Hello Jonas, On Mon, Jun 03, 2024 at 11:58:07AM +0200, Jonas Karlman wrote: > Some initial quick feedback is that the tcpm uclass and driver should > probably take more advantage of the u-boot driver model. > > Few quick thoughts: > - Split out uclass/driver api parts into tcpm-uclass.c > - Take advantage of uclass_driver ops such as post_probe and > per_device_auto/per_device_plat_auto to e.g. handle tcpm_port_init and > tcpm_port/tcpm_dev > - udevice should probably be passed to tcpm_ functions instead of a > tcpm_port, dev_get_uclass_plat/priv could possible be used to get > the tcpm_port data for the udevice > > I have not yet had time to runtime test this but will do and get back > with a more in depth review later this week.
Thanks for the quick thoughts. They do sound very sensible, so I gave them a go. The diff turned out to be quite big (about 20% of all lines changed). It's currently one big cleanup commit, which I still need to split and merge into the tcpm and fusb302 commits. After doing that and some more testing on my side I will send PATCHv2. I expect this to happen within the next 48h, so it might be sensible to wait with further testing / review until that happened. Greetings, -- Sebastian
signature.asc
Description: PGP signature
