Re: [PATCH net-next 2/3] Implementation of RFC 4898 Extended TCP Statistics (Web10G)

[Dominic Hamon <dhamon@twitter.com>]

On Tue, Dec 16, 2014 at 03:18:23PM -0500, David Miller wrote:
> From: rapier <rapier@psc.edu>
> Date: Tue, 16 Dec 2014 15:13:44 -0500
> 
> > On 12/16/14, 3:03 PM, David Miller wrote:
> > 
> >> You shouldn't need to export any symbols.
> > 
> > As a point of clarification - is it acceptable to export symbols for
> > use with in tree modules such as tcp_htcp? We are more than willing to
> > do the work required to bring this in line with best practices.
> 
> I'm saying for data and TCP statistics collection, you shouldn't need
> to add any new symbol exports.
> 
> Keep this in the main kernel, nothing external should be needed.
> 
> Extending tcp_info or similar is the only reasonable way to implement
> this stuff.

There are two aspects of the tcp_estats approach that I think are worth
considering separately:
  - the breadth of instrumentation
  - the access method through kernel module/netlink

I don't see any reason why we can't take the instrumentation parts of
these patches and plumb them in to tcp_info.

I would prefer if we can retain the nested structures and the sysctl
that controls the set of instruments collected as I think this key
feature gives plenty of flexibility to system operators in terms of an
overhead/instrumentation trade-off, however I recognise this complicates
how the socket option is returned, and how ss might report the metrics.

--
Dominic Hamon

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