From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Eilon Greenstein" Subject: Re: [PATCH] bnx2x: EEH recovery fix Date: Tue, 09 Sep 2008 12:34:49 +0300 Message-ID: <1220952889.15079.3.camel@lb-tlvb-eliezer> References: <20080908.124743.30466861.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Cc: "netdev@vger.kernel.org" , "Yitchak Gertner" To: "David Miller" Return-path: Received: from mms1.broadcom.com ([216.31.210.17]:3727 "EHLO mms1.broadcom.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752238AbYIIJfY (ORCPT ); Tue, 9 Sep 2008 05:35:24 -0400 In-Reply-To: <20080908.124743.30466861.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: On Mon, 2008-09-08 at 12:47 -0700, David Miller wrote: > From: "Eilon Greenstein" > Date: Mon, 8 Sep 2008 03:22:32 -0700 > > > The EEH mechanism caused a fatal error when using the bnx2x since > > the driver attempted accessing the device after it was reset by the > > EEH. This patch sets a proper unload when EEH error is detected > > without accessing the device. > > > > Since the current implementation causes a fatal error, I ask that > > you will consider applying this patch. > > What user has run into and reported this failure, and is it > a regression compared to 2.6.26? IBM found it on PPC and it was more obvious with the DMA mapping violation. It is broken in the same way on 2.6.26 as well. > > That is the criteria for my putting this patch in now. > > Since EEH triggers are an exceedingly corner case error condition, > this puts this patch even less valid to go in outside of the merge > window. > If you cannot accept it into 2.6.27, I will re-send it into net-next. Please let me know. Thanks, Eilon