Re: [RFC][PATCH v3 1/3] runtime interpreted power sequences

On Wed, Aug 01, 2012 at 02:55:31PM +0100, Mark Brown wrote:

On Wed, Aug 01, 2012 at 03:38:14PM +0200, Thierry Reding wrote:

quoted

On Wed, Aug 01, 2012 at 02:26:51PM +0100, Mark Brown wrote:

quoted

quoted

This is why __devinit data will only be discarded when this is not
possible.

quoted

That's exactly my point. But I seem to have miserably failed to get that
across. =)

We must be talking at cross purposes then.  What I'm saying is that the
framework shouldn't rely on platform data and should assume that the
kernel might be configured so it can be discarded (by copying most
likely).

Yes. I think this should be solved by the power_seq_build() function
which uses the platform data description of the sequence and builds it
into something for internal use that doesn't rely on anything in the
platform data.

Thierry