense to only show
> >>>those that exist.
> >>>
> >>Could you please try the following patch?
> >>
> >>Though, we already read net dir in create_all_nics and instead we can export
> >>`static map<string, class network *>  nics'
> >>and just iterate through the map to fill vector of tunables net device.
> >Thanks, that appears to have improved it (no intermediate git changes
> >available -->  only custom patch applied):
> >    Good          Wake-on-lan status for device eth1
> 
> thanks for testing; I'll stick Sergey's patch in
> 

Thanks.


>Minor comment: myself I'd probably decide to not use
>the non-throwing operator new and instead class-encapsulate
>the opendir()/closedir() pair.

Encapsulate open/close and let C++ to closedir while destructing 
stack variable is something we probably should have. Thanks.


        Sergey

Attachment: pgpL1lNn1gFky.pgp
Description: PGP signature

_______________________________________________
Power mailing list
[email protected]
http://www.bughost.org/mailman/listinfo/power

Reply via email to