From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755673AbZFAH7A (ORCPT ); Mon, 1 Jun 2009 03:59:00 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752957AbZFAH6w (ORCPT ); Mon, 1 Jun 2009 03:58:52 -0400 Received: from earthlight.etchedpixels.co.uk ([81.2.110.250]:43202 "EHLO www.etchedpixels.co.uk" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752560AbZFAH6w (ORCPT ); Mon, 1 Jun 2009 03:58:52 -0400 Date: Mon, 1 Jun 2009 08:58:14 +0100 From: Alan Cox To: Herbert Xu Cc: David Miller , riel@redhat.com, research@subreption.com, linux-kernel@vger.kernel.org, pageexec@freemail.hu, linux-mm@kvack.org, torvalds@osdl.org, linux-crypto@vger.kernel.org Subject: Re: [PATCH] Use kzfree in crypto API context initialization and key/iv handling Message-ID: <20090601085814.3e010533@lxorguk.ukuu.org.uk> In-Reply-To: <20090601044837.GA9220@gondor.apana.org.au> References: <20090531025720.GC9033@oblivion.subreption.com> <20090530.230213.73434433.davem@davemloft.net> <4A22967C.3080304@redhat.com> <20090531.214623.76344831.davem@davemloft.net> <20090601044837.GA9220@gondor.apana.org.au> X-Mailer: Claws Mail 3.7.0 (GTK+ 2.14.7; 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: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 1 Jun 2009 14:48:37 +1000 Herbert Xu wrote: > On Sun, May 31, 2009 at 09:46:23PM -0700, David Miller wrote: > > > > I don't even want to think about what this does to IPSEC rule creation > > rates, that that matters heavily for cell phone networks where > > hundreds of thousands of nodes come in and out of the server and each > > such entry requires creating 4 IPSEC rules. > > I completely agree. The zeroing of metadata is gratuitous. Zeroing long term keys makes sense but for the short lifepsan keys used on the wire its a bit pointless irrespective of speed (I suspect done properly the performance impact would be close to nil anyway)