1capget(2)                     System Calls Manual                    capget(2)
2
3
4

NAME

6       capget, capset - set/get capabilities of thread(s)
7

LIBRARY

9       Standard C library (libc, -lc)
10

SYNOPSIS

12       #include <linux/capability.h> /* Definition of CAP_* and
13                                        _LINUX_CAPABILITY_* constants */
14       #include <sys/syscall.h>      /* Definition of SYS_* constants */
15       #include <unistd.h>
16
17       int syscall(SYS_capget, cap_user_header_t hdrp,
18                   cap_user_data_t datap);
19       int syscall(SYS_capset, cap_user_header_t hdrp,
20                   const cap_user_data_t datap);
21
22       Note:  glibc provides no wrappers for these system calls, necessitating
23       the use of syscall(2).
24

DESCRIPTION

26       These two system calls are the raw kernel  interface  for  getting  and
27       setting  thread capabilities.  Not only are these system calls specific
28       to Linux, but the kernel API is likely to change and use of these  sys‐
29       tem  calls (in particular the format of the cap_user_*_t types) is sub‐
30       ject to extension with each kernel revision, but old programs will keep
31       working.
32
33       The  portable  interfaces  are  cap_set_proc(3) and cap_get_proc(3); if
34       possible, you should use those interfaces in applications; see NOTES.
35
36   Current details
37       Now that you have been warned, some current kernel details.  The struc‐
38       tures are defined as follows.
39
40           #define _LINUX_CAPABILITY_VERSION_1  0x19980330
41           #define _LINUX_CAPABILITY_U32S_1     1
42
43                   /* V2 added in Linux 2.6.25; deprecated */
44           #define _LINUX_CAPABILITY_VERSION_2  0x20071026
45           #define _LINUX_CAPABILITY_U32S_2     2
46
47                   /* V3 added in Linux 2.6.26 */
48           #define _LINUX_CAPABILITY_VERSION_3  0x20080522
49           #define _LINUX_CAPABILITY_U32S_3     2
50
51           typedef struct __user_cap_header_struct {
52              __u32 version;
53              int pid;
54           } *cap_user_header_t;
55
56           typedef struct __user_cap_data_struct {
57              __u32 effective;
58              __u32 permitted;
59              __u32 inheritable;
60           } *cap_user_data_t;
61
62       The  effective,  permitted, and inheritable fields are bit masks of the
63       capabilities defined in capabilities(7).  Note that  the  CAP_*  values
64       are  bit  indexes  and need to be bit-shifted before ORing into the bit
65       fields.  To define the structures for passing to the system  call,  you
66       have   to   use   the   struct   __user_cap_header_struct   and  struct
67       __user_cap_data_struct names because the typedefs are only pointers.
68
69       Kernels prior to Linux 2.6.25 prefer 32-bit capabilities  with  version
70       _LINUX_CAPABILITY_VERSION_1.   Linux  2.6.25  added  64-bit  capability
71       sets, with version _LINUX_CAPABILITY_VERSION_2.  There was, however, an
72       API  glitch,  and Linux 2.6.26 added _LINUX_CAPABILITY_VERSION_3 to fix
73       the problem.
74
75       Note that 64-bit capabilities use datap[0] and datap[1], whereas 32-bit
76       capabilities use only datap[0].
77
78       On  kernels  that support file capabilities (VFS capabilities support),
79       these system calls behave slightly differently.  This support was added
80       as  an  option in Linux 2.6.24, and became fixed (nonoptional) in Linux
81       2.6.33.
82
83       For capget() calls, one can probe the capabilities of  any  process  by
84       specifying its process ID with the hdrp->pid field value.
85
86       For details on the data, see capabilities(7).
87
88   With VFS capabilities support
89       VFS capabilities employ a file extended attribute (see xattr(7)) to al‐
90       low capabilities to be attached to executables.  This  privilege  model
91       obsoletes kernel support for one process asynchronously setting the ca‐
92       pabilities of another.  That is, on kernels that have VFS  capabilities
93       support, when calling capset(), the only permitted values for hdrp->pid
94       are 0 or, equivalently, the value returned by gettid(2).
95
96   Without VFS capabilities support
97       On older kernels that do not provide VFS capabilities support  capset()
98       can,  if  the  caller has the CAP_SETPCAP capability, be used to change
99       not only the caller's own capabilities, but also  the  capabilities  of
100       other  threads.   The  call  operates on the capabilities of the thread
101       specified by the pid field of hdrp when that is nonzero, or on the  ca‐
102       pabilities  of the calling thread if pid is 0.  If pid refers to a sin‐
103       gle-threaded process, then  pid  can  be  specified  as  a  traditional
104       process ID; operating on a thread of a multithreaded process requires a
105       thread ID of the type returned by gettid(2).   For  capset(),  pid  can
106       also  be:  -1,  meaning  perform  the  change on all threads except the
107       caller and init(1); or a value less than -1, in which case  the  change
108       is applied to all members of the process group whose ID is -pid.
109

RETURN VALUE

111       On  success,  zero is returned.  On error, -1 is returned, and errno is
112       set to indicate the error.
113
114       The calls fail with the error EINVAL, and set the version field of hdrp
115       to  the  kernel preferred value of _LINUX_CAPABILITY_VERSION_?  when an
116       unsupported version value is specified.  In this  way,  one  can  probe
117       what the current preferred capability revision is.
118

ERRORS

120       EFAULT Bad  memory  address.  hdrp must not be NULL.  datap may be NULL
121              only when the user is trying to determine the preferred capabil‐
122              ity version format supported by the kernel.
123
124       EINVAL One of the arguments was invalid.
125
126       EPERM  An attempt was made to add a capability to the permitted set, or
127              to set a capability in the effective set that is not in the per‐
128              mitted set.
129
130       EPERM  An  attempt was made to add a capability to the inheritable set,
131              and either:
132
133              •  that capability was not in the caller's bounding set; or
134
135              •  the capability was not in the caller's permitted set and  the
136                 caller  lacked  the  CAP_SETPCAP  capability in its effective
137                 set.
138
139       EPERM  The caller attempted to use capset() to modify the  capabilities
140              of  a thread other than itself, but lacked sufficient privilege.
141              For kernels supporting VFS capabilities, this is  never  permit‐
142              ted.   For kernels lacking VFS support, the CAP_SETPCAP capabil‐
143              ity is required.  (A bug in kernels before  Linux  2.6.11  meant
144              that  this error could also occur if a thread without this capa‐
145              bility tried to change its own capabilities  by  specifying  the
146              pid  field  as a nonzero value (i.e., the value returned by get‐
147              pid(2)) instead of 0.)
148
149       ESRCH  No such thread.
150

STANDARDS

152       Linux.
153

NOTES

155       The portable interface to the capability querying and setting functions
156       is provided by the libcap library and is available here:
157http://git.kernel.org/cgit/linux/kernel/git/morgan/libcap.git
158

SEE ALSO

160       clone(2), gettid(2), capabilities(7)
161
162
163
164Linux man-pages 6.04              2023-03-30                         capget(2)
Impressum