Hi, On Mon, Jun 25, 2012 at 04:41:36PM +0530, Shubhrajyoti D wrote: > Currently i2c register restore is done always. > Adding conditional restore. > The i2c register restore is done only if the context is lost > or in case of error to be on the safe side. > > Cc: Kevin Hilman <khilman@xxxxxx> > Signed-off-by: Shubhrajyoti D <shubhrajyoti@xxxxxx> how will this ever work with DT ? I say we get rid of the OMAP-specific API and build this "context lost" status directly on dev_pm_info and have something like pm_runtime_lost_context() or something with pm QoS tell you if a device has lost its context. Also, your commit log doesn't really state any problems you might have reached before, or any improvements wrt latency coming out of suspend and so on. IMHO, drivers need a generic way to differentiate if they're resuming from OFF or RET, otherwise we will end up with a bunch of OMAP-specific hackery on all drivers -- balbi
Attachment:
signature.asc
Description: Digital signature
