Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1429854AbdDYKf4 (ORCPT <rfc822;w@1wt.eu>);
        Tue, 25 Apr 2017 06:35:56 -0400
Received: from mail-lf0-f65.google.com ([209.85.215.65]:34970 "EHLO
        mail-lf0-f65.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S980937AbdDYKfn (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Tue, 25 Apr 2017 06:35:43 -0400
Date: Tue, 25 Apr 2017 13:35:33 +0300
From: Maksim Salau <maksim.salau@gmail.com>
To: Florian Fainelli <f.fainelli@gmail.com>
Cc: linux-kernel@vger.kernel.org, arnd@arndb.de,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Alan Stern <stern@rowland.harvard.edu>,
        Mathias Nyman <mathias.nyman@linux.intel.com>,
        Peter Chen <peter.chen@nxp.com>, Roger Quadros <rogerq@ti.com>,
        Baoyou Xie <baoyou.xie@linaro.org>, Sekhar Nori <nsekhar@ti.com>,
        Chris Bainbridge <chris.bainbridge@gmail.com>,
        Wolfram Sang <wsa-dev@sang-engineering.com>,
        linux-usb@vger.kernel.org (open list:USB SUBSYSTEM)
Subject: Re: [PATCH v2] usb: core: Warn if an URB's transfer_buffer is on
 stack
Message-ID: <20170425133533.52ba1cb6@gmail.com>
In-Reply-To: <20170424012939.29929-1-f.fainelli@gmail.com>
References: <20170424012939.29929-1-f.fainelli@gmail.com>
X-Mailer: Claws Mail 3.14.1 (GTK+ 2.24.31; x86_64-redhat-linux-gnu)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org
Content-Length: 1020
Lines: 25

> +			} else if (object_is_on_stack(urb->transfer_buffer)) {
> +				WARN_ONCE(1, "transfer buffer is on stack\n");
> +				ret = -EAGAIN;
>  			} else {

Hi,

Has anyone considered a fail-safe mode? I.e.: if a buffer is on stack,
kmemdup it and continue with a warning. This will give us both: functional
drivers (with possibly decreased efficiency in speed and memory footprint)
and warnings for developers that a particular driver requires attention.

This mode will not affect drivers which obey the rules, but will make
offenders at least functional. My main concern is that not every user is able
to detect and report a problem, which prevents drivers from functioning.
Especially this is a problem for not wide spread devices.
Due to this users a seeing unusable equipment, but developers are not
aware of those, even if fixes are trivial.

Such mode has a also a negative effect: if a developer has a device
with an offending driver, he can miss the warning message, since the driver
just works.

Regards,
Maksim.