[PATCH V6 07/11] nvme: pci: prepare for supporting error recovery from resetting context

[Ming Lei]

Either the admin or normal IO in reset context may be timed out because
controller error happens. When this timeout happens, we may have to
start controller recovery again.

This patch introduces 'reset_lock' and holds this lock when running reset,
so that we may support nested reset in the following patches.

Cc: James Smart <james.smart at broadcom.com>
Cc: Jianchao Wang <jianchao.w.wang at oracle.com>
Cc: Christoph Hellwig <hch at lst.de>
Cc: Sagi Grimberg <sagi at grimberg.me>
Cc: linux-nvme at lists.infradead.org
Cc: Laurence Oberman <loberman at redhat.com>
Signed-off-by: Ming Lei <ming.lei at redhat.com>
---
 drivers/nvme/host/core.c |  2 ++
 drivers/nvme/host/nvme.h |  3 +++
 drivers/nvme/host/pci.c  | 20 +++++++++++++++++---
 3 files changed, 22 insertions(+), 3 deletions(-)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 99b857e5a7a9..3b0cf2fd3f53 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -3430,6 +3430,8 @@ int nvme_init_ctrl(struct nvme_ctrl *ctrl, struct device *dev,
 	INIT_WORK(&ctrl->fw_act_work, nvme_fw_act_work);
 	INIT_WORK(&ctrl->delete_work, nvme_delete_ctrl_work);
 
+	mutex_init(&ctrl->reset_lock);
+
 	ret = ida_simple_get(&nvme_instance_ida, 0, 0, GFP_KERNEL);
 	if (ret < 0)
 		goto out;
diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
index 17d2f7cf3fed..b00a56412bab 100644
--- a/drivers/nvme/host/nvme.h
+++ b/drivers/nvme/host/nvme.h
@@ -151,6 +151,9 @@ struct nvme_ctrl {
 	struct device ctrl_device;
 	struct device *device;	/* char device */
 	struct cdev cdev;
+
+	/* sync reset activities */
+	struct mutex reset_lock;
 	struct work_struct reset_work;
 	struct work_struct delete_work;
 
diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
index 365d1a5ee1eb..50bd1818deb5 100644
--- a/drivers/nvme/host/pci.c
+++ b/drivers/nvme/host/pci.c
@@ -2391,14 +2391,14 @@ static void nvme_remove_dead_ctrl(struct nvme_dev *dev, int status)
 		nvme_put_ctrl(&dev->ctrl);
 }
 
-static void nvme_reset_work(struct work_struct *work)
+static void nvme_reset_dev(struct nvme_dev *dev)
 {
-	struct nvme_dev *dev =
-		container_of(work, struct nvme_dev, ctrl.reset_work);
 	bool was_suspend = !!(dev->ctrl.ctrl_config & NVME_CC_SHN_NORMAL);
 	int result = -ENODEV;
 	enum nvme_ctrl_state new_state = NVME_CTRL_LIVE;
 
+	lockdep_assert_held(&dev->ctrl.reset_lock);
+
 	if (WARN_ON(dev->ctrl.state != NVME_CTRL_RESETTING))
 		goto out;
 
@@ -2474,7 +2474,11 @@ static void nvme_reset_work(struct work_struct *work)
 		new_state = NVME_CTRL_ADMIN_ONLY;
 	} else {
 		nvme_start_queues(&dev->ctrl);
+		mutex_unlock(&dev->ctrl.reset_lock);
+
 		nvme_wait_freeze(&dev->ctrl);
+
+		mutex_lock(&dev->ctrl.reset_lock);
 		/* hit this only when allocate tagset fails */
 		if (nvme_dev_add(dev))
 			new_state = NVME_CTRL_ADMIN_ONLY;
@@ -2498,6 +2502,16 @@ static void nvme_reset_work(struct work_struct *work)
 	nvme_remove_dead_ctrl(dev, result);
 }
 
+static void nvme_reset_work(struct work_struct *work)
+{
+	struct nvme_dev *dev =
+		container_of(work, struct nvme_dev, ctrl.reset_work);
+
+	mutex_lock(&dev->ctrl.reset_lock);
+	nvme_reset_dev(dev);
+	mutex_unlock(&dev->ctrl.reset_lock);
+}
+
 static void nvme_remove_dead_ctrl_work(struct work_struct *work)
 {
 	struct nvme_dev *dev = container_of(work, struct nvme_dev, remove_work);
-- 
2.9.5