From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753504Ab0CWEtw (ORCPT ); Tue, 23 Mar 2010 00:49:52 -0400 Received: from mail-qy0-f182.google.com ([209.85.221.182]:55143 "EHLO mail-qy0-f182.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752829Ab0CWEtv convert rfc822-to-8bit (ORCPT ); Tue, 23 Mar 2010 00:49:51 -0400 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; b=p+bJEjqWjJkXYW2hEYHxH2YKsvt1Rd+pwhfV6KNMQadGizcvDX1VHFVtKF92vzxkOm Qo7rfOrYYDnRW4JQQr+mY9CyS2GD1aI1nb8WoOlSo/XEwbUCRF6ZROkgdKUW2V0UvaMA Q0OlO0vRmAUUZH3lwnUZWkJ3Gjqoz/2hdv+Jo= MIME-Version: 1.0 In-Reply-To: <628d1651003222048k1eaf6929r81fd06e401190e0@mail.gmail.com> References: <20100320143240.GB2942@localhost.localdomain> <4BA7A3CF.8070503@trash.net> <628d1651003221834g543e3e6cl4ea39c0886cb4ba1@mail.gmail.com> <7b6bb4a51003221929i16673523nd5f558ccbaace397@mail.gmail.com> <628d1651003221937q58ff39fel467636ab55531128@mail.gmail.com> <628d1651003222048k1eaf6929r81fd06e401190e0@mail.gmail.com> Date: Tue, 23 Mar 2010 12:49:48 +0800 Message-ID: <7b6bb4a51003222149v5346c3eje94ed1182b207206@mail.gmail.com> Subject: Re: [PATCH] Netfilter: Fix integer overflow in net/ipv6/netfilter/ip6_tables.c From: Xiaotian Feng To: wzt wzt Cc: Jan Engelhardt , Patrick McHardy , linux-kernel@vger.kernel.org, netfilter-devel@vger.kernel.org Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Mar 23, 2010 at 11:48 AM, wzt wzt wrote: > 1、 suppose *len = 35,  sizeof(struct ipt_get_entries) = 36 > 2、 set get.size = 0xffffffff from user space > 3、 sizeof(struct ipt_get_entries) + get.size = 36  + 0xffffffff = 35; > 4、 if (*len != sizeof(struct ipt_get_entries) + get.size) was bypassed. > > you can test with c code: > > #include > > int main(void) > { >        unsigned int arg = 0xffffffff; > >        printf("%u\n", arg + 36); >        if (35 != arg + 36) { >                printf("not over flow.\n"); >                return -1; >        } >        printf("arg over flow.\n"); > >        return 0; > } > You didn't get his point... The key point is that sizeof() result type is size_t, slightly modify you code, try the result. > int main(void) > { >        unsigned int arg = 0xffffffff; > unsigned int foo; >        printf("%lu\n", arg + sizeof(foo)); >        if (sizeof(foo) - 1 != arg + sizeof(foo)) { >                printf("not over flow.\n"); >                return -1; >        } >        printf("arg over flow.\n"); > >        return 0; > } > On Tue, Mar 23, 2010 at 11:04 AM, Jan Engelhardt wrote: >> >> On Tuesday 2010-03-23 03:37, wzt wzt wrote: >>>> And, for the addition overflow, can it be caught by >>>> >>>> "if (*len != sizeof(struct ipt_get_entries) + get.size)"  ??? >>> >>>sizeof(struct ipt_get_entries) + get.size can be overflow as *len, >>>get.size is control by user space with copy_from_user(). >> >> The != should catch it. >> >> For 64-bit environments: >> * + invoked with size_t, unsigned int >>  => right side promoted to size_t, result type is size_t >> * != invoked with int and size_t >>  => left-side promoted to ssize_t (probably; but something as large as size_t) >> * get.size is 32-bit bounded, as is *len, >>  so no overflow to worry about at all unless you make >>  sizeof(X) hilariously big close to 2^64 which is rather unlikely. >> >> For 32-bit environments: >> * Let *len be a number of choice (e.g. 36) >> * Find a sizeof(X)+get.size that equals 36 mod 2^32. >> * Since sizeof(X) is const, get.size must be 0 mod 2^32. >> * So get.size must be a multiple of 2^32 to fool the system. >> * Since get.size itself is only a 32-bit quantity, you cannot >>  represent any value larger than 4294967295. >> >> >> What Was What Was Wanted. >> >