From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: =?utf-8?q?=3CBATV+e33eebbc3bbb69d24b52+5394+infradead=2Eorg+d?= =?utf-8?q?wmw2=40twosheds=2Esrs=2Einfradead=2Eorg=3E?= Received: from twosheds.infradead.org ([2001:8b0:10b:1:21d:7dff:fe04:dbe2]) by Galois.linutronix.de with esmtps (TLS1.2:RSA_AES_256_CBC_SHA256:256) (Exim 4.80) (envelope-from =?utf-8?q?=3CBATV+e33eebbc3bbb69d24b52+5394+infradea?= =?utf-8?q?d=2Eorg+dwmw2=40twosheds=2Esrs=2Einfradead=2Eorg=3E=29?= id 1fOOIE-0006Os-BC for speck@linutronix.de; Thu, 31 May 2018 16:11:07 +0200 Received: from [2001:8b0:10b:1:cad3:ffff:fe76:b9bc] by twosheds.infradead.org with esmtpsa (Exim 4.90_1 #2 (Red Hat Linux)) id 1fOOIC-0004S2-Ew for speck@linutronix.de; Thu, 31 May 2018 14:11:04 +0000 Message-ID: <1527775862.3134.42.camel@infradead.org> Subject: [MODERATED] Re: spectrev1+ From: David Woodhouse In-Reply-To: References: Date: Thu, 31 May 2018 15:11:02 +0100 Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit To: speck@linutronix.de List-ID: On Thu, 2018-05-31 at 14:50 +0200, speck for Jiri Kosina wrote: > > [ GPG doesn't encrypt subjects, so I intentionally didn't make it too  >   verbose, sorry ] > > Hi, > > so, according to the information I have, this is likely to go public on  > 2018-06-12 (it's the one referred to as "Bounds Check Bypass Store" in the  > documents). > > I've specifically asked Intel for permission to allow Dan Carpenter to be  > briefed on this issue, so that he could try to teach smatch about  > detecting such patterns (IIUC it should be in principle as simple as  > changing his spectrev1 check to also look for stores and not just reads)  > some time ago. > > Intel granted such permission (but only limited to this particular issue,  > so we shouldn't really be bringing Dan to this list), but after initial  > response, Dan didn't react any further (despite a few pings), and Intel is  > not going to provide any (semantic/coverity/whatever) patches either. > > Has anyone here tried to teach any semantic analysis tool about this  > code pattern? Yes, we've been working with Synopsis to refine the initial rules that Intel had, although we still have a fair amount of false positives. Norbert, did you subscribe to the list yet?