From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Ball Subject: Re: [PATCH 1/3] mmc: skip detection of nonremovable cards on rescan - card removal detection regression Date: Thu, 03 Mar 2011 21:28:51 -0500 Message-ID: References: <1290921690-6133-1-git-send-email-ohad@wizery.com> <1290921690-6133-2-git-send-email-ohad@wizery.com> <20101217005123.GD17239@void.printf.net> <1299205366.4999.3.camel@maxim-laptop> Mime-Version: 1.0 Content-Type: text/plain Return-path: Received: from void.printf.net ([89.145.121.20]:37071 "EHLO void.printf.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759195Ab1CDCan (ORCPT ); Thu, 3 Mar 2011 21:30:43 -0500 In-Reply-To: <1299205366.4999.3.camel@maxim-laptop> (Maxim Levitsky's message of "Fri, 04 Mar 2011 04:22:46 +0200") Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Maxim Levitsky Cc: Ohad Ben-Cohen , linux-mmc@vger.kernel.org Hi, On Thu, Mar 03 2011, Maxim Levitsky wrote: > This patch breaks the CONFIG_MMC_UNSAFE_RESUME, because it sets fake > non-removeable flag, but the intended use of this option is to assume > that card is not removed _during_ suspend, and it can still be removed > during normal use. > With this commit, card removal isn't detected. I guess we disagree about the semantics of CONFIG_MMC_UNSAFE_RESUME. The Kconfig text says: config MMC_UNSAFE_RESUME bool "Assume MMC/SD cards are non-removable (DANGEROUS)" which I interpret as meaning that you should set MMC_UNSAFE_RESUME *if you're using a card that cannot be removed*. If your card *can* be removed, then it would be extremely foolish to turn on MMC_UNSAFE_RESUME, because then if someone removed your removable card during suspend and modified it in another machine before resuming, you'd get massive filesystem corruption. What am I missing? - Chris. -- Chris Ball One Laptop Per Child