On 02/16/2017 01:01 PM, Andy Lutomirski wrote: > On Thu, Feb 16, 2017 at 10:47 AM, Tyler Hicks <tyhi...@canonical.com> wrote: >> On 02/15/2017 09:14 PM, Andy Lutomirski wrote: >>> On Mon, Feb 13, 2017 at 7:45 PM, Tyler Hicks <tyhi...@canonical.com> wrote: >>>> This patch creates a read-only sysctl containing an ordered list of >>>> seccomp actions that the kernel supports. The ordering, from left to >>>> right, is the lowest action value (kill) to the highest action value >>>> (allow). Currently, a read of the sysctl file would return "kill trap >>>> errno trace allow". The contents of this sysctl file can be useful for >>>> userspace code as well as the system administrator. >>> >>> Would this make more sense as a new seccomp(2) mode a la >>> SECCOMP_HAS_ACTION? Then sandboxy things that have no fs access could >>> use it. >>> >> >> It would make sense for code that needs to check which actions are >> available. It wouldn't make sense for administrators that need to check >> which actions are available unless libseccomp provided a wrapper utility. >> >> Is this a theoretical concern or do you know of a sandboxed piece of >> code that cannot access the sysctl before constructing a seccomp filter? >> > > It's semi-theoretical. But suppose I unshare namespaces, unmount a > bunch of stuff, then ask libseccomp to install a filter. (I've > written code that does exactly that.) libseccomp won't be able to > read the sysctl.
That's a good point. It seems like we might need both mechanisms (SECCOMP_HAS_ACTION for code and actions_avail for humans). Tyler > > --Andy >
signature.asc
Description: OpenPGP digital signature
-- Linux-audit mailing list Linux-audit@redhat.com https://www.redhat.com/mailman/listinfo/linux-audit