fanotify.7: Minor fixes after feedback from Matthew Bobrowski

[thirdparty/man-pages.git] / man7 / capabilities.7

.\" Copyright (c) 2002 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"
.\" %%%LICENSE_START(VERBATIM)
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\"
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\"
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date.  The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein.  The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\"
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\" %%%LICENSE_END
.\"
.\" 6 Aug 2002 - Initial Creation
.\" Modified 2003-05-23, Michael Kerrisk, <mtk.manpages@gmail.com>
.\" Modified 2004-05-27, Michael Kerrisk, <mtk.manpages@gmail.com>
.\" 2004-12-08, mtk Added O_NOATIME for CAP_FOWNER
.\" 2005-08-16, mtk, Added CAP_AUDIT_CONTROL and CAP_AUDIT_WRITE
.\" 2008-07-15, Serge Hallyn <serue@us.bbm.com>
.\"     Document file capabilities, per-process capability
.\"     bounding set, changed semantics for CAP_SETPCAP,
.\"     and other changes in 2.6.2[45].
.\"     Add CAP_MAC_ADMIN, CAP_MAC_OVERRIDE, CAP_SETFCAP.
.\" 2008-07-15, mtk
.\"     Add text describing circumstances in which CAP_SETPCAP
.\"     (theoretically) permits a thread to change the
.\"     capability sets of another thread.
.\"     Add section describing rules for programmatically
.\"     adjusting thread capability sets.
.\"     Describe rationale for capability bounding set.
.\"     Document "securebits" flags.
.\"     Add text noting that if we set the effective flag for one file
.\"     capability, then we must also set the effective flag for all
.\"     other capabilities where the permitted or inheritable bit is set.
.\" 2011-09-07, mtk/Serge hallyn: Add CAP_SYSLOG
.\"
.TH CAPABILITIES 7 2019-03-06 "Linux" "Linux Programmer's Manual"
.SH NAME
capabilities \- overview of Linux capabilities
.SH DESCRIPTION
For the purpose of performing permission checks,
traditional UNIX implementations distinguish two categories of processes:
.I privileged
processes (whose effective user ID is 0, referred to as superuser or root),
and
.I unprivileged
processes (whose effective UID is nonzero).
Privileged processes bypass all kernel permission checks,
while unprivileged processes are subject to full permission
checking based on the process's credentials
(usually: effective UID, effective GID, and supplementary group list).
.PP
Starting with kernel 2.2, Linux divides the privileges traditionally
associated with superuser into distinct units, known as
.IR capabilities ,
which can be independently enabled and disabled.
Capabilities are a per-thread attribute.
.\"
.SS Capabilities list
The following list shows the capabilities implemented on Linux,
and the operations or behaviors that each capability permits:
.TP
.BR CAP_AUDIT_CONTROL " (since Linux 2.6.11)"
Enable and disable kernel auditing; change auditing filter rules;
retrieve auditing status and filtering rules.
.TP
.BR CAP_AUDIT_READ " (since Linux 3.16)"
.\" commit a29b694aa1739f9d76538e34ae25524f9c549d59
.\" commit 3a101b8de0d39403b2c7e5c23fd0b005668acf48
Allow reading the audit log via a multicast netlink socket.
.TP
.BR CAP_AUDIT_WRITE " (since Linux 2.6.11)"
Write records to kernel auditing log.
.\" FIXME Add FAN_ENABLE_AUDIT
.TP
.BR CAP_BLOCK_SUSPEND " (since Linux 3.5)"
Employ features that can block system suspend
.RB ( epoll (7)
.BR EPOLLWAKEUP ,
.IR /proc/sys/wake_lock ).
.TP
.B CAP_CHOWN
Make arbitrary changes to file UIDs and GIDs (see
.BR chown (2)).
.TP
.B CAP_DAC_OVERRIDE
Bypass file read, write, and execute permission checks.
(DAC is an abbreviation of "discretionary access control".)
.TP
.B CAP_DAC_READ_SEARCH
.PD 0
.RS
.IP * 2
Bypass file read permission checks and
directory read and execute permission checks;
.IP *
invoke
.BR open_by_handle_at (2);
.IP *
use the
.BR linkat (2)
.B AT_EMPTY_PATH
flag to create a link to a file referred to by a file descriptor.
.RE
.PD
.TP
.B CAP_FOWNER
.PD 0
.RS
.IP * 2
Bypass permission checks on operations that normally
require the filesystem UID of the process to match the UID of
the file (e.g.,
.BR chmod (2),
.BR utime (2)),
excluding those operations covered by
.B CAP_DAC_OVERRIDE
and
.BR CAP_DAC_READ_SEARCH ;
.IP *
set inode flags (see
.BR ioctl_iflags (2))
on arbitrary files;
.IP *
set Access Control Lists (ACLs) on arbitrary files;
.IP *
ignore directory sticky bit on file deletion;
.IP *
specify
.B O_NOATIME
for arbitrary files in
.BR open (2)
and
.BR fcntl (2).
.RE
.PD
.TP
.B CAP_FSETID
.PD 0
.RS
.IP * 2
Don't clear set-user-ID and set-group-ID mode
bits when a file is modified;
.IP *
set the set-group-ID bit for a file whose GID does not match
the filesystem or any of the supplementary GIDs of the calling process.
.RE
.PD
.TP
.B CAP_IPC_LOCK
.\" FIXME . As at Linux 3.2, there are some strange uses of this capability
.\" in other places; they probably should be replaced with something else.
Lock memory
.RB ( mlock (2),
.BR mlockall (2),
.BR mmap (2),
.BR shmctl (2)).
.TP
.B CAP_IPC_OWNER
Bypass permission checks for operations on System V IPC objects.
.TP
.B CAP_KILL
Bypass permission checks for sending signals (see
.BR kill (2)).
This includes use of the
.BR ioctl (2)
.B KDSIGACCEPT
operation.
.\" FIXME . CAP_KILL also has an effect for threads + setting child
.\"       termination signal to other than SIGCHLD: without this
.\"       capability, the termination signal reverts to SIGCHLD
.\"       if the child does an exec().  What is the rationale
.\"       for this?
.TP
.BR CAP_LEASE " (since Linux 2.4)"
Establish leases on arbitrary files (see
.BR fcntl (2)).
.TP
.B CAP_LINUX_IMMUTABLE
Set the
.B FS_APPEND_FL
and
.B FS_IMMUTABLE_FL
inode flags (see
.BR ioctl_iflags (2)).
.TP
.BR CAP_MAC_ADMIN " (since Linux 2.6.25)"
Allow MAC configuration or state changes.
Implemented for the Smack Linux Security Module (LSM).
.TP
.BR CAP_MAC_OVERRIDE " (since Linux 2.6.25)"
Override Mandatory Access Control (MAC).
Implemented for the Smack LSM.
.TP
.BR CAP_MKNOD " (since Linux 2.4)"
Create special files using
.BR mknod (2).
.TP
.B CAP_NET_ADMIN
Perform various network-related operations:
.PD 0
.RS
.IP * 2
interface configuration;
.IP *
administration of IP firewall, masquerading, and accounting;
.IP *
modify routing tables;
.IP *
bind to any address for transparent proxying;
.IP *
set type-of-service (TOS)
.IP *
clear driver statistics;
.IP *
set promiscuous mode;
.IP *
enabling multicasting;
.IP *
use
.BR setsockopt (2)
to set the following socket options:
.BR SO_DEBUG ,
.BR SO_MARK ,
.BR SO_PRIORITY
(for a priority outside the range 0 to 6),
.BR SO_RCVBUFFORCE ,
and
.BR SO_SNDBUFFORCE .
.RE
.PD
.TP
.B CAP_NET_BIND_SERVICE
Bind a socket to Internet domain privileged ports
(port numbers less than 1024).
.TP
.B CAP_NET_BROADCAST
(Unused)  Make socket broadcasts, and listen to multicasts.
.\" FIXME Since Linux 4.2, there are use cases for netlink sockets
.\"    commit 59324cf35aba5336b611074028777838a963d03b
.TP
.B CAP_NET_RAW
.PD 0
.RS
.IP * 2
Use RAW and PACKET sockets;
.IP *
bind to any address for transparent proxying.
.RE
.PD
.\" Also various IP options and setsockopt(SO_BINDTODEVICE)
.TP
.B CAP_SETGID
.RS
.PD 0
.IP * 2
Make arbitrary manipulations of process GIDs and supplementary GID list;
.IP *
forge GID when passing socket credentials via UNIX domain sockets;
.IP *
write a group ID mapping in a user namespace (see
.BR user_namespaces (7)).
.PD
.RE
.TP
.BR CAP_SETFCAP " (since Linux 2.6.24)"
Set arbitrary capabilities on a file.
.TP
.B CAP_SETPCAP
If file capabilities are supported (i.e., since Linux 2.6.24):
add any capability from the calling thread's bounding set
to its inheritable set;
drop capabilities from the bounding set (via
.BR prctl (2)
.BR PR_CAPBSET_DROP );
make changes to the
.I securebits
flags.
.IP
If file capabilities are not supported (i.e., kernels before Linux 2.6.24):
grant or remove any capability in the
caller's permitted capability set to or from any other process.
(This property of
.B CAP_SETPCAP
is not available when the kernel is configured to support
file capabilities, since
.B CAP_SETPCAP
has entirely different semantics for such kernels.)
.TP
.B CAP_SETUID
.RS
.PD 0
.IP * 2
Make arbitrary manipulations of process UIDs
.RB ( setuid (2),
.BR setreuid (2),
.BR setresuid (2),
.BR setfsuid (2));
.IP *
forge UID when passing socket credentials via UNIX domain sockets;
.IP *
write a user ID mapping in a user namespace (see
.BR user_namespaces (7)).
.PD
.RE
.\" FIXME CAP_SETUID also an effect in exec(); document this.
.TP
.B CAP_SYS_ADMIN
.IR Note :
this capability is overloaded; see
.IR "Notes to kernel developers" ,
below.
.IP
.PD 0
.RS
.IP * 2
Perform a range of system administration operations including:
.BR quotactl (2),
.BR mount (2),
.BR umount (2),
.BR swapon (2),
.BR swapoff (2),
.BR sethostname (2),
and
.BR setdomainname (2);
.IP *
perform privileged
.BR syslog (2)
operations (since Linux 2.6.37,
.BR CAP_SYSLOG
should be used to permit such operations);
.IP *
perform
.B VM86_REQUEST_IRQ
.BR vm86 (2)
command;
.IP *
perform
.B IPC_SET
and
.B IPC_RMID
operations on arbitrary System V IPC objects;
.IP *
override
.B RLIMIT_NPROC
resource limit;
.IP *
perform operations on
.I trusted
and
.I security
Extended Attributes (see
.BR xattr (7));
.IP *
use
.BR lookup_dcookie (2);
.IP *
use
.BR ioprio_set (2)
to assign
.B IOPRIO_CLASS_RT
and (before Linux 2.6.25)
.B IOPRIO_CLASS_IDLE
I/O scheduling classes;
.IP *
forge PID when passing socket credentials via UNIX domain sockets;
.IP *
exceed
.IR /proc/sys/fs/file-max ,
the system-wide limit on the number of open files,
in system calls that open files (e.g.,
.BR accept (2),
.BR execve (2),
.BR open (2),
.BR pipe (2));
.IP *
employ
.B CLONE_*
flags that create new namespaces with
.BR clone (2)
and
.BR unshare (2)
(but, since Linux 3.8,
creating user namespaces does not require any capability);
.IP *
call
.BR perf_event_open (2);
.IP *
access privileged
.I perf
event information;
.IP *
call
.BR setns (2)
(requires
.B CAP_SYS_ADMIN
in the
.I target
namespace);
.IP *
call
.BR fanotify_init (2);
.IP *
call
.BR bpf (2);
.IP *
perform privileged
.B KEYCTL_CHOWN
and
.B KEYCTL_SETPERM
.BR keyctl (2)
operations;
.IP *
perform
.BR madvise (2)
.B MADV_HWPOISON
operation;
.IP *
employ the
.B TIOCSTI
.BR ioctl (2)
to insert characters into the input queue of a terminal other than
the caller's controlling terminal;
.IP *
employ the obsolete
.BR nfsservctl (2)
system call;
.IP *
employ the obsolete
.BR bdflush (2)
system call;
.IP *
perform various privileged block-device
.BR ioctl (2)
operations;
.IP *
perform various privileged filesystem
.BR ioctl (2)
operations;
.IP *
perform privileged
.BR ioctl (2)
operations on the
.IR /dev/random
device (see
.BR random (4));
.IP *
install a
.BR seccomp (2)
filter without first having to set the
.I no_new_privs
thread attribute;
.IP *
modify allow/deny rules for device control groups;
.IP *
employ the
.BR ptrace (2)
.B PTRACE_SECCOMP_GET_FILTER
operation to dump tracee's seccomp filters;
.IP *
employ the
.BR ptrace (2)
.B PTRACE_SETOPTIONS
operation to suspend the tracee's seccomp protections (i.e., the
.B PTRACE_O_SUSPEND_SECCOMP
flag);
.IP *
perform administrative operations on many device drivers.
.RE
.PD
.TP
.B CAP_SYS_BOOT
Use
.BR reboot (2)
and
.BR kexec_load (2).
.TP
.B CAP_SYS_CHROOT
.RS
.PD 0
.IP * 2
Use
.BR chroot (2);
.IP *
change mount namespaces using
.BR setns (2).
.PD
.RE
.TP
.B CAP_SYS_MODULE
.RS
.PD 0
.IP * 2
Load and unload kernel modules
(see
.BR init_module (2)
and
.BR delete_module (2));
.IP *
in kernels before 2.6.25:
drop capabilities from the system-wide capability bounding set.
.PD
.RE
.TP
.B CAP_SYS_NICE
.PD 0
.RS
.IP * 2
Raise process nice value
.RB ( nice (2),
.BR setpriority (2))
and change the nice value for arbitrary processes;
.IP *
set real-time scheduling policies for calling process,
and set scheduling policies and priorities for arbitrary processes
.RB ( sched_setscheduler (2),
.BR sched_setparam (2),
.BR sched_setattr (2));
.IP *
set CPU affinity for arbitrary processes
.RB ( sched_setaffinity (2));
.IP *
set I/O scheduling class and priority for arbitrary processes
.RB ( ioprio_set (2));
.IP *
apply
.BR migrate_pages (2)
to arbitrary processes and allow processes
to be migrated to arbitrary nodes;
.\" FIXME CAP_SYS_NICE also has the following effect for
.\" migrate_pages(2):
.\"     do_migrate_pages(mm, &old, &new,
.\"         capable(CAP_SYS_NICE) ? MPOL_MF_MOVE_ALL : MPOL_MF_MOVE);
.\"
.\" Document this.
.IP *
apply
.BR move_pages (2)
to arbitrary processes;
.IP *
use the
.B MPOL_MF_MOVE_ALL
flag with
.BR mbind (2)
and
.BR move_pages (2).
.RE
.PD
.TP
.B CAP_SYS_PACCT
Use
.BR acct (2).
.TP
.B CAP_SYS_PTRACE
.PD 0
.RS
.IP * 2
Trace arbitrary processes using
.BR ptrace (2);
.IP *
apply
.BR get_robust_list (2)
to arbitrary processes;
.IP *
transfer data to or from the memory of arbitrary processes using
.BR process_vm_readv (2)
and
.BR process_vm_writev (2);
.IP *
inspect processes using
.BR kcmp (2).
.RE
.PD
.TP
.B CAP_SYS_RAWIO
.PD 0
.RS
.IP * 2
Perform I/O port operations
.RB ( iopl (2)
and
.BR ioperm (2));
.IP *
access
.IR /proc/kcore ;
.IP *
employ the
.B FIBMAP
.BR ioctl (2)
operation;
.IP *
open devices for accessing x86 model-specific registers (MSRs, see
.BR msr (4));
.IP *
update
.IR /proc/sys/vm/mmap_min_addr ;
.IP *
create memory mappings at addresses below the value specified by
.IR /proc/sys/vm/mmap_min_addr ;
.IP *
map files in
.IR /proc/bus/pci ;
.IP *
open
.IR /dev/mem
and
.IR /dev/kmem ;
.IP *
perform various SCSI device commands;
.IP *
perform certain operations on
.BR hpsa (4)
and
.BR cciss (4)
devices;
.IP *
perform a range of device-specific operations on other devices.
.RE
.PD
.TP
.B CAP_SYS_RESOURCE
.PD 0
.RS
.IP * 2
Use reserved space on ext2 filesystems;
.IP *
make
.BR ioctl (2)
calls controlling ext3 journaling;
.IP *
override disk quota limits;
.IP *
increase resource limits (see
.BR setrlimit (2));
.IP *
override
.B RLIMIT_NPROC
resource limit;
.IP *
override maximum number of consoles on console allocation;
.IP *
override maximum number of keymaps;
.IP *
allow more than 64hz interrupts from the real-time clock;
.IP *
raise
.I msg_qbytes
limit for a System V message queue above the limit in
.I /proc/sys/kernel/msgmnb
(see
.BR msgop (2)
and
.BR msgctl (2));
.IP *
allow the
.B RLIMIT_NOFILE
resource limit on the number of "in-flight" file descriptors
to be bypassed when passing file descriptors to another process
via a UNIX domain socket (see
.BR unix (7));
.IP *
override the
.I /proc/sys/fs/pipe-size-max
limit when setting the capacity of a pipe using the
.B F_SETPIPE_SZ
.BR fcntl (2)
command.
.IP *
use
.BR F_SETPIPE_SZ
to increase the capacity of a pipe above the limit specified by
.IR /proc/sys/fs/pipe-max-size ;
.IP *
override
.I /proc/sys/fs/mqueue/queues_max
limit when creating POSIX message queues (see
.BR mq_overview (7));
.IP *
employ the
.BR prctl (2)
.B PR_SET_MM
operation;
.IP *
set
.IR /proc/[pid]/oom_score_adj
to a value lower than the value last set by a process with
.BR CAP_SYS_RESOURCE .
.RE
.PD
.TP
.B CAP_SYS_TIME
Set system clock
.RB ( settimeofday (2),
.BR stime (2),
.BR adjtimex (2));
set real-time (hardware) clock.
.TP
.B CAP_SYS_TTY_CONFIG
Use
.BR vhangup (2);
employ various privileged
.BR ioctl (2)
operations on virtual terminals.
.TP
.BR CAP_SYSLOG " (since Linux 2.6.37)"
.RS
.PD 0
.IP * 2
Perform privileged
.BR syslog (2)
operations.
See
.BR syslog (2)
for information on which operations require privilege.
.IP *
View kernel addresses exposed via
.I /proc
and other interfaces when
.IR /proc/sys/kernel/kptr_restrict
has the value 1.
(See the discussion of the
.I kptr_restrict
in
.BR proc (5).)
.PD
.RE
.TP
.BR CAP_WAKE_ALARM " (since Linux 3.0)"
Trigger something that will wake up the system (set
.B CLOCK_REALTIME_ALARM
and
.B CLOCK_BOOTTIME_ALARM
timers).
.\"
.SS Past and current implementation
A full implementation of capabilities requires that:
.IP 1. 3
For all privileged operations,
the kernel must check whether the thread has the required
capability in its effective set.
.IP 2.
The kernel must provide system calls allowing a thread's capability sets to
be changed and retrieved.
.IP 3.
The filesystem must support attaching capabilities to an executable file,
so that a process gains those capabilities when the file is executed.
.PP
Before kernel 2.6.24, only the first two of these requirements are met;
since kernel 2.6.24, all three requirements are met.
.\"
.SS Notes to kernel developers
When adding a new kernel feature that should be governed by a capability,
consider the following points.
.IP * 3
The goal of capabilities is divide the power of superuser into pieces,
such that if a program that has one or more capabilities is compromised,
its power to do damage to the system would be less than the same program
running with root privilege.
.IP *
You have the choice of either creating a new capability for your new feature,
or associating the feature with one of the existing capabilities.
In order to keep the set of capabilities to a manageable size,
the latter option is preferable,
unless there are compelling reasons to take the former option.
(There is also a technical limit:
the size of capability sets is currently limited to 64 bits.)
.IP *
To determine which existing capability might best be associated
with your new feature, review the list of capabilities above in order
to find a "silo" into which your new feature best fits.
One approach to take is to determine if there are other features
requiring capabilities that will always be used along with the new feature.
If the new feature is useless without these other features,
you should use the same capability as the other features.
.IP *
.IR Don't
choose
.B CAP_SYS_ADMIN
if you can possibly avoid it!
A vast proportion of existing capability checks are associated
with this capability (see the partial list above).
It can plausibly be called "the new root",
since on the one hand, it confers a wide range of powers,
and on the other hand,
its broad scope means that this is the capability
that is required by many privileged programs.
Don't make the problem worse.
The only new features that should be associated with
.B CAP_SYS_ADMIN
are ones that
.I closely
match existing uses in that silo.
.IP *
If you have determined that it really is necessary to create
a new capability for your feature,
don't make or name it as a "single-use" capability.
Thus, for example, the addition of the highly specific
.BR CAP_SYS_PACCT
was probably a mistake.
Instead, try to identify and name your new capability as a broader
silo into which other related future use cases might fit.
.\"
.SS Thread capability sets
Each thread has the following capability sets containing zero or more
of the above capabilities:
.TP
.IR Permitted
This is a limiting superset for the effective
capabilities that the thread may assume.
It is also a limiting superset for the capabilities that
may be added to the inheritable set by a thread that does not have the
.B CAP_SETPCAP
capability in its effective set.
.IP
If a thread drops a capability from its permitted set,
it can never reacquire that capability (unless it
.BR execve (2)s
either a set-user-ID-root program, or
a program whose associated file capabilities grant that capability).
.TP
.IR Inheritable
This is a set of capabilities preserved across an
.BR execve (2).
Inheritable capabilities remain inheritable when executing any program,
and inheritable capabilities are added to the permitted set when executing
a program that has the corresponding bits set in the file inheritable set.
.IP
Because inheritable capabilities are not generally preserved across
.BR execve (2)
when running as a non-root user, applications that wish to run helper
programs with elevated capabilities should consider using
ambient capabilities, described below.
.TP
.IR Effective
This is the set of capabilities used by the kernel to
perform permission checks for the thread.
.TP
.IR Bounding " (per-thread since Linux 2.6.25)"
The capability bounding set is a mechanism that can be used
to limit the capabilities that are gained during
.BR execve (2).
.IP
Since Linux 2.6.25, this is a per-thread capability set.
In older kernels, the capability bounding set was a system wide attribute
shared by all threads on the system.
.IP
.IP
For more details on the capability bounding set, see below.
.TP
.IR Ambient " (since Linux 4.3)"
.\" commit 58319057b7847667f0c9585b9de0e8932b0fdb08
This is a set of capabilities that are preserved across an
.BR execve (2)
of a program that is not privileged.
The ambient capability set obeys the invariant that no capability
can ever be ambient if it is not both permitted and inheritable.
.IP
The ambient capability set can be directly modified using
.BR prctl (2).
Ambient capabilities are automatically lowered if either of
the corresponding permitted or inheritable capabilities is lowered.
.IP
Executing a program that changes UID or GID due to the
set-user-ID or set-group-ID bits or executing a program that has
any file capabilities set will clear the ambient set.
Ambient capabilities are added to the permitted set and
assigned to the effective set when
.BR execve (2)
is called.
If ambient capabilities cause a process's permitted and effective
capabilities to increase during an
.BR execve (2),
this does not trigger the secure-execution mode described in
.BR ld.so (8).
.PP
A child created via
.BR fork (2)
inherits copies of its parent's capability sets.
See below for a discussion of the treatment of capabilities during
.BR execve (2).
.PP
Using
.BR capset (2),
a thread may manipulate its own capability sets (see below).
.PP
Since Linux 3.2, the file
.I /proc/sys/kernel/cap_last_cap
.\" commit 73efc0394e148d0e15583e13712637831f926720
exposes the numerical value of the highest capability
supported by the running kernel;
this can be used to determine the highest bit
that may be set in a capability set.
.\"
.SS File capabilities
Since kernel 2.6.24, the kernel supports
associating capability sets with an executable file using
.BR setcap (8).
The file capability sets are stored in an extended attribute (see
.BR setxattr (2)
and
.BR xattr (7))
named
.IR "security.capability" .
Writing to this extended attribute requires the
.BR CAP_SETFCAP
capability.
The file capability sets,
in conjunction with the capability sets of the thread,
determine the capabilities of a thread after an
.BR execve (2).
.PP
The three file capability sets are:
.TP
.IR Permitted " (formerly known as " forced ):
These capabilities are automatically permitted to the thread,
regardless of the thread's inheritable capabilities.
.TP
.IR Inheritable " (formerly known as " allowed ):
This set is ANDed with the thread's inheritable set to determine which
inheritable capabilities are enabled in the permitted set of
the thread after the
.BR execve (2).
.TP
.IR Effective :
This is not a set, but rather just a single bit.
If this bit is set, then during an
.BR execve (2)
all of the new permitted capabilities for the thread are
also raised in the effective set.
If this bit is not set, then after an
.BR execve (2),
none of the new permitted capabilities is in the new effective set.
.IP
Enabling the file effective capability bit implies
that any file permitted or inheritable capability that causes a
thread to acquire the corresponding permitted capability during an
.BR execve (2)
(see the transformation rules described below) will also acquire that
capability in its effective set.
Therefore, when assigning capabilities to a file
.RB ( setcap (8),
.BR cap_set_file (3),
.BR cap_set_fd (3)),
if we specify the effective flag as being enabled for any capability,
then the effective flag must also be specified as enabled
for all other capabilities for which the corresponding permitted or
inheritable flags is enabled.
.\"
.SS File capability extended attribute versioning
To allow extensibility,
the kernel supports a scheme to encode a version number inside the
.I security.capability
extended attribute that is used to implement file capabilities.
These version numbers are internal to the implementation,
and not directly visible to user-space applications.
To date, the following versions are supported:
.TP
.BR VFS_CAP_REVISION_1
This was the original file capability implementation,
which supported 32-bit masks for file capabilities.
.TP
.BR VFS_CAP_REVISION_2 " (since Linux 2.6.25)"
.\" commit e338d263a76af78fe8f38a72131188b58fceb591
This version allows for file capability masks that are 64 bits in size,
and was necessary as the number of supported capabilities grew beyond 32.
The kernel transparently continues to support the execution of files
that have 32-bit version 1 capability masks,
but when adding capabilities to files that did not previously
have capabilities, or modifying the capabilities of existing files,
it automatically uses the version 2 scheme
(or possibly the version 3 scheme, as described below).
.TP
.BR VFS_CAP_REVISION_3 " (since Linux 4.14)"
.\" commit 8db6c34f1dbc8e06aa016a9b829b06902c3e1340
Version 3 file capabilities are provided
to support namespaced file capabilities (described below).
.IP
As with version 2 file capabilities,
version 3 capability masks are 64 bits in size.
But in addition, the root user ID of namespace is encoded in the
.I security.capability
extended attribute.
(A namespace's root user ID is the value that user ID 0
inside that namespace maps to in the initial user namespace.)
.IP
Version 3 file capabilities are designed to coexist
with version 2 capabilities;
that is, on a modern Linux system,
there may be some files with version 2 capabilities
while others have version 3 capabilities.
.PP
Before Linux 4.14,
the only kind of file capability extended attribute
that could be attached to a file was a
.B VFS_CAP_REVISION_2
attribute.
Since Linux 4.14,
the version of the
.I security.capability
extended attribute that is attached to a file
depends on the circumstances in which the attribute was created.
.PP
Starting with Linux 4.14, a
.I security.capability
extended attribute is automatically created as (or converted to)
a version 3
.RB ( VFS_CAP_REVISION_3 )
attribute if both of the following are true:
.IP (1) 4
The thread writing the attribute resides in a noninitial user namespace.
(More precisely: the thread resides in a user namespace other
than the one from which the underlying filesystem was mounted.)
.IP (2)
The thread has the
.BR CAP_SETFCAP
capability over the file inode,
meaning that (a) the thread has the
.B CAP_SETFCAP
capability in its own user namespace;
and (b) the UID and GID of the file inode have mappings in
the writer's user namespace.
.PP
When a
.BR VFS_CAP_REVISION_3
.I security.capability
extended attribute is created, the root user ID of the creating thread's
user namespace is saved in the extended attribute.
.PP
By contrast, creating or modifying a
.I security.capability
extended attribute from a privileged
.RB ( CAP_SETFCAP )
thread that resides in the
namespace where the underlying filesystem was mounted
(this normally means the initial user namespace)
automatically results in the creation of a version 2
.RB ( VFS_CAP_REVISION_2 )
attribute.
.PP
Note that the creation of a version 3
.I security.capability
extended attribute is automatic.
That is to say, when a user-space application writes
.RB ( setxattr (2))
a
.I security.capability
attribute in the version 2 format,
the kernel will automatically create a version 3 attribute
if the attribute is created in the circumstances described above.
Correspondingly, when a version 3
.I security.capability
attribute is retrieved
.RB ( getxattr (2))
by a process that resides inside a user namespace that was created by the
root user ID (or a descendant of that user namespace),
the returned attribute is (automatically)
simplified to appear as a version 2 attribute
(i.e., the returned value is the size of a version 2 attribute and does
not include the root user ID).
These automatic translations mean that no changes are required to
user-space tools (e.g.,
.BR setcap (1)
and
.BR getcap (1))
in order for those tools to be used to create and retrieve version 3
.I security.capability
attributes.
.PP
Note that a file can have either a version 2 or a version 3
.I security.capability
extended attribute associated with it, but not both:
creation or modification of the
.I security.capability
extended attribute will automatically modify the version
according to the circumstances in which the extended attribute is
created or modified.
.\"
.SS Transformation of capabilities during execve()
.PP
During an
.BR execve (2),
the kernel calculates the new capabilities of
the process using the following algorithm:
.PP
.in +4n
.EX
P'(ambient)     = (file is privileged) ? 0 : P(ambient)

P'(permitted)   = (P(inheritable) & F(inheritable)) |
                  (F(permitted) & P(bounding)) | P'(ambient)

P'(effective)   = F(effective) ? P'(permitted) : P'(ambient)

P'(inheritable) = P(inheritable)    [i.e., unchanged]

P'(bounding)    = P(bounding)       [i.e., unchanged]
.EE
.in
.PP
where:
.RS 4
.IP P() 6
denotes the value of a thread capability set before the
.BR execve (2)
.IP P'()
denotes the value of a thread capability set after the
.BR execve (2)
.IP F()
denotes a file capability set
.RE
.PP
Note the following details relating to the above capability
transformation rules:
.IP * 3
The ambient capability set is present only since Linux 4.3.
When determining the transformation of the ambient set during
.BR execve (2),
a privileged file is one that has capabilities or
has the set-user-ID or set-group-ID bit set.
.IP *
Prior to Linux 2.6.25,
the bounding set was a system-wide attribute shared by all threads.
That system-wide value was employed to calculate the new permitted set during
.BR execve (2)
in the same manner as shown above for
.IR P(bounding) .
.PP
.IR Note :
during the capability transitions described above,
file capabilities may be ignored (treated as empty) for the same reasons
that the set-user-ID and set-group-ID bits are ignored; see
.BR execve (2).
File capabilities are similarly ignored if the kernel was booted with the
.I no_file_caps
option.
.PP
.IR Note :
according to the rules above,
if a process with nonzero user IDs performs an
.BR execve (2)
then any capabilities that are present in
its permitted and effective sets will be cleared.
For the treatment of capabilities when a process with a
user ID of zero performs an
.BR execve (2),
see below under
.IR "Capabilities and execution of programs by root" .
.\"
.SS Safety checking for capability-dumb binaries
A capability-dumb binary is an application that has been
marked to have file capabilities, but has not been converted to use the
.BR libcap (3)
API to manipulate its capabilities.
(In other words, this is a traditional set-user-ID-root program
that has been switched to use file capabilities,
but whose code has not been modified to understand capabilities.)
For such applications,
the effective capability bit is set on the file,
so that the file permitted capabilities are automatically
enabled in the process effective set when executing the file.
The kernel recognizes a file which has the effective capability bit set
as capability-dumb for the purpose of the check described here.
.PP
When executing a capability-dumb binary,
the kernel checks if the process obtained all permitted capabilities
that were specified in the file permitted set,
after the capability transformations described above have been performed.
(The typical reason why this might
.I not
occur is that the capability bounding set masked out some
of the capabilities in the file permitted set.)
If the process did not obtain the full set of
file permitted capabilities, then
.BR execve (2)
fails with the error
.BR EPERM .
This prevents possible security risks that could arise when
a capability-dumb application is executed with less privilege that it needs.
Note that, by definition,
the application could not itself recognize this problem,
since it does not employ the
.BR libcap (3)
API.
.\"
.SS Capabilities and execution of programs by root
.\" See cap_bprm_set_creds(), bprm_caps_from_vfs_cap() and
.\" handle_privileged_root() in security/commoncap.c (Linux 5.0 source)
In order to mirror traditional UNIX semantics,
the kernel performs special treatment of file capabilities when
a process with UID 0 (root) executes a program and
when a set-user-ID-root program is executed.
.PP
After having performed any changes to the process effective ID that
were triggered by the set-user-ID mode bit of the binary\(eme.g.,
switching the effective user ID to 0 (root) because
a set-user-ID-root program was executed\(emthe
kernel calculates the file capability sets as follows:
.IP 1. 3
If the real or effective user ID of the process is 0 (root),
then the file inheritable and permitted sets are ignored;
instead they are notionally considered to be all ones
(i.e., all capabilities enabled).
(There is one exception to this behavior, described below in
.IR "Set-user-ID-root programs that have file capabilities" .)
.IP 2.
If the effective user ID of the process is 0 (root) or
the file effective bit is in fact enabled,
then the file effective bit is notionally defined to be one (enabled).
.PP
These notional values for the file's capability sets are then used
as described above to calculate the transformation of the process's
capabilities during
.BR execve (2).
.PP
Thus, when a process with nonzero UIDs
.BR execve (2)s
a set-user-ID-root program that does not have capabilities attached,
or when a process whose real and effective UIDs are zero
.BR execve (2)s
a program, the calculation of the process's new
permitted capabilities simplifies to:
.PP
.in +4n
.EX
P'(permitted)   = P(inheritable) | P(bounding)

P'(effective)   = P'(permitted)
.EE
.in
.PP
Consequently, the process gains all capabilities in its permitted and
effective capability sets,
except those masked out by the capability bounding set.
(In the calculation of P'(permitted),
the P'(ambient) term can be simplified away because it is by
definition a proper subset of P(inheritable).)
.PP
The special treatments of user ID 0 (root) described in this subsection
can be disabled using the securebits mechanism described below.
.\"
.\"
.SS Set-user-ID-root programs that have file capabilities
There is one exception to the behavior described under
.IR "Capabilities and execution of programs by root" .
If (a) the binary that is being executed has capabilities attached and
(b) the real user ID of the process is
.I not
0 (root) and
(c) the effective user ID of the process
.I is
0 (root), then the file capability bits are honored
(i.e., they are not notionally considered to be all ones).
The usual way in which this situation can arise is when executing
a set-UID-root program that also has file capabilities.
When such a program is executed,
the process gains just the capabilities granted by the program
(i.e., not all capabilities,
as would occur when executing a set-user-ID-root program
that does not have any associated file capabilities).
.PP
Note that one can assign empty capability sets to a program file,
and thus it is possible to create a set-user-ID-root program that
changes the effective and saved set-user-ID of the process
that executes the program to 0,
but confers no capabilities to that process.
.\"
.SS Capability bounding set
The capability bounding set is a security mechanism that can be used
to limit the capabilities that can be gained during an
.BR execve (2).
The bounding set is used in the following ways:
.IP * 2
During an
.BR execve (2),
the capability bounding set is ANDed with the file permitted
capability set, and the result of this operation is assigned to the
thread's permitted capability set.
The capability bounding set thus places a limit on the permitted
capabilities that may be granted by an executable file.
.IP *
(Since Linux 2.6.25)
The capability bounding set acts as a limiting superset for
the capabilities that a thread can add to its inheritable set using
.BR capset (2).