Re: [PATCH] bonding: clear header_ops when last slave detached (v2)

[Wengang <wen.gang.wang@oracle.com>]

Hi Eric,

于 2014年11月19日 13:39, Eric Dumazet 写道:
> On Wed, 2014-11-19 at 13:18 +0800, Wengang Wang wrote:
>> When last slave of a bonding master is removed, the bonding then does not work.
>> When packet_snd is called against with a master net_device, it accesses
>> header_ops. In case the header_ops is not valid any longer(say ipoib module
>> unloaded), it will then access an invalid memory address.
>> This patch try to fix this issue by clearing header_ops when last slave
>> detached.
>>
>> Signed-off-by: Wengang Wang <wen.gang.wang@oracle.com>
>> ---
>>   drivers/net/bonding/bond_main.c | 1 +
>>   1 file changed, 1 insertion(+)
> It seems you basically ignored my feedback.
>
>
>
> Some code is doing :
>
> [A] if (dev->header_ops) {
>      ...
> [B]    dev->header_ops->XXX()
>
>
>
> Nothing prevents you doing the clear after [A] , and before [B]
>

Yes, that's true. So the simplest way is move ipoib_header_ops to vmlinux.

thanks for review.
wengang