Re: v3 [PATCH 1/1] dccp: Process incoming Change feature-negotiation

[Wei Yongjun <yjwei@cn.fujitsu.com>]

Gerrit Renker wrote:
>>> +	/*
>>> +	 *	Unidirectional/simultaneous negotiation of SP features (6.3.1)
>>> +	 */
>>> +	entry = dccp_feat_list_lookup(fn, feat, local);
>>> +	if (entry = NULL) {
>>> +		if (!dccp_feat_sp_list_ok(feat, val, len))
>>> +			goto unknown_feature_or_value;
>>>   
>>>       
>> Check for sp feat list should before code "entry =  
>> dccp_feat_list_lookup(fn, feat, local);",
>> here only check for features not register by local endpoint, if the  
>> feature is registed, the validity check is missing?
>>
>>     
> +
> No, in this case the validity check is performed already as part of the socket
> registration routines - which in turn end up calling dccp_feat_sp_list_ok.  
> If a user tries to register invalid SP values on the socket, the attempt
> will fail with EINVAL. If the user does not register values, the feature
> defaults (6.4) are used, which are valid by definition.
> The host is conservative in what it allows to send out.
>   

The socket registration routines check the feature values local set, but 
this place is check
the features we *received* from other endpoints.

I agree with you here do not need chec as your other mail said:
RFC4340 6.6.8
Note that server-priority features do not have value limitations, since
unknown values are handled as a matter of course.

May be this check "if (!dccp_feat_sp_list_ok(feat, val, len))" is too 
strictly for known
sp features but not registed by socket.