Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Tue, 25 Feb 2020 10:35:56 -0800
From:   Kees Cook <keescook@chromium.org>
To:     Daniel Micay <danielmicay@gmail.com>
Cc:     Daniel Axtens <dja@axtens.net>,
        Kernel Hardening <kernel-hardening@lists.openwall.com>,
        Linux-MM <linux-mm@kvack.org>,
        kernel list <linux-kernel@vger.kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH 5/5] [RFC] mm: annotate memory allocation functions with
 their sizes
Message-ID: <202002251035.AD29F84@keescook>
References: <20200120074344.504-1-dja@axtens.net>
 <20200120074344.504-6-dja@axtens.net>
 <CA+DvKQJ6jRHZeZteqY7q-9sU8v3xacSPj65uac3PQfst4cKiMA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CA+DvKQJ6jRHZeZteqY7q-9sU8v3xacSPj65uac3PQfst4cKiMA@mail.gmail.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, Feb 07, 2020 at 03:38:22PM -0500, Daniel Micay wrote:
> There are some uses of ksize in the kernel making use of the real
> usable size of memory allocations rather than only the requested
> amount. It's incorrect when mixed with alloc_size markers, since if a
> number like 14 is passed that's used as the upper bound, rather than a
> rounded size like 16 returned by ksize. It's unlikely to trigger any
> issues with only CONFIG_FORTIFY_SOURCE, but it becomes more likely
> with -fsanitize=object-size or other library-based usage of
> __builtin_object_size.

I think the solution here is to use a macro that does the per-bucket
rounding and applies them to the attributes. Keep the bucket size lists
in sync will likely need some BUILD_BUG_ON()s or similar.

-- 
Kees Cook