Re: [PATCH 1/1] ixgbe: replace rtnl_lock with rcu_read_lock

2016-06-07 Thread Alexander Duyck
I pretty much spelled this out already. The watchdog gets synced and disabled when ixgbe_down is called so you are guaranteed that the number of queues will not be changing while this function is running. What is occurring is that the ixgbe driver can update the netdevs controlled by the ixgbe

Re: [PATCH 1/1] ixgbe: replace rtnl_lock with rcu_read_lock

2016-06-06 Thread Alexander Duyck
Just a quick scan has me wondering what code you are comparing it to? The key bit here that is the reason for taking the RTNL lock is because this section is handled in the watchdog which is not an RTNL protected region, and because it is messing with devices other than the ones controlled by the

Re: [PATCH 1/1] ixgbe: replace rtnl_lock with rcu_read_lock

2016-06-05 Thread Alexander Duyck
On Sun, Jun 5, 2016 at 2:14 AM, wrote: > From: Zhu Yanjun > > > Signed-off-by: Zhu Yanjun > --- > drivers/net/ethernet/intel/ixgbe/ixgbe_main.c |4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git

Re: [PATCH 1/1] ixgbe: replace rtnl_lock with rcu_read_lock

2016-06-05 Thread Eric Dumazet
On Sun, 2016-06-05 at 17:14 +0800, zyjzyj2...@gmail.com wrote: > From: Zhu Yanjun > > > Signed-off-by: Zhu Yanjun > --- You do not explain if this is a bug fix (targeting net tree) or simply an optimization in the slow path (targeting net-next tree)

[PATCH 1/1] ixgbe: replace rtnl_lock with rcu_read_lock

2016-06-05 Thread zyjzyj2000
From: Zhu Yanjun Signed-off-by: Zhu Yanjun --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c |4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c