On Fri, May 12, 2023 at 04:12:42PM -0700, Christoph Hellwig wrote: > On Fri, May 12, 2023 at 03:31:39PM -0600, Jens Axboe wrote: > > That's fine for telling error handling, probing etc apart from "normal" > > IO, but it doesn't solve the problem at hand which is assuming that > > anything inserted into the scheduler has a bio. > > I think this is fine. Passthrough requests should never go to the > scheduler, that's not where they blong. Non-passthrough request always > have a bio. .. except for the flushes generated by block/blk-flush.c. Those do not enter the I/O schedule either. That being said saying every request has to have a bio might be a good thing some day. Totally indepent of the fact that schedulers have no business touching passthrough requests.
