From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: util-linux-owner@vger.kernel.org Received: from imap.thunk.org ([74.207.234.97]:37776 "EHLO imap.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1728309AbeHDVsn (ORCPT ); Sat, 4 Aug 2018 17:48:43 -0400 Date: Sat, 4 Aug 2018 15:46:55 -0400 From: "Theodore Y. Ts'o" To: Sami Kerola Cc: util-linux@vger.kernel.org Subject: Re: [PATCH] libuuid: use kernel crypto api Message-ID: <20180804194655.GD4461@thunk.org> References: <20180804191706.10641-1-kerolasa@iki.fi> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20180804191706.10641-1-kerolasa@iki.fi> Sender: util-linux-owner@vger.kernel.org List-ID: On Sat, Aug 04, 2018 at 08:17:06PM +0100, Sami Kerola wrote: > Try to speed up md5 checksums by using hardware accelerators when they are > present with use of kernel crypto api. > > Reference: http://www.chronox.de/libkcapi/html/ch01s02.html > Signed-off-by: Sami Kerola I'm not sure I see the point of this change. Is there a use case you have in mind where people would be using crypto hash based UUID's where performance is a concern? In in that case where someone was trying to create a huge number of crypto hash based UUID's, how much overhead is there in going through the kernel API if there is *not* a hardware accelerator present? - Ted