diff options
Diffstat (limited to 'drivers/block')
| -rw-r--r-- | drivers/block/Kconfig | 6 | ||||
| -rw-r--r-- | drivers/block/drbd/drbd_main.c | 6 | ||||
| -rw-r--r-- | drivers/block/drbd/drbd_receiver.c | 6 | ||||
| -rw-r--r-- | drivers/block/loop.c | 12 | ||||
| -rw-r--r-- | drivers/block/nbd.c | 10 | ||||
| -rw-r--r-- | drivers/block/rbd.c | 4 | ||||
| -rw-r--r-- | drivers/block/ublk_drv.c | 127 |
7 files changed, 97 insertions, 74 deletions
diff --git a/drivers/block/Kconfig b/drivers/block/Kconfig index 91de9fc29bbe..a2184b428493 100644 --- a/drivers/block/Kconfig +++ b/drivers/block/Kconfig @@ -365,6 +365,12 @@ config BLK_DEV_UBLK definition isn't finalized yet, and might change according to future requirement, so mark is as experimental now. + Say Y if you want to get better performance because task_work_add() + can be used in IO path for replacing io_uring cmd, which will become + shared between IO tasks and ubq daemon, meantime task_work_add() can + can handle batch more effectively, but task_work_add() isn't exported + for module, so ublk has to be built to kernel. + source "drivers/block/rnbd/Kconfig" endif # BLK_DEV diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c index 7213ffd69a16..2f16e1bfb6e7 100644 --- a/drivers/block/drbd/drbd_main.c +++ b/drivers/block/drbd/drbd_main.c @@ -1816,7 +1816,7 @@ int drbd_send(struct drbd_connection *connection, struct socket *sock, /* THINK if (signal_pending) return ... ? */ - iov_iter_kvec(&msg.msg_iter, WRITE, &iov, 1, size); + iov_iter_kvec(&msg.msg_iter, ITER_SOURCE, &iov, 1, size); if (sock == connection->data.socket) { rcu_read_lock(); @@ -2671,7 +2671,7 @@ static int init_submitter(struct drbd_device *device) enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsigned int minor) { struct drbd_resource *resource = adm_ctx->resource; - struct drbd_connection *connection; + struct drbd_connection *connection, *n; struct drbd_device *device; struct drbd_peer_device *peer_device, *tmp_peer_device; struct gendisk *disk; @@ -2790,7 +2790,7 @@ enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsig out_destroy_workqueue: destroy_workqueue(device->submit.wq); out_idr_remove_from_resource: - for_each_connection(connection, resource) { + for_each_connection_safe(connection, n, resource) { peer_device = idr_remove(&connection->peer_devices, vnr); if (peer_device) kref_put(&connection->kref, drbd_destroy_connection); diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c index 24d0a2262982..0e58a3187345 100644 --- a/drivers/block/drbd/drbd_receiver.c +++ b/drivers/block/drbd/drbd_receiver.c @@ -507,7 +507,7 @@ static int drbd_recv_short(struct socket *sock, void *buf, size_t size, int flag struct msghdr msg = { .msg_flags = (flags ? flags : MSG_WAITALL | MSG_NOSIGNAL) }; - iov_iter_kvec(&msg.msg_iter, READ, &iov, 1, size); + iov_iter_kvec(&msg.msg_iter, ITER_DEST, &iov, 1, size); return sock_recvmsg(sock, &msg, msg.msg_flags); } @@ -781,7 +781,7 @@ static struct socket *drbd_wait_for_connect(struct drbd_connection *connection, timeo = connect_int * HZ; /* 28.5% random jitter */ - timeo += prandom_u32_max(2) ? timeo / 7 : -timeo / 7; + timeo += get_random_u32_below(2) ? timeo / 7 : -timeo / 7; err = wait_for_completion_interruptible_timeout(&ad->door_bell, timeo); if (err <= 0) @@ -1004,7 +1004,7 @@ retry: drbd_warn(connection, "Error receiving initial packet\n"); sock_release(s); randomize: - if (prandom_u32_max(2)) + if (get_random_u32_below(2)) goto retry; } } diff --git a/drivers/block/loop.c b/drivers/block/loop.c index ad92192c7d61..1f8f3b87bdfa 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -243,7 +243,7 @@ static int lo_write_bvec(struct file *file, struct bio_vec *bvec, loff_t *ppos) struct iov_iter i; ssize_t bw; - iov_iter_bvec(&i, WRITE, bvec, 1, bvec->bv_len); + iov_iter_bvec(&i, ITER_SOURCE, bvec, 1, bvec->bv_len); file_start_write(file); bw = vfs_iter_write(file, &i, ppos, 0); @@ -286,7 +286,7 @@ static int lo_read_simple(struct loop_device *lo, struct request *rq, ssize_t len; rq_for_each_segment(bvec, rq, iter) { - iov_iter_bvec(&i, READ, &bvec, 1, bvec.bv_len); + iov_iter_bvec(&i, ITER_DEST, &bvec, 1, bvec.bv_len); len = vfs_iter_read(lo->lo_backing_file, &i, &pos, 0); if (len < 0) return len; @@ -392,7 +392,7 @@ static void lo_rw_aio_complete(struct kiocb *iocb, long ret) } static int lo_rw_aio(struct loop_device *lo, struct loop_cmd *cmd, - loff_t pos, bool rw) + loff_t pos, int rw) { struct iov_iter iter; struct req_iterator rq_iter; @@ -448,7 +448,7 @@ static int lo_rw_aio(struct loop_device *lo, struct loop_cmd *cmd, cmd->iocb.ki_flags = IOCB_DIRECT; cmd->iocb.ki_ioprio = IOPRIO_PRIO_VALUE(IOPRIO_CLASS_NONE, 0); - if (rw == WRITE) + if (rw == ITER_SOURCE) ret = call_write_iter(file, &cmd->iocb, &iter); else ret = call_read_iter(file, &cmd->iocb, &iter); @@ -490,12 +490,12 @@ static int do_req_filebacked(struct loop_device *lo, struct request *rq) return lo_fallocate(lo, rq, pos, FALLOC_FL_PUNCH_HOLE); case REQ_OP_WRITE: if (cmd->use_aio) - return lo_rw_aio(lo, cmd, pos, WRITE); + return lo_rw_aio(lo, cmd, pos, ITER_SOURCE); else return lo_write_simple(lo, rq, pos); case REQ_OP_READ: if (cmd->use_aio) - return lo_rw_aio(lo, cmd, pos, READ); + return lo_rw_aio(lo, cmd, pos, ITER_DEST); else return lo_read_simple(lo, rq, pos); default: diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index 5cffd96ef2d7..e379ccc63c52 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -563,7 +563,7 @@ static int nbd_send_cmd(struct nbd_device *nbd, struct nbd_cmd *cmd, int index) u32 nbd_cmd_flags = 0; int sent = nsock->sent, skip = 0; - iov_iter_kvec(&from, WRITE, &iov, 1, sizeof(request)); + iov_iter_kvec(&from, ITER_SOURCE, &iov, 1, sizeof(request)); type = req_to_nbd_cmd_type(req); if (type == U32_MAX) @@ -649,7 +649,7 @@ send_pages: dev_dbg(nbd_to_dev(nbd), "request %p: sending %d bytes data\n", req, bvec.bv_len); - iov_iter_bvec(&from, WRITE, &bvec, 1, bvec.bv_len); + iov_iter_bvec(&from, ITER_SOURCE, &bvec, 1, bvec.bv_len); if (skip) { if (skip >= iov_iter_count(&from)) { skip -= iov_iter_count(&from); @@ -701,7 +701,7 @@ static int nbd_read_reply(struct nbd_device *nbd, int index, int result; reply->magic = 0; - iov_iter_kvec(&to, READ, &iov, 1, sizeof(*reply)); + iov_iter_kvec(&to, ITER_DEST, &iov, 1, sizeof(*reply)); result = sock_xmit(nbd, index, 0, &to, MSG_WAITALL, NULL); if (result < 0) { if (!nbd_disconnected(nbd->config)) @@ -790,7 +790,7 @@ static struct nbd_cmd *nbd_handle_reply(struct nbd_device *nbd, int index, struct iov_iter to; rq_for_each_segment(bvec, req, iter) { - iov_iter_bvec(&to, READ, &bvec, 1, bvec.bv_len); + iov_iter_bvec(&to, ITER_DEST, &bvec, 1, bvec.bv_len); result = sock_xmit(nbd, index, 0, &to, MSG_WAITALL, NULL); if (result < 0) { dev_err(disk_to_dev(nbd->disk), "Receive data failed (result %d)\n", @@ -1267,7 +1267,7 @@ static void send_disconnects(struct nbd_device *nbd) for (i = 0; i < config->num_connections; i++) { struct nbd_sock *nsock = config->socks[i]; - iov_iter_kvec(&from, WRITE, &iov, 1, sizeof(request)); + iov_iter_kvec(&from, ITER_SOURCE, &iov, 1, sizeof(request)); mutex_lock(&nsock->tx_lock); ret = sock_xmit(nbd, i, 1, &from, 0, NULL); if (ret < 0) diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index f9e39301c4af..04453f4a319c 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -7222,8 +7222,10 @@ static int __init rbd_sysfs_init(void) int ret; ret = device_register(&rbd_root_dev); - if (ret < 0) + if (ret < 0) { + put_device(&rbd_root_dev); return ret; + } ret = bus_register(&rbd_bus_type); if (ret < 0) diff --git a/drivers/block/ublk_drv.c b/drivers/block/ublk_drv.c index 5afce6ffaadf..e9de9d846b73 100644 --- a/drivers/block/ublk_drv.c +++ b/drivers/block/ublk_drv.c @@ -57,11 +57,12 @@ #define UBLK_PARAM_TYPE_ALL (UBLK_PARAM_TYPE_BASIC | UBLK_PARAM_TYPE_DISCARD) struct ublk_rq_data { + struct llist_node node; struct callback_head work; }; struct ublk_uring_cmd_pdu { - struct request *req; + struct ublk_queue *ubq; }; /* @@ -119,6 +120,8 @@ struct ublk_queue { struct task_struct *ubq_daemon; char *io_cmd_buf; + struct llist_head io_cmds; + unsigned long io_addr; /* mapped vm address */ unsigned int max_io_sz; bool force_abort; @@ -761,11 +764,31 @@ static inline void __ublk_rq_task_work(struct request *req) ubq_complete_io_cmd(io, UBLK_IO_RES_OK); } +static inline void ublk_forward_io_cmds(struct ublk_queue *ubq) +{ + struct llist_node *io_cmds = llist_del_all(&ubq->io_cmds); + struct ublk_rq_data *data, *tmp; + + io_cmds = llist_reverse_order(io_cmds); + llist_for_each_entry_safe(data, tmp, io_cmds, node) + __ublk_rq_task_work(blk_mq_rq_from_pdu(data)); +} + +static inline void ublk_abort_io_cmds(struct ublk_queue *ubq) +{ + struct llist_node *io_cmds = llist_del_all(&ubq->io_cmds); + struct ublk_rq_data *data, *tmp; + + llist_for_each_entry_safe(data, tmp, io_cmds, node) + __ublk_abort_rq(ubq, blk_mq_rq_from_pdu(data)); +} + static void ublk_rq_task_work_cb(struct io_uring_cmd *cmd) { struct ublk_uring_cmd_pdu *pdu = ublk_get_uring_cmd_pdu(cmd); + struct ublk_queue *ubq = pdu->ubq; - __ublk_rq_task_work(pdu->req); + ublk_forward_io_cmds(ubq); } static void ublk_rq_task_work_fn(struct callback_head *work) @@ -773,8 +796,45 @@ static void ublk_rq_task_work_fn(struct callback_head *work) struct ublk_rq_data *data = container_of(work, struct ublk_rq_data, work); struct request *req = blk_mq_rq_from_pdu(data); + struct ublk_queue *ubq = req->mq_hctx->driver_data; + + ublk_forward_io_cmds(ubq); +} - __ublk_rq_task_work(req); +static void ublk_queue_cmd(struct ublk_queue *ubq, struct request *rq) +{ + struct ublk_rq_data *data = blk_mq_rq_to_pdu(rq); + struct ublk_io *io; + + if (!llist_add(&data->node, &ubq->io_cmds)) + return; + + io = &ubq->ios[rq->tag]; + /* + * If the check pass, we know that this is a re-issued request aborted + * previously in monitor_work because the ubq_daemon(cmd's task) is + * PF_EXITING. We cannot call io_uring_cmd_complete_in_task() anymore + * because this ioucmd's io_uring context may be freed now if no inflight + * ioucmd exists. Otherwise we may cause null-deref in ctx->fallback_work. + * + * Note: monitor_work sets UBLK_IO_FLAG_ABORTED and ends this request(releasing + * the tag). Then the request is re-started(allocating the tag) and we are here. + * Since releasing/allocating a tag implies smp_mb(), finding UBLK_IO_FLAG_ABORTED + * guarantees that here is a re-issued request aborted previously. + */ + if (unlikely(io->flags & UBLK_IO_FLAG_ABORTED)) { + ublk_abort_io_cmds(ubq); + } else if (ublk_can_use_task_work(ubq)) { + if (task_work_add(ubq->ubq_daemon, &data->work, + TWA_SIGNAL_NO_IPI)) + ublk_abort_io_cmds(ubq); + } else { + struct io_uring_cmd *cmd = io->cmd; + struct ublk_uring_cmd_pdu *pdu = ublk_get_uring_cmd_pdu(cmd); + + pdu->ubq = ubq; + io_uring_cmd_complete_in_task(cmd, ublk_rq_task_work_cb); + } } static blk_status_t ublk_queue_rq(struct blk_mq_hw_ctx *hctx, @@ -788,6 +848,7 @@ static blk_status_t ublk_queue_rq(struct blk_mq_hw_ctx *hctx, res = ublk_setup_iod(ubq, rq); if (unlikely(res != BLK_STS_OK)) return BLK_STS_IOERR; + /* With recovery feature enabled, force_abort is set in * ublk_stop_dev() before calling del_gendisk(). We have to * abort all requeued and new rqs here to let del_gendisk() @@ -803,53 +864,15 @@ static blk_status_t ublk_queue_rq(struct blk_mq_hw_ctx *hctx, blk_mq_start_request(bd->rq); if (unlikely(ubq_daemon_is_dying(ubq))) { - fail: __ublk_abort_rq(ubq, rq); return BLK_STS_OK; } - if (ublk_can_use_task_work(ubq)) { - struct ublk_rq_data *data = blk_mq_rq_to_pdu(rq); - enum task_work_notify_mode notify_mode = bd->last ? - TWA_SIGNAL_NO_IPI : TWA_NONE; - - if (task_work_add(ubq->ubq_daemon, &data->work, notify_mode)) - goto fail; - } else { - struct ublk_io *io = &ubq->ios[rq->tag]; - struct io_uring_cmd *cmd = io->cmd; - struct ublk_uring_cmd_pdu *pdu = ublk_get_uring_cmd_pdu(cmd); - - /* - * If the check pass, we know that this is a re-issued request aborted - * previously in monitor_work because the ubq_daemon(cmd's task) is - * PF_EXITING. We cannot call io_uring_cmd_complete_in_task() anymore - * because this ioucmd's io_uring context may be freed now if no inflight - * ioucmd exists. Otherwise we may cause null-deref in ctx->fallback_work. - * - * Note: monitor_work sets UBLK_IO_FLAG_ABORTED and ends this request(releasing - * the tag). Then the request is re-started(allocating the tag) and we are here. - * Since releasing/allocating a tag implies smp_mb(), finding UBLK_IO_FLAG_ABORTED - * guarantees that here is a re-issued request aborted previously. - */ - if ((io->flags & UBLK_IO_FLAG_ABORTED)) - goto fail; - - pdu->req = rq; - io_uring_cmd_complete_in_task(cmd, ublk_rq_task_work_cb); - } + ublk_queue_cmd(ubq, rq); return BLK_STS_OK; } -static void ublk_commit_rqs(struct blk_mq_hw_ctx *hctx) -{ - struct ublk_queue *ubq = hctx->driver_data; - - if (ublk_can_use_task_work(ubq)) - __set_notify_signal(ubq->ubq_daemon); -} - static int ublk_init_hctx(struct blk_mq_hw_ctx *hctx, void *driver_data, unsigned int hctx_idx) { @@ -871,7 +894,6 @@ static int ublk_init_rq(struct blk_mq_tag_set *set, struct request *req, static const struct blk_mq_ops ublk_mq_ops = { .queue_rq = ublk_queue_rq, - .commit_rqs = ublk_commit_rqs, .init_hctx = ublk_init_hctx, .init_request = ublk_init_rq, }; @@ -1164,22 +1186,12 @@ static void ublk_mark_io_ready(struct ublk_device *ub, struct ublk_queue *ubq) } static void ublk_handle_need_get_data(struct ublk_device *ub, int q_id, - int tag, struct io_uring_cmd *cmd) + int tag) { struct ublk_queue *ubq = ublk_get_queue(ub, q_id); struct request *req = blk_mq_tag_to_rq(ub->tag_set.tags[q_id], tag); - if (ublk_can_use_task_work(ubq)) { - struct ublk_rq_data *data = blk_mq_rq_to_pdu(req); - - /* should not fail since we call it just in ubq->ubq_daemon */ - task_work_add(ubq->ubq_daemon, &data->work, TWA_SIGNAL_NO_IPI); - } else { - struct ublk_uring_cmd_pdu *pdu = ublk_get_uring_cmd_pdu(cmd); - - pdu->req = req; - io_uring_cmd_complete_in_task(cmd, ublk_rq_task_work_cb); - } + ublk_queue_cmd(ubq, req); } static int ublk_ch_uring_cmd(struct io_uring_cmd *cmd, unsigned int issue_flags) @@ -1267,7 +1279,7 @@ static int ublk_ch_uring_cmd(struct io_uring_cmd *cmd, unsigned int issue_flags) io->addr = ub_cmd->addr; io->cmd = cmd; io->flags |= UBLK_IO_FLAG_ACTIVE; - ublk_handle_need_get_data(ub, ub_cmd->q_id, ub_cmd->tag, cmd); + ublk_handle_need_get_data(ub, ub_cmd->q_id, ub_cmd->tag); break; default: goto out; @@ -1658,6 +1670,9 @@ static int ublk_ctrl_add_dev(struct io_uring_cmd *cmd) */ ub->dev_info.flags &= UBLK_F_ALL; + if (!IS_BUILTIN(CONFIG_BLK_DEV_UBLK)) + ub->dev_info.flags |= UBLK_F_URING_CMD_COMP_IN_TASK; + /* We are not ready to support zero copy */ ub->dev_info.flags &= ~UBLK_F_SUPPORT_ZERO_COPY; |