If the filesystem dio handler understands IOCB_DIO_CALLER_COMP, we'll get a kiocb->ki_complete() callback with kiocb->dio_complete set. In that case, rather than complete the IO directly through task_work, queue up an intermediate task_work handler that first processes this callback and then immediately completes the request. For XFS, this avoids a punt through a workqueue, which is a lot less efficient and adds latency to lower queue depth (or sync) O_DIRECT writes. Only do this for non-polled IO, as polled IO doesn't need this kind of deferral as it always completes within the task itself. This then avoids a check for deferral in the polled IO completion handler. Reviewed-by: Christoph Hellwig <hch@xxxxxx> Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- io_uring/rw.c | 26 +++++++++++++++++++++++--- 1 file changed, 23 insertions(+), 3 deletions(-) diff --git a/io_uring/rw.c b/io_uring/rw.c index 1bce2208b65c..f19f65b3f0ee 100644 --- a/io_uring/rw.c +++ b/io_uring/rw.c @@ -105,6 +105,7 @@ int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe) } else { rw->kiocb.ki_ioprio = get_current_ioprio(); } + rw->kiocb.dio_complete = NULL; rw->addr = READ_ONCE(sqe->addr); rw->len = READ_ONCE(sqe->len); @@ -285,6 +286,14 @@ static inline int io_fixup_rw_res(struct io_kiocb *req, long res) void io_req_rw_complete(struct io_kiocb *req, struct io_tw_state *ts) { + struct io_rw *rw = io_kiocb_to_cmd(req, struct io_rw); + + if (rw->kiocb.dio_complete) { + long res = rw->kiocb.dio_complete(rw->kiocb.private); + + io_req_set_res(req, io_fixup_rw_res(req, res), 0); + } + io_req_io_end(req); if (req->flags & (REQ_F_BUFFER_SELECTED|REQ_F_BUFFER_RING)) { @@ -300,9 +309,11 @@ static void io_complete_rw(struct kiocb *kiocb, long res) struct io_rw *rw = container_of(kiocb, struct io_rw, kiocb); struct io_kiocb *req = cmd_to_io_kiocb(rw); - if (__io_complete_rw_common(req, res)) - return; - io_req_set_res(req, io_fixup_rw_res(req, res), 0); + if (!rw->kiocb.dio_complete) { + if (__io_complete_rw_common(req, res)) + return; + io_req_set_res(req, io_fixup_rw_res(req, res), 0); + } req->io_task_work.func = io_req_rw_complete; __io_req_task_work_add(req, IOU_F_TWQ_LAZY_WAKE); } @@ -916,6 +927,15 @@ int io_write(struct io_kiocb *req, unsigned int issue_flags) } kiocb->ki_flags |= IOCB_WRITE; + /* + * For non-polled IO, set IOCB_DIO_CALLER_COMP, stating that our handler + * groks deferring the completion to task context. This isn't + * necessary and useful for polled IO as that can always complete + * directly. + */ + if (!(kiocb->ki_flags & IOCB_HIPRI)) + kiocb->ki_flags |= IOCB_DIO_CALLER_COMP; + if (likely(req->file->f_op->write_iter)) ret2 = call_write_iter(req->file, kiocb, &s->iter); else if (req->file->f_op->write) -- 2.40.1