Re: [PATCH net-next-2.6 v2] can: convert protocol handling to RCU

[Oliver Hartkopp <socketcan@hartkopp.net>]

On 06.04.2011 11:27, Kurt Van Dijck wrote:
> On Tue, Apr 05, 2011 at 08:01:16PM +0200, Oliver Hartkopp wrote:

>>  
>> +static struct can_proto *can_try_module_get(int protocol)
>> +{
>> +	struct can_proto *cp;
>> +
>> +	rcu_read_lock();
>> +	cp = rcu_dereference(proto_tab[protocol]);
>> +	if (cp && !try_module_get(cp->prot->owner))
> After the xxx_get, is the 'cp' pointer persistent?

try_module_get() increases the usage counter of the module - therefore it is.
It is protected until module_put(cp->prot->owner) at the end of can_create() .

>>  	/* check for available protocol and correct usage */
>>  
>>  	if (!cp)
>>  		return -EPROTONOSUPPORT;
>>  
>>  	if (cp->type != sock->type) {
> I don't see how this will evaluate to true?
> can_proto_register takes care of it.

This check compares the type of the socket that is to be created with the type
that's defined for this protocol.

E.g. if you would give

s = socket(PF_CAN, SOCK_STREAM, CAN_RAW);

instead of the correct

s = socket(PF_CAN, SOCK_RAW, CAN_RAW);

you will get this error.

>> -		err = -EPROTONOSUPPORT;
>> +		err = -EPROTOTYPE;
>>  		goto errout;
>>  	}
>>  

Regards,
Oliver