>> | --- a/net/dccp/proto.c >> | +++ b/net/dccp/proto.c >> | @@ -547,6 +547,9 @@ static int do_dccp_setsockopt(struct sock *sk, int level, int optname, >> | case DCCP_SOCKOPT_QPOLICY_ID: >> | err = dccp_qpolicy_set_id(sk, val); >> | break; >> | + case DCCP_SOCKOPT_QPOLICY_TXQLEN: >> | + err = dccp_qpolicy_set_txqlen(sk, val); >> | + break; >> Not necessary to declare dccp_qpolicy_set_txqlen to return an int (0): >> err is initialised with err = 0 above, so can simply write >> dccp_qpolicy_set_txqlen(sk, val); >> break; >> or - I can't think of any other place to use this, so a function may not be needed: >> >> dccp_sk(sk)->dccps_qpolicy->qpol_txqlen = len; >> break; >> > That's what I originally did. But then I thought that: > 1. This way is prettier and more consistent. Which is of course very > subjective. > 2. Later we might want to return an error when changing tx_qlen for an > already initialized policy wouldn't make sense or would require > additional code complexity. This is exactly the case you mentioned in > our previous discussion: choosing algorithms based on queue length. > The second point is good -- as it is not immediately obviously from the code, it would be good to have it as a comment. >> | - if(sk->sk_write_queue.qlen >= sysctl_dccp_tx_qlen) >> | + if(sk->sk_write_queue.qlen >= tx_qlen) >> This is good - but should keep in mind to deprecate sysctl_dccp_tx_qlen >> in a future patch, since it is now "unemployed". >> > No, sysctl_dccp_tx_qlen still acts as a default: I still think that it would be good to deprecate the sysctl and replace it with a default value. But it is not important now. The University of Aberdeen is a charity registered in Scotland, No SC013683. -- To unsubscribe from this list: send the line "unsubscribe dccp" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
