On Tue, Jan 16, 2001 at 08:54:53PM +0300, kuznet@xxxxxxxxxxxxx wrote:
> > Recently I noticed that when I simultaneously do 'up' to many network
> > interfaces
> > (many is ~15) netlink drops part of the messages about interface state
> > change and thus
> > my userspace tools don't know that some interfaces are in up state now. The
> > error that
> > I get from netlink socket is "No buffer space available".
> Which means that applications must invalidate stored state
> and to resynchronize doing dumps of all the necessary information.
It means I will resynchronize almost every time :(.
> > sk->receive_queue simultaneously is about 16 only!
> 16 or 116, this is not very essential.
This may be essential. When the sk->receive_queue is bigger, my application
will have a chance to read part of the messages and free space in socket for
> Of course, page size is sort of overkill, but I do not want to estimate
> required room forward. Application must be able to resync in any case.
Resync should be an exception and not the rule IMO.