On Mon, Oct 31, 2022 at 07:50:02AM -0600, Jens Axboe wrote: > > list_add_tail(&e->list, &elv_list); > > spin_unlock(&elv_list_lock); > > What's the idea behind this? Yes it'll be harmless and list ordering > will dictate which one will be found, leaving the other(s) dead, but why > not catch this upfront? I agree likelihood of this ever happening to be > tiny, but seems like a good idea to catch and return BUSY for this case. Because it's just not very useful code bloat here that I stumbled upon. But I can just drop it if you prefer.
