Re: Our (real) problems

[Alan Cox <alan redhat com>]

> Why, is the "broken" signal tied to X? Can't you get the same thing for
> non-X stuff by just pinging the other side of the connection every now
> and then? If you get an EPIPE, it's dead. The remote server doesn't even
> need to ack the ping.

The way this works with TCP is as follows. Reads from broken streams
generate an EOF, writes generate -1, EPIPE and a broken pipe signal

A broken tcp connect will generally be detected on a write (or up to 15 mins
after it if it takes time to time out the attempt) and if keepalives are
enabled "within about 2.5 hours anyway"