From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 4DF0FCDB46E for ; Thu, 12 Oct 2023 15:46:28 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233222AbjJLPq1 (ORCPT ); Thu, 12 Oct 2023 11:46:27 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:47564 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232023AbjJLPq0 (ORCPT ); Thu, 12 Oct 2023 11:46:26 -0400 Received: from verein.lst.de (verein.lst.de [213.95.11.211]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9374CC0 for ; Thu, 12 Oct 2023 08:46:24 -0700 (PDT) Received: by verein.lst.de (Postfix, from userid 2407) id 9FB9668AA6; Thu, 12 Oct 2023 17:46:19 +0200 (CEST) Date: Thu, 12 Oct 2023 17:46:18 +0200 From: Christoph Hellwig To: Keith Busch Cc: Christoph Hellwig , Kanchan Joshi , Kanchan Joshi , axboe@kernel.dk, sagi@grimberg.me, linux-nvme@lists.infradead.org, vincentfu@gmail.com, ankit.kumar@samsung.com, cpgs@samsung.com, stable@vger.kernel.org, Vincent Fu Subject: Re: [PATCH v3] nvme: fix memory corruption for passthrough metadata Message-ID: <20231012154618.GA17670@lst.de> References: <1891546521.01696823881551.JavaMail.epsvc@epcpadp4> <20231010074634.GA6514@lst.de> <20231011050254.GA32444@lst.de> <20231012043652.GA1368@lst.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.17 (2007-11-01) Precedence: bulk List-ID: X-Mailing-List: stable@vger.kernel.org On Thu, Oct 12, 2023 at 09:31:38AM -0600, Keith Busch wrote: > > I don't want that either, but what can we do against a (possibly > > unprivileged) user corrupting data? > > The unpriviledged access is kind of recent. Maybe limit the scope of > decoding to that usage? Let's just drop support for unpriviledged passthrough for now. That's easily backportable and gives us time to sort out what we can do. Probably only allowing it when SGLs are in use, including a flag to force using it.