Merge tag 'for-5.13/drivers-2021-04-27' of git://git.kernel.dk/linux-block

Pull block driver updates from Jens Axboe:

 - MD changes via Song:
        - raid5 POWER fix
        - raid1 failure fix
        - UAF fix for md cluster
        - mddev_find_or_alloc() clean up
        - Fix NULL pointer deref with external bitmap
        - Performance improvement for raid10 discard requests
        - Fix missing information of /proc/mdstat

 - rsxx const qualifier removal (Arnd)

 - Expose allocated brd pages (Calvin)

 - rnbd via Gioh Kim:
        - Change maintainer
        - Change domain address of maintainers' email
        - Add polling IO mode and document update
        - Fix memory leak and some bug detected by static code analysis
          tools
        - Code refactoring

 - Series of floppy cleanups/fixes (Denis)

 - s390 dasd fixes (Julian)

 - kerneldoc fixes (Lee)

 - null_blk double free (Lv)

 - null_blk virtual boundary addition (Max)

 - Remove xsysace driver (Michal)

 - umem driver removal (Davidlohr)

 - ataflop fixes (Dan)

 - Revalidate disk removal (Christoph)

 - Bounce buffer cleanups (Christoph)

 - Mark lightnvm as deprecated (Christoph)

 - mtip32xx init cleanups (Shixin)

 - Various fixes (Tian, Gustavo, Coly, Yang, Zhang, Zhiqiang)

* tag 'for-5.13/drivers-2021-04-27' of git://git.kernel.dk/linux-block: (143 commits)
  async_xor: increase src_offs when dropping destination page
  drivers/block/null_blk/main: Fix a double free in null_init.
  md/raid1: properly indicate failure when ending a failed write request
  md-cluster: fix use-after-free issue when removing rdev
  nvme: introduce generic per-namespace chardev
  nvme: cleanup nvme_configure_apst
  nvme: do not try to reconfigure APST when the controller is not live
  nvme: add 'kato' sysfs attribute
  nvme: sanitize KATO setting
  nvmet: avoid queuing keep-alive timer if it is disabled
  brd: expose number of allocated pages in debugfs
  ataflop: fix off by one in ataflop_probe()
  ataflop: potential out of bounds in do_format()
  drbd: Fix fall-through warnings for Clang
  block/rnbd: Use strscpy instead of strlcpy
  block/rnbd-clt-sysfs: Remove copy buffer overlap in rnbd_clt_get_path_name
  block/rnbd-clt: Remove max_segment_size
  block/rnbd-clt: Generate kobject_uevent when the rnbd device state changes
  block/rnbd-srv: Remove unused arguments of rnbd_srv_rdma_ev
  Documentation/ABI/rnbd-clt: Add description for nr_poll_queues
  ...

1 files changed, 67 insertions, 12 deletions

diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c
index d658c6e8263a..f9f34f6caf5e 100644
--- a/drivers/nvme/target/tcp.c
+++ b/drivers/nvme/target/tcp.c
@@ -29,6 +29,16 @@ static int so_priority;
 module_param(so_priority, int, 0644);
 MODULE_PARM_DESC(so_priority, "nvmet tcp socket optimize priority");
 
+/* Define a time period (in usecs) that io_work() shall sample an activated
+ * queue before determining it to be idle.  This optional module behavior
+ * can enable NIC solutions that support socket optimized packet processing
+ * using advanced interrupt moderation techniques.
+ */
+static int idle_poll_period_usecs;
+module_param(idle_poll_period_usecs, int, 0644);
+MODULE_PARM_DESC(idle_poll_period_usecs,
+		"nvmet tcp io_work poll till idle time period in usecs");
+
 #define NVMET_TCP_RECV_BUDGET		8
 #define NVMET_TCP_SEND_BUDGET		8
 #define NVMET_TCP_IO_WORK_BUDGET	64
@@ -119,6 +129,8 @@ struct nvmet_tcp_queue {
 	struct ahash_request	*snd_hash;
 	struct ahash_request	*rcv_hash;
 
+	unsigned long           poll_end;
+
 	spinlock_t		state_lock;
 	enum nvmet_tcp_queue_state state;
 
@@ -525,11 +537,36 @@ static void nvmet_tcp_queue_response(struct nvmet_req *req)
 	struct nvmet_tcp_cmd *cmd =
 		container_of(req, struct nvmet_tcp_cmd, req);
 	struct nvmet_tcp_queue	*queue = cmd->queue;
+	struct nvme_sgl_desc *sgl;
+	u32 len;
+
+	if (unlikely(cmd == queue->cmd)) {
+		sgl = &cmd->req.cmd->common.dptr.sgl;
+		len = le32_to_cpu(sgl->length);
+
+		/*
+		 * Wait for inline data before processing the response.
+		 * Avoid using helpers, this might happen before
+		 * nvmet_req_init is completed.
+		 */
+		if (queue->rcv_state == NVMET_TCP_RECV_PDU &&
+		    len && len < cmd->req.port->inline_data_size &&
+		    nvme_is_write(cmd->req.cmd))
+			return;
+	}
 
 	llist_add(&cmd->lentry, &queue->resp_list);
 	queue_work_on(queue_cpu(queue), nvmet_tcp_wq, &cmd->queue->io_work);
 }
 
+static void nvmet_tcp_execute_request(struct nvmet_tcp_cmd *cmd)
+{
+	if (unlikely(cmd->flags & NVMET_TCP_F_INIT_FAILED))
+		nvmet_tcp_queue_response(&cmd->req);
+	else
+		cmd->req.execute(&cmd->req);
+}
+
 static int nvmet_try_send_data_pdu(struct nvmet_tcp_cmd *cmd)
 {
 	u8 hdgst = nvmet_tcp_hdgst_len(cmd->queue);
@@ -961,7 +998,7 @@ static int nvmet_tcp_done_recv_pdu(struct nvmet_tcp_queue *queue)
 			le32_to_cpu(req->cmd->common.dptr.sgl.length));
 
 		nvmet_tcp_handle_req_failure(queue, queue->cmd, req);
-		return -EAGAIN;
+		return 0;
 	}
 
 	ret = nvmet_tcp_map_data(queue->cmd);
@@ -1104,10 +1141,8 @@ static int nvmet_tcp_try_recv_data(struct nvmet_tcp_queue *queue)
 		return 0;
 	}
 
-	if (!(cmd->flags & NVMET_TCP_F_INIT_FAILED) &&
-	    cmd->rbytes_done == cmd->req.transfer_len) {
-		cmd->req.execute(&cmd->req);
-	}
+	if (cmd->rbytes_done == cmd->req.transfer_len)
+		nvmet_tcp_execute_request(cmd);
 
 	nvmet_prepare_receive_pdu(queue);
 	return 0;
@@ -1144,9 +1179,9 @@ static int nvmet_tcp_try_recv_ddgst(struct nvmet_tcp_queue *queue)
 		goto out;
 	}
 
-	if (!(cmd->flags & NVMET_TCP_F_INIT_FAILED) &&
-	    cmd->rbytes_done == cmd->req.transfer_len)
-		cmd->req.execute(&cmd->req);
+	if (cmd->rbytes_done == cmd->req.transfer_len)
+		nvmet_tcp_execute_request(cmd);
+
 	ret = 0;
 out:
 	nvmet_prepare_receive_pdu(queue);
@@ -1216,6 +1251,23 @@ static void nvmet_tcp_schedule_release_queue(struct nvmet_tcp_queue *queue)
 	spin_unlock(&queue->state_lock);
 }
 
+static inline void nvmet_tcp_arm_queue_deadline(struct nvmet_tcp_queue *queue)
+{
+	queue->poll_end = jiffies + usecs_to_jiffies(idle_poll_period_usecs);
+}
+
+static bool nvmet_tcp_check_queue_deadline(struct nvmet_tcp_queue *queue,
+		int ops)
+{
+	if (!idle_poll_period_usecs)
+		return false;
+
+	if (ops)
+		nvmet_tcp_arm_queue_deadline(queue);
+
+	return !time_after(jiffies, queue->poll_end);
+}
+
 static void nvmet_tcp_io_work(struct work_struct *w)
 {
 	struct nvmet_tcp_queue *queue =
@@ -1241,9 +1293,10 @@ static void nvmet_tcp_io_work(struct work_struct *w)
 	} while (pending && ops < NVMET_TCP_IO_WORK_BUDGET);
 
 	/*
-	 * We exahusted our budget, requeue our selves
+	 * Requeue the worker if idle deadline period is in progress or any
+	 * ops activity was recorded during the do-while loop above.
 	 */
-	if (pending)
+	if (nvmet_tcp_check_queue_deadline(queue, ops) || pending)
 		queue_work_on(queue_cpu(queue), nvmet_tcp_wq, &queue->io_work);
 }
 
@@ -1434,7 +1487,7 @@ static void nvmet_tcp_state_change(struct sock *sk)
 {
 	struct nvmet_tcp_queue *queue;
 
-	write_lock_bh(&sk->sk_callback_lock);
+	read_lock_bh(&sk->sk_callback_lock);
 	queue = sk->sk_user_data;
 	if (!queue)
 		goto done;
@@ -1452,7 +1505,7 @@ static void nvmet_tcp_state_change(struct sock *sk)
 			queue->idx, sk->sk_state);
 	}
 done:
-	write_unlock_bh(&sk->sk_callback_lock);
+	read_unlock_bh(&sk->sk_callback_lock);
 }
 
 static int nvmet_tcp_set_queue_sock(struct nvmet_tcp_queue *queue)
@@ -1501,6 +1554,8 @@ static int nvmet_tcp_set_queue_sock(struct nvmet_tcp_queue *queue)
 		sock->sk->sk_state_change = nvmet_tcp_state_change;
 		queue->write_space = sock->sk->sk_write_space;
 		sock->sk->sk_write_space = nvmet_tcp_write_space;
+		if (idle_poll_period_usecs)
+			nvmet_tcp_arm_queue_deadline(queue);
 		queue_work_on(queue_cpu(queue), nvmet_tcp_wq, &queue->io_work);
 	}
 	write_unlock_bh(&sock->sk->sk_callback_lock);