freebsd-dev/sys/kern/capabilities.conf
Jonathan Anderson cfb5f76865 Add experimental support for process descriptors
A "process descriptor" file descriptor is used to manage processes
without using the PID namespace. This is required for Capsicum's
Capability Mode, where the PID namespace is unavailable.

New system calls pdfork(2) and pdkill(2) offer the functional equivalents
of fork(2) and kill(2). pdgetpid(2) allows querying the PID of the remote
process for debugging purposes. The currently-unimplemented pdwait(2) will,
in the future, allow querying rusage/exit status. In the interim, poll(2)
may be used to check (and wait for) process termination.

When a process is referenced by a process descriptor, it does not issue
SIGCHLD to the parent, making it suitable for use in libraries---a common
scenario when using library compartmentalisation from within large
applications (such as web browsers). Some observers may note a similarity
to Mach task ports; process descriptors provide a subset of this behaviour,
but in a UNIX style.

This feature is enabled by "options PROCDESC", but as with several other
Capsicum kernel features, is not enabled by default in GENERIC 9.0.

Reviewed by: jhb, kib
Approved by: re (kib), mentor (rwatson)
Sponsored by: Google Inc
2011-08-18 22:51:30 +00:00

746 lines
13 KiB
Plaintext

##
## Copyright (c) 2008-2010 Robert N. M. Watson
## All rights reserved.
##
## This software was developed at the University of Cambridge Computer
## Laboratory with support from a grant from Google, Inc.
##
## 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.
##
## List of system calls enabled in capability mode, one name per line.
##
## Notes:
## - sys_exit(2), abort2(2) and close(2) are very important.
## - Sorted alphabetically, please keep it that way.
##
## $FreeBSD$
##
##
## Allow ACL and MAC label operations by file descriptor, subject to
## capability rights. Allow MAC label operations on the current process but
## we will need to scope __mac_get_pid(2).
##
__acl_aclcheck_fd
__acl_delete_fd
__acl_get_fd
__acl_set_fd
__mac_get_fd
#__mac_get_pid
__mac_get_proc
__mac_set_fd
__mac_set_proc
##
## Allow sysctl(2) as we scope internal to the call; this is a global
## namespace, but there are several critical sysctls required for almost
## anything to run, such as hw.pagesize. For now that policy lives in the
## kernel for performance and simplicity, but perhaps it could move to a
## proxying daemon in userspace.
##
__sysctl
##
## Allow umtx operations as these are scoped by address space.
##
## XXRW: Need to check this very carefully.
##
_umtx_lock
_umtx_op
_umtx_unlock
##
## Allow process termination using abort2(2).
##
abort2
##
## Allow accept(2) since it doesn't manipulate namespaces directly, rather
## relies on existing bindings on a socket, subject to capability rights.
##
accept
##
## Allow AIO operations by file descriptor, subject to capability rights.
##
aio_cancel
aio_error
aio_fsync
aio_read
aio_return
aio_suspend
aio_waitcomplete
aio_write
##
## audit(2) is a global operation, submitting to the global trail, but it is
## controlled by privilege, and it might be useful to be able to submit
## records from sandboxes. For now, disallow, but we may want to think about
## providing some sort of proxy service for this.
##
#audit
##
## Disllow bind(2) for now, even though we support CAP_BIND.
##
## XXXRW: Revisit this.
##
#bind
##
## Allow capability mode and capability system calls.
##
cap_enter
cap_getmode
cap_getrights
cap_new
##
## Allow read-only clock operations.
##
clock_gettime
clock_getres
##
## Always allow file descriptor close(2).
##
close
closefrom
##
## Disallow connect(2) for now, despite CAP_CONNECT.
##
## XXXRW: Revisit this.
##
#connect
##
## cpuset(2) and related calls require scoping by process, but should
## eventually be allowed, at least in the current process case.
##
#cpuset
#cpuset_getaffinity
#cpuset_getid
#cpuset_setaffinity
#cpuset_setid
##
## Always allow dup(2) and dup2(2) manipulation of the file descriptor table.
##
dup
dup2
##
## Allow extended attribute operations by file descriptor, subject to
## capability rights.
##
extattr_delete_fd
extattr_get_fd
extattr_list_fd
extattr_set_fd
##
## Allow changing file flags, mode, and owner by file descriptor, subject to
## capability rights.
##
fchflags
fchmod
fchown
##
## For now, allow fcntl(2), subject to capability rights, but this probably
## needs additional scoping.
##
fcntl
##
## Allow fexecve(2), subject to capability rights. We perform some scoping,
## such as disallowing privilege escalation.
##
fexecve
##
## Allow flock(2), subject to capability rights.
##
flock
##
## Allow fork(2), even though it returns pids -- some applications seem to
## prefer this interface.
##
fork
##
## Allow fpathconf(2), subject to capability rights.
##
fpathconf
##
## Allow various file descriptor-based I/O operations, subject to capability
## rights.
##
freebsd6_ftruncate
freebsd6_lseek
freebsd6_mmap
freebsd6_pread
freebsd6_pwrite
##
## Allow querying file and file system state with fstat(2) and fstatfs(2),
## subject to capability rights.
##
fstat
fstatfs
##
## Allow further file descriptor-based I/O operations, subject to capability
## rights.
##
fsync
ftruncate
##
## Allow futimes(2), subject to capability rights.
##
futimes
##
## Allow querying process audit state, subject to normal access control.
##
getaudit
getaudit_addr
getauid
##
## Allow thread context management with getcontext(2).
##
getcontext
##
## Allow directory I/O on a file descriptor, subject to capability rights.
## Originally we had separate capabilities for directory-specific read
## operations, but on BSD we allow reading the raw directory data, so we just
## rely on CAP_READ and CAP_SEEK now.
##
getdents
getdirentries
##
## Allow querying certain trivial global state.
##
getdomainname
##
## Allow querying current process credential state.
##
getegid
geteuid
##
## Allow querying certain trivial global state.
##
gethostid
gethostname
##
## Allow querying per-process timer.
##
getitimer
##
## Allow querying current process credential state.
##
getgid
getgroups
getlogin
##
## Allow querying certain trivial global state.
##
getpagesize
getpeername
##
## Allow querying certain per-process scheduling, resource limit, and
## credential state.
##
## XXXRW: getpgid(2) needs scoping. It's not clear if it's worth scoping
## getppid(2). getpriority(2) needs scoping. getrusage(2) needs scoping.
## getsid(2) needs scoping.
##
getpgid
getpgrp
getpid
getppid
getpriority
getresgid
getresuid
getrlimit
getrusage
getsid
##
## Allow querying socket state, subject to capability rights.
##
## XXXRW: getsockopt(2) may need more attention.
##
getsockname
getsockopt
##
## Allow querying the global clock.
##
gettimeofday
##
## Allow querying current process credential state.
##
getuid
##
## Disallow ioctl(2) for now, as frequently ioctl(2) operations have global
## scope, but this is a tricky one as it is also required for tty control.
## We do have a capability right for this operation.
##
## XXXRW: This needs to be revisited.
##
#ioctl
##
## Allow querying current process credential state.
##
issetugid
##
## Allow kevent(2), as we will authorize based on capability rights on the
## target descriptor.
##
kevent
##
## Allow message queue operations on file descriptors, subject to capability
## rights.
##
kmq_notify
kmq_setattr
kmq_timedreceive
kmq_timedsend
##
## Allow kqueue(2), we will control use.
##
kqueue
##
## Allow managing per-process timers.
##
ktimer_create
ktimer_delete
ktimer_getoverrun
ktimer_gettime
ktimer_settime
##
## We can't allow ktrace(2) because it relies on a global namespace, but we
## might want to introduce an fktrace(2) of some sort.
##
#ktrace
##
## Allow AIO operations by file descriptor, subject to capability rights.
##
lio_listio
##
## Allow listen(2), subject to capability rights.
##
## XXXRW: One might argue this manipulates a global namespace.
##
listen
##
## Allow I/O-related file descriptors, subject to capability rights.
##
lseek
##
## Allow MAC label operations by file descriptor, subject to capability
## rights.
##
mac_get_fd
mac_set_fd
##
## Allow simple VM operations on the current process.
##
madvise
mincore
minherit
mlock
mlockall
##
## Allow memory mapping a file descriptor, and updating protections, subject
## to capability rights.
##
mmap
mprotect
##
## Allow simple VM operations on the current process.
##
msync
munlock
munlockall
munmap
##
## Allow the current process to sleep.
##
nanosleep
##
## Allow querying the global clock.
##
ntp_gettime
##
## Allow AIO operations by file descriptor, subject to capability rights.
##
oaio_read
oaio_write
##
## Allow simple VM operations on the current process.
##
obreak
##
## Allow AIO operations by file descriptor, subject to capability rights.
##
olio_listio
##
## Operations relative to directory capabilities.
##
faccessat
fstatat
fchmodat
futimesat
mkdirat
rmdirat
mkfifoat
mknodat
openat
renameat
##
## Allow entry into open(2). This system call will fail, since access to the
## global file namespace has been disallowed, but allowing entry into the
## syscall means that an audit trail will be generated (which is also very
## useful for debugging).
##
open
##
## Allow poll(2), which will be scoped by capability rights.
##
## XXXRW: Perhaps we don't need the OpenBSD version?
## XXXRW: We don't yet do that scoping.
##
openbsd_poll
##
## Process descriptor-related system calls are allowed.
##
pdfork
pdgetpid
pdkill
#pdwait4 # not yet implemented
##
## Allow pipe(2).
##
pipe
##
## Allow poll(2), which will be scoped by capability rights.
## XXXRW: We don't yet do that scoping.
##
poll
##
## Allow I/O-related file descriptors, subject to capability rights.
##
pread
preadv
##
## Allow access to profiling state on the current process.
##
profil
##
## Disallow ptrace(2) for now, but we do need debugging facilities in
## capability mode, so we will want to revisit this, possibly by scoping its
## operation.
##
#ptrace
##
## Allow I/O-related file descriptors, subject to capability rights.
##
pwrite
pwritev
read
readv
recv
recvfrom
recvmsg
##
## Allow real-time scheduling primitives to be used.
##
## XXXRW: These require scoping.
##
rtprio
rtprio_thread
##
## Allow simple VM operations on the current process.
##
sbrk
##
## Allow querying trivial global scheduler state.
##
sched_get_priority_max
sched_get_priority_min
##
## Allow various thread/process scheduler operations.
##
## XXXRW: Some of these require further scoping.
##
sched_getparam
sched_getscheduler
sched_rr_getinterval
sched_setparam
sched_setscheduler
sched_yield
##
## Allow I/O-related file descriptors, subject to capability rights.
##
sctp_generic_recvmsg
sctp_generic_sendmsg
sctp_generic_sendmsg_iov
sctp_peeloff
##
## Allow select(2), which will be scoped by capability rights.
##
## XXXRW: But is it?
##
select
##
## Allow I/O-related file descriptors, subject to capability rights. Use of
## explicit addresses here is restricted by the system calls themselves.
##
send
sendfile
sendmsg
sendto
##
## Allow setting per-process audit state, which is controlled separately by
## privileges.
##
setaudit
setaudit_addr
setauid
##
## Allow setting thread context.
##
setcontext
##
## Allow setting current process credential state, which is controlled
## separately by privilege.
##
setegid
seteuid
setgid
##
## Allow use of the process interval timer.
##
setitimer
##
## Allow setpriority(2).
##
## XXXRW: Requires scoping.
##
setpriority
##
## Allow setting current process credential state, which is controlled
## separately by privilege.
##
setregid
setresgid
setresuid
setreuid
##
## Allow setting process resource limits with setrlimit(2).
##
setrlimit
##
## Allow creating a new session with setsid(2).
##
setsid
##
## Allow setting socket options with setsockopt(2), subject to capability
## rights.
##
## XXXRW: Might require scoping.
##
setsockopt
##
## Allow setting current process credential state, which is controlled
## separately by privilege.
##
setuid
##
## shm_open(2) is scoped so as to allow only access to new anonymous objects.
##
shm_open
##
## Allow I/O-related file descriptors, subject to capability rights.
##
shutdown
##
## Allow signal control on current process.
##
sigaction
sigaltstack
sigblock
sigpending
sigprocmask
sigqueue
sigreturn
sigsetmask
sigstack
sigsuspend
sigtimedwait
sigvec
sigwaitinfo
##
## Allow creating new socket pairs with socket(2) and socketpair(2).
##
socket
socketpair
##
## Allow simple VM operations on the current process.
##
## XXXRW: Kernel doesn't implement this, so drop?
##
sstk
##
## Do allow sync(2) for now, but possibly shouldn't.
##
sync
##
## Always allow process termination with sys_exit(2).
##
sys_exit
##
## sysarch(2) does rather diverse things, but is required on at least i386
## in order to configure per-thread data. As such, it's scoped on each
## architecture.
##
sysarch
##
## Allow thread operations operating only on current process.
##
thr_create
thr_exit
thr_kill
##
## Disallow thr_kill2(2), as it may operate beyond the current process.
##
## XXXRW: Requires scoping.
##
#thr_kill2
##
## Allow thread operations operating only on current process.
##
thr_new
thr_self
thr_set_name
thr_suspend
thr_wake
##
## Allow manipulation of the current process umask with umask(2).
##
umask
##
## Allow submitting of process trace entries with utrace(2).
##
utrace
##
## Allow generating UUIDs with uuidgen(2).
##
uuidgen
##
## Allow I/O-related file descriptors, subject to capability rights.
##
write
writev
##
## Allow processes to yield(2).
##
yield