[BNX2]: Fix rtnl deadlock in bnx2_close (afdc08b9) · Commits · 戴 / test

This fixes an rtnl deadlock problem when flush_scheduled_work() is called from bnx2_close(). In rare cases, linkwatch_event() may be on the workqueue from a previous close of a different device and it will try to get the rtnl lock which is already held by dev_close(). The fix is to set a flag if we are in the reset task which is run from the workqueue. bnx2_close() will loop until the flag is cleared. As suggested by Jeff Garzik, the loop is changed to call msleep(1) instead of yield() in the original patch. flush_scheduled_work() is also moved to bnx2_remove_one() before the netdev is freed. Signed-off-by:

Michael Chan <mchan@broadcom.com> Signed-off-by:

David S. Miller <davem@davemloft.net>

drivers/net/bnx2.c

+14 −1

Original line number	Diff line number	Diff line
		@@ -3975,12 +3975,17 @@ bnx2_reset_task(void *data)
		{
		struct bnx2 *bp = data;

		if (!netif_running(bp->dev))
		return;

		bp->in_reset_task = 1;
		bnx2_netif_stop(bp);

		bnx2_init_nic(bp);

		atomic_set(&bp->intr_sem, 1);
		bnx2_netif_start(bp);
		bp->in_reset_task = 0;
		}

		static void
		@@ -4172,7 +4177,13 @@ bnx2_close(struct net_device *dev)
		struct bnx2 *bp = dev->priv;
		u32 reset_code;

		flush_scheduled_work();
		/* Calling flush_scheduled_work() may deadlock because
		* linkwatch_event() may be on the workqueue and it will try to get
		* the rtnl_lock which we are holding.
		*/
		while (bp->in_reset_task)
		msleep(1);

		bnx2_netif_stop(bp);
		del_timer_sync(&bp->timer);
		if (bp->wol)
		@@ -5453,6 +5464,8 @@ bnx2_remove_one(struct pci_dev *pdev)
		struct net_device *dev = pci_get_drvdata(pdev);
		struct bnx2 *bp = dev->priv;

		flush_scheduled_work();

		unregister_netdev(dev);

		if (bp->regview)

drivers/net/bnx2.h

+1 −0

Original line number	Diff line number	Diff line
		@@ -3874,6 +3874,7 @@ struct bnx2 {
		int timer_interval;
		struct timer_list timer;
		struct work_struct reset_task;
		int in_reset_task;

		/* Used to synchronize phy accesses. */
		spinlock_t phy_lock;

Admin message