On 26/07/12 12:42, Mark Brown wrote:
On Thu, Jul 26, 2012 at 12:38:17PM +0100, Lee Jones wrote:
On 26/07/12 12:32, Mark Brown wrote:
Again, this makes no sense. If we're explicitly being asked to use regmap then we should be using regmap or just failing to set up I/O (which is obviously a catastrophic failure).
How much work is there involved in regmap:ing a device, so that dev_get_regmap() doesn't fail?
Trivial if it's on a supported bus, otherwise you just need to write the bus. But why do you care if dev_get_regmap() fails? We only try to use regmap if the driver asked for regmap I/O (or doesn't have registers at all in which case it doesn't matter since we never do any I/O). What you appear to be saying here is that you're using regmap on a device which doesn't have a regmap set up which is clearly never going to work terribly well...
I don't think we want to use regmap at all, but we're forced to by soc-core. How do we over-ride that behavior? By writing some nonsense into codec->control_data?