On 5/14/23 13:18, Takashi Iwai wrote:
The idea is interesting, and it's a definitely good thing to have.
I wonder, though, whether it could be better provided as an extention to the existing snd-dummy driver. The advantage of extending snd-dummy driver would be that it already supports different formats, etc. OTOH, if we create an individual driver, the pro side is the simpleness of the code.
I'm inclined to go with a separate driver, but I'm open about this. Maybe Jaroslav and Mark have some opinions?
About this patch set: the driver name should be a bit more specific, as this isn't a generic virtual driver that is used for general purpose but it's only for testing. And it's only for testing PCM. So, a name like snd-test-pcm would be more appropriate, IMO.
And, we want the coverage of different formats, channels, rates and accesses (interleaved vs non-interleaved). How can we extend somehow more for that?
thanks,
Takashi
Hello Takashi! Thank you for your reply. I fully agree with the naming issue, and I'll change it in the future versions of the patch set in case we choose to have it as a separate driver. I also prefer this option because in my opinion the use cases of these drivers are a little bit different. Also, I believe I can extend the driver to support different formats, channels and accesses in the near future.
Additionally, implementing these changes would be a perfect task for the end of the Linux Kernel Mentorship program I'm going through :) However, I'm open to other views on this, and I'm ready to move the functionality from my driver to the snd-dummy in case we prefer this option.
Thanks again for considering my changes!
Kind regards, Ivan Orlov.