On 9/2/21 12:08 PM, Jens Axboe wrote: > On 9/2/21 12:04 PM, Pavel Begunkov wrote: >> On 9/2/21 6:56 PM, Jens Axboe wrote: >>> If a task exits normally, then the teardown of async workers happens >>> quickly. But if it simply closes the ring fd, then the async teardown >>> ends up waiting for workers to timeout if they are sleeping, which can >>> then take up to 5 seconds (by default). This isn't a big issue as this >>> happens off the workqueue path, but let's be nicer and ensure that we >>> exit as quick as possible. >> >> ring = io_uring_init(); >> ... >> io_uring_close(&ring); // triggers io_ring_ctx_wait_and_kill() >> rint2 = io_uring_init(); >> ... >> >> It looks IO_WQ_BIT_EXIT there will be troublesome. > > I wonder if we can get by with just a wakeup. Let me test... Nah not enough. Would be nice to fix though, pretty annoying that the threads sit there for 5 seconds doing nothing. -- Jens Axboe
