Re: [PATCH] rsocket: Return ECONNRESET when socket in recv is disconnected

[Sreedhar Kodali <srkodali-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org>]

Hi Jason,

Thanks for your explanation - I got it.  On the other mail thread
I have just requested Sean to mark this as R-Sockets limitation or
behavioral difference so users wouldn't confuse when they try
to compare with TCP.

Thank You.

- Sreedhar

On 2014-10-10 23:15, Jason Gunthorpe wrote:
> On Fri, Oct 10, 2014 at 10:37:34AM +0530, Sreedhar Kodali wrote:
>> So the current behavior of returning 0 is wrong as there was no
>> orderly shutdown performed on the other end before a close() is
>> issued
> 
> 'Orderly shutdown' is any call of close() when there is no data in the
> recv queue, or shutdown(SHUT_WR).
> 
>> http://stackoverflow.com/questions/5879560/how-can-i-cause-an-econnreset-in-recv-from-a-client
> 
> This accepted answer is not correct, close on its own does not
> generate ECONNRESET.
> 
> This is trivally shown with strace and a bit of python:
> 
> server:
>  import socket
>  s = socket.socket(socket.AF_INET,socket.SOCK_STREAM)
>  s.bind(("127.0.0.1",9090))
>  s.listen(1)
>  conn, addr = s.accept()
>  conn.recv(1024)
> 
> client:
> 
> import socket
> s = socket.socket(socket.AF_INET,socket.SOCK_STREAM)
> s.connect(("127.0.0.1",9090))
> s.close()
> 
> Result:
> 
> [..]
> accept(3, {sa_family=AF_INET, sin_port=htons(41284),
> sin_addr=inet_addr("127.0.0.1")}, [16]) = 4
> recvfrom(4, "", 1024, 0, NULL, NULL)    = 0
> 
> ECONNRESET signals a special condition that it seems rsockets cannot
> detect, as it doesn't have the one sided close semantics of TCP.
> 
> Jason

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html