lkml.org 
[lkml]   [2011]   [May]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH 3/5] v2 seccomp_filters: Enable ftrace-based system call filtering
From
On Thu, May 26, 2011 at 1:33 PM, Linus Torvalds
<torvalds@linux-foundation.org> wrote:
> On Thu, May 26, 2011 at 10:38 AM, Will Drewry <wad@chromium.org> wrote:
>>>
>>> One option is to just not ever allow execve() from inside a restricted
>>> environment.
>>
>> That'd certainly be fine with me.
>
> So if it ends up being purely a "internal to the process" thing, then
> I'm much happier about it - it not only limits the scope of things
> sufficiently that I don't worry too much about security issues, but it
> makes it very clear that it's about a process going into "lock-down"
> mode on its own.
>
> It also gets rid of all configuration - one of the things that makes
> most security frameworks (look at selinux, but also just ACL's etc)
> such a crazy rats nest is the whole "set up for other processes". If
> it's designed very much to be about just the "self" process (after
> initialization etc), then I think that avoids pretty much all the
> serious issues.
>
> A lot of server processes could probably use it as a way to say "Hey,
> I guarantee that I will only open new files read-only, and will only
> write to the socket that was already opened for me by the accept", and
> explicitly limit their worker threads that way.
>
> If that is really sufficient for some chrome sandboxing, then hey,
> that's all fine.

It adds some hoops, but less than exist today.

> Sometimes limiting yourself (rather than looking for some bigger
> "generic" solution) is the right answer.

I will very happily validate usage and repost with a self-limited
patch series. Doing so makes the change much more explicitly an
expansion of seccomp, which keeps things sane.

Thanks!
will


\
 
 \ /
  Last update: 2011-05-26 20:51    [from the cache]
©2003-2011 Jasper Spaans