On Wed, Oct 4, 2023 at 10:12 AM Bartosz Golaszewski <brgl@xxxxxxxx> wrote: > What is your view on Andy's and Kent's issues with the _new() name > suffix? We have done similar operations in the past, and it is similar to what Uwe is doing for the moment with the .remove() callbacks. Usually the strategy is employed when the work needs to be spread out over a few merge windows so it is a bit of a marker that "this is in transition". There is the horror story of this staying around forever and becoming idiomatic: struct napi_struct (include/linux/netdevice.h) where "napi" means "new API" - yeah that could have been handled better... If there is more moaning about it I will simply squash all the patches into one and call it a day - the end result will be the same and no sign of any *_new suffix anywhere. It was still worth it for reviewing the driver changes on a per-driver basis so then it becomes one of those Schopenhauer ladders that you can toss away after climbing it. Yours, Linus Walleij