On Fri, Jun 01, 2018 at 10:31:13PM -0600, Jens Axboe wrote: > On 6/1/18 8:36 PM, Ming Lei wrote: > >> clear the bit when we need to, potentially even adding a debug check to > >> blk_queue_enter_live() that complains if the ref was not already > >> elevated. Though that would be expensive, compared to the percpu inc > >> now. Not saying the bit is necessarily the best way forward, but I do > >> like it a LOT more than adding an argument to generic_make_request. > > > > We may introduce generic_make_request_with_ref() for blk_queue_split(), > > what do you think of the following patch? > > Functionally it'll work as well. If you are fine with this one, I'm glad to prepare for a normal one, so that we can save one precious bio flag, and the total number is just 13. > > > In DM or MD, there is the following use case: > > > > 1) bio comes from fs, then one clone is made from this bio > > > > 2) the cloned bio is submitted via generic_make_request() from > > another context, such as wq. > > > > 3) then the flag of BIO_QUEUE_ENTERED is inherited and causes > > the queue usage counter isn't held. > > We don't copy flags on a clone. Yeah, you are right, seems I ignored that. Thanks, Ming
