poll and tcp sockets defect

Hello I try to use the following code, but it seems not working, on a
connected socket (sock) i do:

struct pollfd input_fd;
int a=1;
setsockopt(sock,SOL_SOCKET,SO_KEEPALIVE,&a,sizeof( a));
a=1;
setsockopt(sock,SOL_TCP,TCP_KEEPIDLE,&a,sizeof(a)) ;
a=1;
setsockopt(sock,SOL_TCP,TCP_KEEPINTVL,&a,sizeof(a) );
a=1;
setsockopt(sock,SOL_TCP,TCP_KEEPCNT,&a,sizeof(a));
input_fd.fd = sock;
input_fd.events = POLLIN | POLLOUT;
input_fd.revents = 0;
int p=poll(&input_fd, 1,-1);

I pull the LAN plug to the remote computer, but nothing happens. The
kernel seems to be sleeping in poll syscall. Is this desired
behaviour, or it is a defect?
I use linux 2.6.29.4.
Thanks.

Drago <(E-Mail Removed)> wrote:
> Hello I try to use the following code, but it seems not working, on a
> connected socket (sock) i do:

> struct pollfd input_fd;
> int a=1;
> setsockopt(sock,SOL_SOCKET,SO_KEEPALIVE,&a,sizeof( a));
> a=1;
> setsockopt(sock,SOL_TCP,TCP_KEEPIDLE,&a,sizeof(a)) ;
> a=1;
> setsockopt(sock,SOL_TCP,TCP_KEEPINTVL,&a,sizeof(a) );
> a=1;
> setsockopt(sock,SOL_TCP,TCP_KEEPCNT,&a,sizeof(a));
> input_fd.fd = sock;
> input_fd.events = POLLIN | POLLOUT;
> input_fd.revents = 0;
> int p=poll(&input_fd, 1,-1);

> I pull the LAN plug to the remote computer, but nothing happens. The
> kernel seems to be sleeping in poll syscall. Is this desired
> behaviour, or it is a defect?

Are you sure that your setsockopt() calls aren't returning -1 and
setting errno? The code to do it isn't visible above - you are
actually connecting the socket to something?

rick jones
--
The computing industry isn't as much a game of "Follow The Leader" as
it is one of "Ring Around the Rosy" or perhaps "Duck Duck Goose."
- Rick Jones
these opinions are mine, all mine; HP might not want them anyway...
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

Drago <(E-Mail Removed)> wrote:
> Yes, this is just a code fragment, and I removed checking return
> statements to produce more clear source.
> All of the setsockopt calls return 0. The sock is actually connected
> and data is read, until plugging the lan cable.
> After that poll just sleeps.

What does tcpdump show you the system trying to do - for example with
the connection idle but the cable connected, does it show the
keepalive probes being sent when you were expecting them to? I'm
wondering if perhaps there is some silent bounds conformance happening
to the values you provide in the setsockopt() calls for the idle time
before keepalives and how often and how many should be sent.

rick jones
--
No need to believe in either side, or any side. There is no cause.
There's only yourself. The belief is in your own precision. - Joubert
these opinions are mine, all mine; HP might not want them anyway...
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...