engines/xnvme: user space vfio based backend

[fio.git] / engines / io_uring.c

diff --git a/engines/io_uring.c b/engines/io_uring.c
index d0fc61dcac394e69419b78232c86be4aa69a7fab..a9abd11dfc7c9805a7f4b4430df623a5c3187747 100644 (file)
--- a/engines/io_uring.c
+++ b/engines/io_uring.c
@@ -226,7 +226,7 @@ static struct fio_option options[] = {
        {
                .name   = "sqthread_poll",
                .lname  = "Kernel SQ thread polling",
-               .type   = FIO_OPT_INT,
+               .type   = FIO_OPT_STR_SET,
                .off1   = offsetof(struct ioring_options, sqpoll_thread),
                .help   = "Offload submission/completion to kernel thread",
                .category = FIO_OPT_C_ENGINE,
@@ -433,6 +433,10 @@ static int fio_ioring_cmd_prep(struct thread_data *td, struct io_u *io_u)
                ld->prepped = 0;
                sqe->flags |= IOSQE_ASYNC;
        }
+       if (o->fixedbufs) {
+               sqe->uring_cmd_flags = IORING_URING_CMD_FIXED;
+               sqe->buf_index = io_u->index;
+       }
 
        cmd = (struct nvme_uring_cmd *)sqe->cmd;
        return fio_nvme_uring_cmd_prep(cmd, io_u,
@@ -633,12 +637,16 @@ static int fio_ioring_commit(struct thread_data *td)
         */
        if (o->sqpoll_thread) {
                struct io_sq_ring *ring = &ld->sq_ring;
+               unsigned start = *ld->sq_ring.head;
                unsigned flags;
 
                flags = atomic_load_acquire(ring->flags);
                if (flags & IORING_SQ_NEED_WAKEUP)
                        io_uring_enter(ld, ld->queued, 0,
                                        IORING_ENTER_SQ_WAKEUP);
+               fio_ioring_queued(td, start, ld->queued);
+               io_u_mark_submit(td, ld->queued);
+
                ld->queued = 0;
                return 0;
        }
@@ -800,6 +808,14 @@ static int fio_ioring_queue_init(struct thread_data *td)
                        p.flags |= IORING_SETUP_SQ_AFF;
                        p.sq_thread_cpu = o->sqpoll_cpu;
                }
+
+               /*
+                * Submission latency for sqpoll_thread is just the time it
+                * takes to fill in the SQ ring entries, and any syscall if
+                * IORING_SQ_NEED_WAKEUP is set, we don't need to log that time
+                * separately.
+                */
+               td->o.disable_slat = 1;
        }
 
        /*
@@ -872,6 +888,14 @@ static int fio_ioring_cmd_queue_init(struct thread_data *td)
                        p.flags |= IORING_SETUP_SQ_AFF;
                        p.sq_thread_cpu = o->sqpoll_cpu;
                }
+
+               /*
+                * Submission latency for sqpoll_thread is just the time it
+                * takes to fill in the SQ ring entries, and any syscall if
+                * IORING_SQ_NEED_WAKEUP is set, we don't need to log that time
+                * separately.
+                */
+               td->o.disable_slat = 1;
        }
        if (o->cmd_type == FIO_URING_CMD_NVME) {
                p.flags |= IORING_SETUP_SQE128;
@@ -885,9 +909,30 @@ static int fio_ioring_cmd_queue_init(struct thread_data *td)
        p.flags |= IORING_SETUP_CQSIZE;
        p.cq_entries = depth;
 
+       /*
+        * Setup COOP_TASKRUN as we don't need to get IPI interrupted for
+        * completing IO operations.
+        */
+       p.flags |= IORING_SETUP_COOP_TASKRUN;
+
+       /*
+        * io_uring is always a single issuer, and we can defer task_work
+        * runs until we reap events.
+        */
+       p.flags |= IORING_SETUP_SINGLE_ISSUER | IORING_SETUP_DEFER_TASKRUN;
+
 retry:
        ret = syscall(__NR_io_uring_setup, depth, &p);
        if (ret < 0) {
+               if (errno == EINVAL && p.flags & IORING_SETUP_DEFER_TASKRUN) {
+                       p.flags &= ~IORING_SETUP_DEFER_TASKRUN;
+                       p.flags &= ~IORING_SETUP_SINGLE_ISSUER;
+                       goto retry;
+               }
+               if (errno == EINVAL && p.flags & IORING_SETUP_COOP_TASKRUN) {
+                       p.flags &= ~IORING_SETUP_COOP_TASKRUN;
+                       goto retry;
+               }
                if (errno == EINVAL && p.flags & IORING_SETUP_CQSIZE) {
                        p.flags &= ~IORING_SETUP_CQSIZE;
                        goto retry;