Patch "RDMA/ipoib: Fix ABBA deadlock with ipoib_reap_ah()" has been added to the 5.8-stable tree

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



This is a note to let you know that I've just added the patch titled

    RDMA/ipoib: Fix ABBA deadlock with ipoib_reap_ah()

to the 5.8-stable tree which can be found at:
    http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
     rdma-ipoib-fix-abba-deadlock-with-ipoib_reap_ah.patch
and it can be found in the queue-5.8 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.



commit 76fbd441ddafc287520d8edc7fc10048cdb5ae50
Author: Jason Gunthorpe <jgg@xxxxxxxxxx>
Date:   Thu Jun 25 20:42:19 2020 +0300

    RDMA/ipoib: Fix ABBA deadlock with ipoib_reap_ah()
    
    [ Upstream commit 65936bf25f90fe440bb2d11624c7d10fab266639 ]
    
    ipoib_mcast_carrier_on_task() insanely open codes a rtnl_lock() such that
    the only time flush_workqueue() can be called is if it also clears
    IPOIB_FLAG_OPER_UP.
    
    Thus the flush inside ipoib_flush_ah() will deadlock if it gets unlucky
    enough, and lockdep doesn't help us to find it early:
    
              CPU0               CPU1          CPU2
       __ipoib_ib_dev_flush()
          down_read(vlan_rwsem)
    
                             ipoib_vlan_add()
                               rtnl_trylock()
                               down_write(vlan_rwsem)
    
                                          ipoib_mcast_carrier_on_task()
                                             while (!rtnl_trylock())
                                                  msleep(20);
    
          ipoib_flush_ah()
            flush_workqueue(priv->wq)
    
    Clean up the ah_reaper related functions and lifecycle to make sense:
    
     - Start/Stop of the reaper should only be done in open/stop NDOs, not in
       any other places
    
     - cancel and flush of the reaper should only happen in the stop NDO.
       cancel is only functional when combined with IPOIB_STOP_REAPER.
    
     - Non-stop places were flushing the AH's just need to flush out dead AH's
       synchronously and ignore the background task completely. It is fully
       locked and harmless to leave running.
    
    Which ultimately fixes the ABBA deadlock by removing the unnecessary
    flush_workqueue() from the problematic place under the vlan_rwsem.
    
    Fixes: efc82eeeae4e ("IB/ipoib: No longer use flush as a parameter")
    Link: https://lore.kernel.org/r/20200625174219.290842-1-kamalheib1@xxxxxxxxx
    Reported-by: Kamal Heib <kheib@xxxxxxxxxx>
    Tested-by: Kamal Heib <kheib@xxxxxxxxxx>
    Signed-off-by: Jason Gunthorpe <jgg@xxxxxxxxxx>
    Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index 6ee64c25aaff4..494f413dc3c6c 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -670,13 +670,12 @@ int ipoib_send(struct net_device *dev, struct sk_buff *skb,
 	return rc;
 }
 
-static void __ipoib_reap_ah(struct net_device *dev)
+static void ipoib_reap_dead_ahs(struct ipoib_dev_priv *priv)
 {
-	struct ipoib_dev_priv *priv = ipoib_priv(dev);
 	struct ipoib_ah *ah, *tah;
 	unsigned long flags;
 
-	netif_tx_lock_bh(dev);
+	netif_tx_lock_bh(priv->dev);
 	spin_lock_irqsave(&priv->lock, flags);
 
 	list_for_each_entry_safe(ah, tah, &priv->dead_ahs, list)
@@ -687,37 +686,37 @@ static void __ipoib_reap_ah(struct net_device *dev)
 		}
 
 	spin_unlock_irqrestore(&priv->lock, flags);
-	netif_tx_unlock_bh(dev);
+	netif_tx_unlock_bh(priv->dev);
 }
 
 void ipoib_reap_ah(struct work_struct *work)
 {
 	struct ipoib_dev_priv *priv =
 		container_of(work, struct ipoib_dev_priv, ah_reap_task.work);
-	struct net_device *dev = priv->dev;
 
-	__ipoib_reap_ah(dev);
+	ipoib_reap_dead_ahs(priv);
 
 	if (!test_bit(IPOIB_STOP_REAPER, &priv->flags))
 		queue_delayed_work(priv->wq, &priv->ah_reap_task,
 				   round_jiffies_relative(HZ));
 }
 
-static void ipoib_flush_ah(struct net_device *dev)
+static void ipoib_start_ah_reaper(struct ipoib_dev_priv *priv)
 {
-	struct ipoib_dev_priv *priv = ipoib_priv(dev);
-
-	cancel_delayed_work(&priv->ah_reap_task);
-	flush_workqueue(priv->wq);
-	ipoib_reap_ah(&priv->ah_reap_task.work);
+	clear_bit(IPOIB_STOP_REAPER, &priv->flags);
+	queue_delayed_work(priv->wq, &priv->ah_reap_task,
+			   round_jiffies_relative(HZ));
 }
 
-static void ipoib_stop_ah(struct net_device *dev)
+static void ipoib_stop_ah_reaper(struct ipoib_dev_priv *priv)
 {
-	struct ipoib_dev_priv *priv = ipoib_priv(dev);
-
 	set_bit(IPOIB_STOP_REAPER, &priv->flags);
-	ipoib_flush_ah(dev);
+	cancel_delayed_work(&priv->ah_reap_task);
+	/*
+	 * After ipoib_stop_ah_reaper() we always go through
+	 * ipoib_reap_dead_ahs() which ensures the work is really stopped and
+	 * does a final flush out of the dead_ah's list
+	 */
 }
 
 static int recvs_pending(struct net_device *dev)
@@ -846,16 +845,6 @@ int ipoib_ib_dev_stop_default(struct net_device *dev)
 	return 0;
 }
 
-void ipoib_ib_dev_stop(struct net_device *dev)
-{
-	struct ipoib_dev_priv *priv = ipoib_priv(dev);
-
-	priv->rn_ops->ndo_stop(dev);
-
-	clear_bit(IPOIB_FLAG_INITIALIZED, &priv->flags);
-	ipoib_flush_ah(dev);
-}
-
 int ipoib_ib_dev_open_default(struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = ipoib_priv(dev);
@@ -899,10 +888,7 @@ int ipoib_ib_dev_open(struct net_device *dev)
 		return -1;
 	}
 
-	clear_bit(IPOIB_STOP_REAPER, &priv->flags);
-	queue_delayed_work(priv->wq, &priv->ah_reap_task,
-			   round_jiffies_relative(HZ));
-
+	ipoib_start_ah_reaper(priv);
 	if (priv->rn_ops->ndo_open(dev)) {
 		pr_warn("%s: Failed to open dev\n", dev->name);
 		goto dev_stop;
@@ -913,13 +899,20 @@ int ipoib_ib_dev_open(struct net_device *dev)
 	return 0;
 
 dev_stop:
-	set_bit(IPOIB_STOP_REAPER, &priv->flags);
-	cancel_delayed_work(&priv->ah_reap_task);
-	set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags);
-	ipoib_ib_dev_stop(dev);
+	ipoib_stop_ah_reaper(priv);
 	return -1;
 }
 
+void ipoib_ib_dev_stop(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	priv->rn_ops->ndo_stop(dev);
+
+	clear_bit(IPOIB_FLAG_INITIALIZED, &priv->flags);
+	ipoib_stop_ah_reaper(priv);
+}
+
 void ipoib_pkey_dev_check_presence(struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = ipoib_priv(dev);
@@ -1230,7 +1223,7 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv,
 		ipoib_mcast_dev_flush(dev);
 		if (oper_up)
 			set_bit(IPOIB_FLAG_OPER_UP, &priv->flags);
-		ipoib_flush_ah(dev);
+		ipoib_reap_dead_ahs(priv);
 	}
 
 	if (level >= IPOIB_FLUSH_NORMAL)
@@ -1305,7 +1298,7 @@ void ipoib_ib_dev_cleanup(struct net_device *dev)
 	 * the neighbor garbage collection is stopped and reaped.
 	 * That should all be done now, so make a final ah flush.
 	 */
-	ipoib_stop_ah(dev);
+	ipoib_reap_dead_ahs(priv);
 
 	clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
 
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 3cfb682b91b0a..ef60e8e4ae67b 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -1976,6 +1976,8 @@ static void ipoib_ndo_uninit(struct net_device *dev)
 
 	/* no more works over the priv->wq */
 	if (priv->wq) {
+		/* See ipoib_mcast_carrier_on_task() */
+		WARN_ON(test_bit(IPOIB_FLAG_OPER_UP, &priv->flags));
 		flush_workqueue(priv->wq);
 		destroy_workqueue(priv->wq);
 		priv->wq = NULL;



[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux