From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754795AbcGVSDM (ORCPT ); Fri, 22 Jul 2016 14:03:12 -0400 Received: from mga01.intel.com ([192.55.52.88]:5012 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751488AbcGVSDK (ORCPT ); Fri, 22 Jul 2016 14:03:10 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.28,405,1464678000"; d="scan'208";a="143931079" Subject: [PATCH 0/3] x86, pkeys: fix prefetch/pkeys interaction To: linux-kernel@vger.kernel.org Cc: x86@kernel.org, luto@kernel.org, Dave Hansen From: Dave Hansen Date: Fri, 22 Jul 2016 11:03:10 -0700 Message-Id: <20160722180310.99F5BCFF@viggo.jf.intel.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org The first two patches here are useful in any case, I think. But, as for the third: There are no known prefetch errata on processors that support memory protection keys. There have not been any that I can find in any recent generations, either. But, if there were a future erratum, we would need this. Otherwise, apps who hit the theoretical erratum and used pkeys would not be fixed up by the prefetch erratum detection code. It also occurs to me that *if* there were an erratum on a modern CPU, we might not know because we have so many workarounds in place. So, I'm submitting this, but I don't feel that strongly about it. It doesn't fix a real problem, but it's also not that much code, or in any kind of fast path.