Re: [PATCH dovetail] net: skbuff: add missing OOB stubs for !CONFIG_NET_OOB

Hi!

On 3/25/26 15:51, Philippe Gerum wrote:
> Tobias Schaffner <[email protected]> writes:
> 
>> skbuff.c calls skb_mark_oob() and variants behind runtime running_oob()
>> checks but these are only implemented in the CONFIG_NET_OOB section.
>>
>> Add the missing stubs to the !CONFIG_NET_OOB section.
>>
>> Signed-off-by: Tobias Schaffner <[email protected]>
>> ---
>>   include/linux/skbuff.h | 16 ++++++++++++++++
>>   1 file changed, 16 insertions(+)
>>
>> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
>> index 73dd4a58d181..1278d0a58677 100644
>> --- a/include/linux/skbuff.h
>> +++ b/include/linux/skbuff.h
>> @@ -5630,6 +5630,22 @@ static inline struct sk_buff *get_oob_skb(void)
>>   	return __get_oob_skb();
>>   }
>>   
>> +static inline void skb_mark_oob(struct sk_buff *skb)
>> +{
>> +}
>> +
>> +static inline void skb_mark_oob_released(struct sk_buff *skb)
>> +{
>> +}
>> +
>> +static inline void skb_mark_oob_timestamped(struct sk_buff *skb)
>> +{
>> +}
>> +
>> +static inline void skb_mark_oob_managed(struct sk_buff *skb)
>> +{
>> +}
>> +
>>   #endif	/* !CONFIG_NET_OOB */
>>   
>>   #endif	/* __KERNEL__ */
> 
> All current users of skb_oob_mark_*() calls are currently guarded by
> CONFIG_NET_OOB sections, or enable CONFIG_NET_OOB, like the evl
> core. Did you observe a build failure with CONFIG_NET_OOB=n, or is it
> preventive fixup?

I am building without CONFIG_NET_OOB at the moment and the build fails 
because of the unguarded skb_mark_oob calls in net/core/skbuff.c.

Best,
Tobias