On 30/09/20 16:27, Maxim Levitsky wrote: > My patch that switches the direction in scsi_device_find, is supposed to be > completely equavalent, > based on the following train of thought: > > If scsi_device_find finds an exact match it returns only it, as before. > > Otherwise scsi_device_find were to scan from end of the list to the start, > and every time, > it finds a device with same channel/id it would update the target_dev > and return it when it reaches the end of the list. > > If I am not mistaken this means that it would return _first_ device in the > list that matches the channel/id. > This is exactly what new version of scsi_device_find does.
Oh! I missed that subtlety. Thanks, that makes sense. Paolo
