Apply a PAM session to run/task/sysv/services Finit starts, pam_limits
above all, so a service running as a given user picks up that user's
limits the way a login does.
Add a new `pam` setting for the new block format (only), like the
per-service directories, naming a file in /etc/pam.d:
service weston {
user = "weston"
pam = "weston-autologin"
command = "/usr/bin/weston --continue-without-input"
}
pam_close_session() has to be called by a process still holding the
handle, and the handle does not survive exec(). Hence the keeper: it
holds the handle, drops to the service's credentials, and waits for a
parent-death signal before closing the session. Same shape as
systemd's (sd-pam), for the same reason, and one per fork, so the
script hooks open and close their own.
The keeper closes the descriptors it inherited from Finit and only
those. Closing everything would also take out what pam_open_session()
opened for itself, a keyring fd or a lock file, and leave the modules
to close a session with those pulled out from under them. Closing
nothing, as (sd-pam) does, would leave it holding the write end of the
notify pipe for the service's whole lifetime and starve notify = "s6"
services of their ready signal. So the fds open before pam_start()
are snapshotted and exactly those are closed, while the ones PAM opens
after are marked close-on-exec so the daemon does not inherit them
either.
A refused value, a denied account stack, an uninstalled pam.d file,
and a build without PAM support all keep the service from starting
rather than running it with the stacks skipped: one that quietly loses
pam_limits and its private /tmp, with nothing said. Capabilities a
module like pam_cap.so granted are merged into the IAB Finit applies
instead of being replaced by it, which only helps a service that also
sets capabilities, the other arm being a plain setuid() with nothing
left to restore once permitted is empty.
The test sysroot gains pam_permit.so, pam_deny.so and pam_limits.so,
which ldd cannot see, libpam dlopen()s them, and the test skips when
the host has none to stage. The negative cases pin the exit status
rather than only asserting crashed, which serv reports for any early
exit, so a bad command or an unwritable pidfile cannot pass for a
rejected session.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
7.0 KiB
Linux Capabilities
Finit supports Linux capabilities, allowing services to run with minimal required privileges instead of running as root. This significantly improves system security by following the principle of least privilege.
Overview
Linux capabilities divide the traditional root privileges into distinct units that can be independently granted to processes. For example, a web server only needs the capability to bind to privileged ports (< 1024), not full root access.
Finit uses the modern IAB (Inheritable, Ambient, Bounding) API from libcap, which is the same approach used by other modern service managers like dinit.
Basic Usage
Capabilities are specified with the capabilities key, alias caps:
service nginx {
description = "Web server"
runlevel = "2345"
user = "www-data"
group = "www-data"
capabilities = { "^cap_net_bind_service" }
command = "/usr/sbin/nginx -g 'daemon off;'"
}
This example allows nginx to bind to privileged ports (like 80 and 443) while
running as the unprivileged www-data user.
IAB Format
The capability string uses the IAB (Inheritable, Ambient, Bounding) format with the following prefixes:
-
^Ambient (and Inheritable) - Recommended for most use cases- Capabilities survive across
exec()calls - Automatically raised to effective after exec
- Example:
^cap_net_bind_service
- Capabilities survive across
-
%Inheritable only- Requires the executed binary to have matching file capabilities
- Less common, more complex setup
- Example:
%cap_net_admin
-
!Bounding - Block capability from bounding set- Prevents the service from ever acquiring this capability
- Useful for security hardening
- Example:
!cap_sys_admin
Multiple capabilities can be specified as a comma-separated list:
capabilities = { "^cap_net_raw", "^cap_net_admin", "^cap_net_bind_service" }
Common Use Cases
Web Server (Privileged Ports)
Allow a web server to bind to ports 80 and 443 without running as root:
service webserver {
runlevel = "2345"
user = "www-data"
group = "www-data"
capabilities = { "^cap_net_bind_service" }
command = "/usr/sbin/nginx -g 'daemon off;'"
}
Network Monitoring (Raw Sockets)
Allow packet capture without root privileges:
service tcpdump {
runlevel = "2345"
user = "tcpdump"
capabilities = { "^cap_net_raw", "^cap_net_admin" }
command = "/usr/sbin/tcpdump -i eth0 -w /var/log/capture.pcap"
}
NTP Daemon (System Time)
Allow time synchronization without full root:
service ntpd {
runlevel = "2345"
user = "ntp"
capabilities = { "^cap_sys_time", "^cap_sys_nice" }
command = "/usr/sbin/ntpd -n"
}
Available Capabilities
Common capabilities include (see man 7 capabilities for the complete list):
cap_chown- Make arbitrary changes to file UIDs and GIDscap_dac_override- Bypass file read, write, and execute permission checkscap_dac_read_search- Bypass file read permission checkscap_fowner- Bypass permission checks on operations that normally require filesystem UIDcap_kill- Bypass permission checks for sending signalscap_net_admin- Perform various network-related operationscap_net_bind_service- Bind to privileged ports (< 1024)cap_net_raw- Use RAW and PACKET socketscap_setgid- Make arbitrary manipulations of process GIDscap_setuid- Make arbitrary manipulations of process UIDscap_sys_admin- Perform system administration operations (very powerful!)cap_sys_module- Load and unload kernel modulescap_sys_nice- Raise process nice value and change schedulingcap_sys_time- Set system clock
Security Best Practices
-
Use the minimum required capabilities
- Only grant what the service actually needs
- Don't grant
cap_sys_adminunless absolutely necessary
-
Specify a user (preferably non-root)
- The
usersetting is required forcapabilitiesto take effect - For ambient capabilities (
^), use a non-root user (not"root") - Example:
user = "www-data",user = "nginx",user = "tcpdump"
- The
-
Use ambient capabilities (
^)- The
^prefix ensures capabilities survive exec() - Simpler than setting file capabilities on binaries
- The
-
Block dangerous capabilities
- Use
!to explicitly block capabilities you don't want - Example:
!cap_sys_admin,!cap_sys_module
- Use
-
Test with
getpcaps- After starting a service, verify its capabilities:
getpcaps $(pidof nginx) - Should show only the capabilities you granted
- After starting a service, verify its capabilities:
Verification
After configuring a service with capabilities, verify it works correctly:
# Start the service
initctl start webserver
# Check the process capabilities
getpcaps $(pidof nginx)
# Should show something like:
# 12345: cap_net_bind_service=eip
# Verify the user
ps -o user,pid,cmd -p $(pidof nginx)
# Should show the service running as the specified user
Requirements
- Linux kernel 4.3+ (for ambient capabilities support)
- libcap library installed
- Finit built with
--enable-libcap, otherwise acapabilitieslist is ignored, with a warning
Limitations
capabilitiesrequiresuserto be set for it to take effect- Without
user, the service runs as root with full capabilities and thecapabilitieslist is silently ignored - You can use
user = "root", but see below about ambient capabilities
- Without
- For ambient capabilities (
^, recommended), the user must be non-root-
Using
user = "root"with^capabilities will not work effectively, as ambient capabilities are only added to the effective set when euid ≠ 0 -
Use inheritable (
%) or bounding (!) capabilities withuser = "root"if needed -
Finit warns about this when reading the .conf file:
nginx: ambient capabilities ('^') have no effect as root, use a non-root user, or '%' and '!' entries
-
- Services without
capabilitiesuse standard privilege dropping:- Services with a non-root
userhave no special capabilities - Services without
userrun as root with full capabilities
- Services with a non-root
- A capability granted by
pam_cap.soin a PAM session is merged into this set, but only whencapabilitiesis also set. Without it the grant is lost when privileges drop - Some very old binaries may not work correctly with ambient capabilities
- File system capabilities are not managed by Finit (use
setcapfor that)
See Also
- capabilities(7) - Linux capabilities overview
- cap_iab(3) - IAB capability API documentation
- setcap(8) - Set file capabilities
- getcap(8) - Query file capabilities
- capsh(1) - Capability shell wrapper