From mboxrd@z Thu Jan 1 00:00:00 1970 From: Francois Romieu Subject: Re: [PATCH v3 1/1 net-next] net: fec: add napi support to improve proformance Date: Fri, 25 Jan 2013 23:23:06 +0100 Message-ID: <20130125222306.GA2523@electric-eye.fr.zoreil.com> References: <1359077343-8271-1-git-send-email-Frank.Li@freescale.com> <20130125043311.GA4141@ppwaskie-mobl2.amr.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Frank Li , lznuaa@gmail.com, shawn.guo@linaro.org, B38611@freescale.com, davem@davemloft.net, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, s.hauer@pengutronix.de To: "Waskiewicz Jr, Peter P" Return-path: Received: from violet.fr.zoreil.com ([92.243.8.30]:43864 "EHLO violet.fr.zoreil.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753116Ab3AYWww (ORCPT ); Fri, 25 Jan 2013 17:52:52 -0500 Content-Disposition: inline In-Reply-To: <20130125043311.GA4141@ppwaskie-mobl2.amr.corp.intel.com> Sender: netdev-owner@vger.kernel.org List-ID: Waskiewicz Jr, Peter P : [...] > Also, when you're disabling interrupts above, you're doing that in your > HW interrupt handler, you should be using spin_lock_irqsave()/irq_restore(). Does the platform forbid to defer FEC_EIR / FEC_IEVENT write to the napi poll handler and only disable the irq through FEC_EIMR / FEC_IMASK in fec_enet_interrupt so as to remove the spinlock ? (Frank, please keep an empty line between variables declarations and function body). -- Ueimor