mac_do(4): Revamp manual page after MAC/do updates

The new manual page in particular describes MAC/do's new rules syntax
and the jail support, as well as security considerations explaining the
overall design and how to leverage it in the most secure fashion.

Reviewed by:    bapt, otis, Alexander Ziaee <concussious@runbox.com> (in part)
Sponsored by:   The FreeBSD Foundation
Differential Revision:  https://reviews.freebsd.org/D48153

(cherry picked from commit bc201841d13928c2a088fb07ac0a010b36eafa13)
This commit is contained in:
Olivier Certner 2024-12-19 22:13:12 +01:00
parent e286a03736
commit f74cd62b5c
No known key found for this signature in database
GPG key ID: 8CA13040971E2627

View file

@ -1,38 +1,274 @@
.\"-
.\" Copyright (c) 2024 Baptiste Daroussin <bapt@FreeBSD.org>
.\"
.\" SPDX-License-Identifier: BSD-2-Clause
.\"
.Dd May 22, 2024
.\" Copyright (c) 2024 Baptiste Daroussin <bapt@FreeBSD.org>
.\" Copyright (c) 2024 The FreeBSD Foundation
.\"
.\" Portions of this documentation were written by Olivier Certner
.\" <olce@FreeBSD.org> at Kumacom SARL under sponsorship from the FreeBSD
.\" Foundation.
.\"
.Dd December 19, 2024
.Dt MAC_DO 4
.Os
.Sh NAME
.Nm mac_do
.Nd "policy allowing user to execute program as another user"
.Nd "policy allowing unprivileged users to change process credentials"
.Sh SYNOPSIS
To compile the
.Nm
policy into your kernel, place the following lines
in your kernel configruation file:
.Sy mac_do
policy into your kernel, place the following lines in your kernel configuration
file:
.Bd -ragged -offset indent
.Cd "options MAC"
.Cd "options MAC_DO"
.Ed
.Pp
Alternately, to load this policy module at boot time, place the following line
in your kernel configuration file:
.Bd -ragged -offset indent
.Cd "options MAC"
.Ed
.Pp
and in
.Xr loader.conf 5 :
.Bd -literal -offset indent
mac_do_load="YES"
.Ed
.Sh DESCRIPTION
The
.Nm
policy grants users the ability to run processs as other users
according to predefined rules.
policy module allows unprivileged users to change process credentials according
to rules configured by the administrator.
It supports per-jail configuration.
.Pp
The exact set of kernel privileges granted are:
.Bl -inset -compact -offset indent
.It Dv PRIV_CRED_SETGROUPS
.It Dv PRIV_CRED_SETUID
Currently, the
.Nm
policy module only produces effects to processes spwaned from the
.Pa /usr/bin/mdo
executable, please see
.Xr mdo 1
for more details on this program.
.Sh CREDENTIALS RULES
Rules specify which transitions of process credentials
.Nm
will allow, based on current process credentials and the desired final ones.
They are passed by an administrator in the form of a string having the specific
syntax described below in a top-bottom manner.
They have been designed to be able to finely describe the desired target
credentials in a safe and compact way.
.Ss Top-Level List of Rules
At the top, rules are a possibly empty list of individual rules separated by
a semi-colon
.Pq Ql ";" :
.Dl Ao rules Ac \ ⟶\ Oo Ao rule Ac Oo So ";" Sc Ao rule Ac Oc Ns * Oc
They form a disjunction, i.e.,
.Nm
authorizes a credentials transition as soon as at least one rule in the list
matches.
.Pp
One rule is composed of a
.Li Aq from
part
.Pq also called Dq match
and a
.Li Aq to
part
.Pq also called Dq target ,
in this order, separated by a colon
.Pq Ql ":" :
.Dl Ao rule Ac \ ⟶\ Ao from Ac So ":" Sc Ao to Ac
.Ss Rule's Ao from Ac Part
The first part of a rule,
.Li Aq from ,
is matched against the credentials of the process requesting some credentials
transition.
It has the form:
.Dl Ao from Ac \ ⟶\ Ao type Ac So = Sc Ao id Ac
.Pp
.Li Aq type
must be:
.Dl Ao type Ac \ ⟶\ Op So uid Sc | So gid Sc
i.e., one of the literal strings
.Ql uid
or
.Ql gid .
.Li Aq id
must be the numerical ID of a user or group, and is matched with the current
process real ID of the corresponding type.
.Ss Rule's Ao to Ac Part
The second part of a rule,
.Li Aq to ,
is a comma-separated
.Pq Ql ","
non-empty list of target clauses:
.Dl Ao to Ac \ ⟶\ Ao target_clause Ac Oo So "," Sc Ao target_clause Ac Oc Ns *
Target clauses of a given rule also form a disjunction, i.e., the IDs they
specify are alternatives for the target credentials, except in some cases
described below.
.Pp
The next subsections describe the syntax of target clauses, the defaults that
apply and the principle of non-redundancy and non-contradiction in each rule's
.Li Aq to
part.
.Ss Target Clauses
A target clause in a rule's
.Li Aq to
part must be of one of the following forms:
.Dl Ao target_clause Ac \ ⟶\ So any Sc
.Dl Ao target_clause Ac \ ⟶\ Ao flags Ac Ao type Ac So = Sc Ao id Ac
The first form is a compact way to specify that any target credentials are
allowed.
The second form is similar to that of
.Li Aq from
clauses, with the following extensions:
.Bl -bullet -compact
.It
.Li Aq id
may also be a literal
.Ql *
or
.Ql any
or
.Ql "." .
.Ql *
and
.Ql any
both designate any ID for the specified
.Li Aq type ,
and are treated identically.
.Ql "."
designates the process' current IDs for the specified
.Li Aq type ,
as explained below.
.It
.Li Aq flags
may contain at most one of the
.Ql + ,
.Ql -
and
.Ql "!"
characters, and may be non-empty only when
.Li Aq type
is
.Ql gid .
Additionally, if
.Li Aq id
is
.Ql *
or
.Ql any ,
only the
.Ql +
flag may appear.
.El
.Pp
For target clauses of
.Ql gid
type, an absence of flag indicates that the specified group ID is allowed as the
real, effective and/or saved group IDs
.Pq the Do primary Dc groups .
Conversely, the presence of any allowed flag indicates that the specification
concerns supplementary groups.
Each flag has a specific meaning:
.Bl -bullet -compact
.It
.Ql +
indicates that the group ID is allowed as a supplementary group.
.It
.Ql "!"
indicates that the group ID is mandatory, i.e., it must be listed in the
supplementary groups.
.It
.Ql -
indicates that the group ID must not be listed in the supplementary groups.
.El
A specification with
.Ql -
is only useful in conjunction with a
.Ql + Ns
-tagged specification where only one of them has
.Ql "."
as its
.Li Aq id .
Target clauses having the
.Ql "!"
or
.Ql -
flag are
.Dq forcing
clauses, and as such do not take part in the disjunction of the other
target clauses but rather unconditionally apply in their rule.
.Pp
.Ql "."
is a placeholder for IDs that the calling process already has on privilege
check.
For type
.Ql uid ,
it designates any of the process' real, effective or
saved user IDs.
For type
.Ql gid ,
its effect depends on whether flags are present.
If none is present, it designates any of the process' real, effective or saved
group IDs.
If one is present, it designates any of the process' supplementary groups.
.Ss Defaults for the Ao to Ac Part
If the
.Li Aq to
part does not list a target clause with type
.Ql uid ,
any of the current user IDs of the calling process is accepted.
In other words, in this case,
.Nm
behaves as if a target clause of:
.Dl uid=.
had been listed.
.Pp
Similarly, if the
.Li Aq to
part does not list a target clause with type
.Ql gid ,
all the groups of the calling process are assumed to be required.
More precisely, each of the desired real, effective and saved group IDs must be
one of the current real, effective or saved group ID, and all supplementary
groups must be the same as those that are current.
It is as if the
.Li Aq to
part had contained the following two clauses:
.Dl gid=.,!gid=.
.Ss Non-Redundancy and Non-Contradiction in a Ao to Ac Part
No two target clauses of a single rule may express the exact same logical intent
nor contradictory ones.
.Pp
In practice, no two clauses may display the same ID except for group IDs but
only if, each time the same ID appears, it does so with a different flag, or no
flags only once.
Additionally, the specified flags in multiple occurences must not be
contradictory.
For example, the same group ID appearing with both
.Ql +
and
.Ql -
will cause rejection of the rule.
.Ss Parsing Specifics
Any amount of whitespace is allowed around tokens of the above grammar, except
that there may be no spaces between
.Li Aq flags
and
.Li Aq id
in target clauses.
.Pp
For convenience, numerical IDs may be specified as negative integers, which are
then converted to unsigned ones as specified in the C standard for the
.Vt uid_t
and
.Vt gid_t
types, which are both 64-bit unsigned integers.
.Sh RUNTIME CONFIGURATION
The following
.Xr sysctl 8
MIBs are available:
knobs are available:
.Bl -tag -width indent
.It Va security.mac.do.enabled
Enable the
@ -40,39 +276,177 @@ Enable the
policy.
(Default: 1).
.It Va security.mac.do.rules
The set of rules.
The list of credential rules, whose syntax is described in the
.Sx CREDENTIALS RULES
section above.
This list is specific to each jail.
Please see the
.Sx JAIL SUPPORT
section below for more details on the interaction of
.Nm
with jails.
.It Va security.mac.do.print_parse_error
Logs a message on trying to set incorrect rules via the
.Va security.mac.do.rules
.Xr sysctl 8
knob.
.El
.Sh JAIL SUPPORT
.Nm
supports per-jail configuration of rules.
.Pp
By default, at creation, a new jail has no credentials rules, effectively
disabling
.Nm
for its processes.
.Pp
The following jail parameters are defined:
.Bl -tag -width indent
.It Va mac.do
Possible values are:
.Bl -tag -width "'disable'" -compact
.It Ql enable
.Nm
will enforce specific credential rules in the jail.
The
.Va mac.do.rules
jail parameter must also be set in this case.
.It Ql disable
Disables
.Nm
in the jail.
Strictly equivalent to jail creation's default behavior and to setting the rules
to an empty string.
.It Ql inherit
The jail's credentials rules are inherited from the jail's parent
.Pq which may themselves have been inherited .
Modified rules propagate to all children jails configured for inheritance.
.El
.It Va mac.do.rules
The credentials rules for the jail.
It is always equal to the value that can be retrieved by the
.Xr sysctl 8
knob
.Va security.mac.do.rules
described in section
.Sx RUNTIME CONFIGURATION .
If set, and the jail parameter
.Va mac.do
is not so explicitly, the value of the latter will default to
.Ql disable
if empty, else to
.Ql enable .
.El
.Pp
The rules consist of a list of elements separated by
.So , Sc .
Each element is of the form
.Sm off
.Do
.Op Cm uid | Cm gid
.Li =
.Ar fid
.Li :
.Ar tid
.Dc
.Sm on .
Where
.Ar fid
is the uid or gid of the user or group the rule applies to, and
.Ar tid
is the uid of the targetted user.
Two special forms are accepted for
.Ar tid :
.Va any
or
.Va * ,
which allow to target any user.
Each jail must have
.Xr mdo 1
installed at path
.Pa /usr/bin/mdo ,
as this path is currently not configurable.
.Sh EXAMPLES
The following rule:
Here are several examples of single rules matching processes having a real user
ID of 10001:
.Bl -tag -width indent
.It Li uid=10001:uid=10002
Allows the process to switch any of its real, effective or saved user ID to
10002, but keeping the groups it is already in, and with the same
primary/supplementary groups split.
.It Li uid=10001:uid=10002,uid=10003
Same as the first example, but also allows to switch to UID 10003 instead of
10002.
.It Li uid=10001:uid=10002,gid=10002
Same as the first example, but the new primary groups must be set to 10002 and
no supplementary groups should be set.
.It Li uid=10001:uid=10002,gid=10002,+gid=.\&
Same as the previous example, but in addition allowing to retain any current
supplementary groups.
.It Li uid=10001:uid=10002,gid=10002,!gid=.\&
Same as the previous example, but with the additional constraint that all
current supplementary groups must be kept.
.It Li uid=10001:uid=10002,gid=10002,+gid=.,-gid=10001
Same as
.Ql uid=10001:uid=10002,gid=10002,+gid=.\&
above, but 10001 cannot be retained as a supplementary group.
.It Li uid=10001:uid=10002,gid=10002,+gid=.,!gid=10003
Same as
.Ql uid=10001:uid=10002,gid=10002,+gid=.\&
above, with the additional constraint that 10003 must appear in the
supplementary groups.
.It Li uid=10001:uid=10002,gid=*,+gid=*
Same as the first example, but lifting any constraints on groups, allowing the
process to become part of any groups it sees fit.
.El
.Pp
.Dl security.mac.do.rules=uid=1001:80,gid=0:any
.Pp
means the user with the uid 1001 can execute processes as user with uid 80,
all the users which belongs to the group gid 0 can execute processes as any user.
Here are several examples of single rules matching processes having a real group
ID of 10001:
.Bl -tag -width indent
.It Li gid=10001:uid=0
Makes 10001 a more powerful
.Ql wheel
group, allowing its members to switch to root without password.
.It Li gid=10001:gid=10002
Allows the process to enter GID 10002 as a primary group, but only if
giving up all its supplementary groups.
.It Li security.mac.do.rules=gid=10001:gid=10002,+gid=.\&
Same as the previous example, but allows to retain any current supplementary
groups.
.It Li gid=10001:gid=10002,!gid=.\&
Same as the previous example, but with the additional constraint that all
current supplementary groups must be kept.
.El
.Sh SEE ALSO
.Xr mdo 1 ,
.Xr mac 4
.Xr setcred 2 ,
.Xr mac 4 ,
.Xr jail 8 ,
.Xr sysctl 8
.Sh AUTHORS
.An Olivier Certner Aq Mt olce@FreeBSD.org
.An Baptiste Daroussin Aq Mt bapt@FreeBSD.org
.Sh BUGS
Currently,
.Nm
considers only credentials transitions requested through the
.Xr setcred 2
system call.
This system call was in large part created so that
.Nm
can see whole credentials transitions to decide whether to authorize them, which
the traditional UNIX's piecewise approach of successively changing different
parts of them cannot allow.
.Pp
However, calls to traditional or standard credentials-changing functions can be
considered as full transitions on their own, however limited, and as such should
be equally monitored by
.Nm .
Future work will lift this restriction.
.Sh SECURITY CONSIDERATIONS
The threat model for
.Nm
is to consider userland programs as generally untrustable to decide upon which
credentials changes are acceptable.
It is in contrast with the traditional UNIX way to change credentials, in which
specialized programs are installed with the setuid bit, giving them full
administrator privileges so that they are effectively able to establish new
ones.
Vulnerabilities in such credentials-changing programs can have catastrophic
consequences on the integrity of the system.
.Pp
Consequently,
.Nm
does not rely on companion userland programs to decide whether some credentials
transition is acceptable.
Instead, it maintains its own configuration independently from the userland
password and group databases.
Establishing this configuration currently itself relies on userland programs
issuing calls to
.Xr sysctl 3
or
.Xr jail 2 .
It should thus be established near system boot or jail start, before any
possible attacks could happen on the system, and further measures should be
taken to ensure that potential corruptions does not affect the configuration in
subsequent restarts, such as re-establishing pristine state or ensuring that the
boot procedure up to the configuration of
.Nm
can be trusted.