[PATCH net-next 0/2] ldmvsw: port removal stability

[PATCH net-next 0/2] ldmvsw: port removal stability

[Shannon Nelson]

Under heavy reboot stress testing we found a couple of timing issues
when removing the device that could cause the kernel great heartburn,
addressed by these two patches.

Shannon Nelson (1):
  ldmvsw: stop the clean timer at beginning of remove

Thomas Tai (1):
  ldmvsw: unregistering netdev before disable hardware

 drivers/net/ethernet/sun/ldmvsw.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

[PATCH net-next 1/2] ldmvsw: unregistering netdev before disable hardware

[Shannon Nelson]

From: Thomas Tai <thomas.tai@oracle.com>

When running LDom binding/unbinding test, kernel may panic
in ldmvsw_open(). It is more likely that because we're removing
the ldc connection before unregistering the netdev in vsw_port_remove(),
we set up a window of time where one process could be removing the
device while another trying to UP the device. This also sometimes causes
vio handshake error due to opening a device without closing it completely.
We should unregister the netdev before we disable the "hardware".

Orabug: 25980913, 25925306

Signed-off-by: Thomas Tai <thomas.tai@oracle.com>
Signed-off-by: Shannon Nelson <shannon.nelson@oracle.com>
---
 drivers/net/ethernet/sun/ldmvsw.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/sun/ldmvsw.c b/drivers/net/ethernet/sun/ldmvsw.c
index 5a90fed..309747c 100644
--- a/drivers/net/ethernet/sun/ldmvsw.c
+++ b/drivers/net/ethernet/sun/ldmvsw.c
@@ -413,6 +413,7 @@ static int vsw_port_remove(struct vio_dev *vdev)
 		del_timer_sync(&port->vio.timer);
 
 		napi_disable(&port->napi);
+		unregister_netdev(port->dev);
 
 		list_del_rcu(&port->list);
 
@@ -427,7 +428,6 @@ static int vsw_port_remove(struct vio_dev *vdev)
 
 		dev_set_drvdata(&vdev->dev, NULL);
 
-		unregister_netdev(port->dev);
 		free_netdev(port->dev);
 	}
 
-- 
1.7.1

[PATCH net-next 2/2] ldmvsw: stop the clean timer at beginning of remove

[Shannon Nelson]

Stop the clean timer earlier to be sure there's no asynchronous
interference while stopping the port.

Orabug: 25748241

Signed-off-by: Shannon Nelson <shannon.nelson@oracle.com>
---
 drivers/net/ethernet/sun/ldmvsw.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/sun/ldmvsw.c b/drivers/net/ethernet/sun/ldmvsw.c
index 309747c..5b56c24 100644
--- a/drivers/net/ethernet/sun/ldmvsw.c
+++ b/drivers/net/ethernet/sun/ldmvsw.c
@@ -411,6 +411,7 @@ static int vsw_port_remove(struct vio_dev *vdev)
 
 	if (port) {
 		del_timer_sync(&port->vio.timer);
+		del_timer_sync(&port->clean_timer);
 
 		napi_disable(&port->napi);
 		unregister_netdev(port->dev);
@@ -418,7 +419,6 @@ static int vsw_port_remove(struct vio_dev *vdev)
 		list_del_rcu(&port->list);
 
 		synchronize_rcu();
-		del_timer_sync(&port->clean_timer);
 		spin_lock_irqsave(&port->vp->lock, flags);
 		sunvnet_port_rm_txq_common(port);
 		spin_unlock_irqrestore(&port->vp->lock, flags);
-- 
1.7.1

Re: [PATCH net-next 0/2] ldmvsw: port removal stability

[David Miller]

From: Shannon Nelson <shannon.nelson@oracle.com>
Date: Mon, 15 May 2017 10:51:06 -0700

> Under heavy reboot stress testing we found a couple of timing issues
> when removing the device that could cause the kernel great heartburn,
> addressed by these two patches.

These are pretty legit bug fixes so I've applied this series to net,
thanks.