Home Original page

[PING][PATCH] [RFCv2] Document Security process for binutils

Mike Frysinger vapier@gentoo.org
Tue Jan 26 02:46:35 GMT 2021 
On 19 Jan 2021 14:29, Siddhesh Poyarekar wrote:
> On 1/19/21 1:57 PM, Alan Modra wrote:
> > If you are serious about security then "don't run any of binutils as
> > root" is sufficient advice.  I don't think any of this documentation
> > in info files is necessary for binutils, and I'd rather not see more
> > people fuzzing binutils.
> > 
> > As someone who has spent rather a lot of time over the past year
> > responding to asan, ubsan, and fuzzed object file bug reports, I can
> > tell you that the great majority of those reports do not fix real
> > bugs.  By "real bugs", I mean bugs that might conceivably be triggered
> > by real object files created by compilers or assemblers.
> 
> What you said basically implies that running binutils tools in anything 
> other than a fully trusted environments is unsupported, which eliminates 
> all usage of binutils tools where they may be invoked remotely.
> 
> Also, running as root is not the only vector.  For example, one could in 
> theory achieve remote code execution if binutils is invoked on untrusted 
> binaries remotely.  It could either be directly through a service or by 
> chaining with another bug that causes generation or storage of invalid 
> binaries.
> 
> > Yes, we do have libbfd and libopcodes that are used by more than just
> > binutils and gdb, but the number of projects is small.
> 
> Unfortunately that number is not zero and it is conceivable that the 
> libraries may be used in an untrusted context.
> 
> The effect of such documentation is to clearly define usage patterns 
> that will be accepted as CVE-worthy and as a result, limit them 
> considerably.  In that sense, we're on the same team!
> 
> Perhaps explicitly stating that "Bugs in binutils that need tools to be 
> run as root to be locally exploitable will be treated as regular bugs 
> and not as security flaws" is a worthy addition?  Are there any other 
> constraints for considering bugs as security issues that you can think 
> of?  We could keep adding those as we go along.

i'm with Alan here with the current state of the world: it is not safe to
run binutils (or gcc fwiw) on untrusted inputs unless the overall execution
environment has been isolated/secured in someway.  i understand that some
people will find this surprising, but that is the reality of the codebase
today.  i've been telling people this in Gentoo for decades.  i don't see
the situation changing until someone steps up to comprehensively tackle it.

i agree that we should have a document clearly defining the security
posture of the project as people will go looking for it.  but trying to do
embargoes or new branch releases for every bug with exploit possibilities
will be useless drain on an already limited developer pool.  bugs should be
treated as bugs which means using bugzilla to report them.

if you think there should be a better answer here, then realistically you'll
have to get a company/etc... to provide dedicated resources in this space.
imo, we're talking something like isolating all bfd operations into a secure
context (e.g. dropping caps, using seccomp mode1 if possible, etc...).
-mike
More information about the Binutils mailing list