Merge tag 'for-linus-2019-10-18' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:

 - NVMe pull request from Keith that address deadlocks, double resets,
   memory leaks, and other regression.

 - Fixup elv_support_iosched() for bio based devices (Damien)

 - Fixup for the ahci PCS quirk (Dan)

 - Socket O_NONBLOCK handling fix for io_uring (me)

 - Timeout sequence io_uring fixes (yangerkun)

 - MD warning fix for parameter default_layout (Song)

 - blkcg activation fixes (Tejun)

 - blk-rq-qos node deletion fix (Tejun)

* tag 'for-linus-2019-10-18' of git://git.kernel.dk/linux-block:
  nvme-pci: Set the prp2 correctly when using more than 4k page
  io_uring: fix logic error in io_timeout
  io_uring: fix up O_NONBLOCK handling for sockets
  md/raid0: fix warning message for parameter default_layout
  libata/ahci: Fix PCS quirk application
  blk-rq-qos: fix first node deletion of rq_qos_del()
  blkcg: Fix multiple bugs in blkcg_activate_policy()
  io_uring: consider the overflow of sequence for timeout req
  nvme-tcp: fix possible leakage during error flow
  nvmet-loop: fix possible leakage during error flow
  block: Fix elv_support_iosched()
  nvme-tcp: Initialize sk->sk_ll_usec only with NET_RX_BUSY_POLL
  nvme: Wait for reset state when required
  nvme: Prevent resets during paused controller state
  nvme: Restart request timers in resetting state
  nvme: Remove ADMIN_ONLY state
  nvme-pci: Free tagset if no IO queues
  nvme: retain split access workaround for capability reads
  nvme: fix possible deadlock when nvme_update_formats fails

block/blk-cgroup.c

@@ -1362,7 +1362,7 @@ int blkcg_activate_policy(struct request_queue *q,
 			  const struct blkcg_policy *pol)
 {
 	struct blkg_policy_data *pd_prealloc = NULL;
-	struct blkcg_gq *blkg;
+	struct blkcg_gq *blkg, *pinned_blkg = NULL;
 	int ret;
 
 	if (blkcg_policy_enabled(q, pol))
@@ -1370,49 +1370,82 @@ int blkcg_activate_policy(struct request_queue *q,
 
 	if (queue_is_mq(q))
 		blk_mq_freeze_queue(q);
-pd_prealloc:
-	if (!pd_prealloc) {
-		pd_prealloc = pol->pd_alloc_fn(GFP_KERNEL, q, &blkcg_root);
-		if (!pd_prealloc) {
-			ret = -ENOMEM;
-			goto out_bypass_end;
-		}
-	}
-
+retry:
 	spin_lock_irq(&q->queue_lock);
 
-	/* blkg_list is pushed at the head, reverse walk to init parents first */
+	/* blkg_list is pushed at the head, reverse walk to allocate parents first */
 	list_for_each_entry_reverse(blkg, &q->blkg_list, q_node) {
 		struct blkg_policy_data *pd;
 
 		if (blkg->pd[pol->plid])
 			continue;
 
-		pd = pol->pd_alloc_fn(GFP_NOWAIT | __GFP_NOWARN, q, &blkcg_root);
-		if (!pd)
-			swap(pd, pd_prealloc);
+		/* If prealloc matches, use it; otherwise try GFP_NOWAIT */
+		if (blkg == pinned_blkg) {
+			pd = pd_prealloc;
+			pd_prealloc = NULL;
+		} else {
+			pd = pol->pd_alloc_fn(GFP_NOWAIT | __GFP_NOWARN, q,
+					      blkg->blkcg);
+		}
+
 		if (!pd) {
+			/*
+			 * GFP_NOWAIT failed.  Free the existing one and
+			 * prealloc for @blkg w/ GFP_KERNEL.
+			 */
+			if (pinned_blkg)
+				blkg_put(pinned_blkg);
+			blkg_get(blkg);
+			pinned_blkg = blkg;
+
 			spin_unlock_irq(&q->queue_lock);
-			goto pd_prealloc;
+
+			if (pd_prealloc)
+				pol->pd_free_fn(pd_prealloc);
+			pd_prealloc = pol->pd_alloc_fn(GFP_KERNEL, q,
+						       blkg->blkcg);
+			if (pd_prealloc)
+				goto retry;
+			else
+				goto enomem;
 		}
 
 		blkg->pd[pol->plid] = pd;
 		pd->blkg = blkg;
 		pd->plid = pol->plid;
-		if (pol->pd_init_fn)
-			pol->pd_init_fn(pd);
 	}
 
+	/* all allocated, init in the same order */
+	if (pol->pd_init_fn)
+		list_for_each_entry_reverse(blkg, &q->blkg_list, q_node)
+			pol->pd_init_fn(blkg->pd[pol->plid]);
+
 	__set_bit(pol->plid, q->blkcg_pols);
 	ret = 0;
 
 	spin_unlock_irq(&q->queue_lock);
-out_bypass_end:
+out:
 	if (queue_is_mq(q))
 		blk_mq_unfreeze_queue(q);
+	if (pinned_blkg)
+		blkg_put(pinned_blkg);
 	if (pd_prealloc)
 		pol->pd_free_fn(pd_prealloc);
 	return ret;
+
+enomem:
+	/* alloc failed, nothing's initialized yet, free everything */
+	spin_lock_irq(&q->queue_lock);
+	list_for_each_entry(blkg, &q->blkg_list, q_node) {
+		if (blkg->pd[pol->plid]) {
+			pol->pd_free_fn(blkg->pd[pol->plid]);
+			blkg->pd[pol->plid] = NULL;
+		}
+	}
+	spin_unlock_irq(&q->queue_lock);
+	ret = -ENOMEM;
+	goto out;
 }
 EXPORT_SYMBOL_GPL(blkcg_activate_policy);
 

block/blk-rq-qos.h

@@ -108,16 +108,13 @@ static inline void rq_qos_add(struct request_queue *q, struct rq_qos *rqos)
 
 static inline void rq_qos_del(struct request_queue *q, struct rq_qos *rqos)
 {
-	struct rq_qos *cur, *prev = NULL;
-	for (cur = q->rq_qos; cur; cur = cur->next) {
-		if (cur == rqos) {
-			if (prev)
-				prev->next = rqos->next;
-			else
-				q->rq_qos = cur;
+	struct rq_qos **cur;
+
+	for (cur = &q->rq_qos; *cur; cur = &(*cur)->next) {
+		if (*cur == rqos) {
+			*cur = rqos->next;
 			break;
 		}
-		prev = cur;
 	}
 
 	blk_mq_debugfs_unregister_rqos(rqos);

block/elevator.c

@@ -616,7 +616,8 @@ int elevator_switch_mq(struct request_queue *q,
 
 static inline bool elv_support_iosched(struct request_queue *q)
 {
-	if (q->tag_set && (q->tag_set->flags & BLK_MQ_F_NO_SCHED))
+	if (!q->mq_ops ||
+	    (q->tag_set && (q->tag_set->flags & BLK_MQ_F_NO_SCHED)))
 		return false;
 	return true;
 }

drivers/ata/ahci.c

@@ -1600,7 +1600,9 @@ static void ahci_intel_pcs_quirk(struct pci_dev *pdev, struct ahci_host_priv *hp
 	 */
 	if (!id || id->vendor != PCI_VENDOR_ID_INTEL)
 		return;
-	if (((enum board_ids) id->driver_data) < board_ahci_pcs7)
+
+	/* Skip applying the quirk on Denverton and beyond */
+	if (((enum board_ids) id->driver_data) >= board_ahci_pcs7)
 		return;
 
 	/*

drivers/md/raid0.c

@@ -154,7 +154,7 @@ static int create_strip_zones(struct mddev *mddev, struct r0conf **private_conf)
 	} else {
 		pr_err("md/raid0:%s: cannot assemble multi-zone RAID0 with default_layout setting\n",
 		       mdname(mddev));
-		pr_err("md/raid0: please set raid.default_layout to 1 or 2\n");
+		pr_err("md/raid0: please set raid0.default_layout to 1 or 2\n");
 		err = -ENOTSUPP;
 		goto abort;
 	}

drivers/nvme/host/core.c

@@ -116,10 +116,26 @@ static void nvme_queue_scan(struct nvme_ctrl *ctrl)
 	/*
 	 * Only new queue scan work when admin and IO queues are both alive
 	 */
-	if (ctrl->state == NVME_CTRL_LIVE)
+	if (ctrl->state == NVME_CTRL_LIVE && ctrl->tagset)
 		queue_work(nvme_wq, &ctrl->scan_work);
 }
 
+/*
+ * Use this function to proceed with scheduling reset_work for a controller
+ * that had previously been set to the resetting state. This is intended for
+ * code paths that can't be interrupted by other reset attempts. A hot removal
+ * may prevent this from succeeding.
+ */
+int nvme_try_sched_reset(struct nvme_ctrl *ctrl)
+{
+	if (ctrl->state != NVME_CTRL_RESETTING)
+		return -EBUSY;
+	if (!queue_work(nvme_reset_wq, &ctrl->reset_work))
+		return -EBUSY;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(nvme_try_sched_reset);
+
 int nvme_reset_ctrl(struct nvme_ctrl *ctrl)
 {
 	if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_RESETTING))
@@ -137,8 +153,7 @@ int nvme_reset_ctrl_sync(struct nvme_ctrl *ctrl)
 	ret = nvme_reset_ctrl(ctrl);
 	if (!ret) {
 		flush_work(&ctrl->reset_work);
-		if (ctrl->state != NVME_CTRL_LIVE &&
-		    ctrl->state != NVME_CTRL_ADMIN_ONLY)
+		if (ctrl->state != NVME_CTRL_LIVE)
 			ret = -ENETRESET;
 	}
 
@@ -315,15 +330,6 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 
 	old_state = ctrl->state;
 	switch (new_state) {
-	case NVME_CTRL_ADMIN_ONLY:
-		switch (old_state) {
-		case NVME_CTRL_CONNECTING:
-			changed = true;
-			/* FALLTHRU */
-		default:
-			break;
-		}
-		break;
 	case NVME_CTRL_LIVE:
 		switch (old_state) {
 		case NVME_CTRL_NEW:
@@ -339,7 +345,6 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 		switch (old_state) {
 		case NVME_CTRL_NEW:
 		case NVME_CTRL_LIVE:
-		case NVME_CTRL_ADMIN_ONLY:
 			changed = true;
 			/* FALLTHRU */
 		default:
@@ -359,7 +364,6 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 	case NVME_CTRL_DELETING:
 		switch (old_state) {
 		case NVME_CTRL_LIVE:
-		case NVME_CTRL_ADMIN_ONLY:
 		case NVME_CTRL_RESETTING:
 		case NVME_CTRL_CONNECTING:
 			changed = true;
@@ -381,8 +385,10 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 		break;
 	}
 
-	if (changed)
+	if (changed) {
 		ctrl->state = new_state;
+		wake_up_all(&ctrl->state_wq);
+	}
 
 	spin_unlock_irqrestore(&ctrl->lock, flags);
 	if (changed && ctrl->state == NVME_CTRL_LIVE)
@@ -391,6 +397,39 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 }
 EXPORT_SYMBOL_GPL(nvme_change_ctrl_state);
 
+/*
+ * Returns true for sink states that can't ever transition back to live.
+ */
+static bool nvme_state_terminal(struct nvme_ctrl *ctrl)
+{
+	switch (ctrl->state) {
+	case NVME_CTRL_NEW:
+	case NVME_CTRL_LIVE:
+	case NVME_CTRL_RESETTING:
+	case NVME_CTRL_CONNECTING:
+		return false;
+	case NVME_CTRL_DELETING:
+	case NVME_CTRL_DEAD:
+		return true;
+	default:
+		WARN_ONCE(1, "Unhandled ctrl state:%d", ctrl->state);
+		return true;
+	}
+}
+
+/*
+ * Waits for the controller state to be resetting, or returns false if it is
+ * not possible to ever transition to that state.
+ */
+bool nvme_wait_reset(struct nvme_ctrl *ctrl)
+{
+	wait_event(ctrl->state_wq,
+		   nvme_change_ctrl_state(ctrl, NVME_CTRL_RESETTING) ||
+		   nvme_state_terminal(ctrl));
+	return ctrl->state == NVME_CTRL_RESETTING;
+}
+EXPORT_SYMBOL_GPL(nvme_wait_reset);
+
 static void nvme_free_ns_head(struct kref *ref)
 {
 	struct nvme_ns_head *head =
@@ -1306,8 +1345,6 @@ static void nvme_update_formats(struct nvme_ctrl *ctrl)
 		if (ns->disk && nvme_revalidate_disk(ns->disk))
 			nvme_set_queue_dying(ns);
 	up_read(&ctrl->namespaces_rwsem);
-
-	nvme_remove_invalid_namespaces(ctrl, NVME_NSID_ALL);
 }
 
 static void nvme_passthru_end(struct nvme_ctrl *ctrl, u32 effects)
@@ -1323,6 +1360,7 @@ static void nvme_passthru_end(struct nvme_ctrl *ctrl, u32 effects)
 		nvme_unfreeze(ctrl);
 		nvme_mpath_unfreeze(ctrl->subsys);
 		mutex_unlock(&ctrl->subsys->lock);
+		nvme_remove_invalid_namespaces(ctrl, NVME_NSID_ALL);
 		mutex_unlock(&ctrl->scan_lock);
 	}
 	if (effects & NVME_CMD_EFFECTS_CCC)
@@ -2874,7 +2912,6 @@ static int nvme_dev_open(struct inode *inode, struct file *file)
 
 	switch (ctrl->state) {
 	case NVME_CTRL_LIVE:
-	case NVME_CTRL_ADMIN_ONLY:
 		break;
 	default:
 		return -EWOULDBLOCK;
@@ -3168,7 +3205,6 @@ static ssize_t nvme_sysfs_show_state(struct device *dev,
 	static const char *const state_name[] = {
 		[NVME_CTRL_NEW]		= "new",
 		[NVME_CTRL_LIVE]	= "live",
-		[NVME_CTRL_ADMIN_ONLY]	= "only-admin",
 		[NVME_CTRL_RESETTING]	= "resetting",
 		[NVME_CTRL_CONNECTING]	= "connecting",
 		[NVME_CTRL_DELETING]	= "deleting",
@@ -3679,11 +3715,10 @@ static void nvme_scan_work(struct work_struct *work)
 	struct nvme_id_ctrl *id;
 	unsigned nn;
 
-	if (ctrl->state != NVME_CTRL_LIVE)
+	/* No tagset on a live ctrl means IO queues could not created */
+	if (ctrl->state != NVME_CTRL_LIVE || !ctrl->tagset)
 		return;
 
-	WARN_ON_ONCE(!ctrl->tagset);
-
 	if (test_and_clear_bit(NVME_AER_NOTICE_NS_CHANGED, &ctrl->events)) {
 		dev_info(ctrl->device, "rescanning namespaces.\n");
 		nvme_clear_changed_ns_log(ctrl);
@@ -3844,13 +3879,13 @@ static void nvme_fw_act_work(struct work_struct *work)
 		if (time_after(jiffies, fw_act_timeout)) {
 			dev_warn(ctrl->device,
 				"Fw activation timeout, reset controller\n");
-			nvme_reset_ctrl(ctrl);
-			break;
+			nvme_try_sched_reset(ctrl);
+			return;
 		}
 		msleep(100);
 	}
 
-	if (ctrl->state != NVME_CTRL_LIVE)
+	if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_LIVE))
 		return;
 
 	nvme_start_queues(ctrl);
@@ -3870,6 +3905,12 @@ static void nvme_handle_aen_notice(struct nvme_ctrl *ctrl, u32 result)
 		nvme_queue_scan(ctrl);
 		break;
 	case NVME_AER_NOTICE_FW_ACT_STARTING:
+		/*
+		 * We are (ab)using the RESETTING state to prevent subsequent
+		 * recovery actions from interfering with the controller's
+		 * firmware activation.
+		 */
+		if (nvme_change_ctrl_state(ctrl, NVME_CTRL_RESETTING))
 			queue_work(nvme_wq, &ctrl->fw_act_work);
 		break;
 #ifdef CONFIG_NVME_MULTIPATH
@@ -3993,6 +4034,7 @@ int nvme_init_ctrl(struct nvme_ctrl *ctrl, struct device *dev,
 	INIT_WORK(&ctrl->async_event_work, nvme_async_event_work);
 	INIT_WORK(&ctrl->fw_act_work, nvme_fw_act_work);
 	INIT_WORK(&ctrl->delete_work, nvme_delete_ctrl_work);
+	init_waitqueue_head(&ctrl->state_wq);
 
 	INIT_DELAYED_WORK(&ctrl->ka_work, nvme_keep_alive_work);
 	memset(&ctrl->ka_cmd, 0, sizeof(ctrl->ka_cmd));

drivers/nvme/host/fabrics.h

@@ -182,8 +182,7 @@ bool nvmf_ip_options_match(struct nvme_ctrl *ctrl,
 static inline bool nvmf_check_ready(struct nvme_ctrl *ctrl, struct request *rq,
 		bool queue_live)
 {
-	if (likely(ctrl->state == NVME_CTRL_LIVE ||
-		   ctrl->state == NVME_CTRL_ADMIN_ONLY))
+	if (likely(ctrl->state == NVME_CTRL_LIVE))
 		return true;
 	return __nvmf_check_ready(ctrl, rq, queue_live);
 }

drivers/nvme/host/nvme.h

@@ -15,6 +15,7 @@
 #include <linux/sed-opal.h>
 #include <linux/fault-inject.h>
 #include <linux/rcupdate.h>
+#include <linux/wait.h>
 
 #include <trace/events/block.h>
 
@@ -161,7 +162,6 @@ static inline u16 nvme_req_qid(struct request *req)
 enum nvme_ctrl_state {
 	NVME_CTRL_NEW,
 	NVME_CTRL_LIVE,
-	NVME_CTRL_ADMIN_ONLY,    /* Only admin queue live */
 	NVME_CTRL_RESETTING,
 	NVME_CTRL_CONNECTING,
 	NVME_CTRL_DELETING,
@@ -199,6 +199,7 @@ struct nvme_ctrl {
 	struct cdev cdev;
 	struct work_struct reset_work;
 	struct work_struct delete_work;
+	wait_queue_head_t state_wq;
 
 	struct nvme_subsystem *subsys;
 	struct list_head subsys_entry;
@@ -449,6 +450,7 @@ void nvme_complete_rq(struct request *req);
 bool nvme_cancel_request(struct request *req, void *data, bool reserved);
 bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 		enum nvme_ctrl_state new_state);
+bool nvme_wait_reset(struct nvme_ctrl *ctrl);
 int nvme_disable_ctrl(struct nvme_ctrl *ctrl);
 int nvme_enable_ctrl(struct nvme_ctrl *ctrl);
 int nvme_shutdown_ctrl(struct nvme_ctrl *ctrl);
@@ -499,6 +501,7 @@ int nvme_set_queue_count(struct nvme_ctrl *ctrl, int *count);
 void nvme_stop_keep_alive(struct nvme_ctrl *ctrl);
 int nvme_reset_ctrl(struct nvme_ctrl *ctrl);
 int nvme_reset_ctrl_sync(struct nvme_ctrl *ctrl);
+int nvme_try_sched_reset(struct nvme_ctrl *ctrl);
 int nvme_delete_ctrl(struct nvme_ctrl *ctrl);
 
 int nvme_get_log(struct nvme_ctrl *ctrl, u32 nsid, u8 log_page, u8 lsp,

drivers/nvme/host/pci.c

@@ -773,7 +773,8 @@ static blk_status_t nvme_setup_prp_simple(struct nvme_dev *dev,
 		struct bio_vec *bv)
 {
 	struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
-	unsigned int first_prp_len = dev->ctrl.page_size - bv->bv_offset;
+	unsigned int offset = bv->bv_offset & (dev->ctrl.page_size - 1);
+	unsigned int first_prp_len = dev->ctrl.page_size - offset;
 
 	iod->first_dma = dma_map_bvec(dev->dev, bv, rq_dma_dir(req), 0);
 	if (dma_mapping_error(dev->dev, iod->first_dma))
@@ -2263,10 +2264,7 @@ static bool __nvme_disable_io_queues(struct nvme_dev *dev, u8 opcode)
 	return true;
 }
 
-/*
- * return error value only when tagset allocation failed
- */
-static int nvme_dev_add(struct nvme_dev *dev)
+static void nvme_dev_add(struct nvme_dev *dev)
 {
 	int ret;
 
@@ -2296,7 +2294,7 @@ static int nvme_dev_add(struct nvme_dev *dev)
 		if (ret) {
 			dev_warn(dev->ctrl.device,
 				"IO queues tagset allocation failed %d\n", ret);
-			return ret;
+			return;
 		}
 		dev->ctrl.tagset = &dev->tagset;
 	} else {
@@ -2307,7 +2305,6 @@ static int nvme_dev_add(struct nvme_dev *dev)
 	}
 
 	nvme_dbbuf_set(dev);
-	return 0;
 }
 
 static int nvme_pci_enable(struct nvme_dev *dev)
@@ -2467,6 +2464,14 @@ static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown)
 	mutex_unlock(&dev->shutdown_lock);
 }
 
+static int nvme_disable_prepare_reset(struct nvme_dev *dev, bool shutdown)
+{
+	if (!nvme_wait_reset(&dev->ctrl))
+		return -EBUSY;
+	nvme_dev_disable(dev, shutdown);
+	return 0;
+}
+
 static int nvme_setup_prp_pools(struct nvme_dev *dev)
 {
 	dev->prp_page_pool = dma_pool_create("prp list page", dev->dev,
@@ -2490,14 +2495,20 @@ static void nvme_release_prp_pools(struct nvme_dev *dev)
 	dma_pool_destroy(dev->prp_small_pool);
 }
 
+static void nvme_free_tagset(struct nvme_dev *dev)
+{
+	if (dev->tagset.tags)
+		blk_mq_free_tag_set(&dev->tagset);
+	dev->ctrl.tagset = NULL;
+}
+
 static void nvme_pci_free_ctrl(struct nvme_ctrl *ctrl)
 {
 	struct nvme_dev *dev = to_nvme_dev(ctrl);
 
 	nvme_dbbuf_dma_free(dev);
 	put_device(dev->dev);
-	if (dev->tagset.tags)
-		blk_mq_free_tag_set(&dev->tagset);
+	nvme_free_tagset(dev);
 	if (dev->ctrl.admin_q)
 		blk_put_queue(dev->ctrl.admin_q);
 	kfree(dev->queues);
@@ -2508,6 +2519,11 @@ static void nvme_pci_free_ctrl(struct nvme_ctrl *ctrl)
 
 static void nvme_remove_dead_ctrl(struct nvme_dev *dev)
 {
+	/*
+	 * Set state to deleting now to avoid blocking nvme_wait_reset(), which
+	 * may be holding this pci_dev's device lock.
+	 */
+	nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_DELETING);
 	nvme_get_ctrl(&dev->ctrl);
 	nvme_dev_disable(dev, false);
 	nvme_kill_queues(&dev->ctrl);
@@ -2521,7 +2537,6 @@ static void nvme_reset_work(struct work_struct *work)
 		container_of(work, struct nvme_dev, ctrl.reset_work);
 	bool was_suspend = !!(dev->ctrl.ctrl_config & NVME_CC_SHN_NORMAL);
 	int result;
-	enum nvme_ctrl_state new_state = NVME_CTRL_LIVE;
 
 	if (WARN_ON(dev->ctrl.state != NVME_CTRL_RESETTING)) {
 		result = -ENODEV;
@@ -2615,13 +2630,11 @@ static void nvme_reset_work(struct work_struct *work)
 		dev_warn(dev->ctrl.device, "IO queues not created\n");
 		nvme_kill_queues(&dev->ctrl);
 		nvme_remove_namespaces(&dev->ctrl);
-		new_state = NVME_CTRL_ADMIN_ONLY;
+		nvme_free_tagset(dev);
 	} else {
 		nvme_start_queues(&dev->ctrl);
 		nvme_wait_freeze(&dev->ctrl);
-		/* hit this only when allocate tagset fails */
-		if (nvme_dev_add(dev))
-			new_state = NVME_CTRL_ADMIN_ONLY;
+		nvme_dev_add(dev);
 		nvme_unfreeze(&dev->ctrl);
 	}
 
@@ -2629,9 +2642,9 @@ static void nvme_reset_work(struct work_struct *work)
 	 * If only admin queue live, keep it to do further investigation or
 	 * recovery.
 	 */
-	if (!nvme_change_ctrl_state(&dev->ctrl, new_state)) {
+	if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_LIVE)) {
 		dev_warn(dev->ctrl.device,
-			"failed to mark controller state %d\n", new_state);
+			"failed to mark controller live state\n");
 		result = -ENODEV;
 		goto out;
 	}
@@ -2672,7 +2685,7 @@ static int nvme_pci_reg_write32(struct nvme_ctrl *ctrl, u32 off, u32 val)
 
 static int nvme_pci_reg_read64(struct nvme_ctrl *ctrl, u32 off, u64 *val)
 {
-	*val = readq(to_nvme_dev(ctrl)->bar + off);
+	*val = lo_hi_readq(to_nvme_dev(ctrl)->bar + off);
 	return 0;
 }
 
@@ -2836,19 +2849,28 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 static void nvme_reset_prepare(struct pci_dev *pdev)
 {
 	struct nvme_dev *dev = pci_get_drvdata(pdev);
-	nvme_dev_disable(dev, false);
+
+	/*
+	 * We don't need to check the return value from waiting for the reset
+	 * state as pci_dev device lock is held, making it impossible to race
+	 * with ->remove().
+	 */
+	nvme_disable_prepare_reset(dev, false);
+	nvme_sync_queues(&dev->ctrl);
 }
 
 static void nvme_reset_done(struct pci_dev *pdev)
 {
 	struct nvme_dev *dev = pci_get_drvdata(pdev);
-	nvme_reset_ctrl_sync(&dev->ctrl);
+
+	if (!nvme_try_sched_reset(&dev->ctrl))
+		flush_work(&dev->ctrl.reset_work);
 }
 
 static void nvme_shutdown(struct pci_dev *pdev)
 {
 	struct nvme_dev *dev = pci_get_drvdata(pdev);
-	nvme_dev_disable(dev, true);
+	nvme_disable_prepare_reset(dev, true);
 }
 
 /*
@@ -2901,7 +2923,7 @@ static int nvme_resume(struct device *dev)
 
 	if (ndev->last_ps == U32_MAX ||
 	    nvme_set_power_state(ctrl, ndev->last_ps) != 0)
-		nvme_reset_ctrl(ctrl);
+		return nvme_try_sched_reset(&ndev->ctrl);
 	return 0;
 }
 
@@ -2929,17 +2951,14 @@ static int nvme_suspend(struct device *dev)
 	 */
 	if (pm_suspend_via_firmware() || !ctrl->npss ||
 	    !pcie_aspm_enabled(pdev) ||
-	    (ndev->ctrl.quirks & NVME_QUIRK_SIMPLE_SUSPEND)) {
-		nvme_dev_disable(ndev, true);
-		return 0;
-	}
+	    (ndev->ctrl.quirks & NVME_QUIRK_SIMPLE_SUSPEND))
+		return nvme_disable_prepare_reset(ndev, true);
 
 	nvme_start_freeze(ctrl);
 	nvme_wait_freeze(ctrl);
 	nvme_sync_queues(ctrl);
 
-	if (ctrl->state != NVME_CTRL_LIVE &&
-	    ctrl->state != NVME_CTRL_ADMIN_ONLY)
+	if (ctrl->state != NVME_CTRL_LIVE)
 		goto unfreeze;
 
 	ret = nvme_get_power_state(ctrl, &ndev->last_ps);
@@ -2965,9 +2984,8 @@ static int nvme_suspend(struct device *dev)
 		 * Clearing npss forces a controller reset on resume. The
 		 * correct value will be resdicovered then.
 		 */
-		nvme_dev_disable(ndev, true);
+		ret = nvme_disable_prepare_reset(ndev, true);
 		ctrl->npss = 0;
-		ret = 0;
 	}
 unfreeze:
 	nvme_unfreeze(ctrl);
@@ -2977,9 +2995,7 @@ static int nvme_suspend(struct device *dev)
 static int nvme_simple_suspend(struct device *dev)
 {
 	struct nvme_dev *ndev = pci_get_drvdata(to_pci_dev(dev));
-
-	nvme_dev_disable(ndev, true);
-	return 0;
+	return nvme_disable_prepare_reset(ndev, true);
 }
 
 static int nvme_simple_resume(struct device *dev)
@@ -2987,8 +3003,7 @@ static int nvme_simple_resume(struct device *dev)
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct nvme_dev *ndev = pci_get_drvdata(pdev);
 
-	nvme_reset_ctrl(&ndev->ctrl);
-	return 0;
+	return nvme_try_sched_reset(&ndev->ctrl);
 }
 
 static const struct dev_pm_ops nvme_dev_pm_ops = {

drivers/nvme/host/rdma.c

@@ -1701,6 +1701,14 @@ nvme_rdma_timeout(struct request *rq, bool reserved)
 	dev_warn(ctrl->ctrl.device, "I/O %d QID %d timeout\n",
 		 rq->tag, nvme_rdma_queue_idx(queue));
 
+	/*
+	 * Restart the timer if a controller reset is already scheduled. Any
+	 * timed out commands would be handled before entering the connecting
+	 * state.
+	 */
+	if (ctrl->ctrl.state == NVME_CTRL_RESETTING)
+		return BLK_EH_RESET_TIMER;
+
 	if (ctrl->ctrl.state != NVME_CTRL_LIVE) {
 		/*
 		 * Teardown immediately if controller times out while starting

drivers/nvme/host/tcp.c

@@ -1386,7 +1386,9 @@ static int nvme_tcp_alloc_queue(struct nvme_ctrl *nctrl,
 	queue->sock->sk->sk_data_ready = nvme_tcp_data_ready;
 	queue->sock->sk->sk_state_change = nvme_tcp_state_change;
 	queue->sock->sk->sk_write_space = nvme_tcp_write_space;
+#ifdef CONFIG_NET_RX_BUSY_POLL
 	queue->sock->sk->sk_ll_usec = 1;
+#endif
 	write_unlock_bh(&queue->sock->sk->sk_callback_lock);
 
 	return 0;
@@ -2044,6 +2046,14 @@ nvme_tcp_timeout(struct request *rq, bool reserved)
 	struct nvme_tcp_ctrl *ctrl = req->queue->ctrl;
 	struct nvme_tcp_cmd_pdu *pdu = req->pdu;
 
+	/*
+	 * Restart the timer if a controller reset is already scheduled. Any
+	 * timed out commands would be handled before entering the connecting
+	 * state.
+	 */
+	if (ctrl->ctrl.state == NVME_CTRL_RESETTING)
+		return BLK_EH_RESET_TIMER;
+
 	dev_warn(ctrl->ctrl.device,
 		"queue %d: timeout request %#x type %d\n",
 		nvme_tcp_queue_id(req->queue), rq->tag, pdu->hdr.type);
@@ -2126,6 +2136,7 @@ static blk_status_t nvme_tcp_setup_cmd_pdu(struct nvme_ns *ns,
 
 	ret = nvme_tcp_map_data(queue, rq);
 	if (unlikely(ret)) {
+		nvme_cleanup_cmd(rq);
 		dev_err(queue->ctrl->ctrl.device,
 			"Failed to map data (%d)\n", ret);
 		return ret;

drivers/nvme/target/loop.c

@@ -157,8 +157,10 @@ static blk_status_t nvme_loop_queue_rq(struct blk_mq_hw_ctx *hctx,
 		iod->sg_table.sgl = iod->first_sgl;
 		if (sg_alloc_table_chained(&iod->sg_table,
 				blk_rq_nr_phys_segments(req),
-				iod->sg_table.sgl, SG_CHUNK_SIZE))
+				iod->sg_table.sgl, SG_CHUNK_SIZE)) {
+			nvme_cleanup_cmd(req);
 			return BLK_STS_RESOURCE;
+		}
 
 		iod->req.sg = iod->sg_table.sgl;
 		iod->req.sg_cnt = blk_rq_map_sg(req->q, req, iod->sg_table.sgl);

fs/io_uring.c

@@ -322,6 +322,8 @@ struct io_kiocb {
 #define REQ_F_FAIL_LINK		256	/* fail rest of links */
 #define REQ_F_SHADOW_DRAIN	512	/* link-drain shadow req */
 #define REQ_F_TIMEOUT		1024	/* timeout request */
+#define REQ_F_ISREG		2048	/* regular file */
+#define REQ_F_MUST_PUNT		4096	/* must be punted even for NONBLOCK */
 	u64			user_data;
 	u32			result;
 	u32			sequence;
@@ -914,26 +916,26 @@ static int io_iopoll_check(struct io_ring_ctx *ctx, unsigned *nr_events,
 	return ret;
 }
 
-static void kiocb_end_write(struct kiocb *kiocb)
+static void kiocb_end_write(struct io_kiocb *req)
 {
-	if (kiocb->ki_flags & IOCB_WRITE) {
-		struct inode *inode = file_inode(kiocb->ki_filp);
-
 	/*
 	 * Tell lockdep we inherited freeze protection from submission
 	 * thread.
 	 */
-		if (S_ISREG(inode->i_mode))
+	if (req->flags & REQ_F_ISREG) {
+		struct inode *inode = file_inode(req->file);
+
 		__sb_writers_acquired(inode->i_sb, SB_FREEZE_WRITE);
-		file_end_write(kiocb->ki_filp);
 	}
+	file_end_write(req->file);
 }
 
 static void io_complete_rw(struct kiocb *kiocb, long res, long res2)
 {
 	struct io_kiocb *req = container_of(kiocb, struct io_kiocb, rw);
 
-	kiocb_end_write(kiocb);
+	if (kiocb->ki_flags & IOCB_WRITE)
+		kiocb_end_write(req);
 
 	if ((req->flags & REQ_F_LINK) && res != req->result)
 		req->flags |= REQ_F_FAIL_LINK;
@@ -945,7 +947,8 @@ static void io_complete_rw_iopoll(struct kiocb *kiocb, long res, long res2)
 {
 	struct io_kiocb *req = container_of(kiocb, struct io_kiocb, rw);
 
-	kiocb_end_write(kiocb);
+	if (kiocb->ki_flags & IOCB_WRITE)
+		kiocb_end_write(req);
 
 	if ((req->flags & REQ_F_LINK) && res != req->result)
 		req->flags |= REQ_F_FAIL_LINK;
@@ -1059,8 +1062,17 @@ static int io_prep_rw(struct io_kiocb *req, const struct sqe_submit *s,
 	if (!req->file)
 		return -EBADF;
 
-	if (force_nonblock && !io_file_supports_async(req->file))
-		force_nonblock = false;
+	if (S_ISREG(file_inode(req->file)->i_mode))
+		req->flags |= REQ_F_ISREG;
+
+	/*
+	 * If the file doesn't support async, mark it as REQ_F_MUST_PUNT so
+	 * we know to async punt it even if it was opened O_NONBLOCK
+	 */
+	if (force_nonblock && !io_file_supports_async(req->file)) {
+		req->flags |= REQ_F_MUST_PUNT;
+		return -EAGAIN;
+	}
 
 	kiocb->ki_pos = READ_ONCE(sqe->off);
 	kiocb->ki_flags = iocb_flags(kiocb->ki_filp);
@@ -1081,7 +1093,8 @@ static int io_prep_rw(struct io_kiocb *req, const struct sqe_submit *s,
 		return ret;
 
 	/* don't allow async punt if RWF_NOWAIT was requested */
-	if (kiocb->ki_flags & IOCB_NOWAIT)
+	if ((kiocb->ki_flags & IOCB_NOWAIT) ||
+	    (req->file->f_flags & O_NONBLOCK))
 		req->flags |= REQ_F_NOWAIT;
 
 	if (force_nonblock)
@@ -1382,7 +1395,9 @@ static int io_read(struct io_kiocb *req, const struct sqe_submit *s,
 		 * need async punt anyway, so it's more efficient to do it
 		 * here.
 		 */
-		if (force_nonblock && ret2 > 0 && ret2 < read_size)
+		if (force_nonblock && !(req->flags & REQ_F_NOWAIT) &&
+		    (req->flags & REQ_F_ISREG) &&
+		    ret2 > 0 && ret2 < read_size)
 			ret2 = -EAGAIN;
 		/* Catch -EAGAIN return for forced non-blocking submission */
 		if (!force_nonblock || ret2 != -EAGAIN) {
@@ -1447,7 +1462,7 @@ static int io_write(struct io_kiocb *req, const struct sqe_submit *s,
 		 * released so that it doesn't complain about the held lock when
 		 * we return to userspace.
 		 */
-		if (S_ISREG(file_inode(file)->i_mode)) {
+		if (req->flags & REQ_F_ISREG) {
 			__sb_start_write(file_inode(file)->i_sb,
 						SB_FREEZE_WRITE, true);
 			__sb_writers_release(file_inode(file)->i_sb,
@@ -1884,7 +1899,7 @@ static enum hrtimer_restart io_timeout_fn(struct hrtimer *timer)
 
 static int io_timeout(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 {
-	unsigned count, req_dist, tail_index;
+	unsigned count;
 	struct io_ring_ctx *ctx = req->ctx;
 	struct list_head *entry;
 	struct timespec64 ts;
@@ -1907,21 +1922,36 @@ static int io_timeout(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 		count = 1;
 
 	req->sequence = ctx->cached_sq_head + count - 1;
+	/* reuse it to store the count */
+	req->submit.sequence = count;
 	req->flags |= REQ_F_TIMEOUT;
 
 	/*
 	 * Insertion sort, ensuring the first entry in the list is always
 	 * the one we need first.
 	 */
-	tail_index = ctx->cached_cq_tail - ctx->rings->sq_dropped;
-	req_dist = req->sequence - tail_index;
 	spin_lock_irq(&ctx->completion_lock);
 	list_for_each_prev(entry, &ctx->timeout_list) {
 		struct io_kiocb *nxt = list_entry(entry, struct io_kiocb, list);
-		unsigned dist;
+		unsigned nxt_sq_head;
+		long long tmp, tmp_nxt;
 
-		dist = nxt->sequence - tail_index;
-		if (req_dist >= dist)
+		/*
+		 * Since cached_sq_head + count - 1 can overflow, use type long
+		 * long to store it.
+		 */
+		tmp = (long long)ctx->cached_sq_head + count - 1;
+		nxt_sq_head = nxt->sequence - nxt->submit.sequence + 1;
+		tmp_nxt = (long long)nxt_sq_head + nxt->submit.sequence - 1;
+
+		/*
+		 * cached_sq_head may overflow, and it will never overflow twice
+		 * once there is some timeout req still be valid.
+		 */
+		if (ctx->cached_sq_head < nxt_sq_head)
+			tmp += UINT_MAX;
+
+		if (tmp >= tmp_nxt)
 			break;
 	}
 	list_add(&req->list, entry);
@@ -2267,7 +2297,13 @@ static int __io_queue_sqe(struct io_ring_ctx *ctx, struct io_kiocb *req,
 	int ret;
 
 	ret = __io_submit_sqe(ctx, req, s, force_nonblock);
-	if (ret == -EAGAIN && !(req->flags & REQ_F_NOWAIT)) {
+
+	/*
+	 * We async punt it if the file wasn't marked NOWAIT, or if the file
+	 * doesn't support non-blocking read/write attempts
+	 */
+	if (ret == -EAGAIN && (!(req->flags & REQ_F_NOWAIT) ||
+	    (req->flags & REQ_F_MUST_PUNT))) {
 		struct io_uring_sqe *sqe_copy;
 
 		sqe_copy = kmemdup(s->sqe, sizeof(*sqe_copy), GFP_KERNEL);