On Fri, Jan 21, 2022 at 9:53 PM Rafael J. Wysocki rafael@kernel.org wrote:
On Wed, Jan 19, 2022 at 6:33 PM Lucas tanure tanureal@opensource.cirrus.com wrote:
On 1/19/22 16:53, Rafael J. Wysocki wrote:
On Tue, Jan 18, 2022 at 3:53 PM Stefan Binding sbinding@opensource.cirrus.com wrote:
...
Why are you moving it away from platform/x86?
Adding SPI to the mix doesn't seem to be a sufficient reason.
If this were going to be needed on non-x86, that would be a good reason for moving it, but is that actually the case? If so, why isn't that mentioned in the changelog above?
It was a request made by Andy Shevchenko: https://lkml.org/lkml/2021/12/3/347
But he hasn't given any reasons why that'd be better.
My thoughts were that these are related to ACPI handling the serial buses in one place. However, counter arguments might be that the cases of the resources like this are found only on x86 hardware (while ACPI should be agnostic to that) and that the i2c and spi already do ACPI stuff on their own. That said, there are pros and cons and I'm fine with either choice at the end of the day.