From mboxrd@z Thu Jan 1 00:00:00 1970 Reply-To: kernel-hardening@lists.openwall.com Date: Tue, 28 Jun 2011 02:37:12 +0400 From: Solar Designer Message-ID: <20110627223712.GA8685@openwall.com> References: <1308917318-4749-1-git-send-email-segoon@openwall.com> <4E07F1C0.2070305@jp.fujitsu.com> <20110627070300.GA4463@albatros> <4E08324D.9040605@jp.fujitsu.com> <20110627085242.GA6635@albatros> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20110627085242.GA6635@albatros> Subject: [kernel-hardening] Re: [PATCH 1/2] proc: restrict access to /proc/PID/io To: kernel-hardening@lists.openwall.com Cc: Vasiliy Kulikov , KOSAKI Motohiro , linux-kernel@vger.kernel.org, balbir@linux.vnet.ibm.com, akpm@linux-foundation.org, viro@zeniv.linux.org.uk, rientjes@google.com, wilsons@start.ca, security@kernel.org, eparis@redhat.com, Linus Torvalds List-ID: On Mon, Jun 27, 2011 at 12:52:42PM +0400, Vasiliy Kulikov wrote: > As to random bytes - if it is very predictable (e.g. rand() % 10000) one > may restore the original value. But it would still do much harm to good > programs (io stats going up and down!). Adding really random bytes > seems somewhat too complicated for these needs to me. Random noise doesn't help very much even if it's totally unpredictable and even if it's much louder than the signal. It will only increase the number of samples needed to see the signal through the noise. More effective ways to deal with side-channel attacks are to make things appear constant or, better yet, to remove the side-channel altogether if possible. I'd happily break iotop for non-admins on many of my systems, so please give me a way to do it. http://en.wikipedia.org/wiki/Side_channel_attack#Countermeasures Alexander