On Mon, Jul 07, 2025 at 11:35:32AM -0700, Bart Van Assche wrote: > On 7/3/25 2:09 AM, Christoph Hellwig wrote: >> I'm still very doubtful on this whole approach, and I think you are >> ignoring the root cause, which is dm-multipath keeping a q_usage_count >> reference for an unbounded time. It is only supposed to be held over >> I/O, and I/O is expected to time out. > > No. Since queue_if_no_path was introduced, it can queue I/O > indefinitely. The oldest reference I could find to the queue_if_no_path > implementation is from 20 years ago: Thank you ѕo much Bart, without your help I would not be able to git-blame. That still doesn't make it sensible to keep the q usage counter elevator for unlimited time. See nvme multipath for how we can keep bios around forever without elevating the usage counter which is supposed to be transient. Note that dm-multipath should in fact already be doing the right thing in bio based mode as well.
