On 3/30/20 11:10 AM, Amadeusz Sławiński wrote:
Sounds good, but this happens in a loop, so would all the memory previously allocated by denum/dbytes/dmixer_create leak, or is it freed automatically somewhere else?
Well, now that error is propagated, snd_soc_tplg_component_remove() should be called by snd_soc_tplg_component_load() in case of errors while parsing. From quick look it seems like it should be able to free it up correctly by calling remove_enum/bytes/mixer.
I am not sure what you meant by 'should be called', if it's a recommendation for a future change or a description of the existing behavior. Just to be clear, are you saying the existing code will take care of this error flow or that a new patch is needed?
Existing code should handle this properly. No new code is needed.
Sounds good, thanks.