Le 19/05/2017 ? 08:56, Romain Perier a ?crit : > Hello, > > > Le 16/05/2017 ? 13:18, Mark Brown a ?crit : >> On Mon, May 15, 2017 at 04:41:38PM -0500, Rob Herring wrote: >>> On Fri, May 12, 2017 at 03:22:25PM +0200, Romain Perier wrote: >>>> This commit adds a compatible string for everest,es8323. This is an >>>> audio codec that is compatible with es8328 and can be found for example >>>> on the Firefly-RK3288 board. >>> If it is compatible with the es8328, then that should be a fallback and >>> you don't need the driver change. >> While people don't strictly need the driver change it doesn't do any >> harm either and encourages people to get the information into the DT >> that it's a different chip. Thinking about it it might be good to have >> a way for something to validate if fallback compatibles are being listed >> when they should - the drivers could provide the information fairly >> easily I guess, or it could go into the binding docs once we have a >> schema format. >> >> Even if it doesn't currently make a difference to software I'd rather >> get the information in there for mixed signal devices like audio CODECs >> - it's not that unknown to find later that supposedly register identical >> chips have some differences in the analog which we want to care about. > So, what should I do for this patch, finally ? fallback or driver change ? > > Thanks, > Romain To be honest, the doc is not really clear about this, both codecs seem compatible but I think that's preferable to have a driver change for this, just in case (Suppose that we discover a difference later...) Romain
