Re: [PATCH AUTOSEL 5.11 10/38] net: correct sk_acceptq_is_full()

[Eric Dumazet]

On 3/30/21 12:21 AM, Sasha Levin wrote:
> From: liuyacan <yacanliu@xxxxxxx>
> 
> [ Upstream commit f211ac154577ec9ccf07c15f18a6abf0d9bdb4ab ]
> 
> The "backlog" argument in listen() specifies
> the maximom length of pending connections,
> so the accept queue should be considered full
> if there are exactly "backlog" elements.
> 
> Signed-off-by: liuyacan <yacanliu@xxxxxxx>
> Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
> ---
>  include/net/sock.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/net/sock.h b/include/net/sock.h
> index 129d200bccb4..a95f38a4b8c6 100644
> --- a/include/net/sock.h
> +++ b/include/net/sock.h
> @@ -936,7 +936,7 @@ static inline void sk_acceptq_added(struct sock *sk)
>  
>  static inline bool sk_acceptq_is_full(const struct sock *sk)
>  {
> -	return READ_ONCE(sk->sk_ack_backlog) > READ_ONCE(sk->sk_max_ack_backlog);
> +	return READ_ONCE(sk->sk_ack_backlog) >= READ_ONCE(sk->sk_max_ack_backlog);
>  }
>  
>  /*
> 


????

I have not seen this patch going in our trees.

First, there was no Fixes: tag, so this is quite unfortunate.

Second, we already had such wrong patches in the past.

Please look at commits
64a146513f8f12ba204b7bf5cb7e9505594ead42 [NET]: Revert incorrect accept queue backlog changes.
8488df894d05d6fa41c2bd298c335f944bb0e401 [NET]: Fix bugs in "Whether sock accept queue is full" checking

Please revert  this patch, thanks !