On Thu, Apr 17, 2014 at 11:31 AM, Andy Lutomirski <luto@xxxxxxxxxxxxxx> wrote: > On Thu, Apr 17, 2014 at 11:18 AM, Kees Cook <keescook@xxxxxxxxxxxx> wrote: >> On Thu, Apr 17, 2014 at 11:13 AM, Andy Lutomirski <luto@xxxxxxxxxxxxxx> wrote: >>> On Thu, Apr 17, 2014 at 11:05 AM, Kees Cook <keescook@xxxxxxxxxxxx> wrote: >>>> This adds the ability for threads to request seccomp filter >>>> synchronization across their thread group. To support this, >>>> seccomp locking on writes is introduced, along with refactoring >>>> of no_new_privs. Races with thread creation are handled via the >>>> tasklist_list. >>>> >>>> I think all the concerns raised during the discussion[1] of the first >>>> version of this patch have been addressed. However, the races involved >>>> have tricked me before. :) >>>> >>> >>> Would this be easier to use if there were a single syscall to set a >>> seccomp filter and sync threads? That way you wouldn't have to write >>> your filter such that it gives permission to sync threads. >> >> That would be even cleaner, yes. I was hoping to see the new bpf jump >> tables before expanding into new filter calls, with the hope of doing >> it all at the same time. However, I guess we could just include a >> version number in the new call to indicate which filter type it was, >> and include flags (like "threadgroup sync") in there? I'm trying to >> imagine what would be the least painful for future-proofing. > > What's the time frame on the new bpf stuff? If it'll be ready for > 3.16, it might not matter. I don't think seccomp will be extended to deal with that for a while, since it'd require a lot of coordination first. So, for now, I think adding a "flags" field should be sufficient. I'll work that up. -Kees -- Kees Cook Chrome OS Security -- To unsubscribe from this list: send the line "unsubscribe linux-doc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
