minijail0.1

.TH MINIJAIL0 "1" "March 2016" "Chromium OS" "User Commands"
.SH NAME
minijail0 \- sandbox a process
.SH SYNOPSIS
.B minijail0
[\fIOPTION\fR]... <\fIPROGRAM\fR> [\fIargs\fR]...
.SH DESCRIPTION
.PP
Runs PROGRAM inside a sandbox.
.TP
\fB-a <table>\fR
Run using the alternate syscall table named \fItable\fR. Only available on kernels
and architectures that support the \fBPR_ALT_SYSCALL\fR option of \fBprctl\fR(2).
.TP
\fB-b <src>[,<dest>[,<writeable>]]
Bind-mount \fIsrc\fR into the chroot directory at \fIdest\fR, optionally writeable.
The \fIsrc\fR path must be an absolute path.
If \fIdest\fR is not specified, it will default to \fIsrc\fR.
If the destination does not exist, it will be created as a file or directory
based on the \fIsrc\fR type (including missing parent directories).
To create a writable bind-mount set \fIwritable\fR to \fB1\fR. If not specified
it will default to \fB0\fR (read-only).
.TP
\fB-B <mask>\fR
Skip setting securebits in \fImask\fR when restricting capabilities (\fB-c\fR).
\fImask\fR is a hex constant that represents the mask of securebits that will
be preserved.  See \fBcapabilities\fR(7) for the complete list.  By default,
\fBSECURE_NOROOT\fR, \fBSECURE_NO_SETUID_FIXUP\fR, and \fBSECURE_KEEP_CAPS\fR
(together with their respective locks) are set.
\fBSECBIT_NO_CAP_AMBIENT_RAISE\fR (and its respective lock) is never set
because the permitted and inheritable capability sets have already been set
through \fB-c\fR.
.TP
\fB-c <caps>\fR
Restrict capabilities to \fIcaps\fR, which is either a hex constant or a string
that will be passed to \fBcap_from_text\fR(3) (only the effective capability
mask will be considered).  The value will be used as the permitted, effective,
and inheritable sets.  When used in conjunction with \fB-u\fR and \fB-g\fR,
this allows a program to have access to only certain parts of root's default
privileges while running as another user and group ID altogether.  Note that
these capabilities are not inherited by subprocesses of the process given
capabilities unless those subprocesses have POSIX file capabilities or the
\fB--ambient\fR flag is also passed.  See \fBcapabilities\fR(7).
.TP
\fB-C <dir>\fR
Change root (using \fBchroot\fR(2)) to \fIdir\fR.
.TP
\fB-d\fR, \fB--mount-dev\fR
Create a new /dev mount with a minimal set of nodes. Implies \fB-v\fR.
Additional nodes can be bound with the \fB-b\fR or \fB-k\fR options.
The initial set of nodes are: full null tty urandom zero.
Symlinks are also created for: fd ptmx stderr stdin stdout.
.TP
\fB-e[file]\fR
Enter a new network namespace, or if \fIfile\fR is specified, enter an existing
network namespace specified by \fIfile\fR which is typically of the form
/proc/<pid>/ns/net.
.TP
\fB-f <file>\fR
Write the pid of the jailed process to \fIfile\fR.
.TP
\fB-g <group>\fR
Change groups to \fIgroup\fR, which may be either a group name or a numeric
group ID.
.TP
\fB-G\fR
Inherit all the supplementary groups of the user specified with \fB-u\fR. It
is an error to use this option without having specified a \fBuser name\fR to
\fB-u\fR.
.TP
\fB-h\fR
Print a help message.
.TP
\fB-H\fR
Print a help message detailing supported system call names for seccomp_filter.
(Other direct numbers may be specified if minijail0 is not in sync with the
host kernel or something like 32/64-bit compatibility issues exist.)
.TP
\fB-i\fR
Exit immediately after \fBfork\fR(2).  The jailed process will keep running in
the background.

Normally minijail will fork+exec the specified \fIprogram\fR so that it can set
up the right security settings in the new child process. The initial minijail
process will stay resident and wait for the \fIprogram\fR to exit so the script
that ran minijail will correctly block (e.g. standalone scripts). Specifying
\fB-i\fR makes that initial process exit immediately and free up the resources.

This option is recommended for daemons and init services when you want to
background the long running \fIprogram\fR.
.TP
\fB-I\fR
Run \fIprogram\fR as init (pid 1) inside a new pid namespace (implies \fB-p\fR).

Most programs don't expect to run as an init which is why minijail will do it
for you by default. Basically, the \fIprogram\fR needs to reap any processes it
forks to avoid leaving zombies behind. Signal handling needs care since the
kernel will mask all signals that don't have handlers registered (all default
handlers are ignored and cannot be changed).

This means a minijail process (acting as init) will remain resident by default.
While using \fB-I\fR is recommended when possible, strict review is required to
make sure the \fIprogram\fR continues to work as expected.

\fB-i\fR and \fB-I\fR may be safely used together. The \fB-i\fR option controls
the first minijail process outside of the pid namespace while the \fB-I\fR
option controls the minijail process inside of the pid namespace.
.TP
\fB-k <src>,<dest>,<type>[,<flags>[,<data>]]\fR
Mount \fIsrc\fR, a \fItype\fR filesystem, at \fIdest\fR.  If a chroot or pivot
root is active, \fIdest\fR will automatically be placed below that path.

The \fIflags\fR field is optional and may be a mix of \fIMS_XXX\fR or hex
constants separated by \fI|\fR characters.  See \fBmount\fR(2) for details.
\fIMS_NODEV|MS_NOSUID|MS_NOEXEC\fR is the default value (a writable mount
with nodev/nosuid/noexec bits set), and it is strongly recommended that all
mounts have these three bits set whenever possible.  If you need to disable
all three, then specify something like \fIMS_SILENT\fR.

The \fIdata\fR field is optional and is a comma delimited string (see
\fBmount\fR(2) for details).  It is passed directly to the kernel, so all
fields here are filesystem specific.  For \fItmpfs\fR, if no data is specified,
we will default to \fImode=0755,size=10M\fR.  If you want other settings, you
will need to specify them explicitly yourself.

If the mount is not a pseudo filesystem (e.g. proc or sysfs), \fIsrc\fR path
must be an absolute path (e.g. \fI/dev/sda1\fR and not \fIsda1\fR).

If the destination does not exist, it will be created as a directory (including
missing parent directories).
.TP
\fB-K[mode]\fR
Don't mark all existing mounts as MS_PRIVATE.
This option is \fBdangerous\fR as it negates most of the functionality of \fB-v\fR.
You very likely don't need this.

You may specify a mount propagation mode in which case, that will be used
instead of the default MS_PRIVATE.  See the \fBmount\fR(2) man page and the
kernel docs \fIDocumentation/filesystems/sharedsubtree.txt\fR for more
technical details, but a brief guide:

.IP
\[bu] \fBslave\fR Changes in the parent mount namespace will propagate in, but
changes in this mount namespace will not propagate back out.  This is usually
what people want to use.
.IP
\[bu] \fBprivate\fR No changes in either mount namespace will propagate.
This is the default behavior if you don't specify \fB-K\fR.
.IP
\[bu] \fBshared\fR Changes in the parent and this mount namespace will freely
propagate back and forth.  This is not recommended.
.IP
\[bu] \fBunbindable\fR Mark all mounts as unbindable.
.TP
\fB-l\fR
Run inside a new IPC namespace. This option makes the program's System V IPC
namespace independent.
.TP
\fB-L\fR
Report blocked syscalls when using a seccomp filter. On kernels with support for
SECCOMP_RET_LOG, every blocked syscall will be reported through the audit
subsystem (see \fBseccomp\fR(2) for more details on SECCOMP_RET_LOG
availability.) On all other kernels, the first failing syscall will be logged to
syslog. This latter case will also force certain syscalls to be allowed in order
to write to syslog. Note: this option is disabled and ignored for release
builds.
.TP
\fB-m[<uid> <loweruid> <count>[,<uid> <loweruid> <count>]]\fR
Set the uid mapping of a user namespace (implies \fB-pU\fR). Same arguments as
\fBnewuidmap\fR(1). Multiple mappings should be separated by ','. With no mapping,
map the current uid to root inside the user namespace.
.TP
\fB-M[<uid> <loweruid> <count>[,<uid> <loweruid> <count>]]\fR
Set the gid mapping of a user namespace (implies \fB-pU\fR). Same arguments as
\fBnewgidmap\fR(1). Multiple mappings should be separated by ','. With no mapping,
map the current gid to root inside the user namespace.
.TP
\fB-n\fR
Set the process's \fIno_new_privs\fR bit. See \fBprctl\fR(2) and the kernel
source file \fIDocumentation/prctl/no_new_privs.txt\fR for more info.
.TP
\fB-N\fR
Run inside a new cgroup namespace. This option runs the program with a cgroup
view showing the program's cgroup as the root. This is only available on v4.6+
of the Linux kernel.
.TP
\fB-p\fR
Run inside a new PID namespace. This option will make it impossible for the
program to see or affect processes that are not its descendants. This implies
\fB-v\fR and \fB-r\fR, since otherwise the process can see outside its namespace
by inspecting /proc.

If the \fIprogram\fR exits, all of its children will be killed immediately by
the kernel. If you need to daemonize or background things, use the \fB-i\fR
option.

See \fBpid_namespaces\fR(7) for more info.
.TP
\fB-P <dir>\fR
Set \fIdir\fR as the root fs using \fBpivot_root\fR. Implies \fB-v\fR, not
compatible with \fB-C\fR.
.TP
\fB-r\fR
Remount /proc readonly. This implies \fB-v\fR. Remounting /proc readonly means
that even if the process has write access to a system config knob in /proc
(e.g., in /sys/kernel), it cannot change the value.
.TP
\fB-R <rlim_type>,<rlim_cur>,<rlim_max>\fR
Set an rlimit value, see \fBgetrlimit\fR(2) for more details.

\fIrlim_type\fR may be specified using symbolic constants like \fIRLIMIT_AS\fR.

\fIrlim_cur\fR and \fIrlim_max\fR are specified either with a number (decimal or
hex starting with \fI0x\fR), or with the string \fIunlimited\fR (which will
translate to \fIRLIM_INFINITY\fR).
.TP
\fB-s\fR
Enable \fBseccomp\fR(2) in mode 1, which restricts the child process to a very
small set of system calls.
You most likely do not want to use this with the seccomp filter mode (\fB-S\fR)
as they are completely different (even though they have similar names).
.TP
\fB-S <arch-specific seccomp_filter policy file>\fR
Enable \fBseccomp\fR(2) in mode 13 which restricts the child process to a set of
system calls defined in the policy file. Note that system call names may be
different based on the runtime environment; see \fBminijail0\fR(5) for more
details.
.TP
\fB-t[size]\fR
Mounts a tmpfs filesystem on /tmp. /tmp must exist already (e.g. in the chroot).
The filesystem has a default size of "64M", overridden with an optional
argument. It has standard /tmp permissions (1777), and is mounted
nodev/noexec/nosuid. Implies \fB-v\fR.
.TP
\fB-T <type>\fR
Assume binary's ELF linkage type is \fItype\fR, which must be either 'static'
or 'dynamic'. Either setting will prevent minijail0 from manually parsing the
ELF header to determine the type. Type 'static' can be used to avoid preload
hooking, and will force minijail0 to instead set everything up before the
program is executed. Type 'dynamic' will force minijail0 to preload
\fIlibminijailpreload.so\fR to setup hooks, but will fail on actually
statically-linked binaries.
.TP
\fB-u <user>\fR
Change users to \fIuser\fR, which may be either a user name or a numeric user
ID.
.TP
\fB-U\fR
Enter a new user namespace (implies \fB-p\fR).
.TP
\fB-v\fR
Run inside a new VFS namespace. This option makes the program's mountpoints
independent of the rest of the system's.
.TP
\fB-V <file>\fR
Enter the VFS namespace specified by \fIfile\fR.
.TP
\fB-w\fR
Create and join a new anonymous session keyring.  See \fBkeyrings\fR(7) for more
details.
.TP
\fB-y\fR
Keep the current user's supplementary groups.
.TP
\fB-Y\fR
Synchronize seccomp filters across thread group.
.TP
\fB-z\fR
Don't forward any signals to the jailed process.  For example, when not using
\fB-i\fR, sending \fBSIGINT\fR (e.g., CTRL-C on the terminal), will kill the
minijail0 process, not the jailed process.
.TP
\fB--ambient\fR
Raise ambient capabilities to match the mask specified by \fB-c\fR.  Since
ambient capabilities are preserved across \fBexecve\fR(2), this allows for
process trees to have a restricted set of capabilities, even if they are
capability-dumb binaries.  See \fBcapabilities\fR(7).
.TP
\fB--uts[=hostname]\fR
Create a new UTS/hostname namespace, and optionally set the hostname in the new
namespace to \fIhostname\fR.
.TP
\fB--logging=<system>\fR
Use \fIsystem\fR as the logging system. \fIsystem\fR must be one of
\fBsyslog\fR (the default) or \fBstderr\fR.
.TP
\fB--profile <profile>\fR
Choose from one of the available sandboxing profiles, which are simple way to
get a standardized environment. See the
.BR "SANDBOXING PROFILES"
section below for the full list of supported values for \fIprofile\fR.
.TP
\fB--preload-library <file path>\fR
Allows overriding the default path of \fI/lib/libminijailpreload.so\fR.  This
is only really useful for testing.
\fB--seccomp-bpf-binary <arch-specific BPF binary>\fR
This is similar to \fB-S\fR, but
instead of using a policy file, \fB--secomp-bpf-binary\fR expects a
arch-and-kernel-version-specific pre-compiled BPF binary (such as the ones
produced by \fBparse_seccomp_policy\fR).  Note that the filter might be
different based on the runtime environment; see \fBminijail0\fR(5) for more
details.
.SH SANDBOXING PROFILES
The following sandboxing profiles are supported:
.TP
\fBminimalistic-mountns\fR
Set up a minimalistic mount namespace.  Equivalent to \fB-v -P /var/empty
-b / -b /proc -b /dev/log -t -r --mount-dev\fR.
.SH IMPLEMENTATION
This program is broken up into two parts: \fBminijail0\fR (the frontend) and a helper
library called \fBlibminijailpreload\fR.  Some jailings can only be achieved
from the process to which they will actually apply:

.IP
\[bu] capability use (without using ambient capabilities): non-ambient
capabilities are not inherited across \fBexecve\fR(2) unless the file being
executed has POSIX file capabilities.  Ambient capabilities (the
\fB--ambient\fR flag) fix capability inheritance across \fBexecve\fR(2) to
avoid the need for file capabilities.

\[bu] seccomp: a meaningful seccomp filter policy should disallow
\fBexecve\fR(2), to prevent a compromised process from executing a different
binary.  However, this would prevent the seccomp policy from being applied
before \fBexecve\fR(2).
.RE

To this end, \fBlibminijailpreload\fR is forcibly loaded into all
dynamically-linked target programs by default; we pass the specific
restrictions in an environment variable which the preloaded library looks for.
The forcibly-loaded library then applies the restrictions to the newly-loaded
program.

This behavior can be disabled by the use of the \fB-T static\fR flag.  There
are other cases in which the use of this flag might be useful:

.IP
\[bu] When \fIprogram\fR is linked against a different version of \fBlibc.so\fR
than \fBlibminijailpreload.so\fR.

\[bu] When \fBexecve\fR(2) has side-effects that interact badly with the
jailing process.  If the system uses SELinux, \fBexecve\fR(2) can cause an
automatic domain transition, which would then require that the target domain
allows the operations to jail \fIprogram\fR.
.RE

.SH AUTHOR
The Chromium OS Authors <chromiumos-dev@chromium.org>
.SH COPYRIGHT
Copyright \(co 2011 The Chromium OS Authors
License BSD-like.
.SH "SEE ALSO"
\fBlibminijail.h\fR \fBminijail0\fR(5) \fBseccomp\fR(2)