Re: [PATCH 4/5] man-pages: cap_rights_limit.2: limit FD rights for Capsicum

From: David Drysdale
Date: Mon Jun 30 2014 - 11:35:33 EST


On Mon, Jun 30, 2014 at 07:53:57AM -0700, Andy Lutomirski wrote:
> On Mon, Jun 30, 2014 at 3:28 AM, David Drysdale <drysdale@xxxxxxxxxx> wrote:
> > Signed-off-by: David Drysdale <drysdale@xxxxxxxxxx>
> > ---
> > man2/cap_rights_limit.2 | 171 ++++++++++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 171 insertions(+)
> > create mode 100644 man2/cap_rights_limit.2
> >
> > diff --git a/man2/cap_rights_limit.2 b/man2/cap_rights_limit.2
> > new file mode 100644
> > index 000000000000..3484ee1076aa
> > --- /dev/null
> > +++ b/man2/cap_rights_limit.2
> > @@ -0,0 +1,171 @@
> > +.\"
> > +.\" Copyright (c) 2008-2010 Robert N. M. Watson
> > +.\" Copyright (c) 2012-2013 The FreeBSD Foundation
> > +.\" Copyright (c) 2013-2014 Google, Inc.
> > +.\" All rights reserved.
> > +.\"
> > +.\" %%%LICENSE_START(BSD_2_CLAUSE)
> > +.\" Redistribution and use in source and binary forms, with or without
> > +.\" modification, are permitted provided that the following conditions
> > +.\" are met:
> > +.\" 1. Redistributions of source code must retain the above copyright
> > +.\" notice, this list of conditions and the following disclaimer.
> > +.\" 2. Redistributions in binary form must reproduce the above copyright
> > +.\" notice, this list of conditions and the following disclaimer in the
> > +.\" documentation and/or other materials provided with the distribution.
> > +.\"
> > +.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
> > +.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
> > +.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
> > +.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
> > +.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
> > +.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
> > +.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
> > +.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
> > +.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
> > +.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
> > +.\" SUCH DAMAGE.
> > +.\" %%%LICENSE_END
> > +.\"
> > +.TH CAP_RIGHTS_LIMIT 2 2014-05-07 "Linux" "Linux Programmer's Manual"
> > +.SH NAME
> > +cap_rights_limit \- limit Capsicum capability rights
> > +.SH SYNOPSIS
> > +.nf
> > +.B #include <sys/capsicum.h>
> > +.sp
> > +.BI "int cap_rights_limit(int " fd ", const struct cap_rights *" rights ,
> > +.BI " unsigned int " fcntls ,
> > +.BI " int " nioctls ", unsigned int *" ioctls );
>
> Am I missing the docs for struct cap_rights somewhere?

There's a little bit of discussion in rights.7 (mail 3/5 of the
man-pages set), but there isn't a structure description.

I was trying to keep the structure opaque to userspace, which would
be expected to manipulate the rights with various utility functions
rather than directly.

But I now realize this leaves a gap -- the description of this syscall
doesn't include a full description of its ABI.

So I'll add in a description of the structure to this page -- basically:

struct cap_rights {
__u64 cr_rights[2];
};

with a slightly complicated scheme to encode rights into the bitmask
array. (The encoding scheme is taken from the FreeBSD implementation,
which I've tried to stick to unless there's good reason to change.)

> > +.SH DESCRIPTION
> > +When a file descriptor is created by a function such as
> > +.BR accept (2),
> > +.BR accept4 (2),
> > +.BR creat (2),
> > +.BR epoll_create (2),
> > +.BR eventfd (2),
> > +.BR mq_open (2),
> > +.BR open (2),
> > +.BR openat (2),
> > +.BR pdfork (2),
> > +.BR pipe (2),
> > +.BR pipe2 (2),
> > +.BR signalfd (2),
> > +.BR socket (2),
> > +.BR socketpair (2)
> > +or
> > +.BR timerfd_create (2),
> > +it implicitly has all Capsicum capability rights.
> > +Those rights can be reduced (but never expanded) by using the
> > +.BR cap_rights_limit ()
> > +system call.
> > +Once Capsicum capability rights are reduced, operations on the file descriptor
> > +.I fd
> > +will be limited to those permitted by the remainder of the arguments.
> > +.PP
> > +The
> > +.I rights
> > +argument describes the primary rights for the file descriptor, and
> > +should be prepared using
> > +.BR cap_rights_init (3)
> > +family of functions. The complete list of primary rights can be found in the
> > +.BR rights (7)
> > +manual page.
> > +.PP
> > +If a file descriptor is granted the
> > +.B CAP_FCNTL
> > +primary capability right, the list of allowed
> > +.BR fcntl (2)
> > +commands can be selectively reduced (but never expanded) with the
> > +.I fcntls
> > +argument. The following flags may be specified in the
> > +.I fcntls
> > +argument:
> > +.TP
> > +.B CAP_FCNTL_GETFL
> > +Permit
> > +.B F_GETFL
> > +command.
> > +.TP
> > +.B CAP_FCNTL_SETFL
> > +Permit
> > +.B F_SETFL
> > +command.
> > +.TP
> > +.B CAP_FCNTL_GETOWN
> > +Permit
> > +.B F_GETOWN
> > +command.
> > +.TP
> > +.B CAP_FCNTL_SETOWN
> > +Permit
> > +.B F_SETOWN
> > +command.
> > +.PP
> > +A value of
> > +.B CAP_FCNTL_ALL
> > +for the
> > +.I fcntls
> > +argument leaves the set of allowed
> > +.BR fcntl (2)
> > +commands unchanged.
>
> What about the locking fcntl operations? (Arguably the old crappy
> POSIX lock operations should be flat-out disallowed on capability fds,
> but I see nothing wrong with selectively allowing the new open file
> description locks.)
>
> --Andy

The locking operations are policed against a separate CAP_FLOCK right,
consistently with flock(2). I'll try to improve the wording -- there
are actually a few fcntl operations that are covered by different
rights because they're analogous to other functionality (e.g.
F_GETPIPE_SZ/F_SETPIPE_SZ needs CAP_GETSOCKOPT/CAP_SETSOCKOPT).


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/