summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorBjoern A. Zeeb <bz@FreeBSD.org>2009-02-07 13:19:08 +0000
committerBjoern A. Zeeb <bz@FreeBSD.org>2009-02-07 13:19:08 +0000
commitda0bebf915b0494f2279c47f43f36c94f7d2fff5 (patch)
tree13c112004187522aaadada0f39d777757c44b187
parent75b76418f23497322f5811ce92d1ca971abb80f9 (diff)
Notes
-rw-r--r--UPDATING6
-rw-r--r--etc/defaults/rc.conf5
-rw-r--r--lib/libc/sys/cpuset_getaffinity.24
-rw-r--r--lib/libc/sys/jail.241
-rw-r--r--lib/libkvm/kvm_proc.c5
-rw-r--r--share/man/man4/ddb.412
-rw-r--r--sys/compat/freebsd32/freebsd32.h18
-rw-r--r--sys/compat/freebsd32/freebsd32_misc.c61
-rw-r--r--sys/compat/freebsd32/syscalls.master2
-rw-r--r--sys/kern/kern_cpuset.c115
-rw-r--r--sys/kern/kern_exit.c5
-rw-r--r--sys/kern/kern_fork.c6
-rw-r--r--sys/kern/kern_jail.c947
-rw-r--r--sys/kern/uipc_socket.c4
-rw-r--r--sys/net/if.c2
-rw-r--r--sys/net/rtsock.c161
-rw-r--r--sys/netinet/in_pcb.c163
-rw-r--r--sys/netinet/raw_ip.c46
-rw-r--r--sys/netinet/sctp_pcb.c35
-rw-r--r--sys/netinet/sctp_usrreq.c77
-rw-r--r--sys/netinet/tcp_usrreq.c12
-rw-r--r--sys/netinet/udp_usrreq.c14
-rw-r--r--sys/netinet6/in6_pcb.c111
-rw-r--r--sys/netinet6/in6_src.c21
-rw-r--r--sys/netinet6/raw_ip6.c13
-rw-r--r--sys/netinet6/udp6_usrreq.c14
-rw-r--r--sys/security/mac_bsdextended/mac_bsdextended.c1
-rw-r--r--sys/sys/cpuset.h6
-rw-r--r--sys/sys/jail.h89
-rw-r--r--sys/sys/param.h2
-rw-r--r--usr.bin/cpuset/cpuset.114
-rw-r--r--usr.bin/cpuset/cpuset.c20
-rw-r--r--usr.sbin/jail/Makefile6
-rw-r--r--usr.sbin/jail/jail.875
-rw-r--r--usr.sbin/jail/jail.c252
-rw-r--r--usr.sbin/jexec/Makefile2
-rw-r--r--usr.sbin/jexec/jexec.835
-rw-r--r--usr.sbin/jexec/jexec.c182
-rw-r--r--usr.sbin/jls/Makefile2
-rw-r--r--usr.sbin/jls/jls.834
-rw-r--r--usr.sbin/jls/jls.c214
41 files changed, 2509 insertions, 325 deletions
diff --git a/UPDATING b/UPDATING
index dc818e9778ce..091e550fcdf8 100644
--- a/UPDATING
+++ b/UPDATING
@@ -8,6 +8,12 @@ Items affecting the ports and packages system can be found in
/usr/ports/UPDATING. Please read that file before running
portupgrade.
+20090207:
+ Multi-IPv4/v6/no-IP jail support was merged to STABLE.
+ You need to rebuild jls(8) and to use the new features
+ jail(8), jexec(8) and cpuset(1) with a new kernel.
+ __FreeBSD_version was bumped to 701103.
+
20090119:
NTFS has been removed from GENERIC kernel on amd64 to match
GENERIC on i386. Should not cause any issues since mount_ntfs(8)
diff --git a/etc/defaults/rc.conf b/etc/defaults/rc.conf
index c44c1685e4a3..4da177a93448 100644
--- a/etc/defaults/rc.conf
+++ b/etc/defaults/rc.conf
@@ -620,7 +620,7 @@ jail_sysvipc_allow="NO" # Allow SystemV IPC use from within a jail
#
#jail_example_rootdir="/usr/jail/default" # Jail's root directory
#jail_example_hostname="default.domain.com" # Jail's hostname
-#jail_example_ip="192.168.0.10" # Jail's IP number
+#jail_example_ip="192.0.2.10" # Jail's IP number
#jail_example_interface="" # Interface to create the IP alias on
#jail_example_fib="0" # routing table for setfib(1)
#jail_example_exec_start="/bin/sh /etc/rc" # command to execute in jail for starting
@@ -629,10 +629,11 @@ jail_sysvipc_allow="NO" # Allow SystemV IPC use from within a jail
# specified using a trailing number
#jail_example_exec_stop="/bin/sh /etc/rc.shutdown" # command to execute in jail for stopping
#jail_example_devfs_enable="NO" # mount devfs in the jail
+#jail_example_devfs_ruleset="ruleset_name" # devfs ruleset to apply to jail -
+ # usually you want "devfsrules_jail".
#jail_example_fdescfs_enable="NO" # mount fdescfs in the jail
#jail_example_procfs_enable="NO" # mount procfs in jail
#jail_example_mount_enable="NO" # mount/umount jail's fs
-#jail_example_devfs_ruleset="ruleset_name" # devfs ruleset to apply to jail
#jail_example_fstab="" # fstab(5) for mount/umount
#jail_example_flags="-l -U root" # flags for jail(8)
diff --git a/lib/libc/sys/cpuset_getaffinity.2 b/lib/libc/sys/cpuset_getaffinity.2
index 87250ec8d662..b065ac362da0 100644
--- a/lib/libc/sys/cpuset_getaffinity.2
+++ b/lib/libc/sys/cpuset_getaffinity.2
@@ -25,7 +25,7 @@
.\"
.\" $FreeBSD$
.\"
-.Dd March 29, 2008
+.Dd November 29, 2008
.Dt CPUSET 2
.Os
.Sh NAME
@@ -46,7 +46,7 @@
and
.Fn cpuset_setaffinity
allow the manipulation of sets of CPUs available to processes, threads,
-interrupts and other resources.
+interrupts, jails and other resources.
These functions may manipulate sets of CPUs that contain many processes
or per-object anonymous masks that effect only a single object.
.Pp
diff --git a/lib/libc/sys/jail.2 b/lib/libc/sys/jail.2
index 5cba4bf93d4d..d06f946aa150 100644
--- a/lib/libc/sys/jail.2
+++ b/lib/libc/sys/jail.2
@@ -8,7 +8,7 @@
.\"
.\" $FreeBSD$
.\"
-.Dd April 8, 2003
+.Dd January 6, 2009
.Dt JAIL 2
.Os
.Sh NAME
@@ -32,15 +32,20 @@ The argument is a pointer to a structure describing the prison:
.Bd -literal -offset indent
struct jail {
u_int32_t version;
- char *path;
- char *hostname;
- u_int32_t ip_number;
+ char *path;
+ char *hostname;
+ char *jailname;
+ unsigned int ip4s;
+ unsigned int ip6s;
+ struct in_addr *ip4;
+ struct in6_addr *ip6;
};
.Ed
.Pp
.Dq Li version
defines the version of the API in use.
-It should be set to zero at this time.
+.Dv JAIL_API_VERSION
+is defined for the current version.
.Pp
The
.Dq Li path
@@ -54,8 +59,24 @@ This can be changed
from the inside of the prison.
.Pp
The
-.Dq Li ip_number
-can be set to the IP number assigned to the prison.
+.Dq Li jailname
+pointer is an optional name that can be assigned to the jail
+for example for managment purposes.
+.Pp
+The
+.Dq Li ip4s
+and
+.Dq Li ip6s
+give the numbers of IPv4 and IPv6 addresses that will be passed
+via their respective pointers.
+.Pp
+The
+.Dq Li ip4
+and
+.Dq Li ip6
+pointers can be set to an arrays of IPv4 and IPv6 addresses to be assigned to
+the prison, or NULL if none.
+IPv4 addresses must be in network byte order.
.Pp
The
.Fn jail_attach
@@ -97,6 +118,12 @@ or, if present, the per-jail
.Pp
All IP activity will be forced to happen to/from the IP number specified,
which should be an alias on one of the network interfaces.
+All connections to/from the loopback address
+.Pf ( Li 127.0.0.1
+for IPv4,
+.Li ::1
+for IPv6) will be changed to be to/from the primary address
+of the jail for the given address family.
.Pp
It is possible to identify a process as jailed by examining
.Dq Li /proc/<pid>/status :
diff --git a/lib/libkvm/kvm_proc.c b/lib/libkvm/kvm_proc.c
index f945bb3f0588..7434b20ed6f7 100644
--- a/lib/libkvm/kvm_proc.c
+++ b/lib/libkvm/kvm_proc.c
@@ -54,10 +54,11 @@ __FBSDID("$FreeBSD$");
#include <sys/_lock.h>
#include <sys/_mutex.h>
#include <sys/_task.h>
-#define _WANT_PRISON /* make jail.h give us 'struct prison' */
-#include <sys/jail.h>
+#include <sys/cpuset.h>
#include <sys/user.h>
#include <sys/proc.h>
+#define _WANT_PRISON /* make jail.h give us 'struct prison' */
+#include <sys/jail.h>
#include <sys/exec.h>
#include <sys/stat.h>
#include <sys/sysent.h>
diff --git a/share/man/man4/ddb.4 b/share/man/man4/ddb.4
index 7e923f59b2e0..dd278d22babb 100644
--- a/share/man/man4/ddb.4
+++ b/share/man/man4/ddb.4
@@ -60,7 +60,7 @@
.\"
.\" $FreeBSD$
.\"
-.Dd December 26, 2007
+.Dd November 29, 2008
.Dt DDB 4
.Os
.Sh NAME
@@ -539,6 +539,16 @@ If the
.Ar addr
is given, displays details about the given GEOM object (class, geom, provider
or consumer).
+.\"
+.Pp
+.It Ic show Cm jails
+Show the list of
+.Xr jail 8
+instances.
+In addition to what
+.Xr jls 8
+shows, also list kernel internal details.
+.\"
.Pp
.It Ic show Cm map Ns Oo Li / Ns Cm f Oc Ar addr
Prints the VM map at
diff --git a/sys/compat/freebsd32/freebsd32.h b/sys/compat/freebsd32/freebsd32.h
index 0c77bab0a6cb..08d6510cf9f5 100644
--- a/sys/compat/freebsd32/freebsd32.h
+++ b/sys/compat/freebsd32/freebsd32.h
@@ -153,6 +153,24 @@ struct stat32 {
unsigned int :(8 / 2) * (16 - (int)sizeof(struct timespec32));
};
+struct jail32_v0 {
+ u_int32_t version;
+ uint32_t path;
+ uint32_t hostname;
+ u_int32_t ip_number;
+};
+
+struct jail32 {
+ uint32_t version;
+ uint32_t path;
+ uint32_t hostname;
+ uint32_t jailname;
+ uint32_t ip4s;
+ uint32_t ip6s;
+ uint32_t ip4;
+ uint32_t ip6;
+};
+
struct sigaction32 {
u_int32_t sa_u;
int sa_flags;
diff --git a/sys/compat/freebsd32/freebsd32_misc.c b/sys/compat/freebsd32/freebsd32_misc.c
index 79e28bbc5e23..806d79fb9ad7 100644
--- a/sys/compat/freebsd32/freebsd32_misc.c
+++ b/sys/compat/freebsd32/freebsd32_misc.c
@@ -36,6 +36,7 @@ __FBSDID("$FreeBSD$");
#include <sys/fcntl.h>
#include <sys/filedesc.h>
#include <sys/imgact.h>
+#include <sys/jail.h>
#include <sys/kernel.h>
#include <sys/limits.h>
#include <sys/lock.h>
@@ -1983,6 +1984,66 @@ done2:
}
int
+freebsd32_jail(struct thread *td, struct freebsd32_jail_args *uap)
+{
+ uint32_t version;
+ int error;
+ struct jail j;
+
+ error = copyin(uap->jail, &version, sizeof(uint32_t));
+ if (error)
+ return (error);
+ switch (version) {
+ case 0:
+ {
+ /* FreeBSD single IPv4 jails. */
+ struct jail32_v0 j32_v0;
+
+ bzero(&j, sizeof(struct jail));
+ error = copyin(uap->jail, &j32_v0, sizeof(struct jail32_v0));
+ if (error)
+ return (error);
+ CP(j32_v0, j, version);
+ PTRIN_CP(j32_v0, j, path);
+ PTRIN_CP(j32_v0, j, hostname);
+ j.ip4s = j32_v0.ip_number;
+ break;
+ }
+
+ case 1:
+ /*
+ * Version 1 was used by multi-IPv4 jail implementations
+ * that never made it into the official kernel.
+ */
+ return (EINVAL);
+
+ case 2: /* JAIL_API_VERSION */
+ {
+ /* FreeBSD multi-IPv4/IPv6,noIP jails. */
+ struct jail32 j32;
+
+ error = copyin(uap->jail, &j32, sizeof(struct jail32));
+ if (error)
+ return (error);
+ CP(j32, j, version);
+ PTRIN_CP(j32, j, path);
+ PTRIN_CP(j32, j, hostname);
+ PTRIN_CP(j32, j, jailname);
+ CP(j32, j, ip4s);
+ CP(j32, j, ip6s);
+ PTRIN_CP(j32, j, ip4);
+ PTRIN_CP(j32, j, ip6);
+ break;
+ }
+
+ default:
+ /* Sci-Fi jails are not supported, sorry. */
+ return (EINVAL);
+ }
+ return (kern_jail(td, &j));
+}
+
+int
freebsd32_sigaction(struct thread *td, struct freebsd32_sigaction_args *uap)
{
struct sigaction32 s32;
diff --git a/sys/compat/freebsd32/syscalls.master b/sys/compat/freebsd32/syscalls.master
index 35a3414e5caa..313d2f670110 100644
--- a/sys/compat/freebsd32/syscalls.master
+++ b/sys/compat/freebsd32/syscalls.master
@@ -587,7 +587,7 @@
off_t *sbytes, int flags); }
337 AUE_NULL NOPROTO { int kldsym(int fileid, int cmd, \
void *data); }
-338 AUE_JAIL NOPROTO { int jail(struct jail *jail); }
+338 AUE_JAIL STD { int freebsd32_jail(struct jail32 *jail); }
339 AUE_NULL UNIMPL pioctl
340 AUE_SIGPROCMASK NOPROTO { int sigprocmask(int how, \
const sigset_t *set, sigset_t *oset); }
diff --git a/sys/kern/kern_cpuset.c b/sys/kern/kern_cpuset.c
index a7b2cdbab9d9..3a0134cb99db 100644
--- a/sys/kern/kern_cpuset.c
+++ b/sys/kern/kern_cpuset.c
@@ -53,6 +53,7 @@ __FBSDID("$FreeBSD$");
#include <sys/limits.h>
#include <sys/bus.h>
#include <sys/interrupt.h>
+#include <sys/jail.h> /* Must come after sys/proc.h */
#include <vm/uma.h>
@@ -208,7 +209,7 @@ cpuset_rel_complete(struct cpuset *set)
* Find a set based on an id. Returns it with a ref.
*/
static struct cpuset *
-cpuset_lookup(cpusetid_t setid)
+cpuset_lookup(cpusetid_t setid, struct thread *td)
{
struct cpuset *set;
@@ -221,6 +222,28 @@ cpuset_lookup(cpusetid_t setid)
if (set)
cpuset_ref(set);
mtx_unlock_spin(&cpuset_lock);
+
+ KASSERT(td != NULL, ("[%s:%d] td is NULL", __func__, __LINE__));
+ if (set != NULL && jailed(td->td_ucred)) {
+ struct cpuset *rset, *jset;
+ struct prison *pr;
+
+ rset = cpuset_refroot(set);
+
+ pr = td->td_ucred->cr_prison;
+ mtx_lock(&pr->pr_mtx);
+ cpuset_ref(pr->pr_cpuset);
+ jset = pr->pr_cpuset;
+ mtx_unlock(&pr->pr_mtx);
+
+ if (jset->cs_id != rset->cs_id) {
+ cpuset_rel(set);
+ set = NULL;
+ }
+ cpuset_rel(jset);
+ cpuset_rel(rset);
+ }
+
return (set);
}
@@ -414,12 +437,38 @@ cpuset_which(cpuwhich_t which, id_t id, struct proc **pp, struct thread **tdp,
set = cpuset_refbase(curthread->td_cpuset);
thread_unlock(curthread);
} else
- set = cpuset_lookup(id);
+ set = cpuset_lookup(id, curthread);
if (set) {
*setp = set;
return (0);
}
return (ESRCH);
+ case CPU_WHICH_JAIL:
+ {
+ /* Find `set' for prison with given id. */
+ struct prison *pr;
+
+ sx_slock(&allprison_lock);
+ pr = prison_find(id);
+ sx_sunlock(&allprison_lock);
+ if (pr == NULL)
+ return (ESRCH);
+ if (jailed(curthread->td_ucred)) {
+ if (curthread->td_ucred->cr_prison == pr) {
+ cpuset_ref(pr->pr_cpuset);
+ set = pr->pr_cpuset;
+ }
+ } else {
+ cpuset_ref(pr->pr_cpuset);
+ set = pr->pr_cpuset;
+ }
+ mtx_unlock(&pr->pr_mtx);
+ if (set) {
+ *setp = set;
+ return (0);
+ }
+ return (ESRCH);
+ }
default:
return (EINVAL);
}
@@ -668,6 +717,59 @@ cpuset_thread0(void)
}
/*
+ * Create a cpuset, which would be cpuset_create() but
+ * mark the new 'set' as root.
+ *
+ * We are not going to reparent the td to it. Use cpuset_reparentproc() for that.
+ *
+ * In case of no error, returns the set in *setp locked with a reference.
+ */
+int
+cpuset_create_root(struct thread *td, struct cpuset **setp)
+{
+ struct cpuset *root;
+ struct cpuset *set;
+ int error;
+
+ KASSERT(td != NULL, ("[%s:%d] invalid td", __func__, __LINE__));
+ KASSERT(setp != NULL, ("[%s:%d] invalid setp", __func__, __LINE__));
+
+ thread_lock(td);
+ root = cpuset_refroot(td->td_cpuset);
+ thread_unlock(td);
+
+ error = cpuset_create(setp, td->td_cpuset, &root->cs_mask);
+ cpuset_rel(root);
+ if (error)
+ return (error);
+
+ KASSERT(*setp != NULL, ("[%s:%d] cpuset_create returned invalid data",
+ __func__, __LINE__));
+
+ /* Mark the set as root. */
+ set = *setp;
+ set->cs_flags |= CPU_SET_ROOT;
+
+ return (0);
+}
+
+int
+cpuset_setproc_update_set(struct proc *p, struct cpuset *set)
+{
+ int error;
+
+ KASSERT(p != NULL, ("[%s:%d] invalid proc", __func__, __LINE__));
+ KASSERT(set != NULL, ("[%s:%d] invalid set", __func__, __LINE__));
+
+ cpuset_ref(set);
+ error = cpuset_setproc(p->p_pid, set, NULL);
+ if (error)
+ return (error);
+ cpuset_rel(set);
+ return (0);
+}
+
+/*
* This is called once the final set of system cpus is known. Modifies
* the root set and all children and mark the root readonly.
*/
@@ -732,7 +834,7 @@ cpuset_setid(struct thread *td, struct cpuset_setid_args *uap)
*/
if (uap->which != CPU_WHICH_PID)
return (EINVAL);
- set = cpuset_lookup(uap->setid);
+ set = cpuset_lookup(uap->setid, td);
if (set == NULL)
return (ESRCH);
error = cpuset_setproc(uap->id, set, NULL);
@@ -771,6 +873,7 @@ cpuset_getid(struct thread *td, struct cpuset_getid_args *uap)
PROC_UNLOCK(p);
break;
case CPU_WHICH_CPUSET:
+ case CPU_WHICH_JAIL:
break;
}
switch (uap->level) {
@@ -831,6 +934,7 @@ cpuset_getaffinity(struct thread *td, struct cpuset_getaffinity_args *uap)
thread_unlock(ttd);
break;
case CPU_WHICH_CPUSET:
+ case CPU_WHICH_JAIL:
break;
}
if (uap->level == CPU_LEVEL_ROOT)
@@ -857,6 +961,7 @@ cpuset_getaffinity(struct thread *td, struct cpuset_getaffinity_args *uap)
PROC_SUNLOCK(p);
break;
case CPU_WHICH_CPUSET:
+ case CPU_WHICH_JAIL:
CPU_COPY(&set->cs_mask, mask);
break;
}
@@ -934,6 +1039,7 @@ cpuset_setaffinity(struct thread *td, struct cpuset_setaffinity_args *uap)
PROC_UNLOCK(p);
break;
case CPU_WHICH_CPUSET:
+ case CPU_WHICH_JAIL:
break;
}
if (uap->level == CPU_LEVEL_ROOT)
@@ -953,7 +1059,8 @@ cpuset_setaffinity(struct thread *td, struct cpuset_setaffinity_args *uap)
error = cpuset_setproc(uap->id, NULL, mask);
break;
case CPU_WHICH_CPUSET:
- error = cpuset_which(CPU_WHICH_CPUSET, uap->id, &p,
+ case CPU_WHICH_JAIL:
+ error = cpuset_which(uap->which, uap->id, &p,
&ttd, &set);
if (error == 0) {
error = cpuset_modify(set, mask);
diff --git a/sys/kern/kern_exit.c b/sys/kern/kern_exit.c
index 11682c0a13be..37380d5b755b 100644
--- a/sys/kern/kern_exit.c
+++ b/sys/kern/kern_exit.c
@@ -52,6 +52,7 @@ __FBSDID("$FreeBSD$");
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/pioctl.h>
+#include <sys/jail.h>
#include <sys/tty.h>
#include <sys/wait.h>
#include <sys/vmmeter.h>
@@ -461,6 +462,10 @@ retry:
p->p_xstat = rv;
p->p_xthread = td;
+ /* In case we are jailed tell the prison that we are gone. */
+ if (jailed(p->p_ucred))
+ prison_proc_free(p->p_ucred->cr_prison);
+
#ifdef KDTRACE_HOOKS
/*
* Tell the DTrace fasttrap provider about the exit if it
diff --git a/sys/kern/kern_fork.c b/sys/kern/kern_fork.c
index 71fe803c4653..a9c7fcc64a58 100644
--- a/sys/kern/kern_fork.c
+++ b/sys/kern/kern_fork.c
@@ -54,6 +54,7 @@ __FBSDID("$FreeBSD$");
#include <sys/mutex.h>
#include <sys/priv.h>
#include <sys/proc.h>
+#include <sys/jail.h>
#include <sys/pioctl.h>
#include <sys/resourcevar.h>
#include <sys/sched.h>
@@ -442,6 +443,11 @@ again:
__rangeof(struct proc, p_startzero, p_endzero));
p2->p_ucred = crhold(td->td_ucred);
+
+ /* In case we are jailed tell the prison that we exist. */
+ if (jailed(p2->p_ucred))
+ prison_proc_hold(p2->p_ucred->cr_prison);
+
PROC_UNLOCK(p2);
/*
diff --git a/sys/kern/kern_jail.c b/sys/kern/kern_jail.c
index a3594a289e1c..79bb990a5b45 100644
--- a/sys/kern/kern_jail.c
+++ b/sys/kern/kern_jail.c
@@ -1,5 +1,7 @@
/*-
- * Copyright (c) 1999 Poul-Henning Kamp. All rights reserved.
+ * Copyright (c) 1999 Poul-Henning Kamp.
+ * Copyright (c) 2008 Bjoern A. Zeeb.
+ * All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
@@ -26,6 +28,9 @@
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
+#include "opt_ddb.h"
+#include "opt_inet.h"
+#include "opt_inet6.h"
#include "opt_mac.h"
#include <sys/param.h>
@@ -51,6 +56,12 @@ __FBSDID("$FreeBSD$");
#include <sys/vnode.h>
#include <net/if.h>
#include <netinet/in.h>
+#ifdef DDB
+#include <ddb/ddb.h>
+#ifdef INET6
+#include <netinet6/in6_var.h>
+#endif /* INET6 */
+#endif /* DDB */
#include <security/mac/mac_framework.h>
@@ -67,7 +78,7 @@ SYSCTL_INT(_security_jail, OID_AUTO, set_hostname_allowed, CTLFLAG_RW,
int jail_socket_unixiproute_only = 1;
SYSCTL_INT(_security_jail, OID_AUTO, socket_unixiproute_only, CTLFLAG_RW,
&jail_socket_unixiproute_only, 0,
- "Processes in jail are limited to creating UNIX/IPv4/route sockets only");
+ "Processes in jail are limited to creating UNIX/IP/route sockets only");
int jail_sysvipc_allowed = 0;
SYSCTL_INT(_security_jail, OID_AUTO, sysvipc_allowed, CTLFLAG_RW,
@@ -94,6 +105,11 @@ SYSCTL_INT(_security_jail, OID_AUTO, mount_allowed, CTLFLAG_RW,
&jail_mount_allowed, 0,
"Processes in jail can mount/unmount jail-friendly file systems");
+int jail_max_af_ips = 255;
+SYSCTL_INT(_security_jail, OID_AUTO, jail_max_af_ips, CTLFLAG_RW,
+ &jail_max_af_ips, 0,
+ "Number of IP addresses a jail may have at most per address family");
+
/* allprison, lastprid, and prisoncount are protected by allprison_lock. */
struct prisonlist allprison;
struct sx allprison_lock;
@@ -119,6 +135,12 @@ struct prison_service {
static void init_prison(void *);
static void prison_complete(void *context, int pending);
static int sysctl_jail_list(SYSCTL_HANDLER_ARGS);
+#ifdef INET
+static int _prison_check_ip4(struct prison *, struct in_addr *);
+#endif
+#ifdef INET6
+static int _prison_check_ip6(struct prison *, struct in6_addr *);
+#endif
static void
init_prison(void *data __unused)
@@ -130,6 +152,278 @@ init_prison(void *data __unused)
SYSINIT(prison, SI_SUB_INTRINSIC, SI_ORDER_ANY, init_prison, NULL);
+#ifdef INET
+static int
+qcmp_v4(const void *ip1, const void *ip2)
+{
+ in_addr_t iaa, iab;
+
+ /*
+ * We need to compare in HBO here to get the list sorted as expected
+ * by the result of the code. Sorting NBO addresses gives you
+ * interesting results. If you do not understand, do not try.
+ */
+ iaa = ntohl(((const struct in_addr *)ip1)->s_addr);
+ iab = ntohl(((const struct in_addr *)ip2)->s_addr);
+
+ /*
+ * Do not simply return the difference of the two numbers, the int is
+ * not wide enough.
+ */
+ if (iaa > iab)
+ return (1);
+ else if (iaa < iab)
+ return (-1);
+ else
+ return (0);
+}
+#endif
+
+#ifdef INET6
+static int
+qcmp_v6(const void *ip1, const void *ip2)
+{
+ const struct in6_addr *ia6a, *ia6b;
+ int i, rc;
+
+ ia6a = (const struct in6_addr *)ip1;
+ ia6b = (const struct in6_addr *)ip2;
+
+ rc = 0;
+ for (i=0; rc == 0 && i < sizeof(struct in6_addr); i++) {
+ if (ia6a->s6_addr[i] > ia6b->s6_addr[i])
+ rc = 1;
+ else if (ia6a->s6_addr[i] < ia6b->s6_addr[i])
+ rc = -1;
+ }
+ return (rc);
+}
+#endif
+
+#if defined(INET) || defined(INET6)
+static int
+prison_check_conflicting_ips(struct prison *p)
+{
+ struct prison *pr;
+ int i;
+
+ sx_assert(&allprison_lock, SX_LOCKED);
+
+ if (p->pr_ip4s == 0 && p->pr_ip6s == 0)
+ return (0);
+
+ LIST_FOREACH(pr, &allprison, pr_list) {
+ /*
+ * Skip 'dying' prisons to avoid problems when
+ * restarting multi-IP jails.
+ */
+ if (pr->pr_state == PRISON_STATE_DYING)
+ continue;
+
+ /*
+ * We permit conflicting IPs if there is no
+ * more than 1 IP on eeach jail.
+ * In case there is one duplicate on a jail with
+ * more than one IP stop checking and return error.
+ */
+#ifdef INET
+ if ((p->pr_ip4s >= 1 && pr->pr_ip4s > 1) ||
+ (p->pr_ip4s > 1 && pr->pr_ip4s >= 1)) {
+ for (i = 0; i < p->pr_ip4s; i++) {
+ if (_prison_check_ip4(pr, &p->pr_ip4[i]))
+ return (EINVAL);
+ }
+ }
+#endif
+#ifdef INET6
+ if ((p->pr_ip6s >= 1 && pr->pr_ip6s > 1) ||
+ (p->pr_ip6s > 1 && pr->pr_ip6s >= 1)) {
+ for (i = 0; i < p->pr_ip6s; i++) {
+ if (_prison_check_ip6(pr, &p->pr_ip6[i]))
+ return (EINVAL);
+ }
+ }
+#endif
+ }
+
+ return (0);
+}
+
+static int
+jail_copyin_ips(struct jail *j)
+{
+#ifdef INET
+ struct in_addr *ip4;
+#endif
+#ifdef INET6
+ struct in6_addr *ip6;
+#endif
+ int error, i;
+
+ /*
+ * Copy in addresses, check for duplicate addresses and do some
+ * simple 0 and broadcast checks. If users give other bogus addresses
+ * it is their problem.
+ *
+ * IP addresses are all sorted but ip[0] to preserve the primary IP
+ * address as given from userland. This special IP is used for
+ * unbound outgoing connections as well for "loopback" traffic.
+ */
+#ifdef INET
+ ip4 = NULL;
+#endif
+#ifdef INET6
+ ip6 = NULL;
+#endif
+#ifdef INET
+ if (j->ip4s > 0) {
+ ip4 = (struct in_addr *)malloc(j->ip4s * sizeof(struct in_addr),
+ M_PRISON, M_WAITOK | M_ZERO);
+ error = copyin(j->ip4, ip4, j->ip4s * sizeof(struct in_addr));
+ if (error)
+ goto e_free_ip;
+ /* Sort all but the first IPv4 address. */
+ if (j->ip4s > 1)
+ qsort((ip4 + 1), j->ip4s - 1,
+ sizeof(struct in_addr), qcmp_v4);
+
+ /*
+ * We do not have to care about byte order for these checks
+ * so we will do them in NBO.
+ */
+ for (i=0; i<j->ip4s; i++) {
+ if (ip4[i].s_addr == htonl(INADDR_ANY) ||
+ ip4[i].s_addr == htonl(INADDR_BROADCAST)) {
+ error = EINVAL;
+ goto e_free_ip;
+ }
+ if ((i+1) < j->ip4s &&
+ (ip4[0].s_addr == ip4[i+1].s_addr ||
+ ip4[i].s_addr == ip4[i+1].s_addr)) {
+ error = EINVAL;
+ goto e_free_ip;
+ }
+ }
+
+ j->ip4 = ip4;
+ } else
+ j->ip4 = NULL;
+#endif
+#ifdef INET6
+ if (j->ip6s > 0) {
+ ip6 = (struct in6_addr *)malloc(j->ip6s * sizeof(struct in6_addr),
+ M_PRISON, M_WAITOK | M_ZERO);
+ error = copyin(j->ip6, ip6, j->ip6s * sizeof(struct in6_addr));
+ if (error)
+ goto e_free_ip;
+ /* Sort all but the first IPv6 address. */
+ if (j->ip6s > 1)
+ qsort((ip6 + 1), j->ip6s - 1,
+ sizeof(struct in6_addr), qcmp_v6);
+ for (i=0; i<j->ip6s; i++) {
+ if (IN6_IS_ADDR_UNSPECIFIED(&ip6[i])) {
+ error = EINVAL;
+ goto e_free_ip;
+ }
+ if ((i+1) < j->ip6s &&
+ (IN6_ARE_ADDR_EQUAL(&ip6[0], &ip6[i+1]) ||
+ IN6_ARE_ADDR_EQUAL(&ip6[i], &ip6[i+1]))) {
+ error = EINVAL;
+ goto e_free_ip;
+ }
+ }
+
+ j->ip6 = ip6;
+ } else
+ j->ip6 = NULL;
+#endif
+ return (0);
+
+e_free_ip:
+#ifdef INET6
+ free(ip6, M_PRISON);
+#endif
+#ifdef INET
+ free(ip4, M_PRISON);
+#endif
+ return (error);
+}
+#endif /* INET || INET6 */
+
+static int
+jail_handle_ips(struct jail *j)
+{
+#if defined(INET) || defined(INET6)
+ int error;
+#endif
+
+ /*
+ * Finish conversion for older versions, copyin and setup IPs.
+ */
+ switch (j->version) {
+ case 0:
+ {
+#ifdef INET
+ /* FreeBSD single IPv4 jails. */
+ struct in_addr *ip4;
+
+ if (j->ip4s == INADDR_ANY || j->ip4s == INADDR_BROADCAST)
+ return (EINVAL);
+ ip4 = (struct in_addr *)malloc(sizeof(struct in_addr),
+ M_PRISON, M_WAITOK | M_ZERO);
+
+ /*
+ * Jail version 0 still used HBO for the IPv4 address.
+ */
+ ip4->s_addr = htonl(j->ip4s);
+ j->ip4s = 1;
+ j->ip4 = ip4;
+ break;
+#else
+ return (EINVAL);
+#endif
+ }
+
+ case 1:
+ /*
+ * Version 1 was used by multi-IPv4 jail implementations
+ * that never made it into the official kernel.
+ * We should never hit this here; jail() should catch it.
+ */
+ return (EINVAL);
+
+ case 2: /* JAIL_API_VERSION */
+ /* FreeBSD multi-IPv4/IPv6,noIP jails. */
+#if defined(INET) || defined(INET6)
+#ifdef INET
+ if (j->ip4s > jail_max_af_ips)
+ return (EINVAL);
+#else
+ if (j->ip4s != 0)
+ return (EINVAL);
+#endif
+#ifdef INET6
+ if (j->ip6s > jail_max_af_ips)
+ return (EINVAL);
+#else
+ if (j->ip6s != 0)
+ return (EINVAL);
+#endif
+ error = jail_copyin_ips(j);
+ if (error)
+ return (error);
+#endif
+ break;
+
+ default:
+ /* Sci-Fi jails are not supported, sorry. */
+ return (EINVAL);
+ }
+
+ return (0);
+}
+
+
/*
* struct jail_args {
* struct jail *jail;
@@ -138,23 +432,73 @@ SYSINIT(prison, SI_SUB_INTRINSIC, SI_ORDER_ANY, init_prison, NULL);
int
jail(struct thread *td, struct jail_args *uap)
{
+ uint32_t version;
+ int error;
+ struct jail j;
+
+ error = copyin(uap->jail, &version, sizeof(uint32_t));
+ if (error)
+ return (error);
+
+ switch (version) {
+ case 0:
+ /* FreeBSD single IPv4 jails. */
+ {
+ struct jail_v0 j0;
+
+ bzero(&j, sizeof(struct jail));
+ error = copyin(uap->jail, &j0, sizeof(struct jail_v0));
+ if (error)
+ return (error);
+ j.version = j0.version;
+ j.path = j0.path;
+ j.hostname = j0.hostname;
+ j.ip4s = j0.ip_number;
+ break;
+ }
+
+ case 1:
+ /*
+ * Version 1 was used by multi-IPv4 jail implementations
+ * that never made it into the official kernel.
+ */
+ return (EINVAL);
+
+ case 2: /* JAIL_API_VERSION */
+ /* FreeBSD multi-IPv4/IPv6,noIP jails. */
+ error = copyin(uap->jail, &j, sizeof(struct jail));
+ if (error)
+ return (error);
+ break;
+
+ default:
+ /* Sci-Fi jails are not supported, sorry. */
+ return (EINVAL);
+ }
+ return (kern_jail(td, &j));
+}
+
+int
+kern_jail(struct thread *td, struct jail *j)
+{
struct nameidata nd;
struct prison *pr, *tpr;
struct prison_service *psrv;
- struct jail j;
struct jail_attach_args jaa;
int vfslocked, error, tryprid;
- error = copyin(uap->jail, &j, sizeof(j));
+ KASSERT(j != NULL, ("%s: j is NULL", __func__));
+
+ /* Handle addresses - convert old structs, copyin, check IPs. */
+ error = jail_handle_ips(j);
if (error)
return (error);
- if (j.version != 0)
- return (EINVAL);
- MALLOC(pr, struct prison *, sizeof(*pr), M_PRISON, M_WAITOK | M_ZERO);
+ /* Allocate struct prison and fill it with life. */
+ pr = malloc(sizeof(*pr), M_PRISON, M_WAITOK | M_ZERO);
mtx_init(&pr->pr_mtx, "jail mutex", NULL, MTX_DEF);
pr->pr_ref = 1;
- error = copyinstr(j.path, &pr->pr_path, sizeof(pr->pr_path), 0);
+ error = copyinstr(j->path, &pr->pr_path, sizeof(pr->pr_path), NULL);
if (error)
goto e_killmtx;
NDINIT(&nd, LOOKUP, MPSAFE | FOLLOW | LOCKLEAF, UIO_SYSSPACE,
@@ -167,10 +511,25 @@ jail(struct thread *td, struct jail_args *uap)
VOP_UNLOCK(nd.ni_vp, 0, td);
NDFREE(&nd, NDF_ONLY_PNBUF);
VFS_UNLOCK_GIANT(vfslocked);
- error = copyinstr(j.hostname, &pr->pr_host, sizeof(pr->pr_host), 0);
+ error = copyinstr(j->hostname, &pr->pr_host, sizeof(pr->pr_host), NULL);
if (error)
goto e_dropvnref;
- pr->pr_ip = j.ip_number;
+ if (j->jailname != NULL) {
+ error = copyinstr(j->jailname, &pr->pr_name,
+ sizeof(pr->pr_name), NULL);
+ if (error)
+ goto e_dropvnref;
+ }
+ if (j->ip4s > 0) {
+ pr->pr_ip4 = j->ip4;
+ pr->pr_ip4s = j->ip4s;
+ }
+#ifdef INET6
+ if (j->ip6s > 0) {
+ pr->pr_ip6 = j->ip6;
+ pr->pr_ip6s = j->ip6s;
+ }
+#endif
pr->pr_linux = NULL;
pr->pr_securelevel = securelevel;
if (prison_service_slots == 0)
@@ -180,8 +539,29 @@ jail(struct thread *td, struct jail_args *uap)
M_PRISON, M_ZERO | M_WAITOK);
}
- /* Determine next pr_id and add prison to allprison list. */
+ /*
+ * Pre-set prison state to ALIVE upon cration. This is needed so we
+ * can later attach the process to it, etc (avoiding another extra
+ * state for ther process of creation, complicating things).
+ */
+ pr->pr_state = PRISON_STATE_ALIVE;
+
+ /* Allocate a dedicated cpuset for each jail. */
+ error = cpuset_create_root(td, &pr->pr_cpuset);
+ if (error)
+ goto e_dropvnref;
+
sx_xlock(&allprison_lock);
+ /* Make sure we cannot run into problems with ambiguous bind()ings. */
+#if defined(INET) || defined(INET6)
+ error = prison_check_conflicting_ips(pr);
+ if (error) {
+ sx_xunlock(&allprison_lock);
+ goto e_dropcpuset;
+ }
+#endif
+
+ /* Determine next pr_id and add prison to allprison list. */
tryprid = lastprid + 1;
if (tryprid == JAIL_MAX)
tryprid = 1;
@@ -192,7 +572,7 @@ next:
if (tryprid == JAIL_MAX) {
sx_xunlock(&allprison_lock);
error = EAGAIN;
- goto e_dropvnref;
+ goto e_dropcpuset;
}
goto next;
}
@@ -223,15 +603,23 @@ e_dropprref:
psrv->ps_destroy(psrv, pr);
}
sx_sunlock(&allprison_lock);
+e_dropcpuset:
+ cpuset_rel(pr->pr_cpuset);
e_dropvnref:
if (pr->pr_slots != NULL)
- FREE(pr->pr_slots, M_PRISON);
+ free(pr->pr_slots, M_PRISON);
vfslocked = VFS_LOCK_GIANT(pr->pr_root->v_mount);
vrele(pr->pr_root);
VFS_UNLOCK_GIANT(vfslocked);
e_killmtx:
mtx_destroy(&pr->pr_mtx);
- FREE(pr, M_PRISON);
+ free(pr, M_PRISON);
+#ifdef INET6
+ free(j->ip6, M_PRISON);
+#endif
+#ifdef INET
+ free(j->ip4, M_PRISON);
+#endif
return (error);
}
@@ -267,10 +655,27 @@ jail_attach(struct thread *td, struct jail_attach_args *uap)
sx_sunlock(&allprison_lock);
return (EINVAL);
}
+
+ /*
+ * Do not allow a process to attach to a prison that is not
+ * considered to be "ALIVE".
+ */
+ if (pr->pr_state != PRISON_STATE_ALIVE) {
+ mtx_unlock(&pr->pr_mtx);
+ sx_sunlock(&allprison_lock);
+ return (EINVAL);
+ }
pr->pr_ref++;
mtx_unlock(&pr->pr_mtx);
sx_sunlock(&allprison_lock);
+ /*
+ * Reparent the newly attached process to this jail.
+ */
+ error = cpuset_setproc_update_set(p, pr->pr_cpuset);
+ if (error)
+ goto e_unref;
+
vfslocked = VFS_LOCK_GIANT(pr->pr_root->v_mount);
vn_lock(pr->pr_root, LK_EXCLUSIVE | LK_RETRY, td);
if ((error = change_dir(pr->pr_root, td)) != 0)
@@ -290,12 +695,14 @@ jail_attach(struct thread *td, struct jail_attach_args *uap)
crcopy(newcred, oldcred);
newcred->cr_prison = pr;
p->p_ucred = newcred;
+ prison_proc_hold(pr);
PROC_UNLOCK(p);
crfree(oldcred);
return (0);
e_unlock:
VOP_UNLOCK(pr->pr_root, 0, td);
VFS_UNLOCK_GIANT(vfslocked);
+e_unref:
mtx_lock(&pr->pr_mtx);
pr->pr_ref--;
mtx_unlock(&pr->pr_mtx);
@@ -325,10 +732,10 @@ prison_find(int prid)
}
void
-prison_free(struct prison *pr)
+prison_free_locked(struct prison *pr)
{
- mtx_lock(&pr->pr_mtx);
+ mtx_assert(&pr->pr_mtx, MA_OWNED);
pr->pr_ref--;
if (pr->pr_ref == 0) {
mtx_unlock(&pr->pr_mtx);
@@ -339,6 +746,14 @@ prison_free(struct prison *pr)
mtx_unlock(&pr->pr_mtx);
}
+void
+prison_free(struct prison *pr)
+{
+
+ mtx_lock(&pr->pr_mtx);
+ prison_free_locked(pr);
+}
+
static void
prison_complete(void *context, int pending)
{
@@ -356,103 +771,404 @@ prison_complete(void *context, int pending)
psrv->ps_destroy(psrv, pr);
}
sx_sunlock(&allprison_lock);
+
+ cpuset_rel(pr->pr_cpuset);
+
if (pr->pr_slots != NULL)
- FREE(pr->pr_slots, M_PRISON);
+ free(pr->pr_slots, M_PRISON);
vfslocked = VFS_LOCK_GIANT(pr->pr_root->v_mount);
vrele(pr->pr_root);
VFS_UNLOCK_GIANT(vfslocked);
mtx_destroy(&pr->pr_mtx);
- if (pr->pr_linux != NULL)
- FREE(pr->pr_linux, M_PRISON);
- FREE(pr, M_PRISON);
+ free(pr->pr_linux, M_PRISON);
+#ifdef INET6
+ free(pr->pr_ip6, M_PRISON);
+#endif
+#ifdef INET
+ free(pr->pr_ip4, M_PRISON);
+#endif
+ free(pr, M_PRISON);
}
void
-prison_hold(struct prison *pr)
+prison_hold_locked(struct prison *pr)
{
- mtx_lock(&pr->pr_mtx);
+ mtx_assert(&pr->pr_mtx, MA_OWNED);
KASSERT(pr->pr_ref > 0,
("Trying to hold dead prison (id=%d).", pr->pr_id));
pr->pr_ref++;
+}
+
+void
+prison_hold(struct prison *pr)
+{
+
+ mtx_lock(&pr->pr_mtx);
+ prison_hold_locked(pr);
mtx_unlock(&pr->pr_mtx);
}
-u_int32_t
-prison_getip(struct ucred *cred)
+void
+prison_proc_hold(struct prison *pr)
+{
+
+ mtx_lock(&pr->pr_mtx);
+ KASSERT(pr->pr_state == PRISON_STATE_ALIVE,
+ ("Cannot add a process to a non-alive prison (id=%d).", pr->pr_id));
+ pr->pr_nprocs++;
+ mtx_unlock(&pr->pr_mtx);
+}
+
+void
+prison_proc_free(struct prison *pr)
{
- return (cred->cr_prison->pr_ip);
+ mtx_lock(&pr->pr_mtx);
+ KASSERT(pr->pr_state == PRISON_STATE_ALIVE && pr->pr_nprocs > 0,
+ ("Trying to kill a process in a dead prison (id=%d).", pr->pr_id));
+ pr->pr_nprocs--;
+ if (pr->pr_nprocs == 0)
+ pr->pr_state = PRISON_STATE_DYING;
+ mtx_unlock(&pr->pr_mtx);
}
+
+#ifdef INET
+/*
+ * Pass back primary IPv4 address of this jail.
+ *
+ * If not jailed return success but do not alter the address. Caller has to
+ * make sure to intialize it correctly (INADDR_ANY).
+ *
+ * Returns 0 on success, 1 on error. Address returned in NBO.
+ */
int
-prison_ip(struct ucred *cred, int flag, u_int32_t *ip)
+prison_getip4(struct ucred *cred, struct in_addr *ia)
{
- u_int32_t tmp;
+
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
if (!jailed(cred))
+ /* Do not change address passed in. */
return (0);
- if (flag)
- tmp = *ip;
- else
- tmp = ntohl(*ip);
- if (tmp == INADDR_ANY) {
- if (flag)
- *ip = cred->cr_prison->pr_ip;
- else
- *ip = htonl(cred->cr_prison->pr_ip);
+
+ if (cred->cr_prison->pr_ip4 == NULL)
+ return (1);
+
+ ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr;
+ return (0);
+}
+
+/*
+ * Make sure our (source) address is set to something meaningful to this
+ * jail.
+ *
+ * Returns 0 on success, 1 on error. Address passed in in NBO and returned
+ * in NBO.
+ */
+int
+prison_local_ip4(struct ucred *cred, struct in_addr *ia)
+{
+ struct in_addr ia0;
+
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
+
+ if (!jailed(cred))
+ return (0);
+ if (cred->cr_prison->pr_ip4 == NULL)
+ return (1);
+
+ ia0.s_addr = ntohl(ia->s_addr);
+ if (ia0.s_addr == INADDR_LOOPBACK) {
+ ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr;
return (0);
}
- if (tmp == INADDR_LOOPBACK) {
- if (flag)
- *ip = cred->cr_prison->pr_ip;
- else
- *ip = htonl(cred->cr_prison->pr_ip);
+
+ /*
+ * In case there is only 1 IPv4 address, bind directly.
+ */
+ if (ia0.s_addr == INADDR_ANY && cred->cr_prison->pr_ip4s == 1) {
+ ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr;
return (0);
}
- if (cred->cr_prison->pr_ip != tmp)
+
+ if (ia0.s_addr == INADDR_ANY || prison_check_ip4(cred, ia))
+ return (0);
+
+ return (1);
+}
+
+/*
+ * Rewrite destination address in case we will connect to loopback address.
+ *
+ * Returns 0 on success, 1 on error. Address passed in in NBO and returned
+ * in NBO.
+ */
+int
+prison_remote_ip4(struct ucred *cred, struct in_addr *ia)
+{
+
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
+
+ if (!jailed(cred))
+ return (0);
+ if (cred->cr_prison->pr_ip4 == NULL)
return (1);
+ if (ntohl(ia->s_addr) == INADDR_LOOPBACK) {
+ ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr;
+ return (0);
+ }
+
+ /*
+ * Return success because nothing had to be changed.
+ */
return (0);
}
-void
-prison_remote_ip(struct ucred *cred, int flag, u_int32_t *ip)
+/*
+ * Check if given address belongs to the jail referenced by cred.
+ *
+ * Returns 1 if address belongs to jail, 0 if not. Address passed in in NBO.
+ */
+static int
+_prison_check_ip4(struct prison *pr, struct in_addr *ia)
+{
+ int i, a, z, d;
+
+ if (pr->pr_ip4 == NULL)
+ return (0);
+
+ /*
+ * Check the primary IP.
+ */
+ if (pr->pr_ip4[0].s_addr == ia->s_addr)
+ return (1);
+
+ /*
+ * All the other IPs are sorted so we can do a binary search.
+ */
+ a = 0;
+ z = pr->pr_ip4s - 2;
+ while (a <= z) {
+ i = (a + z) / 2;
+ d = qcmp_v4(&pr->pr_ip4[i+1], ia);
+ if (d > 0)
+ z = i - 1;
+ else if (d < 0)
+ a = i + 1;
+ else
+ return (1);
+ }
+ return (0);
+}
+
+int
+prison_check_ip4(struct ucred *cred, struct in_addr *ia)
{
- u_int32_t tmp;
+
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
if (!jailed(cred))
- return;
- if (flag)
- tmp = *ip;
- else
- tmp = ntohl(*ip);
- if (tmp == INADDR_LOOPBACK) {
- if (flag)
- *ip = cred->cr_prison->pr_ip;
+ return (1);
+
+ return (_prison_check_ip4(cred->cr_prison, ia));
+}
+#endif
+
+#ifdef INET6
+/*
+ * Pass back primary IPv6 address for this jail.
+ *
+ * If not jailed return success but do not alter the address. Caller has to
+ * make sure to intialize it correctly (IN6ADDR_ANY_INIT).
+ *
+ * Returns 0 on success, 1 on error.
+ */
+int
+prison_getip6(struct ucred *cred, struct in6_addr *ia6)
+{
+
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(ia6 != NULL, ("%s: ia6 is NULL", __func__));
+
+ if (!jailed(cred))
+ return (0);
+ if (cred->cr_prison->pr_ip6 == NULL)
+ return (1);
+ bcopy(&cred->cr_prison->pr_ip6[0], ia6, sizeof(struct in6_addr));
+ return (0);
+}
+
+/*
+ * Make sure our (source) address is set to something meaningful to this jail.
+ *
+ * v6only should be set based on (inp->inp_flags & IN6P_IPV6_V6ONLY != 0)
+ * when needed while binding.
+ *
+ * Returns 0 on success, 1 on error.
+ */
+int
+prison_local_ip6(struct ucred *cred, struct in6_addr *ia6, int v6only)
+{
+
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(ia6 != NULL, ("%s: ia6 is NULL", __func__));
+
+ if (!jailed(cred))
+ return (0);
+ if (cred->cr_prison->pr_ip6 == NULL)
+ return (1);
+ if (IN6_IS_ADDR_LOOPBACK(ia6)) {
+ bcopy(&cred->cr_prison->pr_ip6[0], ia6,
+ sizeof(struct in6_addr));
+ return (0);
+ }
+
+ /*
+ * In case there is only 1 IPv6 address, and v6only is true, then
+ * bind directly.
+ */
+ if (v6only != 0 && IN6_IS_ADDR_UNSPECIFIED(ia6) &&
+ cred->cr_prison->pr_ip6s == 1) {
+ bcopy(&cred->cr_prison->pr_ip6[0], ia6,
+ sizeof(struct in6_addr));
+ return (0);
+ }
+ if (IN6_IS_ADDR_UNSPECIFIED(ia6) || prison_check_ip6(cred, ia6))
+ return (0);
+ return (1);
+}
+
+/*
+ * Rewrite destination address in case we will connect to loopback address.
+ *
+ * Returns 0 on success, 1 on error.
+ */
+int
+prison_remote_ip6(struct ucred *cred, struct in6_addr *ia6)
+{
+
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(ia6 != NULL, ("%s: ia6 is NULL", __func__));
+
+ if (!jailed(cred))
+ return (0);
+ if (cred->cr_prison->pr_ip6 == NULL)
+ return (1);
+ if (IN6_IS_ADDR_LOOPBACK(ia6)) {
+ bcopy(&cred->cr_prison->pr_ip6[0], ia6,
+ sizeof(struct in6_addr));
+ return (0);
+ }
+
+ /*
+ * Return success because nothing had to be changed.
+ */
+ return (0);
+}
+
+/*
+ * Check if given address belongs to the jail referenced by cred.
+ *
+ * Returns 1 if address belongs to jail, 0 if not.
+ */
+static int
+_prison_check_ip6(struct prison *pr, struct in6_addr *ia6)
+{
+ int i, a, z, d;
+
+ if (pr->pr_ip6 == NULL)
+ return (0);
+
+ /*
+ * Check the primary IP.
+ */
+ if (IN6_ARE_ADDR_EQUAL(&pr->pr_ip6[0], ia6))
+ return (1);
+
+ /*
+ * All the other IPs are sorted so we can do a binary search.
+ */
+ a = 0;
+ z = pr->pr_ip6s - 2;
+ while (a <= z) {
+ i = (a + z) / 2;
+ d = qcmp_v6(&pr->pr_ip6[i+1], ia6);
+ if (d > 0)
+ z = i - 1;
+ else if (d < 0)
+ a = i + 1;
else
- *ip = htonl(cred->cr_prison->pr_ip);
- return;
+ return (1);
}
- return;
+ return (0);
+}
+
+int
+prison_check_ip6(struct ucred *cred, struct in6_addr *ia6)
+{
+
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(ia6 != NULL, ("%s: ia6 is NULL", __func__));
+
+ if (!jailed(cred))
+ return (1);
+
+ return (_prison_check_ip6(cred->cr_prison, ia6));
}
+#endif
+/*
+ * Check if given address belongs to the jail referenced by cred (wrapper to
+ * prison_check_ip[46]).
+ *
+ * Returns 1 if address belongs to jail, 0 if not. IPv4 Address passed in in
+ * NBO.
+ */
int
prison_if(struct ucred *cred, struct sockaddr *sa)
{
+#ifdef INET
struct sockaddr_in *sai;
+#endif
+#ifdef INET6
+ struct sockaddr_in6 *sai6;
+#endif
int ok;
- sai = (struct sockaddr_in *)sa;
- if ((sai->sin_family != AF_INET) && jail_socket_unixiproute_only)
- ok = 1;
- else if (sai->sin_family != AF_INET)
- ok = 0;
- else if (cred->cr_prison->pr_ip != ntohl(sai->sin_addr.s_addr))
- ok = 1;
- else
- ok = 0;
+ KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
+ KASSERT(sa != NULL, ("%s: sa is NULL", __func__));
+
+ ok = 0;
+ switch(sa->sa_family)
+ {
+#ifdef INET
+ case AF_INET:
+ sai = (struct sockaddr_in *)sa;
+ if (prison_check_ip4(cred, &sai->sin_addr))
+ ok = 1;
+ break;
+
+#endif
+#ifdef INET6
+ case AF_INET6:
+ sai6 = (struct sockaddr_in6 *)sa;
+ if (prison_check_ip6(cred, (struct in6_addr *)&sai6->sin6_addr))
+ ok = 1;
+ break;
+
+#endif
+ default:
+ if (!jail_socket_unixiproute_only)
+ ok = 1;
+ }
return (ok);
}
@@ -668,6 +1384,7 @@ prison_priv_check(struct ucred *cred, int priv)
* processes in the same jail. Likewise for signalling.
*/
case PRIV_SCHED_DIFFCRED:
+ case PRIV_SCHED_CPUSET:
case PRIV_SIGNAL_DIFFCRED:
case PRIV_SIGNAL_SUGID:
@@ -963,6 +1680,8 @@ sysctl_jail_list(SYSCTL_HANDLER_ARGS)
{
struct xprison *xp, *sxp;
struct prison *pr;
+ char *p;
+ size_t len;
int count, error;
if (jailed(req->td->td_ucred))
@@ -974,21 +1693,54 @@ sysctl_jail_list(SYSCTL_HANDLER_ARGS)
return (0);
}
- sxp = xp = malloc(sizeof(*xp) * count, M_TEMP, M_WAITOK | M_ZERO);
+ len = sizeof(*xp) * count;
+ LIST_FOREACH(pr, &allprison, pr_list) {
+#ifdef INET
+ len += pr->pr_ip4s * sizeof(struct in_addr);
+#endif
+#ifdef INET6
+ len += pr->pr_ip6s * sizeof(struct in6_addr);
+#endif
+ }
+
+ sxp = xp = malloc(len, M_TEMP, M_WAITOK | M_ZERO);
LIST_FOREACH(pr, &allprison, pr_list) {
xp->pr_version = XPRISON_VERSION;
xp->pr_id = pr->pr_id;
- xp->pr_ip = pr->pr_ip;
+ xp->pr_state = pr->pr_state;
+ xp->pr_cpusetid = pr->pr_cpuset->cs_id;
strlcpy(xp->pr_path, pr->pr_path, sizeof(xp->pr_path));
mtx_lock(&pr->pr_mtx);
strlcpy(xp->pr_host, pr->pr_host, sizeof(xp->pr_host));
+ strlcpy(xp->pr_name, pr->pr_name, sizeof(xp->pr_name));
mtx_unlock(&pr->pr_mtx);
- xp++;
+#ifdef INET
+ xp->pr_ip4s = pr->pr_ip4s;
+#endif
+#ifdef INET6
+ xp->pr_ip6s = pr->pr_ip6s;
+#endif
+ p = (char *)(xp + 1);
+#ifdef INET
+ if (pr->pr_ip4s > 0) {
+ bcopy(pr->pr_ip4, (struct in_addr *)p,
+ pr->pr_ip4s * sizeof(struct in_addr));
+ p += (pr->pr_ip4s * sizeof(struct in_addr));
+ }
+#endif
+#ifdef INET6
+ if (pr->pr_ip6s > 0) {
+ bcopy(pr->pr_ip6, (struct in6_addr *)p,
+ pr->pr_ip6s * sizeof(struct in6_addr));
+ p += (pr->pr_ip6s * sizeof(struct in6_addr));
+ }
+#endif
+ xp = (struct xprison *)p;
}
sx_sunlock(&allprison_lock);
- error = SYSCTL_OUT(req, sxp, sizeof(*sxp) * count);
+ error = SYSCTL_OUT(req, sxp, len);
free(sxp, M_TEMP);
return (error);
}
@@ -1008,3 +1760,60 @@ sysctl_jail_jailed(SYSCTL_HANDLER_ARGS)
}
SYSCTL_PROC(_security_jail, OID_AUTO, jailed, CTLTYPE_INT | CTLFLAG_RD,
NULL, 0, sysctl_jail_jailed, "I", "Process in jail?");
+
+#ifdef DDB
+DB_SHOW_COMMAND(jails, db_show_jails)
+{
+ struct prison *pr;
+#ifdef INET
+ struct in_addr ia;
+#endif
+#ifdef INET6
+ char ip6buf[INET6_ADDRSTRLEN];
+#endif
+ const char *state;
+#if defined(INET) || defined(INET6)
+ int i;
+#endif
+
+ db_printf(
+ " JID pr_ref pr_nprocs pr_ip4s pr_ip6s\n");
+ db_printf(
+ " Hostname Path\n");
+ db_printf(
+ " Name State\n");
+ db_printf(
+ " Cpusetid\n");
+ db_printf(
+ " IP Address(es)\n");
+ LIST_FOREACH(pr, &allprison, pr_list) {
+ db_printf("%6d %6d %9d %7d %7d\n",
+ pr->pr_id, pr->pr_ref, pr->pr_nprocs,
+ pr->pr_ip4s, pr->pr_ip6s);
+ db_printf("%6s %-29.29s %.74s\n",
+ "", pr->pr_host, pr->pr_path);
+ if (pr->pr_state < 0 || pr->pr_state >= (int)((sizeof(
+ prison_states) / sizeof(struct prison_state))))
+ state = "(bogus)";
+ else
+ state = prison_states[pr->pr_state].state_name;
+ db_printf("%6s %-29.29s %.74s\n",
+ "", (pr->pr_name[0] != '\0') ? pr->pr_name : "", state);
+ db_printf("%6s %-6d\n",
+ "", pr->pr_cpuset->cs_id);
+#ifdef INET
+ for (i=0; i < pr->pr_ip4s; i++) {
+ ia.s_addr = pr->pr_ip4[i].s_addr;
+ db_printf("%6s %s\n", "", inet_ntoa(ia));
+ }
+#endif
+#ifdef INET6
+ for (i=0; i < pr->pr_ip6s; i++)
+ db_printf("%6s %s\n",
+ "", ip6_sprintf(ip6buf, &pr->pr_ip6[i]));
+#endif /* INET6 */
+ if (db_pager_quit)
+ break;
+ }
+}
+#endif /* DDB */
diff --git a/sys/kern/uipc_socket.c b/sys/kern/uipc_socket.c
index 75083beabb46..7eaef1f80f17 100644
--- a/sys/kern/uipc_socket.c
+++ b/sys/kern/uipc_socket.c
@@ -98,6 +98,7 @@
__FBSDID("$FreeBSD$");
#include "opt_inet.h"
+#include "opt_inet6.h"
#include "opt_mac.h"
#include "opt_zero.h"
#include "opt_compat.h"
@@ -347,6 +348,9 @@ socreate(int dom, struct socket **aso, int type, int proto,
if (jailed(cred) && jail_socket_unixiproute_only &&
prp->pr_domain->dom_family != PF_LOCAL &&
prp->pr_domain->dom_family != PF_INET &&
+#ifdef INET6
+ prp->pr_domain->dom_family != PF_INET6 &&
+#endif
prp->pr_domain->dom_family != PF_ROUTE) {
return (EPROTONOSUPPORT);
}
diff --git a/sys/net/if.c b/sys/net/if.c
index ec286651c1a4..faa5f4872dd6 100644
--- a/sys/net/if.c
+++ b/sys/net/if.c
@@ -2197,7 +2197,7 @@ again:
struct sockaddr *sa = ifa->ifa_addr;
if (jailed(curthread->td_ucred) &&
- prison_if(curthread->td_ucred, sa))
+ !prison_if(curthread->td_ucred, sa))
continue;
addrs++;
#ifdef COMPAT_43
diff --git a/sys/net/rtsock.c b/sys/net/rtsock.c
index c0c5dc4f2f35..3e70d1cbc231 100644
--- a/sys/net/rtsock.c
+++ b/sys/net/rtsock.c
@@ -30,10 +30,13 @@
* $FreeBSD$
*/
#include "opt_sctp.h"
+#include "opt_inet.h"
+#include "opt_inet6.h"
+
#include <sys/param.h>
#include <sys/domain.h>
-#include <sys/kernel.h>
#include <sys/jail.h>
+#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/priv.h>
@@ -51,6 +54,9 @@
#include <net/route.h>
#include <netinet/in.h>
+#ifdef INET6
+#include <netinet6/scope6_var.h>
+#endif
#ifdef SCTP
extern void sctp_addr_change(struct ifaddr *ifa, int cmd);
@@ -306,6 +312,136 @@ static struct pr_usrreqs route_usrreqs = {
.pru_close = rts_close,
};
+#ifndef _SOCKADDR_UNION_DEFINED
+#define _SOCKADDR_UNION_DEFINED
+/*
+ * The union of all possible address formats we handle.
+ */
+union sockaddr_union {
+ struct sockaddr sa;
+ struct sockaddr_in sin;
+ struct sockaddr_in6 sin6;
+};
+#endif /* _SOCKADDR_UNION_DEFINED */
+
+static int
+rtm_get_jailed(struct rt_addrinfo *info, struct ifnet *ifp,
+ struct rtentry *rt, union sockaddr_union *saun, struct ucred *cred)
+{
+
+ switch (info->rti_info[RTAX_DST]->sa_family) {
+#ifdef INET
+ case AF_INET:
+ {
+ struct in_addr ia;
+
+ /*
+ * 1. Check if the returned address is part of the jail.
+ */
+ ia = ((struct sockaddr_in *)rt->rt_ifa->ifa_addr)->sin_addr;
+ if (prison_check_ip4(cred, &ia) != 0) {
+ info->rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr;
+
+ } else {
+ struct ifaddr *ifa;
+ int found;
+
+ found = 0;
+
+ /*
+ * 2. Try to find an address on the given outgoing
+ * interface that belongs to the jail.
+ */
+ TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
+ struct sockaddr *sa;
+ sa = ifa->ifa_addr;
+ if (sa->sa_family != AF_INET)
+ continue;
+ ia = ((struct sockaddr_in *)sa)->sin_addr;
+ if (prison_check_ip4(cred, &ia) != 0) {
+ found = 1;
+ break;
+ }
+ }
+ if (!found) {
+ /*
+ * 3. As a last resort return the 'default'
+ * jail address.
+ */
+ if (prison_getip4(cred, &ia) != 0)
+ return (ESRCH);
+ }
+ bzero(&saun->sin, sizeof(struct sockaddr_in));
+ saun->sin.sin_len = sizeof(struct sockaddr_in);
+ saun->sin.sin_family = AF_INET;
+ saun->sin.sin_addr.s_addr = ia.s_addr;
+ info->rti_info[RTAX_IFA] =
+ (struct sockaddr *)&saun->sin;
+ }
+ break;
+ }
+#endif
+#ifdef INET6
+ case AF_INET6:
+ {
+ struct in6_addr ia6;
+
+ /*
+ * 1. Check if the returned address is part of the jail.
+ */
+ bcopy(&((struct sockaddr_in6 *)rt->rt_ifa->ifa_addr)->sin6_addr,
+ &ia6, sizeof(struct in6_addr));
+ if (prison_check_ip6(cred, &ia6) != 0) {
+ info->rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr;
+ } else {
+ struct ifaddr *ifa;
+ int found;
+
+ found = 0;
+
+ /*
+ * 2. Try to find an address on the given outgoing
+ * interface that belongs to the jail.
+ */
+ TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
+ struct sockaddr *sa;
+ sa = ifa->ifa_addr;
+ if (sa->sa_family != AF_INET6)
+ continue;
+ bcopy(&((struct sockaddr_in6 *)sa)->sin6_addr,
+ &ia6, sizeof(struct in6_addr));
+ if (prison_check_ip6(cred, &ia6) != 0) {
+ found = 1;
+ break;
+ }
+ }
+ if (!found) {
+ /*
+ * 3. As a last resort return the 'default'
+ * jail address.
+ */
+ if (prison_getip6(cred, &ia6) != 0)
+ return (ESRCH);
+ }
+ bzero(&saun->sin6, sizeof(struct sockaddr_in6));
+ saun->sin6.sin6_len = sizeof(struct sockaddr_in6);
+ saun->sin6.sin6_family = AF_INET6;
+ bcopy(&ia6, &saun->sin6.sin6_addr,
+ sizeof(struct in6_addr));
+ if (sa6_recoverscope(&saun->sin6) != 0)
+ return (ESRCH);
+ info->rti_info[RTAX_IFA] =
+ (struct sockaddr *)&saun->sin6;
+ }
+ break;
+ }
+#endif
+ default:
+ return (ESRCH);
+ }
+ return (0);
+}
+
/*ARGSUSED*/
static int
route_output(struct mbuf *m, struct socket *so)
@@ -317,7 +453,7 @@ route_output(struct mbuf *m, struct socket *so)
struct rt_addrinfo info;
int len, error = 0;
struct ifnet *ifp = NULL;
- struct sockaddr_in jail;
+ union sockaddr_union saun;
#define senderr(e) { error = e; goto flush;}
if (m == NULL || ((m->m_len < sizeof(long)) &&
@@ -459,16 +595,17 @@ route_output(struct mbuf *m, struct socket *so)
info.rti_info[RTAX_IFP] =
ifp->if_addr->ifa_addr;
if (jailed(so->so_cred)) {
- bzero(&jail, sizeof(jail));
- jail.sin_family = PF_INET;
- jail.sin_len = sizeof(jail);
- jail.sin_addr.s_addr =
- htonl(prison_getip(so->so_cred));
- info.rti_info[RTAX_IFA] =
- (struct sockaddr *)&jail;
- } else
+ error = rtm_get_jailed(
+ &info, ifp, rt, &saun,
+ so->so_cred);
+ if (error != 0) {
+ RT_UNLOCK(rt);
+ senderr(ESRCH);
+ }
+ } else {
info.rti_info[RTAX_IFA] =
rt->rt_ifa->ifa_addr;
+ }
if (ifp->if_flags & IFF_POINTOPOINT)
info.rti_info[RTAX_BRD] =
rt->rt_ifa->ifa_dstaddr;
@@ -1147,7 +1284,7 @@ sysctl_iflist(int af, struct walkarg *w)
if (af && af != ifa->ifa_addr->sa_family)
continue;
if (jailed(curthread->td_ucred) &&
- prison_if(curthread->td_ucred, ifa->ifa_addr))
+ !prison_if(curthread->td_ucred, ifa->ifa_addr))
continue;
info.rti_info[RTAX_IFA] = ifa->ifa_addr;
info.rti_info[RTAX_NETMASK] = ifa->ifa_netmask;
@@ -1195,7 +1332,7 @@ sysctl_ifmalist(int af, struct walkarg *w)
if (af && af != ifma->ifma_addr->sa_family)
continue;
if (jailed(curproc->p_ucred) &&
- prison_if(curproc->p_ucred, ifma->ifma_addr))
+ !prison_if(curproc->p_ucred, ifma->ifma_addr))
continue;
info.rti_info[RTAX_IFA] = ifma->ifma_addr;
info.rti_info[RTAX_GATEWAY] =
diff --git a/sys/netinet/in_pcb.c b/sys/netinet/in_pcb.c
index d3b4cd2513ca..04776a9f4123 100644
--- a/sys/netinet/in_pcb.c
+++ b/sys/netinet/in_pcb.c
@@ -279,7 +279,7 @@ in_pcbbind_setup(struct inpcb *inp, struct sockaddr *nam, in_addr_t *laddrp,
struct in_addr laddr;
u_short lport = 0;
int wild = 0, reuseport = (so->so_options & SO_REUSEPORT);
- int error, prison = 0;
+ int error;
int dorandom;
/*
@@ -308,9 +308,8 @@ in_pcbbind_setup(struct inpcb *inp, struct sockaddr *nam, in_addr_t *laddrp,
if (sin->sin_family != AF_INET)
return (EAFNOSUPPORT);
#endif
- if (sin->sin_addr.s_addr != INADDR_ANY)
- if (prison_ip(cred, 0, &sin->sin_addr.s_addr))
- return(EINVAL);
+ if (prison_local_ip4(cred, &sin->sin_addr))
+ return (EINVAL);
if (sin->sin_port != *lportp) {
/* Don't allow the port to change. */
if (*lportp != 0)
@@ -345,14 +344,11 @@ in_pcbbind_setup(struct inpcb *inp, struct sockaddr *nam, in_addr_t *laddrp,
priv_check_cred(cred, PRIV_NETINET_RESERVEDPORT,
0))
return (EACCES);
- if (jailed(cred))
- prison = 1;
if (!IN_MULTICAST(ntohl(sin->sin_addr.s_addr)) &&
priv_check_cred(inp->inp_cred,
PRIV_NETINET_REUSEPORT, 0) != 0) {
t = in_pcblookup_local(pcbinfo, sin->sin_addr,
- lport, prison ? 0 : INPLOOKUP_WILDCARD,
- cred);
+ lport, INPLOOKUP_WILDCARD, cred);
/*
* XXX
* This entire block sorely needs a rewrite.
@@ -369,10 +365,10 @@ in_pcbbind_setup(struct inpcb *inp, struct sockaddr *nam, in_addr_t *laddrp,
t->inp_cred->cr_uid))
return (EADDRINUSE);
}
- if (prison && prison_ip(cred, 0, &sin->sin_addr.s_addr))
+ if (prison_local_ip4(cred, &sin->sin_addr))
return (EADDRNOTAVAIL);
t = in_pcblookup_local(pcbinfo, sin->sin_addr,
- lport, prison ? 0 : wild, cred);
+ lport, wild, cred);
if (t && (t->inp_vflag & INP_TIMEWAIT)) {
/*
* XXXRW: If an incpb has had its timewait
@@ -404,9 +400,8 @@ in_pcbbind_setup(struct inpcb *inp, struct sockaddr *nam, in_addr_t *laddrp,
u_short first, last;
int count;
- if (laddr.s_addr != INADDR_ANY)
- if (prison_ip(cred, 0, &laddr.s_addr))
- return (EINVAL);
+ if (prison_local_ip4(cred, &laddr))
+ return (EINVAL);
if (inp->inp_flags & INP_HIGHPORT) {
first = ipport_hifirstauto; /* sysctl */
@@ -490,7 +485,7 @@ in_pcbbind_setup(struct inpcb *inp, struct sockaddr *nam, in_addr_t *laddrp,
wild, cred));
}
}
- if (prison_ip(cred, 0, &laddr.s_addr))
+ if (prison_local_ip4(cred, &laddr))
return (EINVAL);
*laddrp = laddr.s_addr;
*lportp = lport;
@@ -559,7 +554,7 @@ in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr,
struct route sro;
int error;
- KASSERT(laddr != NULL, ("%s: null laddr", __func__));
+ KASSERT(laddr != NULL, ("%s: laddr NULL", __func__));
error = 0;
ia = NULL;
@@ -611,7 +606,7 @@ in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr,
if (sa->sa_family != AF_INET)
continue;
sin = (struct sockaddr_in *)sa;
- if (htonl(prison_getip(cred)) == sin->sin_addr.s_addr) {
+ if (prison_check_ip4(cred, &sin->sin_addr)) {
ia = (struct in_ifaddr *)ifa;
break;
}
@@ -622,7 +617,8 @@ in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr,
}
/* 3. As a last resort return the 'default' jail address. */
- laddr->s_addr = htonl(prison_getip(cred));
+ if (prison_getip4(cred, laddr) != 0)
+ error = EADDRNOTAVAIL;
goto done;
}
@@ -647,7 +643,7 @@ in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr,
/* Jailed. */
/* 1. Check if the iface address belongs to the jail. */
sin = (struct sockaddr_in *)sro.ro_rt->rt_ifa->ifa_addr;
- if (htonl(prison_getip(cred)) == sin->sin_addr.s_addr) {
+ if (prison_check_ip4(cred, &sin->sin_addr)) {
ia = (struct in_ifaddr *)sro.ro_rt->rt_ifa;
laddr->s_addr = ia->ia_addr.sin_addr.s_addr;
goto done;
@@ -663,7 +659,7 @@ in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr,
if (sa->sa_family != AF_INET)
continue;
sin = (struct sockaddr_in *)sa;
- if (htonl(prison_getip(cred)) == sin->sin_addr.s_addr) {
+ if (prison_check_ip4(cred, &sin->sin_addr)) {
ia = (struct in_ifaddr *)ifa;
break;
}
@@ -674,7 +670,8 @@ in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr,
}
/* 3. As a last resort return the 'default' jail address. */
- laddr->s_addr = htonl(prison_getip(cred));
+ if (prison_getip4(cred, laddr) != 0)
+ error = EADDRNOTAVAIL;
goto done;
}
@@ -724,8 +721,7 @@ in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr,
if (sa->sa_family != AF_INET)
continue;
sin = (struct sockaddr_in *)sa;
- if (htonl(prison_getip(cred)) ==
- sin->sin_addr.s_addr) {
+ if (prison_check_ip4(cred, &sin->sin_addr)) {
ia = (struct in_ifaddr *)ifa;
break;
}
@@ -737,7 +733,8 @@ in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr,
}
/* 3. As a last resort return the 'default' jail address. */
- laddr->s_addr = htonl(prison_getip(cred));
+ if (prison_getip4(cred, laddr) != 0)
+ error = EADDRNOTAVAIL;
goto done;
}
@@ -770,7 +767,7 @@ in_pcbconnect_setup(struct inpcb *inp, struct sockaddr *nam,
struct sockaddr_in *sin = (struct sockaddr_in *)nam;
struct in_ifaddr *ia;
struct inpcb *oinp;
- struct in_addr laddr, faddr;
+ struct in_addr laddr, faddr, jailia;
u_short lport, fport;
int error;
@@ -802,9 +799,17 @@ in_pcbconnect_setup(struct inpcb *inp, struct sockaddr *nam,
* and the primary interface supports broadcast,
* choose the broadcast address for that interface.
*/
- if (faddr.s_addr == INADDR_ANY)
- faddr = IA_SIN(TAILQ_FIRST(&in_ifaddrhead))->sin_addr;
- else if (faddr.s_addr == (u_long)INADDR_BROADCAST &&
+ if (faddr.s_addr == INADDR_ANY) {
+ if (cred != NULL && jailed(cred)) {
+ if (prison_getip4(cred, &jailia) != 0)
+ return (EADDRNOTAVAIL);
+ faddr.s_addr = jailia.s_addr;
+ } else {
+ faddr =
+ IA_SIN(TAILQ_FIRST(&in_ifaddrhead))->
+ sin_addr;
+ }
+ } else if (faddr.s_addr == (u_long)INADDR_BROADCAST &&
(TAILQ_FIRST(&in_ifaddrhead)->ia_ifp->if_flags &
IFF_BROADCAST))
faddr = satosin(&TAILQ_FIRST(
@@ -1114,6 +1119,7 @@ in_pcblookup_local(struct inpcbinfo *pcbinfo, struct in_addr laddr,
0, pcbinfo->ipi_hashmask)];
LIST_FOREACH(inp, head, inp_hash) {
#ifdef INET6
+ /* XXX inp locking */
if ((inp->inp_vflag & INP_IPV4) == 0)
continue;
#endif
@@ -1121,9 +1127,11 @@ in_pcblookup_local(struct inpcbinfo *pcbinfo, struct in_addr laddr,
inp->inp_laddr.s_addr == laddr.s_addr &&
inp->inp_lport == lport) {
/*
- * Found.
+ * Found?
*/
- return (inp);
+ if (cred == NULL ||
+ inp->inp_cred->cr_prison == cred->cr_prison)
+ return (inp);
}
}
/*
@@ -1153,7 +1161,11 @@ in_pcblookup_local(struct inpcbinfo *pcbinfo, struct in_addr laddr,
*/
LIST_FOREACH(inp, &phd->phd_pcblist, inp_portlist) {
wildcard = 0;
+ if (cred != NULL &&
+ inp->inp_cred->cr_prison != cred->cr_prison)
+ continue;
#ifdef INET6
+ /* XXX inp locking */
if ((inp->inp_vflag & INP_IPV4) == 0)
continue;
/*
@@ -1186,9 +1198,8 @@ in_pcblookup_local(struct inpcbinfo *pcbinfo, struct in_addr laddr,
if (wildcard < matchwild) {
match = inp;
matchwild = wildcard;
- if (matchwild == 0) {
+ if (matchwild == 0)
break;
- }
}
}
}
@@ -1206,7 +1217,7 @@ in_pcblookup_hash(struct inpcbinfo *pcbinfo, struct in_addr faddr,
struct ifnet *ifp)
{
struct inpcbhead *head;
- struct inpcb *inp;
+ struct inpcb *inp, *tmpinp;
u_short fport = fport_arg, lport = lport_arg;
INP_INFO_LOCK_ASSERT(pcbinfo);
@@ -1214,60 +1225,108 @@ in_pcblookup_hash(struct inpcbinfo *pcbinfo, struct in_addr faddr,
/*
* First look for an exact match.
*/
+ tmpinp = NULL;
head = &pcbinfo->ipi_hashbase[INP_PCBHASH(faddr.s_addr, lport, fport,
pcbinfo->ipi_hashmask)];
LIST_FOREACH(inp, head, inp_hash) {
#ifdef INET6
+ /* XXX inp locking */
if ((inp->inp_vflag & INP_IPV4) == 0)
continue;
#endif
if (inp->inp_faddr.s_addr == faddr.s_addr &&
inp->inp_laddr.s_addr == laddr.s_addr &&
inp->inp_fport == fport &&
- inp->inp_lport == lport)
- return (inp);
+ inp->inp_lport == lport) {
+ /*
+ * XXX We should be able to directly return
+ * the inp here, without any checks.
+ * Well unless both bound with SO_REUSEPORT?
+ */
+ if (jailed(inp->inp_cred))
+ return (inp);
+ if (tmpinp == NULL)
+ tmpinp = inp;
+ }
}
+ if (tmpinp != NULL)
+ return (tmpinp);
/*
* Then look for a wildcard match, if requested.
*/
- if (wildcard) {
- struct inpcb *local_wild = NULL;
+ if (wildcard == INPLOOKUP_WILDCARD) {
+ struct inpcb *local_wild = NULL, *local_exact = NULL;
#ifdef INET6
struct inpcb *local_wild_mapped = NULL;
#endif
+ struct inpcb *jail_wild = NULL;
+ int injail;
+
+ /*
+ * Order of socket selection - we always prefer jails.
+ * 1. jailed, non-wild.
+ * 2. jailed, wild.
+ * 3. non-jailed, non-wild.
+ * 4. non-jailed, wild.
+ */
head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport,
0, pcbinfo->ipi_hashmask)];
LIST_FOREACH(inp, head, inp_hash) {
#ifdef INET6
+ /* XXX inp locking */
if ((inp->inp_vflag & INP_IPV4) == 0)
continue;
#endif
- if (inp->inp_faddr.s_addr == INADDR_ANY &&
- inp->inp_lport == lport) {
- if (ifp && ifp->if_type == IFT_FAITH &&
- (inp->inp_flags & INP_FAITH) == 0)
+ if (inp->inp_faddr.s_addr != INADDR_ANY ||
+ inp->inp_lport != lport)
+ continue;
+
+ /* XXX inp locking */
+ if (ifp && ifp->if_type == IFT_FAITH &&
+ (inp->inp_flags & INP_FAITH) == 0)
+ continue;
+
+ injail = jailed(inp->inp_cred);
+ if (injail) {
+ if (!prison_check_ip4(inp->inp_cred, &laddr))
continue;
- if (inp->inp_laddr.s_addr == laddr.s_addr)
+ } else {
+ if (local_exact != NULL)
+ continue;
+ }
+
+ if (inp->inp_laddr.s_addr == laddr.s_addr) {
+ if (injail)
return (inp);
- else if (inp->inp_laddr.s_addr == INADDR_ANY) {
+ else
+ local_exact = inp;
+ } else if (inp->inp_laddr.s_addr == INADDR_ANY) {
#ifdef INET6
- if (INP_CHECK_SOCKAF(inp->inp_socket,
- AF_INET6))
- local_wild_mapped = inp;
+ /* XXX inp locking, NULL check */
+ if (inp->inp_vflag & INP_IPV6PROTO)
+ local_wild_mapped = inp;
+ else
+#endif /* INET6 */
+ if (injail)
+ jail_wild = inp;
else
-#endif
local_wild = inp;
- }
}
- }
+ } /* LIST_FOREACH */
+ if (jail_wild != NULL)
+ return (jail_wild);
+ if (local_exact != NULL)
+ return (local_exact);
+ if (local_wild != NULL)
+ return (local_wild);
#ifdef INET6
- if (local_wild == NULL)
+ if (local_wild_mapped != NULL)
return (local_wild_mapped);
-#endif
- return (local_wild);
- }
+#endif /* defined(INET6) */
+ } /* if (wildcard == INPLOOKUP_WILDCARD) */
+
return (NULL);
}
diff --git a/sys/netinet/raw_ip.c b/sys/netinet/raw_ip.c
index d9295c531a36..b8f8d51fad15 100644
--- a/sys/netinet/raw_ip.c
+++ b/sys/netinet/raw_ip.c
@@ -265,10 +265,9 @@ rip_input(struct mbuf *m, int off)
continue;
if (inp->inp_faddr.s_addr != ip->ip_src.s_addr)
continue;
- if (jailed(inp->inp_cred) &&
- (htonl(prison_getip(inp->inp_cred)) !=
- ip->ip_dst.s_addr)) {
- continue;
+ if (jailed(inp->inp_cred)) {
+ if (!prison_check_ip4(inp->inp_cred, &ip->ip_dst))
+ continue;
}
if (last) {
struct mbuf *n;
@@ -296,10 +295,9 @@ rip_input(struct mbuf *m, int off)
if (inp->inp_faddr.s_addr &&
inp->inp_faddr.s_addr != ip->ip_src.s_addr)
continue;
- if (jailed(inp->inp_cred) &&
- (htonl(prison_getip(inp->inp_cred)) !=
- ip->ip_dst.s_addr)) {
- continue;
+ if (jailed(inp->inp_cred)) {
+ if (!prison_check_ip4(inp->inp_cred, &ip->ip_dst))
+ continue;
}
if (last) {
struct mbuf *n;
@@ -360,11 +358,15 @@ rip_output(struct mbuf *m, struct socket *so, u_long dst)
ip->ip_off = 0;
ip->ip_p = inp->inp_ip_p;
ip->ip_len = m->m_pkthdr.len;
- if (jailed(inp->inp_cred))
- ip->ip_src.s_addr =
- htonl(prison_getip(inp->inp_cred));
- else
+ if (jailed(inp->inp_cred)) {
+ if (prison_getip4(inp->inp_cred, &ip->ip_src)) {
+ INP_RUNLOCK(inp);
+ m_freem(m);
+ return (EPERM);
+ }
+ } else {
ip->ip_src = inp->inp_laddr;
+ }
ip->ip_dst.s_addr = dst;
ip->ip_ttl = inp->inp_ip_ttl;
} else {
@@ -374,13 +376,10 @@ rip_output(struct mbuf *m, struct socket *so, u_long dst)
}
INP_RLOCK(inp);
ip = mtod(m, struct ip *);
- if (jailed(inp->inp_cred)) {
- if (ip->ip_src.s_addr !=
- htonl(prison_getip(inp->inp_cred))) {
- INP_RUNLOCK(inp);
- m_freem(m);
- return (EPERM);
- }
+ if (!prison_check_ip4(inp->inp_cred, &ip->ip_src)) {
+ INP_RUNLOCK(inp);
+ m_freem(m);
+ return (EPERM);
}
/*
@@ -788,13 +787,8 @@ rip_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
if (nam->sa_len != sizeof(*addr))
return (EINVAL);
- if (jailed(td->td_ucred)) {
- if (addr->sin_addr.s_addr == INADDR_ANY)
- addr->sin_addr.s_addr =
- htonl(prison_getip(td->td_ucred));
- if (htonl(prison_getip(td->td_ucred)) != addr->sin_addr.s_addr)
- return (EADDRNOTAVAIL);
- }
+ if (!prison_check_ip4(td->td_ucred, &addr->sin_addr))
+ return (EADDRNOTAVAIL);
if (TAILQ_EMPTY(&ifnet) ||
(addr->sin_family != AF_INET && addr->sin_family != AF_IMPLINK) ||
diff --git a/sys/netinet/sctp_pcb.c b/sys/netinet/sctp_pcb.c
index ef3a939531a2..3412efeb81fd 100644
--- a/sys/netinet/sctp_pcb.c
+++ b/sys/netinet/sctp_pcb.c
@@ -2553,7 +2553,6 @@ sctp_inpcb_bind(struct socket *so, struct sockaddr *addr,
struct inpcb *ip_inp;
int port_reuse_active = 0;
int bindall;
- int prison = 0;
uint16_t lport;
int error;
uint32_t vrf_id;
@@ -2580,9 +2579,6 @@ sctp_inpcb_bind(struct socket *so, struct sockaddr *addr,
if (p == NULL)
panic("null proc/thread");
#endif
- if (p && jailed(p->td_ucred)) {
- prison = 1;
- }
if (addr != NULL) {
switch (addr->sa_family) {
case AF_INET:
@@ -2600,18 +2596,13 @@ sctp_inpcb_bind(struct socket *so, struct sockaddr *addr,
}
sin = (struct sockaddr_in *)addr;
lport = sin->sin_port;
- if (prison) {
- /*
- * For INADDR_ANY and LOOPBACK the
- * prison_ip() call will transmute
- * the ip address to the proper
- * value (i.e. the IP address owned
- * by the jail).
- */
- if (prison_ip(p->td_ucred, 0, &sin->sin_addr.s_addr)) {
- SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_PCB, EINVAL);
- return (EINVAL);
- }
+ /*
+ * For LOOPBACK the prison_local_ip4() call will transmute the ip address
+ * to the proper value.
+ */
+ if (p && prison_local_ip4(p->td_ucred, &sin->sin_addr) != 0) {
+ SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_PCB, EINVAL);
+ return (EINVAL);
}
if (sin->sin_addr.s_addr != INADDR_ANY) {
bindall = 0;
@@ -2634,12 +2625,16 @@ sctp_inpcb_bind(struct socket *so, struct sockaddr *addr,
return (EINVAL);
}
lport = sin6->sin6_port;
+
/*
- * Jail checks for IPv6 should go HERE! i.e.
- * add the prison_ip() equivilant in this
- * postion to transmute the addresses to the
- * proper one jailed.
+ * For LOOPBACK the prison_local_ip6() call will transmute the ipv6 address
+ * to the proper value.
*/
+ if (p && prison_local_ip6(p->td_ucred, &sin6->sin6_addr,
+ (SCTP_IPV6_V6ONLY(inp) != 0)) != 0) {
+ SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_PCB, EINVAL);
+ return (EINVAL);
+ }
if (!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) {
bindall = 0;
/* KAME hack: embed scopeid */
diff --git a/sys/netinet/sctp_usrreq.c b/sys/netinet/sctp_usrreq.c
index 397f809e9594..1b9e276842ab 100644
--- a/sys/netinet/sctp_usrreq.c
+++ b/sys/netinet/sctp_usrreq.c
@@ -507,11 +507,10 @@ sctp_attach(struct socket *so, int proto, struct thread *p)
struct inpcb *ip_inp;
int error;
uint32_t vrf_id = SCTP_DEFAULT_VRFID;
-
#ifdef IPSEC
uint32_t flags;
-
#endif
+
inp = (struct sctp_inpcb *)so->so_pcb;
if (inp != 0) {
SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EINVAL);
@@ -3949,12 +3948,8 @@ sctp_setopt(struct socket *so, int optname, void *optval, size_t optsize,
struct sctp_getaddresses *addrs;
size_t sz;
struct thread *td;
- int prison = 0;
td = (struct thread *)p;
- if (jailed(td->td_ucred)) {
- prison = 1;
- }
SCTP_CHECK_AND_CAST(addrs, optval, struct sctp_getaddresses,
optsize);
if (addrs->addr->sa_family == AF_INET) {
@@ -3964,10 +3959,12 @@ sctp_setopt(struct socket *so, int optname, void *optval, size_t optsize,
error = EINVAL;
break;
}
- if (prison && prison_ip(td->td_ucred, 0, &(((struct sockaddr_in *)(addrs->addr))->sin_addr.s_addr))) {
+ if (td != NULL && prison_local_ip4(td->td_ucred, &(((struct sockaddr_in *)(addrs->addr))->sin_addr))) {
SCTP_LTRACE_ERR_RET(inp, stcb, NULL, SCTP_FROM_SCTP_USRREQ, EADDRNOTAVAIL);
error = EADDRNOTAVAIL;
+ break;
}
+#ifdef INET6
} else if (addrs->addr->sa_family == AF_INET6) {
sz = sizeof(struct sctp_getaddresses) - sizeof(struct sockaddr) + sizeof(struct sockaddr_in6);
if (optsize < sz) {
@@ -3975,7 +3972,16 @@ sctp_setopt(struct socket *so, int optname, void *optval, size_t optsize,
error = EINVAL;
break;
}
- /* JAIL XXXX Add else here for V6 */
+ if (td != NULL && prison_local_ip6(td->td_ucred, &(((struct sockaddr_in6 *)(addrs->addr))->sin6_addr),
+ (SCTP_IPV6_V6ONLY(inp) != 0)) != 0) {
+ SCTP_LTRACE_ERR_RET(inp, stcb, NULL, SCTP_FROM_SCTP_USRREQ, EADDRNOTAVAIL);
+ error = EADDRNOTAVAIL;
+ break;
+ }
+#endif
+ } else {
+ error = EAFNOSUPPORT;
+ break;
}
sctp_bindx_add_address(so, inp, addrs->addr,
addrs->sget_assoc_id, vrf_id,
@@ -3987,12 +3993,9 @@ sctp_setopt(struct socket *so, int optname, void *optval, size_t optsize,
struct sctp_getaddresses *addrs;
size_t sz;
struct thread *td;
- int prison = 0;
td = (struct thread *)p;
- if (jailed(td->td_ucred)) {
- prison = 1;
- }
+
SCTP_CHECK_AND_CAST(addrs, optval, struct sctp_getaddresses, optsize);
if (addrs->addr->sa_family == AF_INET) {
sz = sizeof(struct sctp_getaddresses) - sizeof(struct sockaddr) + sizeof(struct sockaddr_in);
@@ -4001,10 +4004,12 @@ sctp_setopt(struct socket *so, int optname, void *optval, size_t optsize,
error = EINVAL;
break;
}
- if (prison && prison_ip(td->td_ucred, 0, &(((struct sockaddr_in *)(addrs->addr))->sin_addr.s_addr))) {
+ if (td != NULL && prison_local_ip4(td->td_ucred, &(((struct sockaddr_in *)(addrs->addr))->sin_addr))) {
SCTP_LTRACE_ERR_RET(inp, stcb, NULL, SCTP_FROM_SCTP_USRREQ, EADDRNOTAVAIL);
error = EADDRNOTAVAIL;
+ break;
}
+#ifdef INET6
} else if (addrs->addr->sa_family == AF_INET6) {
sz = sizeof(struct sctp_getaddresses) - sizeof(struct sockaddr) + sizeof(struct sockaddr_in6);
if (optsize < sz) {
@@ -4012,7 +4017,16 @@ sctp_setopt(struct socket *so, int optname, void *optval, size_t optsize,
error = EINVAL;
break;
}
- /* JAIL XXXX Add else here for V6 */
+ if (td != NULL && prison_local_ip6(td->td_ucred, &(((struct sockaddr_in6 *)(addrs->addr))->sin6_addr),
+ (SCTP_IPV6_V6ONLY(inp) != 0)) != 0) {
+ SCTP_LTRACE_ERR_RET(inp, stcb, NULL, SCTP_FROM_SCTP_USRREQ, EADDRNOTAVAIL);
+ error = EADDRNOTAVAIL;
+ break;
+ }
+#endif
+ } else {
+ error = EAFNOSUPPORT;
+ break;
}
sctp_bindx_delete_address(so, inp, addrs->addr,
addrs->sget_assoc_id, vrf_id,
@@ -4104,13 +4118,34 @@ sctp_connect(struct socket *so, struct sockaddr *addr, struct thread *p)
SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EINVAL);
return EINVAL;
}
- if ((addr->sa_family == AF_INET6) && (addr->sa_len != sizeof(struct sockaddr_in6))) {
- SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EINVAL);
- return (EINVAL);
- }
- if ((addr->sa_family == AF_INET) && (addr->sa_len != sizeof(struct sockaddr_in))) {
- SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EINVAL);
- return (EINVAL);
+#ifdef INET6
+ if (addr->sa_family == AF_INET6) {
+ struct sockaddr_in6 *sin6p;
+ if (addr->sa_len != sizeof(struct sockaddr_in6)) {
+ SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EINVAL);
+ return (EINVAL);
+ }
+ sin6p = (struct sockaddr_in6 *)addr;
+ if (p != NULL && prison_remote_ip6(p->td_ucred, &sin6p->sin6_addr) != 0) {
+ SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EINVAL);
+ return (EINVAL);
+ }
+ } else
+#endif
+ if (addr->sa_family == AF_INET) {
+ struct sockaddr_in *sinp;
+ if (addr->sa_len != sizeof(struct sockaddr_in)) {
+ SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EINVAL);
+ return (EINVAL);
+ }
+ sinp = (struct sockaddr_in *)addr;
+ if (p != NULL && prison_remote_ip4(p->td_ucred, &sinp->sin_addr) != 0) {
+ SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EINVAL);
+ return (EINVAL);
+ }
+ } else {
+ SCTP_LTRACE_ERR_RET(inp, NULL, NULL, SCTP_FROM_SCTP_USRREQ, EAFNOSUPPORT);
+ return (EAFNOSUPPORT);
}
SCTP_INP_INCR_REF(inp);
SCTP_ASOC_CREATE_LOCK(inp);
diff --git a/sys/netinet/tcp_usrreq.c b/sys/netinet/tcp_usrreq.c
index b57f25c7a430..bac4a1d66c07 100644
--- a/sys/netinet/tcp_usrreq.c
+++ b/sys/netinet/tcp_usrreq.c
@@ -462,8 +462,8 @@ tcp_usr_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
if (sinp->sin_family == AF_INET
&& IN_MULTICAST(ntohl(sinp->sin_addr.s_addr)))
return (EAFNOSUPPORT);
- if (jailed(td->td_ucred))
- prison_remote_ip(td->td_ucred, 0, &sinp->sin_addr.s_addr);
+ if (prison_remote_ip4(td->td_ucred, &sinp->sin_addr) != 0)
+ return (EINVAL);
TCPDEBUG0;
INP_INFO_WLOCK(&tcbinfo);
@@ -528,6 +528,10 @@ tcp6_usr_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
in6_sin6_2_sin(&sin, sin6p);
inp->inp_vflag |= INP_IPV4;
inp->inp_vflag &= ~INP_IPV6;
+ if (prison_remote_ip4(td->td_ucred, &sin.sin_addr) != 0) {
+ error = EINVAL;
+ goto out;
+ }
if ((error = tcp_connect(tp, (struct sockaddr *)&sin, td)) != 0)
goto out;
error = tcp_output_connect(so, nam);
@@ -536,6 +540,10 @@ tcp6_usr_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
inp->inp_vflag &= ~INP_IPV4;
inp->inp_vflag |= INP_IPV6;
inp->inp_inc.inc_isipv6 = 1;
+ if (prison_remote_ip6(td->td_ucred, &sin6p->sin6_addr) != 0) {
+ error = EINVAL;
+ goto out;
+ }
if ((error = tcp6_connect(tp, nam, td)) != 0)
goto out;
error = tcp_output_connect(so, nam);
diff --git a/sys/netinet/udp_usrreq.c b/sys/netinet/udp_usrreq.c
index 9661550c26fb..ff328c5d2b6f 100644
--- a/sys/netinet/udp_usrreq.c
+++ b/sys/netinet/udp_usrreq.c
@@ -938,9 +938,10 @@ udp_output(struct inpcb *inp, struct mbuf *m, struct sockaddr *addr,
* Jail may rewrite the destination address, so let it do
* that before we use it.
*/
- if (jailed(td->td_ucred))
- prison_remote_ip(td->td_ucred, 0,
- &sin->sin_addr.s_addr);
+ if (prison_remote_ip4(td->td_ucred, &sin->sin_addr) != 0) {
+ error = EINVAL;
+ goto release;
+ }
/*
* If a local address or port hasn't yet been selected, or if
@@ -1187,8 +1188,11 @@ udp_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
return (EISCONN);
}
sin = (struct sockaddr_in *)nam;
- if (jailed(td->td_ucred))
- prison_remote_ip(td->td_ucred, 0, &sin->sin_addr.s_addr);
+ if (prison_remote_ip4(td->td_ucred, &sin->sin_addr) != 0) {
+ INP_WUNLOCK(inp);
+ INP_INFO_WUNLOCK(&udbinfo);
+ return (EAFNOSUPPORT);
+ }
error = in_pcbconnect(inp, nam, td->td_ucred);
if (error == 0)
soisconnected(so);
diff --git a/sys/netinet6/in6_pcb.c b/sys/netinet6/in6_pcb.c
index 81c7bc1a3bb7..fb53de33204d 100644
--- a/sys/netinet6/in6_pcb.c
+++ b/sys/netinet6/in6_pcb.c
@@ -145,6 +145,10 @@ in6_pcbbind(register struct inpcb *inp, struct sockaddr *nam,
if ((error = sa6_embedscope(sin6, ip6_use_defzone)) != 0)
return(error);
+ if (prison_local_ip6(cred, &sin6->sin6_addr,
+ ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0)
+ return (EINVAL);
+
lport = sin6->sin6_port;
if (IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr)) {
/*
@@ -219,6 +223,9 @@ in6_pcbbind(register struct inpcb *inp, struct sockaddr *nam,
return (EADDRINUSE);
}
}
+ if (prison_local_ip6(cred, &sin6->sin6_addr,
+ ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0)
+ return (EADDRNOTAVAIL);
t = in6_pcblookup_local(pcbinfo, &sin6->sin6_addr,
lport, wild, cred);
if (t && (reuseport & ((t->inp_vflag & INP_TIMEWAIT) ?
@@ -251,6 +258,9 @@ in6_pcbbind(register struct inpcb *inp, struct sockaddr *nam,
}
inp->in6p_laddr = sin6->sin6_addr;
}
+ if (prison_local_ip6(cred, &inp->in6p_laddr,
+ ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0)
+ return (EINVAL);
if (lport == 0) {
int e;
if ((e = in6_pcbsetport(&inp->in6p_laddr, inp, cred)) != 0)
@@ -309,6 +319,8 @@ in6_pcbladdr(register struct inpcb *inp, struct sockaddr *nam,
if (IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr))
sin6->sin6_addr = in6addr_loopback;
}
+ if (prison_remote_ip6(inp->inp_cred, &sin6->sin6_addr) != 0)
+ return (EADDRNOTAVAIL);
/*
* XXX: in6_selectsrc might replace the bound local address
@@ -690,15 +702,16 @@ in6_pcblookup_local(struct inpcbinfo *pcbinfo, struct in6_addr *laddr,
head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport,
0, pcbinfo->ipi_hashmask)];
LIST_FOREACH(inp, head, inp_hash) {
+ /* XXX inp locking */
if ((inp->inp_vflag & INP_IPV6) == 0)
continue;
if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) &&
IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr) &&
inp->inp_lport == lport) {
- /*
- * Found.
- */
- return (inp);
+ /* Found. */
+ if (cred == NULL ||
+ inp->inp_cred->cr_prison == cred->cr_prison)
+ return (inp);
}
}
/*
@@ -728,6 +741,10 @@ in6_pcblookup_local(struct inpcbinfo *pcbinfo, struct in6_addr *laddr,
*/
LIST_FOREACH(inp, &phd->phd_pcblist, inp_portlist) {
wildcard = 0;
+ if (cred != NULL &&
+ inp->inp_cred->cr_prison != cred->cr_prison)
+ continue;
+ /* XXX inp locking */
if ((inp->inp_vflag & INP_IPV6) == 0)
continue;
if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr))
@@ -737,7 +754,7 @@ in6_pcblookup_local(struct inpcbinfo *pcbinfo, struct in6_addr *laddr,
if (IN6_IS_ADDR_UNSPECIFIED(laddr))
wildcard++;
else if (!IN6_ARE_ADDR_EQUAL(
- &inp->in6p_laddr, laddr))
+ &inp->in6p_laddr, laddr))
continue;
} else {
if (!IN6_IS_ADDR_UNSPECIFIED(laddr))
@@ -746,9 +763,8 @@ in6_pcblookup_local(struct inpcbinfo *pcbinfo, struct in6_addr *laddr,
if (wildcard < matchwild) {
match = inp;
matchwild = wildcard;
- if (matchwild == 0) {
+ if (matchwild == 0)
break;
- }
}
}
}
@@ -831,11 +847,11 @@ in6_rtchange(struct inpcb *inp, int errno)
*/
struct inpcb *
in6_pcblookup_hash(struct inpcbinfo *pcbinfo, struct in6_addr *faddr,
- u_int fport_arg, struct in6_addr *laddr, u_int lport_arg,
- int wildcard, struct ifnet *ifp)
+ u_int fport_arg, struct in6_addr *laddr, u_int lport_arg, int wildcard,
+ struct ifnet *ifp)
{
struct inpcbhead *head;
- register struct inpcb *inp;
+ struct inpcb *inp, *tmpinp;
u_short fport = fport_arg, lport = lport_arg;
int faith;
@@ -849,10 +865,12 @@ in6_pcblookup_hash(struct inpcbinfo *pcbinfo, struct in6_addr *faddr,
/*
* First look for an exact match.
*/
+ tmpinp = NULL;
head = &pcbinfo->ipi_hashbase[
INP_PCBHASH(faddr->s6_addr32[3] /* XXX */, lport, fport,
pcbinfo->ipi_hashmask)];
LIST_FOREACH(inp, head, inp_hash) {
+ /* XXX inp locking */
if ((inp->inp_vflag & INP_IPV6) == 0)
continue;
if (IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, faddr) &&
@@ -860,32 +878,79 @@ in6_pcblookup_hash(struct inpcbinfo *pcbinfo, struct in6_addr *faddr,
inp->inp_fport == fport &&
inp->inp_lport == lport) {
/*
- * Found.
+ * XXX We should be able to directly return
+ * the inp here, without any checks.
+ * Well unless both bound with SO_REUSEPORT?
*/
- return (inp);
+ if (jailed(inp->inp_cred))
+ return (inp);
+ if (tmpinp == NULL)
+ tmpinp = inp;
}
}
- if (wildcard) {
- struct inpcb *local_wild = NULL;
+ if (tmpinp != NULL)
+ return (tmpinp);
+ /*
+ * Then look for a wildcard match, if requested.
+ */
+ if (wildcard == INPLOOKUP_WILDCARD) {
+ struct inpcb *local_wild = NULL, *local_exact = NULL;
+ struct inpcb *jail_wild = NULL;
+ int injail;
+
+ /*
+ * Order of socket selection - we always prefer jails.
+ * 1. jailed, non-wild.
+ * 2. jailed, wild.
+ * 3. non-jailed, non-wild.
+ * 4. non-jailed, wild.
+ */
head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport,
0, pcbinfo->ipi_hashmask)];
LIST_FOREACH(inp, head, inp_hash) {
+ /* XXX inp locking */
if ((inp->inp_vflag & INP_IPV6) == 0)
continue;
- if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) &&
- inp->inp_lport == lport) {
- if (faith && (inp->inp_flags & INP_FAITH) == 0)
+
+ if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) ||
+ inp->inp_lport != lport) {
+ continue;
+ }
+
+ /* XXX inp locking */
+ if (faith && (inp->inp_flags & INP_FAITH) == 0)
+ continue;
+
+ injail = jailed(inp->inp_cred);
+ if (injail) {
+ if (!prison_check_ip6(inp->inp_cred, laddr))
continue;
- if (IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr,
- laddr))
+ } else {
+ if (local_exact != NULL)
+ continue;
+ }
+
+ if (IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr)) {
+ if (injail)
return (inp);
- else if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr))
+ else
+ local_exact = inp;
+ } else if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) {
+ if (injail)
+ jail_wild = inp;
+ else
local_wild = inp;
}
- }
- return (local_wild);
- }
+ } /* LIST_FOREACH */
+
+ if (jail_wild != NULL)
+ return (jail_wild);
+ if (local_exact != NULL)
+ return (local_exact);
+ if (local_wild != NULL)
+ return (local_wild);
+ } /* if (wildcard == INPLOOKUP_WILDCARD) */
/*
* Not found.
diff --git a/sys/netinet6/in6_src.c b/sys/netinet6/in6_src.c
index ec813402a359..e0358b0a4010 100644
--- a/sys/netinet6/in6_src.c
+++ b/sys/netinet6/in6_src.c
@@ -79,6 +79,7 @@ __FBSDID("$FreeBSD$");
#include <sys/sysctl.h>
#include <sys/errno.h>
#include <sys/time.h>
+#include <sys/jail.h>
#include <sys/kernel.h>
#include <sys/sx.h>
@@ -224,6 +225,11 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
if (*errorp != 0)
return (NULL);
}
+ if (cred != NULL && prison_local_ip6(cred, &srcsock.sin6_addr,
+ (inp != NULL && (inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) {
+ *errorp = EADDRNOTAVAIL;
+ return (NULL);
+ }
ia6 = (struct in6_ifaddr *)ifa_ifwithaddr((struct sockaddr *)(&srcsock));
if (ia6 == NULL ||
@@ -241,6 +247,11 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
* Otherwise, if the socket has already bound the source, just use it.
*/
if (inp != NULL && !IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) {
+ if (cred != NULL && prison_local_ip6(cred, &inp->in6p_laddr,
+ ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) {
+ *errorp = EADDRNOTAVAIL;
+ return (NULL);
+ }
return (&inp->in6p_laddr);
}
@@ -292,6 +303,12 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
if (!ip6_use_deprecated && IFA6_IS_DEPRECATED(ia))
continue;
+ if (cred != NULL &&
+ prison_local_ip6(cred, &ia->ia_addr.sin6_addr,
+ (inp != NULL &&
+ (inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0)
+ continue;
+
/* Rule 1: Prefer same address */
if (IN6_ARE_ADDR_EQUAL(&dst, &ia->ia_addr.sin6_addr)) {
ia_best = ia;
@@ -765,6 +782,10 @@ in6_pcbsetport(struct in6_addr *laddr, struct inpcb *inp, struct ucred *cred)
INP_INFO_WLOCK_ASSERT(pcbinfo);
INP_WLOCK_ASSERT(inp);
+ if (prison_local_ip6(cred, laddr,
+ ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0)
+ return(EINVAL);
+
/* XXX: this is redundant when called from in6_pcbbind */
if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0)
wild = INPLOOKUP_WILDCARD;
diff --git a/sys/netinet6/raw_ip6.c b/sys/netinet6/raw_ip6.c
index 9a35dc1473a6..53ee71fd8266 100644
--- a/sys/netinet6/raw_ip6.c
+++ b/sys/netinet6/raw_ip6.c
@@ -67,6 +67,7 @@ __FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/errno.h>
+#include <sys/jail.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
@@ -154,6 +155,7 @@ rip6_input(struct mbuf **mp, int *offp, int proto)
INP_INFO_RLOCK(&ripcbinfo);
LIST_FOREACH(in6p, &ripcb, inp_list) {
+ /* XXX inp locking */
if ((in6p->in6p_vflag & INP_IPV6) == 0)
continue;
if (in6p->in6p_ip6_nxt &&
@@ -165,6 +167,10 @@ rip6_input(struct mbuf **mp, int *offp, int proto)
if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) &&
!IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, &ip6->ip6_src))
continue;
+ if (jailed(in6p->inp_cred)) {
+ if (!prison_check_ip6(in6p->inp_cred, &ip6->ip6_dst))
+ continue;
+ }
INP_RLOCK(in6p);
if (in6p->in6p_cksum != -1) {
rip6stat.rip6s_isum++;
@@ -391,6 +397,11 @@ rip6_output(m, va_alist)
error = EADDRNOTAVAIL;
goto bad;
}
+ if (jailed(in6p->inp_cred))
+ if (prison_getip6(in6p->inp_cred, in6a) != 0) {
+ error = EPERM;
+ goto bad;
+ }
ip6->ip6_src = *in6a;
if (oifp && scope_ambiguous) {
@@ -649,6 +660,8 @@ rip6_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
if (nam->sa_len != sizeof(*addr))
return (EINVAL);
+ if (!prison_check_ip6(td->td_ucred, &addr->sin6_addr))
+ return (EADDRNOTAVAIL);
if (TAILQ_EMPTY(&ifnet) || addr->sin6_family != AF_INET6)
return (EADDRNOTAVAIL);
if ((error = sa6_embedscope(addr, ip6_use_defzone)) != 0)
diff --git a/sys/netinet6/udp6_usrreq.c b/sys/netinet6/udp6_usrreq.c
index bb7075198de8..121dd0bcb9a9 100644
--- a/sys/netinet6/udp6_usrreq.c
+++ b/sys/netinet6/udp6_usrreq.c
@@ -71,6 +71,7 @@ __FBSDID("$FreeBSD$");
#include "opt_mac.h"
#include <sys/param.h>
+#include <sys/jail.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/mbuf.h>
@@ -845,6 +846,12 @@ udp6_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
goto out;
}
in6_sin6_2_sin(&sin, sin6_p);
+ if (td && jailed(td->td_ucred))
+ if (prison_remote_ip4(td->td_ucred,
+ &sin.sin_addr) != 0) {
+ error = EAFNOSUPPORT;
+ goto out;
+ }
error = in_pcbconnect(inp, (struct sockaddr *)&sin,
td->td_ucred);
if (error == 0) {
@@ -859,6 +866,13 @@ udp6_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
error = EISCONN;
goto out;
}
+ if (td && jailed(td->td_ucred)) {
+ struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam;
+ if (prison_remote_ip6(td->td_ucred, &sin6->sin6_addr) != 0) {
+ error = EAFNOSUPPORT;
+ goto out;
+ }
+ }
error = in6_pcbconnect(inp, nam, td->td_ucred);
if (error == 0) {
if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0) {
diff --git a/sys/security/mac_bsdextended/mac_bsdextended.c b/sys/security/mac_bsdextended/mac_bsdextended.c
index e220851227c8..490882b0b627 100644
--- a/sys/security/mac_bsdextended/mac_bsdextended.c
+++ b/sys/security/mac_bsdextended/mac_bsdextended.c
@@ -53,6 +53,7 @@
#include <sys/mount.h>
#include <sys/mutex.h>
#include <sys/priv.h>
+#include <sys/proc.h>
#include <sys/systm.h>
#include <sys/vnode.h>
#include <sys/sysctl.h>
diff --git a/sys/sys/cpuset.h b/sys/sys/cpuset.h
index 581a1e776a46..e0539a864410 100644
--- a/sys/sys/cpuset.h
+++ b/sys/sys/cpuset.h
@@ -131,6 +131,8 @@ typedef struct _cpuset {
#define CPU_WHICH_TID 1 /* Specifies a thread id. */
#define CPU_WHICH_PID 2 /* Specifies a process id. */
#define CPU_WHICH_CPUSET 3 /* Specifies a set id. */
+/* CPU_WHICH_IRQ */
+#define CPU_WHICH_JAIL 5 /* Specifies a jail id. */
/*
* Reserved cpuset identifiers.
@@ -167,11 +169,15 @@ struct cpuset {
#define CPU_SET_RDONLY 0x0002 /* No modification allowed. */
extern cpuset_t *cpuset_root;
+struct proc;
+struct thread;
struct cpuset *cpuset_thread0(void);
struct cpuset *cpuset_ref(struct cpuset *);
void cpuset_rel(struct cpuset *);
int cpuset_setthread(lwpid_t id, cpuset_t *);
+int cpuset_create_root(struct thread *, struct cpuset **);
+int cpuset_setproc_update_set(struct proc *, struct cpuset *);
#else
__BEGIN_DECLS
diff --git a/sys/sys/jail.h b/sys/sys/jail.h
index b8972f83d64d..d21a9579564d 100644
--- a/sys/sys/jail.h
+++ b/sys/sys/jail.h
@@ -13,21 +13,74 @@
#ifndef _SYS_JAIL_H_
#define _SYS_JAIL_H_
-struct jail {
+#ifdef _KERNEL
+struct jail_v0 {
u_int32_t version;
char *path;
char *hostname;
u_int32_t ip_number;
};
+#endif
+
+struct jail {
+ uint32_t version;
+ char *path;
+ char *hostname;
+ char *jailname;
+ uint32_t ip4s;
+ uint32_t ip6s;
+ struct in_addr *ip4;
+ struct in6_addr *ip6;
+};
+#define JAIL_API_VERSION 2
+
+/*
+ * For all xprison structs, always keep the pr_version an int and
+ * the first variable so userspace can easily distinguish them.
+ */
+#ifndef _KERNEL
+struct xprison_v1 {
+ int pr_version;
+ int pr_id;
+ char pr_path[MAXPATHLEN];
+ char pr_host[MAXHOSTNAMELEN];
+ u_int32_t pr_ip;
+};
+#endif
struct xprison {
int pr_version;
int pr_id;
+ int pr_state;
+ cpusetid_t pr_cpusetid;
char pr_path[MAXPATHLEN];
char pr_host[MAXHOSTNAMELEN];
- u_int32_t pr_ip;
+ char pr_name[MAXHOSTNAMELEN];
+ uint32_t pr_ip4s;
+ uint32_t pr_ip6s;
+#if 0
+ /*
+ * sizeof(xprison) will be malloced + size needed for all
+ * IPv4 and IPv6 addesses. Offsets are based numbers of addresses.
+ */
+ struct in_addr pr_ip4[];
+ struct in6_addr pr_ip6[];
+#endif
+};
+#define XPRISON_VERSION 3
+
+static const struct prison_state {
+ int pr_state;
+ const char * state_name;
+} prison_states[] = {
+#define PRISON_STATE_INVALID 0
+ { PRISON_STATE_INVALID, "INVALID" },
+#define PRISON_STATE_ALIVE 1
+ { PRISON_STATE_ALIVE, "ALIVE" },
+#define PRISON_STATE_DYING 2
+ { PRISON_STATE_DYING, "DYING" },
};
-#define XPRISON_VERSION 1
+
#ifndef _KERNEL
@@ -48,6 +101,8 @@ MALLOC_DECLARE(M_PRISON);
#endif
#endif /* _KERNEL */
+struct cpuset;
+
/*
* This structure describes a prison. It is pointed to by all struct
* ucreds's of the inmates. pr_ref keeps track of them and is used to
@@ -65,15 +120,22 @@ struct prison {
LIST_ENTRY(prison) pr_list; /* (a) all prisons */
int pr_id; /* (c) prison id */
int pr_ref; /* (p) refcount */
+ int pr_state; /* (p) prison state */
+ int pr_nprocs; /* (p) process count */
char pr_path[MAXPATHLEN]; /* (c) chroot path */
+ struct cpuset *pr_cpuset; /* (p) cpuset */
struct vnode *pr_root; /* (c) vnode to rdir */
char pr_host[MAXHOSTNAMELEN]; /* (p) jail hostname */
- u_int32_t pr_ip; /* (c) ip addr host */
+ char pr_name[MAXHOSTNAMELEN]; /* (c) admin jail name */
void *pr_linux; /* (p) linux abi */
int pr_securelevel; /* (p) securelevel */
struct task pr_task; /* (d) destroy task */
struct mtx pr_mtx;
void **pr_slots; /* (p) additional data */
+ int pr_ip4s; /* (c) number of v4 IPs */
+ struct in_addr *pr_ip4; /* (c) v4 IPs of jail */
+ int pr_ip6s; /* (c) number of v6 IPs */
+ struct in6_addr *pr_ip6; /* (c) v6 IPs of jail */
};
#endif /* _KERNEL || _WANT_PRISON */
@@ -101,6 +163,8 @@ struct ucred;
struct mount;
struct sockaddr;
struct statfs;
+struct thread;
+int kern_jail(struct thread *, struct jail *);
int jailed(struct ucred *cred);
void getcredhostname(struct ucred *cred, char *, size_t);
int prison_check(struct ucred *cred1, struct ucred *cred2);
@@ -109,12 +173,23 @@ void prison_enforce_statfs(struct ucred *cred, struct mount *mp,
struct statfs *sp);
struct prison *prison_find(int prid);
void prison_free(struct prison *pr);
-u_int32_t prison_getip(struct ucred *cred);
+void prison_free_locked(struct prison *pr);
void prison_hold(struct prison *pr);
+void prison_hold_locked(struct prison *pr);
+void prison_proc_hold(struct prison *);
+void prison_proc_free(struct prison *);
+int prison_getip4(struct ucred *cred, struct in_addr *ia);
+int prison_local_ip4(struct ucred *cred, struct in_addr *ia);
+int prison_remote_ip4(struct ucred *cred, struct in_addr *ia);
+int prison_check_ip4(struct ucred *cred, struct in_addr *ia);
+#ifdef INET6
+int prison_getip6(struct ucred *, struct in6_addr *);
+int prison_local_ip6(struct ucred *, struct in6_addr *, int);
+int prison_remote_ip6(struct ucred *, struct in6_addr *);
+int prison_check_ip6(struct ucred *, struct in6_addr *);
+#endif
int prison_if(struct ucred *cred, struct sockaddr *sa);
-int prison_ip(struct ucred *cred, int flag, u_int32_t *ip);
int prison_priv_check(struct ucred *cred, int priv);
-void prison_remote_ip(struct ucred *cred, int flags, u_int32_t *ip);
/*
* Kernel jail services.
diff --git a/sys/sys/param.h b/sys/sys/param.h
index 488acc928281..4e5ea67d41b5 100644
--- a/sys/sys/param.h
+++ b/sys/sys/param.h
@@ -57,7 +57,7 @@
* is created, otherwise 1.
*/
#undef __FreeBSD_version
-#define __FreeBSD_version 701102 /* Master, propagated to newvers */
+#define __FreeBSD_version 701103 /* Master, propagated to newvers */
#ifndef LOCORE
#include <sys/types.h>
diff --git a/usr.bin/cpuset/cpuset.1 b/usr.bin/cpuset/cpuset.1
index 12a42f6285f1..0310793c4f37 100644
--- a/usr.bin/cpuset/cpuset.1
+++ b/usr.bin/cpuset/cpuset.1
@@ -25,7 +25,7 @@
.\"
.\" $FreeBSD$
.\"
-.Dd March 2, 2008
+.Dd November 29, 2008
.Dt CPUSET 1
.Os
.Sh NAME
@@ -43,10 +43,10 @@
.Nm
.Op Fl cr
.Op Fl l Ar cpu-list
-.Op Fl p Ar pid | Fl r Ar tid | Fl s Ar setid
+.Op Fl j Ar jailid | Fl p Ar pid | Fl r Ar tid | Fl s Ar setid | Fl x Ar irq
.Nm
.Op Fl cgir
-.Op Fl p Ar pid | Fl r Ar tid | Fl s Ar setid
+.Op Fl j Ar jailid | Fl p Ar pid | Fl r Ar tid | Fl s Ar setid | Fl x Ar irq
.Sh DESCRIPTION
The
.Nm
@@ -56,8 +56,8 @@ about processor binding, sets, and available processors in the system.
.Pp
.Nm
requires a target to modify or query.
-The target may be specified as a command, process id, thread id, or a
-cpuset id.
+The target may be specified as a command, process id, thread id, a
+cpuset id, an irq or a jail id.
Using
.Fl g
the target's set id or mask may be queried.
@@ -111,6 +111,8 @@ the id of the target.
When used with the
.Fl g
option print the id rather than the valid mask of the target.
+.It Fl j Ar jailid
+Specifies a jail id as the target of the operation.
.It Fl l Ar cpu-list
Specifies a list of CPUs to apply to a target.
Specification may include
@@ -124,6 +126,8 @@ The requested operation should reference the root set available via the
target specifier.
.It Fl t Ar tid
Specifies a thread id as the target of the operation.
+.It Fl x Ar irq
+Specifies an irq as the target of the operation.
.El
.Sh EXIT STATUS
.Ex -std
diff --git a/usr.bin/cpuset/cpuset.c b/usr.bin/cpuset/cpuset.c
index c31ebb7fb04f..03fe64a4a466 100644
--- a/usr.bin/cpuset/cpuset.c
+++ b/usr.bin/cpuset/cpuset.c
@@ -46,6 +46,7 @@ __FBSDID("$FreeBSD$");
int cflag;
int gflag;
int iflag;
+int jflag;
int lflag;
int pflag;
int rflag;
@@ -149,7 +150,7 @@ printset(cpuset_t *mask)
printf("\n");
}
-const char *whichnames[] = { NULL, "tid", "pid", "cpuset" };
+const char *whichnames[] = { NULL, "tid", "pid", "cpuset", "N/A", "jail" };
const char *levelnames[] = { NULL, " root", " cpuset", "" };
static void
@@ -194,7 +195,7 @@ main(int argc, char *argv[])
level = CPU_LEVEL_WHICH;
which = CPU_WHICH_PID;
id = pid = tid = setid = -1;
- while ((ch = getopt(argc, argv, "cgil:p:rs:t:")) != -1) {
+ while ((ch = getopt(argc, argv, "cgij:l:p:rs:t:")) != -1) {
switch (ch) {
case 'c':
if (rflag)
@@ -208,6 +209,11 @@ main(int argc, char *argv[])
case 'i':
iflag = 1;
break;
+ case 'j':
+ jflag = 1;
+ which = CPU_WHICH_JAIL;
+ id = atoi(optarg);
+ break;
case 'l':
lflag = 1;
parselist(optarg, &mask);
@@ -243,7 +249,7 @@ main(int argc, char *argv[])
if (argc || lflag)
usage();
/* Only one identity specifier. */
- if (sflag + pflag + tflag > 1)
+ if (jflag + sflag + pflag + tflag > 1)
usage();
if (iflag)
printsetid();
@@ -257,7 +263,7 @@ main(int argc, char *argv[])
* The user wants to run a command with a set and possibly cpumask.
*/
if (argc) {
- if (pflag | rflag | tflag)
+ if (pflag | rflag | tflag | jflag)
usage();
if (sflag) {
if (cpuset_setid(CPU_WHICH_PID, -1, setid))
@@ -283,7 +289,7 @@ main(int argc, char *argv[])
if (!lflag && !sflag)
usage();
/* You can only set a mask on a thread. */
- if (tflag && (sflag || pflag))
+ if (tflag && (sflag | pflag | jflag))
usage();
if (pflag && sflag) {
if (cpuset_setid(CPU_WHICH_PID, pid, setid))
@@ -313,8 +319,8 @@ usage(void)
fprintf(stderr,
" cpuset [-l cpu-list] [-s setid] -p pid\n");
fprintf(stderr,
- " cpuset [-cr] [-l cpu-list] [-p pid | -t tid | -s setid]\n");
+ " cpuset [-cr] [-l cpu-list] [-j jailid | -p pid | -t tid | -s setid]\n");
fprintf(stderr,
- " cpuset [-cgir] [-p pid | -t tid | -s setid]\n");
+ " cpuset [-cgir] [-j jailid | -p pid | -t tid | -s setid]\n");
exit(1);
}
diff --git a/usr.sbin/jail/Makefile b/usr.sbin/jail/Makefile
index 812cbc96c765..e92ced013a3a 100644
--- a/usr.sbin/jail/Makefile
+++ b/usr.sbin/jail/Makefile
@@ -1,5 +1,7 @@
# $FreeBSD$
+.include <bsd.own.mk>
+
PROG= jail
MAN= jail.8
DPADD= ${LIBUTIL}
@@ -7,4 +9,8 @@ LDADD= -lutil
WARNS?= 6
+.if ${MK_INET6_SUPPORT} != "no"
+CFLAGS+= -DINET6
+.endif
+
.include <bsd.prog.mk>
diff --git a/usr.sbin/jail/jail.8 b/usr.sbin/jail/jail.8
index 147c210e04f6..3a0767e66144 100644
--- a/usr.sbin/jail/jail.8
+++ b/usr.sbin/jail/jail.8
@@ -33,7 +33,7 @@
.\"
.\" $FreeBSD$
.\"
-.Dd April 5, 2007
+.Dd January 24, 2009
.Dt JAIL 8
.Os
.Sh NAME
@@ -41,11 +41,12 @@
.Nd "imprison process and its descendants"
.Sh SYNOPSIS
.Nm
-.Op Fl i
+.Op Fl hi
+.Op Fl n Ar jailname
.Op Fl J Ar jid_file
.Op Fl s Ar securelevel
.Op Fl l u Ar username | Fl U Ar username
-.Ar path hostname ip-number command ...
+.Ar path hostname [ip[,..]] command ...
.Sh DESCRIPTION
The
.Nm
@@ -53,8 +54,28 @@ utility imprisons a process and all future descendants.
.Pp
The options are as follows:
.Bl -tag -width ".Fl u Ar username"
+.It Fl h
+Resolve
+.Va hostname
+and add all IP addresses returned by the resolver
+to the list of
+.Va ip-addresses
+for this prison.
+This may affect default address selection for outgoing IPv4 connections
+of prisons.
+The address first returned by the resolver for each address family
+will be used as primary address.
+See
+.Va ip-addresses
+further down for details.
.It Fl i
Output the jail identifier of the newly created jail.
+.It Fl n Ar jailname
+Assign and administrative name to the jail that can be used for management
+or auditing purposes.
+The system will
+.Sy not enforce
+the name to be unique.
.It Fl J Ar jid_file
Write a
.Ar jid_file
@@ -92,8 +113,14 @@ should run.
Directory which is to be the root of the prison.
.It Ar hostname
Hostname of the prison.
-.It Ar ip-number
-IP number assigned to the prison.
+.It Ar ip-addresses
+None, one or more IPv4 and IPv6 addresses assigned to the prison.
+The first address of each address family that was assigned to the jail will
+be used as the source address in case source address selection on unbound
+sockets cannot find a better match.
+It is only possible to start multiple jails with the same IP address,
+if none of the jails has more than this single overlapping IP address
+assigned to itself for the address family in question.
.It Ar command
Pathname of the program which is to be executed.
.El
@@ -179,7 +206,7 @@ is to disable IP services on the host system that listen on all local
IP addresses for a service.
If a network service is present in the host environment that binds all
available IP addresses rather than specific IP addresses, it may service
-requests sent to jail IP addresses.
+requests sent to jail IP addresses if the jail did not bind the port.
This means changing
.Xr inetd 8
to only listen on the
@@ -455,6 +482,29 @@ pkill -j 3
or:
.Pp
.Dl "killall -j 3"
+.Ss "Jails and File Systems"
+It is not possible to
+.Xr mount 8
+or
+.Xr umount 8
+any file system inside a jail unless the file system is marked
+jail-friendly.
+See
+.Va security.jail.mount_allowed
+in the
+.Va "Sysctl MIB Entries"
+section.
+.Pp
+Multiple jails sharing the same file system can influence each other.
+For example a user in one jail can fill the file system also
+leaving no space for processes in the other jail.
+Trying to use
+.Xr quota 1
+to prevent this will not work either as the file system quotas
+are not aware of jails but only look at the user and group IDs.
+This means the same user ID in two jails share the same file
+system quota.
+One would need to use one file system per jail to make this working.
.Ss "Sysctl MIB Entries"
Certain aspects of the jail containments environment may be modified from
the host environment using
@@ -555,6 +605,9 @@ command can be used to find file system types available for mount from within
a jail.
This functionality is disabled by default, but can be enabled by setting this
MIB entry to 1.
+.It Va security.jail.jail_max_af_ips
+This MIB entry determines how may address per address family a prison
+may have. The default is 255.
.El
.Pp
The read-only sysctl variable
@@ -586,6 +639,7 @@ and
.Xr pgrep 1 ,
.Xr pkill 1 ,
.Xr ps 1 ,
+.Xr quota 1 ,
.Xr chroot 2 ,
.Xr jail 2 ,
.Xr jail_attach 2 ,
@@ -604,7 +658,8 @@ and
.Xr sendmail 8 ,
.Xr shutdown 8 ,
.Xr sysctl 8 ,
-.Xr syslogd 8
+.Xr syslogd 8 ,
+.Xr umount 8
.Sh HISTORY
The
.Nm
@@ -622,6 +677,12 @@ who contributed it to
.An Robert Watson
wrote the extended documentation, found a few bugs, added
a few new features, and cleaned up the userland jail environment.
+.Pp
+.An Bjoern A. Zeeb
+added multi-IP jail support for IPv4 and IPv6 based on a patch
+originally done by
+.An Pawel Jakub Dawidek
+for IPv4.
.Sh BUGS
Jail currently lacks the ability to allow access to
specific jail information via
diff --git a/usr.sbin/jail/jail.c b/usr.sbin/jail/jail.c
index e21fcc847d49..be337b7c8d31 100644
--- a/usr.sbin/jail/jail.c
+++ b/usr.sbin/jail/jail.c
@@ -12,10 +12,14 @@ __FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/jail.h>
+#include <sys/queue.h>
+#include <sys/socket.h>
#include <sys/sysctl.h>
+#include <sys/types.h>
#include <netinet/in.h>
#include <arpa/inet.h>
+#include <netdb.h>
#include <err.h>
#include <errno.h>
@@ -25,12 +29,34 @@ __FBSDID("$FreeBSD$");
#include <pwd.h>
#include <stdio.h>
#include <stdlib.h>
+#include <strings.h>
#include <string.h>
#include <unistd.h>
-static void usage(void);
+static void usage(void);
+static int add_addresses(struct addrinfo *);
+static struct in_addr *copy_addr4(void);
+#ifdef INET6
+static struct in6_addr *copy_addr6(void);
+#endif
+
extern char **environ;
+struct addr4entry {
+ STAILQ_ENTRY(addr4entry) addr4entries;
+ struct in_addr ip4;
+ int count;
+};
+struct addr6entry {
+ STAILQ_ENTRY(addr6entry) addr6entries;
+#ifdef INET6
+ struct in6_addr ip6;
+#endif
+ int count;
+};
+STAILQ_HEAD(addr4head, addr4entry) addr4 = STAILQ_HEAD_INITIALIZER(addr4);
+STAILQ_HEAD(addr6head, addr6entry) addr6 = STAILQ_HEAD_INITIALIZER(addr6);
+
#define GET_USER_INFO do { \
pwd = getpwnam(username); \
if (pwd == NULL) { \
@@ -53,22 +79,26 @@ main(int argc, char **argv)
login_cap_t *lcap = NULL;
struct jail j;
struct passwd *pwd = NULL;
- struct in_addr in;
gid_t groups[NGROUPS];
- int ch, i, iflag, Jflag, lflag, ngroups, securelevel, uflag, Uflag;
- char path[PATH_MAX], *ep, *username, *JidFile;
+ int ch, error, i, ngroups, securelevel;
+ int hflag, iflag, Jflag, lflag, uflag, Uflag;
+ char path[PATH_MAX], *jailname, *ep, *username, *JidFile, *ip;
static char *cleanenv;
const char *shell, *p = NULL;
long ltmp;
FILE *fp;
+ struct addrinfo hints, *res0;
- iflag = Jflag = lflag = uflag = Uflag = 0;
+ hflag = iflag = Jflag = lflag = uflag = Uflag = 0;
securelevel = -1;
- username = JidFile = cleanenv = NULL;
+ jailname = username = JidFile = cleanenv = NULL;
fp = NULL;
- while ((ch = getopt(argc, argv, "ils:u:U:J:")) != -1) {
+ while ((ch = getopt(argc, argv, "hiln:s:u:U:J:")) != -1) {
switch (ch) {
+ case 'h':
+ hflag = 1;
+ break;
case 'i':
iflag = 1;
break;
@@ -76,6 +106,9 @@ main(int argc, char **argv)
JidFile = optarg;
Jflag = 1;
break;
+ case 'n':
+ jailname = optarg;
+ break;
case 's':
ltmp = strtol(optarg, &ep, 0);
if (*ep || ep == optarg || ltmp > INT_MAX || !ltmp)
@@ -111,13 +144,62 @@ main(int argc, char **argv)
err(1, "realpath: %s", argv[0]);
if (chdir(path) != 0)
err(1, "chdir: %s", path);
+ /* Initialize struct jail. */
memset(&j, 0, sizeof(j));
- j.version = 0;
+ j.version = JAIL_API_VERSION;
j.path = path;
j.hostname = argv[1];
- if (inet_aton(argv[2], &in) == 0)
- errx(1, "Could not make sense of ip-number: %s", argv[2]);
- j.ip_number = ntohl(in.s_addr);
+ if (jailname != NULL)
+ j.jailname = jailname;
+
+ /* Handle IP addresses. If requested resolve hostname too. */
+ bzero(&hints, sizeof(struct addrinfo));
+ hints.ai_protocol = IPPROTO_TCP;
+ hints.ai_socktype = SOCK_STREAM;
+ if (JAIL_API_VERSION < 2)
+ hints.ai_family = PF_INET;
+ else
+ hints.ai_family = PF_UNSPEC;
+ /* Handle hostname. */
+ if (hflag != 0) {
+ error = getaddrinfo(j.hostname, NULL, &hints, &res0);
+ if (error != 0)
+ errx(1, "failed to handle hostname: %s",
+ gai_strerror(error));
+ error = add_addresses(res0);
+ freeaddrinfo(res0);
+ if (error != 0)
+ errx(1, "failed to add addresses.");
+ }
+ /* Handle IP addresses. */
+ hints.ai_flags = AI_NUMERICHOST;
+ ip = strtok(argv[2], ",");
+ while (ip != NULL) {
+ error = getaddrinfo(ip, NULL, &hints, &res0);
+ if (error != 0)
+ errx(1, "failed to handle ip: %s", gai_strerror(error));
+ error = add_addresses(res0);
+ freeaddrinfo(res0);
+ if (error != 0)
+ errx(1, "failed to add addresses.");
+ ip = strtok(NULL, ",");
+ }
+ /* Count IP addresses and add them to struct jail. */
+ if (!STAILQ_EMPTY(&addr4)) {
+ j.ip4s = STAILQ_FIRST(&addr4)->count;
+ j.ip4 = copy_addr4();
+ if (j.ip4s > 0 && j.ip4 == NULL)
+ errx(1, "copy_addr4()");
+ }
+#ifdef INET6
+ if (!STAILQ_EMPTY(&addr6)) {
+ j.ip6s = STAILQ_FIRST(&addr6)->count;
+ j.ip6 = copy_addr6();
+ if (j.ip6s > 0 && j.ip6 == NULL)
+ errx(1, "copy_addr6()");
+ }
+#endif
+
if (Jflag) {
fp = fopen(JidFile, "w");
if (fp == NULL)
@@ -125,7 +207,7 @@ main(int argc, char **argv)
}
i = jail(&j);
if (i == -1)
- err(1, "jail");
+ err(1, "syscall failed with");
if (iflag) {
printf("%d\n", i);
fflush(stdout);
@@ -183,8 +265,148 @@ usage(void)
{
(void)fprintf(stderr, "%s%s%s\n",
- "usage: jail [-i] [-J jid_file] [-s securelevel] [-l -u ",
- "username | -U username]",
- " path hostname ip-number command ...");
+ "usage: jail [-hi] [-n jailname] [-J jid_file] ",
+ "[-s securelevel] [-l -u username | -U username] ",
+ "path hostname [ip[,..]] command ...");
exit(1);
}
+
+static int
+add_addresses(struct addrinfo *res0)
+{
+ int error;
+ struct addrinfo *res;
+ struct addr4entry *a4p;
+ struct sockaddr_in *sai;
+#ifdef INET6
+ struct addr6entry *a6p;
+ struct sockaddr_in6 *sai6;
+#endif
+ int count;
+
+ error = 0;
+ for (res = res0; res && error == 0; res = res->ai_next) {
+ switch (res->ai_family) {
+ case AF_INET:
+ sai = (struct sockaddr_in *)(void *)res->ai_addr;
+ STAILQ_FOREACH(a4p, &addr4, addr4entries) {
+ if (bcmp(&sai->sin_addr, &a4p->ip4,
+ sizeof(struct in_addr)) == 0) {
+ err(1, "Ignoring duplicate IPv4 address.");
+ break;
+ }
+ }
+ a4p = (struct addr4entry *) malloc(
+ sizeof(struct addr4entry));
+ if (a4p == NULL) {
+ error = 1;
+ break;
+ }
+ bzero(a4p, sizeof(struct addr4entry));
+ bcopy(&sai->sin_addr, &a4p->ip4,
+ sizeof(struct in_addr));
+ if (!STAILQ_EMPTY(&addr4))
+ count = STAILQ_FIRST(&addr4)->count;
+ else
+ count = 0;
+ STAILQ_INSERT_TAIL(&addr4, a4p, addr4entries);
+ STAILQ_FIRST(&addr4)->count = count + 1;
+ break;
+#ifdef INET6
+ case AF_INET6:
+ sai6 = (struct sockaddr_in6 *)(void *)res->ai_addr;
+ STAILQ_FOREACH(a6p, &addr6, addr6entries) {
+ if (bcmp(&sai6->sin6_addr, &a6p->ip6,
+ sizeof(struct in6_addr)) == 0) {
+ err(1, "Ignoring duplicate IPv6 address.");
+ break;
+ }
+ }
+ a6p = (struct addr6entry *) malloc(
+ sizeof(struct addr6entry));
+ if (a6p == NULL) {
+ error = 1;
+ break;
+ }
+ bzero(a6p, sizeof(struct addr6entry));
+ bcopy(&sai6->sin6_addr, &a6p->ip6,
+ sizeof(struct in6_addr));
+ if (!STAILQ_EMPTY(&addr6))
+ count = STAILQ_FIRST(&addr6)->count;
+ else
+ count = 0;
+ STAILQ_INSERT_TAIL(&addr6, a6p, addr6entries);
+ STAILQ_FIRST(&addr6)->count = count + 1;
+ break;
+#endif
+ default:
+ err(1, "Address family %d not supported. Ignoring.\n",
+ res->ai_family);
+ break;
+ }
+ }
+
+ return (error);
+}
+
+static struct in_addr *
+copy_addr4(void)
+{
+ size_t len;
+ struct in_addr *ip4s, *p, ia;
+ struct addr4entry *a4p;
+
+ if (STAILQ_EMPTY(&addr4))
+ return NULL;
+
+ len = STAILQ_FIRST(&addr4)->count * sizeof(struct in_addr);
+
+ ip4s = p = (struct in_addr *)malloc(len);
+ if (ip4s == NULL)
+ return (NULL);
+
+ bzero(p, len);
+
+ while (!STAILQ_EMPTY(&addr4)) {
+ a4p = STAILQ_FIRST(&addr4);
+ STAILQ_REMOVE_HEAD(&addr4, addr4entries);
+ ia.s_addr = a4p->ip4.s_addr;
+ bcopy(&ia, p, sizeof(struct in_addr));
+ p++;
+ free(a4p);
+ }
+
+ return (ip4s);
+}
+
+#ifdef INET6
+static struct in6_addr *
+copy_addr6(void)
+{
+ size_t len;
+ struct in6_addr *ip6s, *p;
+ struct addr6entry *a6p;
+
+ if (STAILQ_EMPTY(&addr6))
+ return NULL;
+
+ len = STAILQ_FIRST(&addr6)->count * sizeof(struct in6_addr);
+
+ ip6s = p = (struct in6_addr *)malloc(len);
+ if (ip6s == NULL)
+ return (NULL);
+
+ bzero(p, len);
+
+ while (!STAILQ_EMPTY(&addr6)) {
+ a6p = STAILQ_FIRST(&addr6);
+ STAILQ_REMOVE_HEAD(&addr6, addr6entries);
+ bcopy(&a6p->ip6, p, sizeof(struct in6_addr));
+ p++;
+ free(a6p);
+ }
+
+ return (ip6s);
+}
+#endif
+
diff --git a/usr.sbin/jexec/Makefile b/usr.sbin/jexec/Makefile
index 2bf817cb84b6..049ccd46e2e2 100644
--- a/usr.sbin/jexec/Makefile
+++ b/usr.sbin/jexec/Makefile
@@ -6,4 +6,6 @@ DPADD= ${LIBUTIL}
LDADD= -lutil
WARNS?= 6
+CFLAGS+= -DSUPPORT_OLD_XPRISON
+
.include <bsd.prog.mk>
diff --git a/usr.sbin/jexec/jexec.8 b/usr.sbin/jexec/jexec.8
index 7dbdffedabba..bdda23d0209b 100644
--- a/usr.sbin/jexec/jexec.8
+++ b/usr.sbin/jexec/jexec.8
@@ -25,7 +25,7 @@
.\"
.\" $FreeBSD$
.\"
-.Dd April 19, 2006
+.Dd November 29, 2008
.Dt JEXEC 8
.Os
.Sh NAME
@@ -34,17 +34,36 @@
.Sh SYNOPSIS
.Nm
.Op Fl u Ar username | Fl U Ar username
+.Op Fl n Ar jailname
.Ar jid command ...
.Sh DESCRIPTION
The
.Nm
utility executes
.Ar command
-inside the jail identified by
-.Ar jid .
+inside the jail identified by either
+.Ar jailname
+or
+.Ar jid
+or both.
+.Pp
+If the jail cannot be identified uniquely by the given parameters,
+an error message is printed.
+.Nm
+will also check the state of the jail (once supported) to be
+.Dv ALIVE
+and ignore jails in other states.
+The mandatory argument
+.Ar jid
+is the unique jail identifier as given by
+.Xr jls 8 .
+In case you only want to match on other criteria, give an empty string.
.Pp
The following options are available:
.Bl -tag -width indent
+.It Fl n Ar jailname
+The name of the jail, if given upon creation of the jail.
+This is not the hostname of the jail.
.It Fl u Ar username
The user name from host environment as whom the
.Ar command
@@ -63,3 +82,13 @@ The
.Nm
utility was added in
.Fx 5.1 .
+.Sh BUGS
+If the jail is not identified by
+.Ar jid
+there is a possible race in between the lookup of the jail
+and executing the command inside the jail.
+Giving a
+.Ar jid
+has a similar race as another process can stop the jail and
+start another one after the user looked up the
+.Ar jid .
diff --git a/usr.sbin/jexec/jexec.c b/usr.sbin/jexec/jexec.c
index a4c0ff832d42..eb8992b2696b 100644
--- a/usr.sbin/jexec/jexec.c
+++ b/usr.sbin/jexec/jexec.c
@@ -1,5 +1,6 @@
/*-
* Copyright (c) 2003 Mike Barcroft <mike@FreeBSD.org>
+ * Copyright (c) 2008 Bjoern A. Zeeb <bz@FreeBSD.org>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
@@ -29,16 +30,171 @@
#include <sys/param.h>
#include <sys/jail.h>
+#include <netinet/in.h>
#include <err.h>
#include <errno.h>
#include <login_cap.h>
#include <stdio.h>
#include <stdlib.h>
+#include <string.h>
#include <pwd.h>
#include <unistd.h>
static void usage(void);
+#ifdef SUPPORT_OLD_XPRISON
+static
+char *lookup_xprison_v1(void *p, char *end, int *id)
+{
+ struct xprison_v1 *xp;
+
+ if (id == NULL)
+ errx(1, "Internal error. Invalid ID pointer.");
+
+ if ((char *)p + sizeof(struct xprison_v1) > end)
+ errx(1, "Invalid length for jail");
+
+ xp = (struct xprison_v1 *)p;
+
+ *id = xp->pr_id;
+ return ((char *)(xp + 1));
+}
+#endif
+
+static
+char *lookup_xprison_v3(void *p, char *end, int *id, char *jailname)
+{
+ struct xprison *xp;
+ char *q;
+ int ok;
+
+ if (id == NULL)
+ errx(1, "Internal error. Invalid ID pointer.");
+
+ if ((char *)p + sizeof(struct xprison) > end)
+ errx(1, "Invalid length for jail");
+
+ xp = (struct xprison *)p;
+ ok = 1;
+
+ /* Jail state and name. */
+ if (xp->pr_state < 0 || xp->pr_state >=
+ (int)((sizeof(prison_states) / sizeof(struct prison_state))))
+ errx(1, "Invalid jail state.");
+ else if (xp->pr_state != PRISON_STATE_ALIVE)
+ ok = 0;
+ if (jailname != NULL) {
+ if (xp->pr_name[0] == '\0')
+ ok = 0;
+ else if (strcmp(jailname, xp->pr_name) != 0)
+ ok = 0;
+ }
+
+ q = (char *)(xp + 1);
+ /* IPv4 addresses. */
+ q += (xp->pr_ip4s * sizeof(struct in_addr));
+ if ((char *)q > end)
+ errx(1, "Invalid length for jail");
+ /* IPv6 addresses. */
+ q += (xp->pr_ip6s * sizeof(struct in6_addr));
+ if ((char *)q > end)
+ errx(1, "Invalid length for jail");
+
+ if (ok)
+ *id = xp->pr_id;
+ return (q);
+}
+
+static int
+lookup_jail(int jid, char *jailname)
+{
+ size_t i, j, len;
+ void *p, *q;
+ int version, id, xid, count;
+
+ if (sysctlbyname("security.jail.list", NULL, &len, NULL, 0) == -1)
+ err(1, "sysctlbyname(): security.jail.list");
+
+ j = len;
+ for (i = 0; i < 4; i++) {
+ if (len <= 0)
+ exit(0);
+ p = q = malloc(len);
+ if (p == NULL)
+ err(1, "malloc()");
+
+ if (sysctlbyname("security.jail.list", q, &len, NULL, 0) == -1) {
+ if (errno == ENOMEM) {
+ free(p);
+ p = NULL;
+ len += j;
+ continue;
+ }
+ err(1, "sysctlbyname(): security.jail.list");
+ }
+ break;
+ }
+ if (p == NULL)
+ err(1, "sysctlbyname(): security.jail.list");
+ if (len < sizeof(int))
+ errx(1, "This is no prison. Kernel and userland out of sync?");
+ version = *(int *)p;
+ if (version > XPRISON_VERSION)
+ errx(1, "Sci-Fi prison. Kernel/userland out of sync?");
+
+ count = 0;
+ xid = -1;
+ for (; q != NULL && (char *)q + sizeof(int) < (char *)p + len;) {
+ version = *(int *)q;
+ if (version > XPRISON_VERSION)
+ errx(1, "Sci-Fi prison. Kernel/userland out of sync?");
+ id = -1;
+ switch (version) {
+#ifdef SUPPORT_OLD_XPRISON
+ case 1:
+ if (jailname != NULL)
+ errx(1, "Version 1 prisons did not "
+ "support jail names.");
+ q = lookup_xprison_v1(q, (char *)p + len, &id);
+ break;
+ case 2:
+ errx(1, "Version 2 was used by multi-IPv4 jail "
+ "implementations that never made it into the "
+ "official kernel.");
+ /* NOTREACHED */
+ break;
+#endif
+ case 3:
+ q = lookup_xprison_v3(q, (char *)p + len, &id, jailname);
+ break;
+ default:
+ errx(1, "Prison unknown. Kernel/userland out of sync?");
+ /* NOTREACHED */
+ break;
+ }
+ /* Possible match. */
+ if (id > 0) {
+ /* Do we have a jail ID to match as well? */
+ if (jid > 0) {
+ if (jid == id) {
+ xid = id;
+ count++;
+ }
+ } else {
+ xid = id;
+ count++;
+ }
+ }
+ }
+
+ free(p);
+
+ if (count != 1)
+ errx(1, "Could not uniquely identify the jail.");
+
+ return (xid);
+}
+
#define GET_USER_INFO do { \
pwd = getpwnam(username); \
if (pwd == NULL) { \
@@ -63,12 +219,17 @@ main(int argc, char *argv[])
struct passwd *pwd = NULL;
gid_t groups[NGROUPS];
int ch, ngroups, uflag, Uflag;
- char *username;
+ char *jailname, *username;
+
ch = uflag = Uflag = 0;
- username = NULL;
+ jailname = username = NULL;
+ jid = -1;
- while ((ch = getopt(argc, argv, "u:U:")) != -1) {
+ while ((ch = getopt(argc, argv, "i:n:u:U:")) != -1) {
switch (ch) {
+ case 'n':
+ jailname = optarg;
+ break;
case 'u':
username = optarg;
uflag = 1;
@@ -85,11 +246,22 @@ main(int argc, char *argv[])
argv += optind;
if (argc < 2)
usage();
+ if (strlen(argv[0]) > 0) {
+ jid = (int)strtol(argv[0], NULL, 10);
+ if (errno)
+ err(1, "Unable to parse jail ID.");
+ }
+ if (jid <= 0 && jailname == NULL) {
+ fprintf(stderr, "Neither jail ID nor jail name given.\n");
+ usage();
+ }
if (uflag && Uflag)
usage();
if (uflag)
GET_USER_INFO;
- jid = (int)strtol(argv[0], NULL, 10);
+ jid = lookup_jail(jid, jailname);
+ if (jid <= 0)
+ errx(1, "Cannot identify jail.");
if (jail_attach(jid) == -1)
err(1, "jail_attach(): %d", jid);
if (chdir("/") == -1)
@@ -117,6 +289,6 @@ usage(void)
fprintf(stderr, "%s%s\n",
"usage: jexec [-u username | -U username]",
- " jid command ...");
+ " [-n jailname] jid command ...");
exit(1);
}
diff --git a/usr.sbin/jls/Makefile b/usr.sbin/jls/Makefile
index 3968946ce642..01294bd43538 100644
--- a/usr.sbin/jls/Makefile
+++ b/usr.sbin/jls/Makefile
@@ -4,4 +4,6 @@ PROG= jls
MAN= jls.8
WARNS?= 6
+CFLAGS+= -DSUPPORT_OLD_XPRISON
+
.include <bsd.prog.mk>
diff --git a/usr.sbin/jls/jls.8 b/usr.sbin/jls/jls.8
index 4dc68c77ad28..aff9848afaab 100644
--- a/usr.sbin/jls/jls.8
+++ b/usr.sbin/jls/jls.8
@@ -25,20 +25,44 @@
.\"
.\" $FreeBSD$
.\"
-.Dd April 8, 2003
+.Dd November 29, 2008
.Dt JLS 8
.Os
.Sh NAME
.Nm jls
-.Nd "list active jails"
+.Nd "list jails"
.Sh SYNOPSIS
.Nm
+.Op Fl av
.Sh DESCRIPTION
The
.Nm
-utility lists all active jails.
-Each jail is represented by one row which contains the following columns:
-jail identifier (JID), IP address, hostname, and path.
+utility lists all jails.
+By default only active jails are listed.
+.Pp
+The options are as follows:
+.Bl -tag -width ".Fl a"
+.It Fl a
+Show jails in all states, not only active ones.
+.It Fl v
+Show more verbose information.
+This also lists cpusets, jail state, multi-IP, etc. instead of the
+classic single-IP jail output.
+.El
+.Pp
+Each jail is represented by rows which, depending on
+.Fl v ,
+contain the following columns:
+.Bl -item -offset indent -compact
+.It
+jail identifier (JID), hostname and path
+.It
+jail state and name
+.It
+jail cpuset
+.It
+followed by one IP adddress per line.
+.El
.Sh SEE ALSO
.Xr jail 2 ,
.Xr jail 8 ,
diff --git a/usr.sbin/jls/jls.c b/usr.sbin/jls/jls.c
index 23a39468a9d1..95f17d434b2d 100644
--- a/usr.sbin/jls/jls.c
+++ b/usr.sbin/jls/jls.c
@@ -1,5 +1,6 @@
/*-
* Copyright (c) 2003 Mike Barcroft <mike@FreeBSD.org>
+ * Copyright (c) 2008 Bjoern A. Zeeb <bz@FreeBSD.org>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
@@ -27,57 +28,230 @@
*/
#include <sys/param.h>
+#include <sys/types.h>
#include <sys/jail.h>
#include <sys/sysctl.h>
+#include <sys/socket.h>
+#include <netinet/in.h>
#include <arpa/inet.h>
#include <err.h>
#include <errno.h>
#include <limits.h>
#include <stdio.h>
#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#define FLAG_A 0x00001
+#define FLAG_V 0x00002
+
+#ifdef SUPPORT_OLD_XPRISON
+static
+char *print_xprison_v1(void *p, char *end, unsigned flags)
+{
+ struct xprison_v1 *xp;
+ struct in_addr in;
+
+ if ((char *)p + sizeof(struct xprison_v1) > end)
+ errx(1, "Invalid length for jail");
+
+ xp = (struct xprison_v1 *)p;
+ if (flags & FLAG_V) {
+ printf("%6d %-29.29s %.74s\n",
+ xp->pr_id, xp->pr_host, xp->pr_path);
+ /* We are not printing an empty line here for state and name. */
+ /* We are not printing an empty line here for cpusetid. */
+ /* IPv4 address. */
+ in.s_addr = htonl(xp->pr_ip);
+ printf("%6s %-15.15s\n", "", inet_ntoa(in));
+ } else {
+ printf("%6d %-15.15s %-29.29s %.74s\n",
+ xp->pr_id, inet_ntoa(in), xp->pr_host, xp->pr_path);
+ }
+
+ return ((char *)(xp + 1));
+}
+#endif
+
+static
+char *print_xprison_v3(void *p, char *end, unsigned flags)
+{
+ struct xprison *xp;
+ struct in_addr *iap, in;
+ struct in6_addr *ia6p;
+ char buf[INET6_ADDRSTRLEN];
+ const char *state;
+ char *q;
+ uint32_t i;
+
+ if ((char *)p + sizeof(struct xprison) > end)
+ errx(1, "Invalid length for jail");
+ xp = (struct xprison *)p;
+
+ if (xp->pr_state < 0 || xp->pr_state >= (int)
+ ((sizeof(prison_states) / sizeof(struct prison_state))))
+ state = "(bogus)";
+ else
+ state = prison_states[xp->pr_state].state_name;
+
+ /* See if we should print non-ACTIVE jails. No? */
+ if ((flags & FLAG_A) == 0 && strcmp(state, "ALIVE")) {
+ q = (char *)(xp + 1);
+ q += (xp->pr_ip4s * sizeof(struct in_addr));
+ if (q > end)
+ errx(1, "Invalid length for jail");
+ q += (xp->pr_ip6s * sizeof(struct in6_addr));
+ if (q > end)
+ errx(1, "Invalid length for jail");
+ return (q);
+ }
+
+ if (flags & FLAG_V)
+ printf("%6d %-29.29s %.74s\n",
+ xp->pr_id, xp->pr_host, xp->pr_path);
+
+ /* Jail state and name. */
+ if (flags & FLAG_V)
+ printf("%6s %-29.29s %.74s\n",
+ "", (xp->pr_name[0] != '\0') ? xp->pr_name : "", state);
+
+ /* cpusetid. */
+ if (flags & FLAG_V)
+ printf("%6s %-6d\n",
+ "", xp->pr_cpusetid);
+
+ q = (char *)(xp + 1);
+ /* IPv4 addresses. */
+ iap = (struct in_addr *)(void *)q;
+ q += (xp->pr_ip4s * sizeof(struct in_addr));
+ if (q > end)
+ errx(1, "Invalid length for jail");
+ in.s_addr = 0;
+ for (i = 0; i < xp->pr_ip4s; i++) {
+ if (i == 0 || flags & FLAG_V)
+ in.s_addr = iap[i].s_addr;
+ if (flags & FLAG_V)
+ printf("%6s %-15.15s\n", "", inet_ntoa(in));
+ }
+ /* IPv6 addresses. */
+ ia6p = (struct in6_addr *)(void *)q;
+ q += (xp->pr_ip6s * sizeof(struct in6_addr));
+ if (q > end)
+ errx(1, "Invalid length for jail");
+ for (i = 0; i < xp->pr_ip6s; i++) {
+ if (flags & FLAG_V) {
+ inet_ntop(AF_INET6, &ia6p[i], buf, sizeof(buf));
+ printf("%6s %s\n", "", buf);
+ }
+ }
+
+ /* If requested print the old style single line version. */
+ if (!(flags & FLAG_V))
+ printf("%6d %-15.15s %-29.29s %.74s\n",
+ xp->pr_id, (in.s_addr) ? inet_ntoa(in) : "",
+ xp->pr_host, xp->pr_path);
+
+ return (q);
+}
+
+static void
+usage(void)
+{
+
+ (void)fprintf(stderr, "usage: jls [-av]\n");
+ exit(1);
+}
int
-main(void)
+main(int argc, char *argv[])
{
- struct xprison *sxp, *xp;
- struct in_addr in;
- size_t i, len;
+ int ch, version;
+ unsigned flags;
+ size_t i, j, len;
+ void *p, *q;
+
+ flags = 0;
+ while ((ch = getopt(argc, argv, "av")) != -1) {
+ switch (ch) {
+ case 'a':
+ flags |= FLAG_A;
+ break;
+ case 'v':
+ flags |= FLAG_V;
+ break;
+ default:
+ usage();
+ }
+ }
+ argc -= optind;
+ argv += optind;
if (sysctlbyname("security.jail.list", NULL, &len, NULL, 0) == -1)
err(1, "sysctlbyname(): security.jail.list");
+ j = len;
for (i = 0; i < 4; i++) {
if (len <= 0)
exit(0);
- sxp = xp = malloc(len);
- if (sxp == NULL)
+ p = q = malloc(len);
+ if (p == NULL)
err(1, "malloc()");
- if (sysctlbyname("security.jail.list", xp, &len, NULL, 0) == -1) {
+ if (sysctlbyname("security.jail.list", q, &len, NULL, 0) == -1) {
if (errno == ENOMEM) {
- free(sxp);
- sxp = NULL;
+ free(p);
+ p = NULL;
+ len += j;
continue;
}
err(1, "sysctlbyname(): security.jail.list");
}
break;
}
- if (sxp == NULL)
+ if (p == NULL)
err(1, "sysctlbyname(): security.jail.list");
- if (len < sizeof(*xp) || len % sizeof(*xp) ||
- xp->pr_version != XPRISON_VERSION)
- errx(1, "Kernel and userland out of sync");
+ if (len < sizeof(int))
+ errx(1, "This is no prison. Kernel and userland out of sync?");
+ version = *(int *)p;
+ if (version > XPRISON_VERSION)
+ errx(1, "Sci-Fi prison. Kernel/userland out of sync?");
- printf(" JID IP Address Hostname Path\n");
- for (i = 0; i < len / sizeof(*xp); i++) {
- in.s_addr = ntohl(xp->pr_ip);
- printf("%6d %-15.15s %-29.29s %.74s\n",
- xp->pr_id, inet_ntoa(in), xp->pr_host, xp->pr_path);
- xp++;
+ if (flags & FLAG_V) {
+ printf(" JID Hostname Path\n");
+ printf(" Name State\n");
+ printf(" CPUSetID\n");
+ printf(" IP Address(es)\n");
+ } else {
+ printf(" JID IP Address Hostname"
+ " Path\n");
+ }
+ for (; q != NULL && (char *)q + sizeof(int) < (char *)p + len;) {
+ version = *(int *)q;
+ if (version > XPRISON_VERSION)
+ errx(1, "Sci-Fi prison. Kernel/userland out of sync?");
+ switch (version) {
+#ifdef SUPPORT_OLD_XPRISON
+ case 1:
+ q = print_xprison_v1(q, (char *)p + len, flags);
+ break;
+ case 2:
+ errx(1, "Version 2 was used by multi-IPv4 jail "
+ "implementations that never made it into the "
+ "official kernel.");
+ /* NOTREACHED */
+ break;
+#endif
+ case 3:
+ q = print_xprison_v3(q, (char *)p + len, flags);
+ break;
+ default:
+ errx(1, "Prison unknown. Kernel/userland out of sync?");
+ /* NOTREACHED */
+ break;
+ }
}
- free(sxp);
+
+ free(p);
exit(0);
}