tech-net archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: [patch] bug fix & TCP networking performance improvements

On Fri, Apr 22, 2011 at 01:40:35PM +0200, Matthias Drochner wrote:
> said:
> > The test that's in there is for "reclaimed" not for "expired", and
> > reclamation can lag expiration if there is not pressure on the VTW
> > table
> Perhaps I'm not understanding it correctly, but as I see the
> entries are marked as expired/reclaimed by special values (0/-1)
> in tv_sec/tv_usec. Then, the netstat display should show either
> the same integral or the same fractional part for all affected
> time differences since the base time is the same for all.
> It doesn't...

Ok, I see what you're saying.

Expired and reclaimed are different states.  The entries that you showed
to me had expired (hence the negative remaining lifetime in netstat's
output), but they had not been reclaimed.  I don't think the operator
will usually care to see either expired or reclaimed entries, so I plan
to hide them all from view.

> Btw, I think that it would be better to use microuptime(9)
> rather than microtime(9) for these interval things, to make
> it robust against time changes. Or getmicrouptime(9) to save
> some cycles - accuracy shouldn't be an issue.



David Young             OJC Technologies      Urbana, IL * (217) 344-0444 x24

Home | Main Index | Thread Index | Old Index