[linux-yocto] [PATCH] meta: add features/rfkill

Darren Hart dvhart at linux.intel.com
Wed Aug 21 13:20:20 PDT 2013


On Wed, 2013-08-21 at 16:12 -0400, Bruce Ashfield wrote:
> On 13-08-21 01:21 PM, Darren Hart wrote:
> > On Wed, 2013-08-21 at 17:20 +0100, Ross Burton wrote:
> >> Add standalone CONFIG_RFKILL feature for cross-BSP/driver use.
> >>
> >
> > I try to be cautious when it comes to single line CONFIG fragments (you
> > can see as how this doesn't scale very well).
> >
> > How is this going to be used? Is the need broad enough to justify the
> > overhead in the kernel meta-data rather than a single norfkill.cfg in a
> > BSP layer recipe space?
> 
> My suggestion is this, turn it on for the "standard" kernel type, which
> is where we collect our policy for typical boards. Yes, this sort of has
> a hardware binding, but it isn't that strong.
> 
> We turn on CONFIG_NET in the base configuration already, and netfilter
> and friends in the standard kernel configuration. So adding this into
> the standard.cfg file would be acceptable as a general policy config
> change.
> 
> The standard kernel config wouldn't impact the tiny kernel config, since
> it doesn't inherit it, and that's what Darren would want. preempt-rt would
> need to have it turned on as well .. which leads me to my next statement:
> 
> That being said, I *still* think it is ok in the features/ subdirectory,
> since we don't typically trigger them from recipes directly, but instead
> include them from other configuration fragments. preempt-rt and the
> standard kernel would include it, and it would be clear that they are
> enabling the option. What I wouldn't want to see is a BSP or set of
> BSPs needing to enable it explicitly via KERNEL_FEATURES from their
> layers.
> 

Happy to defer to you on this one. Pull it.

-- 
Darren Hart
Intel Open Source Technology Center
Yocto Project - Linux Kernel





More information about the linux-yocto mailing list