Quoting Yong Wu (2022-01-14 23:39:52)
On Fri, 2022-01-14 at 15:30 -0600, Stephen Boyd wrote:
This sounds more correct to me. I'm not an expert on runtime PM though as I always have to read the code to remember how it works. if the device isn't ready for runtime PM until the component bind function is called then runtime PM shouldn't be enabled on the component device.
Anyway, We should update the SMI driver for this case. I prepare a patch into this patchset or I send it independently? which way is better?
I can roll it into this patch. It needs to be combined otherwise it breaks the bisectability of the series.