Re: [PATCH net-next] net: Make gro complete function to return void

[David Ahern <dsahern@kernel.org>]

On 5/30/23 9:39 AM, Parav Pandit wrote:
>> From: David Ahern <dsahern@kernel.org>
>> Sent: Tuesday, May 30, 2023 11:26 AM
>>
>> On 5/29/23 7:44 AM, Parav Pandit wrote:
>>> diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c index
>>> 45dda7889387..88f9b0081ee7 100644
>>> --- a/net/ipv4/tcp_offload.c
>>> +++ b/net/ipv4/tcp_offload.c
>>> @@ -296,7 +296,7 @@ struct sk_buff *tcp_gro_receive(struct list_head
>> *head, struct sk_buff *skb)
>>>  	return pp;
>>>  }
>>>
>>> -int tcp_gro_complete(struct sk_buff *skb)
>>> +void tcp_gro_complete(struct sk_buff *skb)
>>>  {
>>>  	struct tcphdr *th = tcp_hdr(skb);
>>>
>>> @@ -311,8 +311,6 @@ int tcp_gro_complete(struct sk_buff *skb)
>>>
>>>  	if (skb->encapsulation)
>>>  		skb->inner_transport_header = skb->transport_header;
>>> -
>>> -	return 0;
>>>  }
>>>  EXPORT_SYMBOL(tcp_gro_complete);
>>
>> tcp_gro_complete seems fairly trivial. Any reason not to make it an inline and
>> avoid another function call in the datapath?
>>
> Sounds good to me.
> With inline it should mostly improve the perf, but I do not have any of the 3 adapters which are calling this API to show perf results.
> 
> Since, it is a different change touching the performance, I prefer to do follow up patch that bnx owners can test.
> Is that ok?

sure.