From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: skge vs sk98lin Date: Mon, 14 May 2007 09:01:30 -0700 Message-ID: <20070514090130.1e3bb3f7@freepuppy> References: <20070508120034.GB1785@mathom.us> <20070508101518.288d202a@freekitty> <20070514142024.GF1785@mathom.us> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org To: Michael Stone Return-path: Received: from smtp.linux-foundation.org ([207.189.120.12]:58238 "EHLO smtp.osdl.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757270AbXENQBt (ORCPT ); Mon, 14 May 2007 12:01:49 -0400 In-Reply-To: <20070514142024.GF1785@mathom.us> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Mon, 14 May 2007 10:20:26 -0400 Michael Stone wrote: > On Tue, May 08, 2007 at 10:15:18AM -0700, Stephen Hemminger wrote: > >On Tue, 08 May 2007 08:00:36 -0400 > >Michael Stone wrote: > >> 2.6.21 seems to have fixed the stability issues I was seeing when using > >> the skge driver with the older sk98xx dual port fiber cards. There is > >> still one more lingering oddness: if I have *two* dual port cards in a > >> system, say eth2-5, I see traffic on eth2, eth3, and eth5, but nothing > >> on eth4. This seems to be consistent accross a couple of systems I've > >> tested; only the first card's second interface sees packets (e.g., with > >> tcpdump). If I reboot with the sk98lin driver on the same kernel I see > >> all traffic, as expected. > > > >Are the statistics changing? ie. ethtool -S eth4 and ifconfig eth4 > > Ok, I ran this for a lot longer and it seems less consistent than it did > at first. It now appears that there is no correlation between a NIC not > working and its order. In some cases the NICs both come up ok, and see > traffic for a while. I left one example running over the weekend, and > neither interface on one card was seeing traffic, and the interface > counters weren't increasing. In another case, one interface on one NIC > worked for a short period of time, then stopped. The interface counters > would increase by a few packets every 30 seconds or so. Any kind of > ethtool operation (e.g., ethtool -A eth4 autoneg off rx off tx off) > would cause the interface to start seeing traffic again, then eventually > go back to seeing only a few packets every few seconds. > > Mike Stone Sounds like a missed interrupt getting the driver stuck. skge uses NAPI and therefore needs level triggered irq's. What is the contents of /proc/interrupts? -- Stephen Hemminger