On Wed, May 20, 2020 at 12:02:15AM -0700, Jeff Kirsher wrote: > This series contains the initial implementation of the Virtual Bus, > virtbus_device, virtbus_driver, updates to 'ice' and 'i40e' to use the new > Virtual Bus. > > The primary purpose of the Virtual bus is to put devices on it and hook the > devices up to drivers. This will allow drivers, like the RDMA drivers, to > hook up to devices via this Virtual bus. > > The associated irdma driver designed to use this new interface, is still > in RFC currently and was sent in a separate series. The latest RFC > series follows this series, named "Intel RDMA Driver Updates 2020-05-19". > > This series currently builds against net-next tree. > > Revision history: > v2: Made changes based on community feedback, like Pierre-Louis's and > Jason's comments to update virtual bus interface. > v3: Updated the virtual bus interface based on feedback from Jason and > Greg KH. Also updated the initial ice driver patch to handle the > virtual bus changes and changes requested by Jason and Greg KH. > v4: Updated the kernel documentation based on feedback from Greg KH. > Also added PM interface updates to satisfy the sound driver > requirements. Added the sound driver changes that makes use of the > virtual bus. Why didn't you change patch 2 like I asked you to? And I still have no idea why you all are not using the virtual bus in the "ice" driver implementation. Why is it even there if you don't need it? I thought that was the whole reason you wrote this code, not for the sound drivers. How can you get away with just using a virtual device but not the bus? What does that help out with? What "bus" do those devices belong to? Again, please fix up patch 2 to only add virtual device/bus support to, right now it is just too much of a mess with all of the other functionality you are adding in there to be able to determine if you are using the new api correctly. And again, didn't I ask for this last time? greg k-h
