On Fri, 10 Jun 2011 12:59:54 +0100 Mark Brown broonie@opensource.wolfsonmicro.com wrote:
On Fri, Jun 10, 2011 at 02:56:10PM +0300, Jarkko Nikula wrote:
I was thinking to return -EBUSY here but went thinking is it really worth of trouble? Machine drivers are calling this from their hw_params callback with the same configuration always (I suppose) for both playback and capture and it sounded kind of overkill to check what is existing configuration or skip the -EBUSY error.
One fairly common case is selecting the system clock rate to a fixed multiple of the sample rate so applications that try to set up different playback and record rates trigger issues.
Umm, yeah. I'll cook another version which returns -EBUSY but still trying to keep this regression fix minimal. Anyway it's only pandora that can trigger the regression in mainline.