On Wed, 2017-12-06 at 00:28 +0800, Ming Lei wrote: > This is still a workaround for RESTART, see my comment before: > > https://marc.info/?l=linux-block&m=151217500929341&w=2 A quote from that e-mail: "The theory about using BLK_MQ_S_SCHED_RESTART in current way is that we mark it after requests are added to hctx->dispatch". Reading that makes me wonder whether you understand the purpose of the BLK_MQ_S_SCHED_RESTART flag? That flag is not set after requests are added to the dispatch list but after requests have been *removed*. The purpose of that flag is to detect whether another thread has run the queue after requests were removed from the dispatch list and before these were readded. If so, the queue needs to be rerun. Bart.
