Re: [PATCH] define convenient securebits masks for prctl users (v2)

[James Morris <jmorris@namei.org>]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 5741 bytes --]

On Sat, 14 Nov 2009, Michael Kerrisk wrote:

> On Thu, Oct 29, 2009 at 10:51 PM, James Morris <jmorris@namei.org> wrote:
> > On Thu, 29 Oct 2009, Serge E. Hallyn wrote:
> >
> >> Hi James, would you mind taking the following into
> >> security-testing?
> >
> >
> > Applied to
> > git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6#next
> 
> It doesn't look like this change is in 2.6.32-rc7. Is it planned to
> push this out for 2.6.32?

No, this is being queued for the next kernel (2.6.33).

Currently, only fixes to regressions can be upstreamed for 2.6.32.



> 
> Cheers,
> 
> Michael
> 
> >> The securebits are used by passing them to prctl with the
> >> PR_{S,G}ET_SECUREBITS commands.  But the defines must be
> >> shifted to be used in prctl, which begs to be confused and
> >> misused by userspace.  So define some more convenient
> >> values for userspace to specify.  This way userspace does
> >>
> >>       prctl(PR_SET_SECUREBITS, SECBIT_NOROOT);
> >>
> >> instead of
> >>
> >>       prctl(PR_SET_SECUREBITS, 1 << SECURE_NOROOT);
> >>
> >> (Thanks to Michael for the idea)
> >>
> >> This patch also adds include/linux/securebits to the installed headers.
> >> Then perhaps it can be included by glibc's sys/prctl.h.
> >>
> >> Changelog:
> >>       Oct 29: Stephen Rothwell points out that issecure can
> >>               be under __KERNEL__.
> >>       Oct 14: (Suggestions by Michael Kerrisk):
> >>               1. spell out SETUID in SECBIT_NO_SETUID*
> >>               2. SECBIT_X_LOCKED does not imply SECBIT_X
> >>               3. add definitions for keepcaps
> >>         Oct 14: As suggested by Michael Kerrisk, don't
> >>               use SB_* as that convention is already in
> >>               use.  Use SECBIT_ prefix instead.
> >>
> >> Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
> >> Acked-by: Andrew G. Morgan <morgan@kernel.org>
> >> Acked-by: Michael Kerrisk <mtk.manpages@gmail.com>
> >> Cc: Ulrich Drepper <drepper@redhat.com>
> >> Cc: James Morris <jmorris@namei.org>
> >> ---
> >>  include/linux/Kbuild       |    1 +
> >>  include/linux/securebits.h |   24 ++++++++++++++++++------
> >>  2 files changed, 19 insertions(+), 6 deletions(-)
> >>
> >> diff --git a/include/linux/Kbuild b/include/linux/Kbuild
> >> index 1feed71..5a53857 100644
> >> --- a/include/linux/Kbuild
> >> +++ b/include/linux/Kbuild
> >> @@ -330,6 +330,7 @@ unifdef-y += scc.h
> >>  unifdef-y += sched.h
> >>  unifdef-y += screen_info.h
> >>  unifdef-y += sdla.h
> >> +unifdef-y += securebits.h
> >>  unifdef-y += selinux_netlink.h
> >>  unifdef-y += sem.h
> >>  unifdef-y += serial_core.h
> >> diff --git a/include/linux/securebits.h b/include/linux/securebits.h
> >> index d2c5ed8..3340617 100644
> >> --- a/include/linux/securebits.h
> >> +++ b/include/linux/securebits.h
> >> @@ -1,6 +1,15 @@
> >>  #ifndef _LINUX_SECUREBITS_H
> >>  #define _LINUX_SECUREBITS_H 1
> >>
> >> +/* Each securesetting is implemented using two bits. One bit specifies
> >> +   whether the setting is on or off. The other bit specify whether the
> >> +   setting is locked or not. A setting which is locked cannot be
> >> +   changed from user-level. */
> >> +#define issecure_mask(X)     (1 << (X))
> >> +#ifdef __KERNEL__
> >> +#define issecure(X)          (issecure_mask(X) & current_cred_xxx(securebits))
> >> +#endif
> >> +
> >>  #define SECUREBITS_DEFAULT 0x00000000
> >>
> >>  /* When set UID 0 has no special privileges. When unset, we support
> >> @@ -12,6 +21,9 @@
> >>  #define SECURE_NOROOT                        0
> >>  #define SECURE_NOROOT_LOCKED         1  /* make bit-0 immutable */
> >>
> >> +#define SECBIT_NOROOT                (issecure_mask(SECURE_NOROOT))
> >> +#define SECBIT_NOROOT_LOCKED (issecure_mask(SECURE_NOROOT_LOCKED))
> >> +
> >>  /* When set, setuid to/from uid 0 does not trigger capability-"fixup".
> >>     When unset, to provide compatiblility with old programs relying on
> >>     set*uid to gain/lose privilege, transitions to/from uid 0 cause
> >> @@ -19,6 +31,10 @@
> >>  #define SECURE_NO_SETUID_FIXUP               2
> >>  #define SECURE_NO_SETUID_FIXUP_LOCKED        3  /* make bit-2 immutable */
> >>
> >> +#define SECBIT_NO_SETUID_FIXUP       (issecure_mask(SECURE_NO_SETUID_FIXUP))
> >> +#define SECBIT_NO_SETUID_FIXUP_LOCKED \
> >> +                     (issecure_mask(SECURE_NO_SETUID_FIXUP_LOCKED))
> >> +
> >>  /* When set, a process can retain its capabilities even after
> >>     transitioning to a non-root user (the set-uid fixup suppressed by
> >>     bit 2). Bit-4 is cleared when a process calls exec(); setting both
> >> @@ -27,12 +43,8 @@
> >>  #define SECURE_KEEP_CAPS             4
> >>  #define SECURE_KEEP_CAPS_LOCKED              5  /* make bit-4 immutable */
> >>
> >> -/* Each securesetting is implemented using two bits. One bit specifies
> >> -   whether the setting is on or off. The other bit specify whether the
> >> -   setting is locked or not. A setting which is locked cannot be
> >> -   changed from user-level. */
> >> -#define issecure_mask(X)     (1 << (X))
> >> -#define issecure(X)          (issecure_mask(X) & current_cred_xxx(securebits))
> >> +#define SECBIT_KEEP_CAPS     (issecure_mask(SECURE_KEEP_CAPS))
> >> +#define SECBIT_KEEP_CAPS_LOCKED (issecure_mask(SECURE_KEEP_CAPS_LOCKED))
> >>
> >>  #define SECURE_ALL_BITS              (issecure_mask(SECURE_NOROOT) | \
> >>                                issecure_mask(SECURE_NO_SETUID_FIXUP) | \
> >> --
> >> 1.6.1
> >>
> >
> > --
> > James Morris
> > <jmorris@namei.org>
> >
> 
> 
> 
> -- 
> Michael Kerrisk
> Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
> Author of "The Linux Programming Interface" http://blog.man7.org/
> 

-- 
James Morris
<jmorris@namei.org>