From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH 2/2] net: socionext: reset hardware in ndo_stop Date: Mon, 16 Apr 2018 13:47:04 -0400 (EDT) Message-ID: <20180416.134704.356952030127566216.davem@davemloft.net> References: <1523864399-27296-1-git-send-email-jassisinghbrar@gmail.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, masahisa.kojima@linaro.org, ard.biesheuvel@linaro.org, jaswinder.singh@linaro.org To: jassisinghbrar@gmail.com Return-path: Received: from shards.monkeyblade.net ([184.105.139.130]:44588 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752788AbeDPRrF (ORCPT ); Mon, 16 Apr 2018 13:47:05 -0400 In-Reply-To: <1523864399-27296-1-git-send-email-jassisinghbrar@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: From: jassisinghbrar@gmail.com Date: Mon, 16 Apr 2018 13:09:59 +0530 > From: Masahisa KOJIMA > > When the interface is down, head/tail of the descriptor > ring address is set to 0 in netsec_netdev_stop(). > But netsec hardware still keeps the previous descriptor > ring address, so there is inconsistency between driver > and hardware after interface is up at a later time. > To address this inconsistency, add netsec_reset_hardware() > when the interface is down. > > In addition, to minimize the reset process, > add flag to decide whether driver loads the netsec microcode. > Even if driver resets the netsec hardware, netsec microcode > keeps resident on RAM, so it is ok we only load the microcode > at initialization. > > This patch is critical for installation over network. > > Signed-off-by: Masahisa KOJIMA > Fixes: 533dd11a12f6 ("net: socionext: Add Synquacer NetSec driver") > Signed-off-by: Jassi Brar Applied.