EHOSTDOWN on unconnected UDP6

[J. Hannken-Illjes]

Running NSD I get sometimes "recvmmsg failed: Host is down" errors
for this socket: "udp6 *.53 *.*"

I never saw this behaviour on the udp4 socket.

Why do we set this "Host is down" error on unconnected udp6 sockets?

--
J. Hannken-Illjes - ***@eis.cs.tu-bs.de - TU Braunschweig (Germany)


--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Michael van Elst]

Post by J. Hannken-Illjes
Why do we set this "Host is down" error on unconnected udp6 sockets?

EHOSTDOWN is the result of a received "address unreachable" ICMPv6 packet,
there is no corresponding IPv4 message.

[J. Hannken-Illjes]

Post by Michael van Elst

Post by J. Hannken-Illjes
Why do we set this "Host is down" error on unconnected udp6 sockets?

EHOSTDOWN is the result of a received "address unreachable" ICMPv6 packet,
there is no corresponding IPv4 message.

Ok, but why do we deliver this error asynchronously to an
udp6 socket that is bound but not connected.

How should the server know which of the messages
it sent failed?

This behaviour looks quite useless ...

--
J. Hannken-Illjes - ***@eis.cs.tu-bs.de - TU Braunschweig

[Michael van Elst]

Post by J. Hannken-Illjes

Post by Michael van Elst
EHOSTDOWN is the result of a received "address unreachable" ICMPv6 packet,
there is no corresponding IPv4 message.

Ok, but why do we deliver this error asynchronously to an
udp6 socket that is bound but not connected.
How should the server know which of the messages
it sent failed?

If I read the code correctly, the socket destination address
is set in the PCB when you send a message, and is searched
later when a notification should be delivered to the socket.

So you get the error for the last destination.