Skip to contentSkip to conversation

nvme-pci: Fix rapid add remove sequence

PATCH New Keith Busch 4 participants 5 messages Jan 24, 2019 – Feb 5, 2019 NVMe

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 nvme-pci: Fix rapid add remove sequence

A surprise removal may fail to tear down request queues if it is racing with the initial asynchronous probe. If that happens, the remove path won't see the queue resources to tear down, and the controller reset path may create a new request queue on a removed device, but will not be able to make forward progress, deadlocking the pci removal.

Protect setting up non-blocking resources from a shutdown by holding the same mutex, and transition to the CONNECTING state after these resources are initialized so the probe path may see the dead controller state before dispatching new IO.

Reported-by Alex Gagniuc <Alex_Gagniuc at Dellteam.com>
Signed-off-by Keith Busch <keith.busch at intel.com>
---
 drivers/nvme/host/pci.c | 22 ++++++++++++----------
 1 file changed, 12 insertions(+), 10 deletions(-)
drivers/nvme/host/pci.c +12 -10
index 9bc585415d9b..022ea1ee63f8 100644
@@ -2557,16 +2557,7 @@ static void nvme_reset_work(struct work_struct *work)
 	if (dev->ctrl.ctrl_config & NVME_CC_ENABLE)
 		nvme_dev_disable(dev, false);
 
-	/*
-	 * Introduce CONNECTING state from nvme-fc/rdma transports to mark the
-	 * initializing procedure here.
-	 */
-	if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_CONNECTING)) {
-		dev_warn(dev->ctrl.device,
-			"failed to mark controller CONNECTING\n");
-		goto out;
-	}
-
+	mutex_lock(&dev->shutdown_lock);
 	result = nvme_pci_enable(dev);
 	if (result)
 		goto out;
@@ -2585,6 +2576,17 @@ static void nvme_reset_work(struct work_struct *work)
 	 */
 	dev->ctrl.max_hw_sectors = NVME_MAX_KB_SZ << 1;
 	dev->ctrl.max_segments = NVME_MAX_SEGS;
+	mutex_unlock(&dev->shutdown_lock);
+
+	/*
+	 * Introduce CONNECTING state from nvme-fc/rdma transports to mark the
+	 * initializing procedure here.
+	 */
+	if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_CONNECTING)) {
+		dev_warn(dev->ctrl.device,
+			"failed to mark controller CONNECTING\n");
+		goto out;
+	}
 
 	result = nvme_init_identify(&dev->ctrl);
 	if (result)
Signature · 1 line
2.14.4
#1 lore raw

PATCH 1/1 nvme-pci: Fix rapid add remove sequence

Quoted 6 lines · Keith Busch

A surprise removal may fail to tear down request queues if it is racing with the initial asynchronous probe. If that happens, the remove path won't see the queue resources to tear down, and the controller reset path may create a new request queue on a removed device, but will not be able to make forward progress, deadlocking the pci removal.

Quoted 8 lines

Protect setting up non-blocking resources from a shutdown by holding the same mutex, and transition to the CONNECTING state after these resources are initialized so the probe path may see the dead controller state before dispatching new IO.

Reported-by Alex Gagniuc <Alex_Gagniuc at Dellteam.com>
Signed-off-by Keith Busch <keith.busch at intel.com>
Tested-by Alex Gagniuc <mr.nuke.me at gmail.com>
Quoted 45 lines
---
  drivers/nvme/host/pci.c | 22 ++++++++++++----------
  1 file changed, 12 insertions(+), 10 deletions(-)
drivers/nvme/host/pci.c +12 -10
index 9bc585415d9b..022ea1ee63f8 100644
@@ -2557,16 +2557,7 @@ static void nvme_reset_work(struct work_struct *work)
  	if (dev->ctrl.ctrl_config & NVME_CC_ENABLE)
  		nvme_dev_disable(dev, false);
 
-	/*
-	 * Introduce CONNECTING state from nvme-fc/rdma transports to mark the
-	 * initializing procedure here.
-	 */
-	if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_CONNECTING)) {
-		dev_warn(dev->ctrl.device,
-			"failed to mark controller CONNECTING\n");
-		goto out;
-	}
-
+	mutex_lock(&dev->shutdown_lock);
  	result = nvme_pci_enable(dev);
  	if (result)
  		goto out;
@@ -2585,6 +2576,17 @@ static void nvme_reset_work(struct work_struct *work)
  	 */
  	dev->ctrl.max_hw_sectors = NVME_MAX_KB_SZ << 1;
  	dev->ctrl.max_segments = NVME_MAX_SEGS;
+	mutex_unlock(&dev->shutdown_lock);
+
+	/*
+	 * Introduce CONNECTING state from nvme-fc/rdma transports to mark the
+	 * initializing procedure here.
+	 */
+	if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_CONNECTING)) {
+		dev_warn(dev->ctrl.device,
+			"failed to mark controller CONNECTING\n");
+		goto out;
+	}
 
  	result = nvme_init_identify(&dev->ctrl);
  	if (result)
maintainer
#2 lore raw

PATCH 1/1 nvme-pci: Fix rapid add remove sequence

In reply to Keith Busch
Quoted 6 lines

A surprise removal may fail to tear down request queues if it is racing with the initial asynchronous probe. If that happens, the remove path won't see the queue resources to tear down, and the controller reset path may create a new request queue on a removed device, but will not be able to make forward progress, deadlocking the pci removal.

Doesn't pci removal flush the reset work before making forward progress? Perhaps what is needed that it will flush it earlier instead of serializing with the shutdown lock?

#3 lore raw

PATCH 1/1 nvme-pci: Fix rapid add remove sequence

Quoted 10 lines · Sagi Grimberg
Quoted 6 lines

A surprise removal may fail to tear down request queues if it is racing with the initial asynchronous probe. If that happens, the remove path won't see the queue resources to tear down, and the controller reset path may create a new request queue on a removed device, but will not be able to make forward progress, deadlocking the pci removal.

Doesn't pci removal flush the reset work before making forward progress? Perhaps what is needed that it will flush it earlier instead of serializing with the shutdown lock?

Removal does flush reset work, but doesn't help this particular issue. It's pretty timing sensitive to trigger.

Before flushing reset on an surprise removal, we do an ungraceful device teardown first in order to unblock any IO that reset work is waiting on.

In this case that Alex discovered, though, the surprise removal happens just before the nvme driver has set up the admin and io tagsets, so removal doesn't find any tagsets to kill, and proceeds with flushing the reset work.

The reset work, though, just allocated brand new tagsets right after that, so it looks like they are good to use, so dispatches an admin command to a device that's gone.

You might expect the nvme_timeout() work to trigger 60 seconds later, but we can't use that when the pci device is not in a normal channel state. I wouldn't want to wait 60 seconds either, so the removal task needs to handle get things unblocked.