Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751663AbdGaWHE (ORCPT ); Mon, 31 Jul 2017 18:07:04 -0400 Received: from mail-io0-f179.google.com ([209.85.223.179]:35278 "EHLO mail-io0-f179.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751595AbdGaWHB (ORCPT ); Mon, 31 Jul 2017 18:07:01 -0400 MIME-Version: 1.0 In-Reply-To: References: <20170731065016.2947796-1-arnd@arndb.de> From: Kees Cook Date: Mon, 31 Jul 2017 15:06:59 -0700 X-Google-Sender-Auth: lzAtgdamUOG_zpcb1dOQzyt1nZo Message-ID: Subject: Re: [PATCH] infiniband: avoid overflow warning To: Arnd Bergmann Cc: Moni Shoua , Doug Ledford , Sean Hefty , Hal Rosenstock , Daniel Micay , "Kalderon, Michal" , Ariel Elior , "David S. Miller" , Bart Van Assche , Parav Pandit , Noa Osherovich , linux-rdma , Linux Kernel Mailinglist Content-Type: text/plain; charset="UTF-8" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2289 Lines: 50 On Mon, Jul 31, 2017 at 2:52 PM, Arnd Bergmann wrote: > On Mon, Jul 31, 2017 at 11:18 PM, Kees Cook wrote: >> On Mon, Jul 31, 2017 at 2:10 PM, Arnd Bergmann wrote: >>> On Mon, Jul 31, 2017 at 10:58 PM, Kees Cook wrote: >>>> On Mon, Jul 31, 2017 at 12:30 AM, Arnd Bergmann wrote: >>>>> On Mon, Jul 31, 2017 at 9:08 AM, Moni Shoua wrote: >>>>>>> break; >>>>>>> default: >>>>>>> return -EINVAL; >>>>>> what happens if you replace 16 with sizeof(struct in6_addr)? >>>>> >>>>> Same thing: the problem is that gcc already knows the size of the structure we >>>>> pass in here, and it is in fact shorter. >>>> >>>> So gcc is ignoring both the cast (to 16 byte struct in6_addr) and the >>>> caller's actual 128 byte struct sockaddr_storage, and looking only at >>>> struct sockaddr? That seems really weird. >>> >>> Using a sockaddr_storage on the stack would address the warning, but >>> the question was about just changing the hardcoded 16 to a sizeof() >>> operation, and that has no effect. >> >> Right, I didn't mean that; I was curious why the fortify macro >> resulted in an error at all. The callers are casting from struct >> sockaddr_storage (large enough) to struct sockaddr (not large enough), >> and then the inline is casting back to sockaddr_in6 (large enough). I >> would have expected fortify to check either sockaddr_storage or >> sockaddr_in6, but not sockaddr. > > To clarify: this happens in inetaddr_event(), which has a sockaddr_in > on the stack, not a sockaddr_storage. I tried casting the sockaddr_in > pointer to sockaddr_storage, but that did not help. Changing the Oooh, I see now. Yeah, addr_event() sees it directly as struct sockaddr and even with the resulting inlining into inetaddr_event(), the dead-code analysis doesn't eliminate the AF_INET6 case, which is a shame. > type of the stack variable to sockaddr_storage does help. That seems like an unfortunate waste of stack space for a false positive. :) I think your original fix is fine. (In fact, I think it's actually more robust since there isn't a hard-coded "16" -- not that it'll ever change, of course.) -Kees -- Kees Cook Pixel Security