On Wed, Oct 09, 2013 at 03:59:17PM +0100, Mark Brown wrote:
On Wed, Oct 09, 2013 at 04:12:42PM +0200, Markus Pargmann wrote:
On Wed, Oct 09, 2013 at 02:20:04PM +0100, Mark Brown wrote:
I'm not sure what you mean here. How would the error be handled without calling _STOP?
The START error should be handled in mxs_saif and there shouldn't be a _STOP call for mxs_saif_trigger afterwards to handle the failed START.
How would the framework know that the error has been handled - what tells it that the error returned is for an error that has been handled as opposed to an error that has not been handled?
I finally understand what you mean. Yes, RFCv1 fixes the behavior for the components without _START failures. But it doesn't fix it for the failing component.
For mxs-saif, we still have to store if the last START failed to handle the following STOP correctly which is not very different from handling double STOPs correctly. I don't have an idea how to fix RFCv1 without using some state as in RFCv2. Perhaps we should simply fix it in mxs-saif and other drivers then?
Regards,
Markus