From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 569D9C433EF for ; Mon, 18 Jul 2022 23:09:19 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S236456AbiGRXJR (ORCPT ); Mon, 18 Jul 2022 19:09:17 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:35706 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230165AbiGRXJH (ORCPT ); Mon, 18 Jul 2022 19:09:07 -0400 Received: from relay5-d.mail.gandi.net (relay5-d.mail.gandi.net [217.70.183.197]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 4818D764E for ; Mon, 18 Jul 2022 16:09:05 -0700 (PDT) Received: (Authenticated sender: joao@overdrivepizza.com) by mail.gandi.net (Postfix) with ESMTPA id 9D6581C0002; Mon, 18 Jul 2022 23:08:59 +0000 (UTC) MIME-Version: 1.0 Date: Mon, 18 Jul 2022 16:08:59 -0700 From: Joao Moreira To: Sami Tolvanen Cc: Thomas Gleixner , Peter Zijlstra , "Torvalds, Linus" , LKML , the arch/x86 maintainers , Tim Chen , Josh Poimboeuf , "Cooper, Andrew" , Pawan Gupta , Johannes Wikner , Alyssa Milburn , Jann Horn , "H.J. Lu" , "Moreira, Joao" , "Nuzman, Joseph" , Steven Rostedt , "Gross, Jurgen" , Masami Hiramatsu , Alexei Starovoitov , Daniel Borkmann , Peter Collingbourne Subject: Re: [patch 00/38] x86/retbleed: Call depth tracking mitigation In-Reply-To: References: <20220716230344.239749011@linutronix.de> <87wncauslw.ffs@tglx> <87tu7euska.ffs@tglx> <87o7xmup5t.ffs@tglx> <87lesqukm5.ffs@tglx> <2f7f899cb75b79b08b0662ff4d2cb877@overdrivepizza.com> Message-ID: X-Sender: joao@overdrivepizza.com Content-Type: text/plain; charset=US-ASCII; format=flowed Content-Transfer-Encoding: 7bit Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > The problem with this is that the cmpl instruction contains the full > type hash, which means that any instruction that's FIXED_OFFSET from > the cmpl is a valid indirect call target as far as KCFI is concerned. > -6 was chosen specifically to make the ud2 the only possible target. Ugh. The bitter truth. I'll think a bit further.