igb: close/suspend race in netif_device_detach

Similar to ixgbe, when an interface is part of a namespace it is possible that igb_close() may be called while __igb_shutdown() is running which ends up in a double free WARN and/or a BUG in free_msi_irqs().

Extend the rtnl_lock() to protect the call to netif_device_detach() and igb_clear_interrupt_scheme() in __igb_shutdown() and check for netif_device_present() to avoid calling igb_clear_interrupt_scheme() a second time in igb_close().

Also extend the rtnl lock in igb_resume() to netif_device_attach().

Signed-off-by Todd Fujinaka <[email protected]>

---
 drivers/net/ethernet/intel/igb/igb_main.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

drivers/net/ethernet/intel/igb/igb_main.c +12 -9

index 4feca69..8b1b4dd 100644
@@ -3275,7 +3275,9 @@ static int __igb_close(struct net_device *netdev, bool suspending)
 
 int igb_close(struct net_device *netdev)
 {
-	return __igb_close(netdev, false);
+	if (netif_device_present(netdev))
+		return __igb_close(netdev, false);
+	return 0;
 }
 
 /**

@@ -7537,6 +7539,7 @@ static int __igb_shutdown(struct pci_dev *pdev, bool *enable_wake,
 	int retval = 0;
 #endif
 
+	rtnl_lock();
 	netif_device_detach(netdev);
 
 	if (netif_running(netdev))

@@ -7545,6 +7548,7 @@ static int __igb_shutdown(struct pci_dev *pdev, bool *enable_wake,
 	igb_ptp_suspend(adapter);
 
 	igb_clear_interrupt_scheme(adapter);
+	rtnl_unlock();
 
 #ifdef CONFIG_PM
 	retval = pci_save_state(pdev);

@@ -7663,16 +7667,15 @@ static int igb_resume(struct device *dev)
 
 	wr32(E1000_WUS, ~0);
 
-	if (netdev->flags & IFF_UP) {
-		rtnl_lock();
+	rtnl_lock();
+	if (!err && netif_running(netdev))
 		err = __igb_open(netdev, true);
-		rtnl_unlock();
-		if (err)
-			return err;
-	}
 
-	netif_device_attach(netdev);
-	return 0;
+	if (!err)
+		netif_device_attach(netdev);
+	rtnl_unlock();
+
+	return err;
 }
 
 static int igb_runtime_idle(struct device *dev)

From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org] On Behalf Of Todd Fujinaka
Sent: Tuesday, November 15, 2016 8:54 AM
To: intel-wired-lan at lists.osuosl.org
Subject: [Intel-wired-lan] [PATCH] [PATCH v2] igb: close/suspend race in netif_device_detach

Similar to ixgbe, when an interface is part of a namespace it is possible that igb_close() may be called while __igb_shutdown() is running which ends up in a double free WARN and/or a BUG in free_msi_irqs().

Extend the rtnl_lock() to protect the call to netif_device_detach() and igb_clear_interrupt_scheme() in __igb_shutdown() and check for netif_device_present() to avoid calling igb_clear_interrupt_scheme() a second time in igb_close().

Also extend the rtnl lock in igb_resume() to netif_device_attach().

Signed-off-by Todd Fujinaka <[email protected]>

---
 drivers/net/ethernet/intel/igb/igb_main.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org] On Behalf Of Brown, Aaron F
Sent: Wednesday, November 30, 2016 12:26 PM
To: Fujinaka, Todd <[email protected]>; intel-wired-
lan at lists.osuosl.org
Subject: Re: [Intel-wired-lan] [PATCH] [PATCH v2] igb: close/suspend race in netif_device_detach

Quoted 17 lines

From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org] On Behalf Of Todd Fujinaka
Sent: Tuesday, November 15, 2016 8:54 AM
To: intel-wired-lan at lists.osuosl.org
Subject: [Intel-wired-lan] [PATCH] [PATCH v2] igb: close/suspend race in netif_device_detach

Similar to ixgbe, when an interface is part of a namespace it is possible that igb_close() may be called while __igb_shutdown() is running which ends up in a double free WARN and/or a BUG in free_msi_irqs().

Extend the rtnl_lock() to protect the call to netif_device_detach() and igb_clear_interrupt_scheme() in __igb_shutdown() and check for netif_device_present() to avoid calling igb_clear_interrupt_scheme() a second time in igb_close().

Quoted 6 lines

Also extend the rtnl lock in igb_resume() to netif_device_attach().

Signed-off-by Todd Fujinaka <[email protected]>

---
 drivers/net/ethernet/intel/igb/igb_main.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

Tested-by Aaron Brown <[email protected]>

From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org] On Behalf Of Brown, Aaron F
Sent: Tuesday, December 6, 2016 2:14 PM
To: Fujinaka, Todd <[email protected]>; intel-wired-
lan at lists.osuosl.org
Subject: Re: [Intel-wired-lan] [PATCH] [PATCH v2] igb: close/suspend race in netif_device_detach

Quoted 9 lines

From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org] On Behalf Of Brown, Aaron F
Sent: Wednesday, November 30, 2016 12:26 PM
To: Fujinaka, Todd <[email protected]>; intel-wired-
lan at lists.osuosl.org
Subject: Re: [Intel-wired-lan] [PATCH] [PATCH v2] igb: close/suspend race in netif_device_detach

Quoted 1 line

From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org]

On

Quoted 23 lines

Quoted 22 lines

Behalf Of Todd Fujinaka
Sent: Tuesday, November 15, 2016 8:54 AM
To: intel-wired-lan at lists.osuosl.org
Subject: [Intel-wired-lan] [PATCH] [PATCH v2] igb: close/suspend race in netif_device_detach

Similar to ixgbe, when an interface is part of a namespace it is possible that igb_close() may be called while __igb_shutdown() is running which ends up in a double free WARN and/or a BUG in free_msi_irqs().

Extend the rtnl_lock() to protect the call to netif_device_detach() and igb_clear_interrupt_scheme() in __igb_shutdown() and check for netif_device_present() to avoid calling igb_clear_interrupt_scheme() a second time in igb_close().

Also extend the rtnl lock in igb_resume() to netif_device_attach().

Signed-off-by Todd Fujinaka <[email protected]>

---
 drivers/net/ethernet/intel/igb/igb_main.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

Quoted 1 line

Tested-by Aaron Brown <[email protected]>

Patchwork did not pick this Tested-by, probably because I left out the colon. Try again...
Tested-by: Aaron Brown <aaron.f.brown at intel.com<