Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 8 Oct 2004 10:23:14 -0400
From:      "Gerrit Nagelhout" <gnagelhout@sandvine.com>
To:        "David Gilbert" <dgilbert@dclg.ca>, <freebsd-current@freebsd.org>
Subject:   RE: Packet passing performance study on exotic hardware.
Message-ID:  <A8535F8D62F3644997E91F4F66E341FC1F1CB9@exchange.sandvine.com>

next in thread | raw e-mail | index | archive | help
David Gilbert Wrote:
>=20
>   - with polling and excessive packets, it doesn't "receive" the full
>     load of packets.  In netstat -w, they show as input "errors"
>     although the number of "errors" isn't strictly related to the
>     number of dropped packets.  It's just some large number that
>     generally increases with the number of dropped packets.
>=20

In em_update_stats_counters, errors is calculated as follows:

	/* Rx Errors */
	ifp->if_ierrors =3D
	adapter->dropped_pkts +
	adapter->stats.rxerrc +
	adapter->stats.crcerrs +
	adapter->stats.algnerrc +
	adapter->stats.rlec + adapter->stats.rnbc +=20
	adapter->stats.mpc + adapter->stats.cexterr;

The extra errors you are talking about come from doing mpc + rnbc. =20
Only mpc (missed packet count) (and the other errors of course) should=20
be used.=20
Rnbc is an indication of the internal buffer filling up, but not yet=20
dropping packets, and should not be included.


Gerrit



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?A8535F8D62F3644997E91F4F66E341FC1F1CB9>