On 10/4/17 11:04 AM, Jiri Pirko wrote:
> Wed, Oct 04, 2017 at 05:35:46PM CEST, dsah...@gmail.com wrote:
>> On 10/3/17 11:38 PM, Jiri Pirko wrote:
>>> Wed, Oct 04, 2017 at 06:58:52AM CEST, dsah...@gmail.com wrote:
>>>> A number of bond_enslave errors are logged using the netdev_err API.
>>>> Return those messages to userspace via the extack facility.
>>>>
>>>> Signed-off-by: David Ahern <dsah...@gmail.com>
>>>> ---
>>>> drivers/net/bonding/bond_main.c | 12 ++++++++++++
>>>> 1 file changed, 12 insertions(+)
>>>>
>>>> diff --git a/drivers/net/bonding/bond_main.c
>>>> b/drivers/net/bonding/bond_main.c
>>>> index bc92307c2082..6688dc9154e0 100644
>>>> --- a/drivers/net/bonding/bond_main.c
>>>> +++ b/drivers/net/bonding/bond_main.c
>>>> @@ -1348,12 +1348,15 @@ int bond_enslave(struct net_device *bond_dev,
>>>> struct net_device *slave_dev,
>>>>
>>>> /* already in-use? */
>>>> if (netdev_is_rx_handler_busy(slave_dev)) {
>>>> + NL_SET_ERR_MSG(extack,
>>>> + "Device is in use and cannot be enslaved");
>>>
>>> Please don't do this kind of wrapping. Just let the string be on the
>>> same line.
>>>
>>
>> Ok, I will do that for bonding only since it is the existing style.
>
> I don't believe you need to do this wrap for any code. Just don't wrap.
> General code stype says no wrap for strings :)
>
I do not break / wrap strings; they need to be searchable. I assumed you
meant this is preferred for bonding:
NL_SET_ERR_MSG(extack, "Device is in use and cannot be enslaved");
over what I have done:
NL_SET_ERR_MSG(extack,
"Device is in use and cannot be enslaved");
