Hi (again), Quoting Daniele Bellucci <bellucda@tiscali.it>: > > |you do it that way: > | > |retval = sock->ops->bind(sock, (struct sockaddr *) knetlog->addr_in, > |sizeof(struct sockaddr)); > | > | > |whereas I use IPv6 and therefore do so: > | > | > |struct sockaddr_in6 sin; > |... > |error = sock->ops->bind(sock,(struct sockaddr*)&sin,sizeof(sin)); > > i'm sorry but i never played with IPv6 ;( nm, it doesnt very much difference here. just that the address is a bit bigger ;) > > |my socket works very fine, but I'm wondering if passing sizeof(sin) instead > |of sizeof(struct sockaddr) is wise. My question was more porting the size expected by bind. Is it mandatory sizeof(struct sockaddr) or sizeof(struct_we_use) ? I just do like kHTTPd (+ IPv6) > > Why should it be wise? > > |Could it cause some troubles at some point? > > i don't think so. what does bind() use this size for? if the given size is wrong, wouldnt it fool some pointers? > |khttpd might also be used as example of threads and sockets... > |(from 2.4 if i recall correctly, but has been given up in 2.6) > > Be aware, my kernel module doesn't compile with 2.6. > I'm with 2.4.20 -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
