Re: Routing Socket and New Addresses
On Sun, 27 Jan 2002, Justin C.Walker wrote: It is and it is :-}. At least, Stevens discusses it in Unix Network Programming, v1, 2e (sec. 20.3). Different systems, alas, treat this case differently. My section 20.3 is on UDP Datagram Trunctation...did you mean 17.3 (Routing Sockets: Reading and Writing). I can't find any mention of this behaviour in either place mind you. Thanks, Andrew To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Routing Socket and New Addresses
On Monday, January 28, 2002, at 05:03 PM, Andrew wrote: On Sun, 27 Jan 2002, Justin C.Walker wrote: It is and it is :-}. At least, Stevens discusses it in Unix Network Programming, v1, 2e (sec. 20.3). Different systems, alas, treat this case differently. My section 20.3 is on UDP Datagram Trunctation...did you mean 17.3 (Routing Sockets: Reading and Writing). I can't find any mention of this behaviour in either place mind you. Maybe I misunderstood the original message; I thought this thread dealt with the observed truncation of packets when read from a socket. 20.3 was what I intended, as it covers the observed behavior (at least, that's my story, and I'm sticking with it :-]). Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics| If you're not confused, | You're not paying attention *--*---* To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Routing Socket and New Addresses
On Wed, 23 Jan 2002, Andrew wrote: configured I get a RTM_NEWADDR message. The bit I'm confused with is the struct sockaddr associated with RTA_IFA (that I assumed would hold the IP of the interface) has an sa_family value of AF_IMPLINK. If I cast it to a struct sockaddr_in then s_addr is 0. Well it turns out that this was a combination of a bug in my code (*red face*) and something else. The something else I haven't quite worked out but it seems that if I don't read the packet with one read call then the packet is lost. Is this correct behaviour? I guess if the buffer is small and a number of packets (such as wehn an interface goes up or down) then it might happen but there shouldnt be that much time between read calls. I'll play with it a bit more tonight. Thanks, Andrew To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Routing Socket and New Addresses
On Sun, 27 Jan 2002, Andrew wrote: but it seems that if I don't read the packet with one read call then the packet is lost. Is this correct behaviour? I guess if the buffer is small Well it seems that if you dont get the entire packet in one read it is lost forever. It also seems that no matter how many bytes you try and read you only ever get one packet (though I haven't confirmed that for sure). If this is how it really works then perhaps it should be documented somewhere. Thanks, Andrew To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Routing Socket and New Addresses
On Sunday, January 27, 2002, at 08:03 AM, Andrew wrote: On Sun, 27 Jan 2002, Andrew wrote: but it seems that if I don't read the packet with one read call then the packet is lost. Is this correct behaviour? I guess if the buffer is small Well it seems that if you dont get the entire packet in one read it is lost forever. It also seems that no matter how many bytes you try and read you only ever get one packet (though I haven't confirmed that for sure). If this is how it really works then perhaps it should be documented somewhere. It is and it is :-}. At least, Stevens discusses it in Unix Network Programming, v1, 2e (sec. 20.3). Different systems, alas, treat this case differently. Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics|Men are from Earth. |Women are from Earth. | Deal with it. *--*---* To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message