From mboxrd@z Thu Jan 1 00:00:00 1970 Reply-To: kernel-hardening@lists.openwall.com Message-ID: <510ADDAB.3010500@linux.vnet.ibm.com> Date: Thu, 31 Jan 2013 16:10:03 -0500 From: Corey Bryant MIME-Version: 1.0 References: <510A8F11.6050908@linux.vnet.ibm.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [kernel-hardening] Secure Open Source Project Guide To: kernel-hardening@lists.openwall.com Cc: Kees Cook , Anthony Liguori , Frank Novak , George Wilson , Joel Schopp , Kevin Wolf , Warren Grunbok II List-ID: On 01/31/2013 01:37 PM, Kees Cook wrote: > On Thu, Jan 31, 2013 at 7:34 AM, Corey Bryant wrote: >> In light of events like this http://lwn.net/Articles/535149/ "China, GitHub >> and the man-in-the-middle (Greatfire)", we are thinking that a guide for >> securing open source projects is needed. For example, recommending pull >> requests or commits be PGP signed are a few things we've discussed that >> could defend against a MITM attack inserting malicious code. >> >> Does anyone have any thoughts as to where we could publish such a guide? >> Perhaps the Linux Foundation? >> >> I believe we have the resources on this mailing list to work through the >> details and put together a succinct guide that we could take to a wider >> audience. > > Yeah, sounds good. I think we could easily use the kernel-security > wiki to work on it initially, and if it needs a different home in the > end, we can move it then. > > -Kees > > -- > Kees Cook > Chrome OS Security > > > Does it make sense to get everyone edit access to the wiki? If not I can set up a page for it and get input from folks here on the mailing list as it progresses and update the wiki myself. We should probably start by gathering a list of ideas to include in the guide. Some initial ideas that come to mind are: * Secure programming practices (Secure "Programming for Linux and Unix HOWTO" is a good reference for Linux though probably out of date) * Performing secure code reviews and detecting common vulnerabilities * Ensuring code is reviewed by trusted parties and proper patch tagging is used * Signing of releases, pull requests, patches, commits, etc by trusted parties * Removing vulnerabilities with automated tooling (Static/Dynamic analysis, Fuzzing) Any thoughts? -- Regards, Corey Bryant