Re: [RFC][PATCH 0/9] Network receive deadlock prevention for NBD

On Sun, Aug 13, 2006 at 01:06:21PM +0400, Evgeniy Polyakov (johnpol@2ka.mipt.ru) wrote:

On Sat, Aug 12, 2006 at 05:46:07PM -0700, David Miller (davem@davemloft.net) wrote:

quoted

From: Evgeniy Polyakov <redacted>
Date: Sat, 12 Aug 2006 13:37:06 +0400

quoted

Does it? I though it is possible to only have 64k of working sockets per
device in TCP.

Where does this limit come from?

You think there is something magic about 64K local ports,
but if remote IP addresses in the TCP socket IDs are all
different, number of possible TCP sockets is only limited
by "number of client IPs * 64K" and ram :-)

I talked about working sockets, but not about how many of them system
can have at all :)

working -> bound.

-- 
	Evgeniy Polyakov