Re: [PATCH] ethtool: check the ethtool_ops is NULL in dev_ethtool

[Daniel Borkmann <dborkman@redhat.com>]

On 02/18/2014 02:20 AM, Wang Weidong wrote:
> On 2014/2/18 1:09, Daniel Borkmann wrote:
>> On 02/17/2014 12:31 PM, Wang Weidong wrote:
>>> some drivers maybe not implement the ethtool_ops with only
>>> set NULL. So when call the ethtool cmds will lead to a
>>> 'NULL pointer dereference'.
>>>
>>> So add a checking in dev_ethtool.
>>>
>>> Signed-off-by: Wang Weidong <wangweidong1@huawei.com>
>>> ---
>>>    net/core/ethtool.c | 3 +++
>>>    1 file changed, 3 insertions(+)
>>>
>>> diff --git a/net/core/ethtool.c b/net/core/ethtool.c
>>> index 30071de..f418dcb 100644
>>> --- a/net/core/ethtool.c
>>> +++ b/net/core/ethtool.c
>>> @@ -1499,6 +1499,9 @@ int dev_ethtool(struct net *net, struct ifreq *ifr)
>>>            if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
>>>                return -EPERM;
>>>        }
>>> +
>>
>> You have a trailing whitespace/tab in the line above. Please
>> use checkpatch for detecting such things.
>>
> Sorry for that. I will fix it soon.
>
>> Can you be more specific with "some drivers"? Any driver that
>> is in the mainline tree?
>>
> No. My team implements a driver without considering the ethtool_ops.
> So I got the problem.

If it's code that is out of the mainline tree, then why should the
kernel support that? Submit your driver to the tree first, and then
such a change could be considered. Or, even better, let them implement
ethtool ops/stubs.

>>> +    if (!dev->ethtool_ops)
>>> +        return -EOPNOTSUPP;
>>>
>>>        if (dev->ethtool_ops->begin) {
>>>            rc = dev->ethtool_ops->begin(dev);
>>>
>>
>>
>
>