From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 4808 invoked by alias); 5 Jun 2009 15:32:33 -0000 Received: (qmail 4794 invoked by uid 22791); 5 Jun 2009 15:32:30 -0000 X-SWARE-Spam-Status: No, hits=-1.7 required=5.0 tests=AWL,BAYES_00,SPF_HELO_PASS,SPF_PASS X-Spam-Check-By: sourceware.org Received: from connor.medar.com (HELO connor.medar.com) (206.181.246.34) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Fri, 05 Jun 2009 15:32:26 +0000 Received: from [89.194.102.40] (win0040 [89.194.102.40]) by skynet.wtccorp.com (Postfix) with ESMTP id 0221B3F for ; Fri, 5 Jun 2009 11:32:24 -0400 (EDT) Subject: PowerPC Ethernet issue From: Robert Morse To: ecos-devel@ecos.sourceware.org Content-Type: text/plain Date: Fri, 05 Jun 2009 15:32:00 -0000 Message-Id: <1244215787.30295.28.camel@win0040> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Mailing-List: contact ecos-devel-help@ecos.sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Post: List-Help: , Sender: ecos-devel-owner@ecos.sourceware.org X-SW-Source: 2009-06/txt/msg00013.txt.bz2 Hi all, We are working on a PowerPC ppc405 (AMCC 405ep) based board. We are using the following drivers from ECOS v2. A slight change to the hal/powerpc/.. directory to support our devices. A change to devs/eth/powerpc/ppc405 directory to support dual ethernet. Then a new devs/eth/phy to support the national DP8348C part. Looking at the data sheet for the 405EP processor, the MAC has a bit that tells the ethernet channel if it is FULL or HALF duplex mode. And on initialization or any time a Ethernet Reset is required we do set it to the correct Duplex. My current issue is what happens if someone unplugs changes the duplex and plugs back in. I see nothing that will update the duplex. Looking at the situation, it seems like I would need to enable/configure the interrupts for the PHYs. and when a change of duplex is detected cause a reset of the MAC. Has anyone done anything like this before. Looking through the ethernet drivers and phy drivers I do not see anyone doing this (I admit not all). Does my idea seem sound, or am I off in the woods somewhere. I know about this duplex issue from another design. Where we had a commercial rtos, that I found out they did not do this and caused a lot of issues, I thought the device driver should of handled it, but they said it was up to the application to control the bit. So I do want to repeat the same issue. Thanks Robert