[Maverick][GIT PULL] replacement of hacks with Yama

[Tim Gardner]

On 06/29/2010 03:12 PM, Kees Cook wrote:
> This replaces the individual symlink/hardlink/ptrace patches with the Yama
> LSM that is being upstreamed currently.  It includes a SAUCE patch to glue
> it into place on Ubuntu, since there is no upstream way yet to sensibly
> chain arbitrary LSMs.  The PTRACE exception tracking patch is the
> kernel-side of the fix for LP: #589841, but since it is intended for
> upstream, I did not include the bug tags in the patch directly.
>
> The following changes since commit 0719a918ccdaabb2188e94c94c4edceba5b56f03:
>
>    UBUNTU: Ubuntu-2.6.35-6.9 (2010-06-28 12:41:55 -0700)
>
> are available in the git repository at:
>    git://kernel.ubuntu.com/kees/linux-2.6.git maverick-yama
>
> Kees Cook (8):
>        Revert "ptrace: limit scope to attach only (allow read)"
>        Revert "UBUNTU: SAUCE: ptrace: restrict ptrace scope to children"
>        Revert "UBUNTU: SAUCE: fs: block hardlinks to non-accessible sources"
>        Revert "UBUNTU: SAUCE: fs: block cross-uid sticky symlinks"
>        security: Yama LSM
>        security: create task_free security callback
>        Yama: add PTRACE exception tracking and interface
>        SAUCE: security: unconditionally chain to Yama LSM
>
>   Documentation/Yama.txt   |   91 +++++++++++
>   include/linux/prctl.h    |    6 +
>   include/linux/security.h |   13 +-
>   kernel/fork.c            |    1 +
>   kernel/ptrace.c          |   25 ---
>   kernel/sysctl.c          |   28 ----
>   security/Kconfig         |    6 +
>   security/Makefile        |    2 +
>   security/apparmor/lsm.c  |    3 -
>   security/capability.c    |   16 ++
>   security/commoncap.c     |   68 --------
>   security/security.c      |   44 +++++
>   security/yama/Kconfig    |   13 ++
>   security/yama/Makefile   |    3 +
>   security/yama/yama_lsm.c |  404 ++++++++++++++++++++++++++++++++++++++++++++++
>   15 files changed, 594 insertions(+), 129 deletions(-)
>   create mode 100644 Documentation/Yama.txt
>   create mode 100644 security/yama/Kconfig
>   create mode 100644 security/yama/Makefile
>   create mode 100644 security/yama/yama_lsm.c
>

These commits all have '(cherry picked from commit *)' comments in the 
commit log, but as the objects are not from Linus' tree, they are not 
relevant. Can you amend these commit log messages?

I've noted a couple of upstream suggestions for your patch set inclusion 
in Morris' tree. Will you be including those so that we don't drift from 
upstream?

Given that this is an LSM, and LSMs don't stack or chain (I think), how 
is it going to interact with AppArmor ?

rtg
-- 
Tim Gardner tim.gardner at canonical.com