1CAPGET(2) Linux Programmer's Manual CAPGET(2)
2
6 capget, capset - set/get capabilities of thread(s)
7
9 #undef _POSIX_SOURCE
10 #include <sys/capability.h>
11 int capget(cap_user_header_t hdrp, cap_user_data_t datap);
13 int capset(cap_user_header_t hdrp, const cap_user_data_t datap);
15
17 As of Linux 2.2, the power of the superuser (root) has been partitioned
18 into a set of discrete capabilities. Each thread has a set of effec‐
19 tive capabilities identifying which capabilities (if any) it may cur‐
20 rently exercise. Each thread also has a set of inheritable capabili‐
21 ties that may be passed through an execve(2) call, and a set of permit‐
22 ted capabilities that it can make effective or inheritable.
23 These two functions are the raw kernel interface for getting and set‐
25 ting thread capabilities. Not only are these system calls specific to
26 Linux, but the kernel API is likely to change and use of these func‐
27 tions (in particular the format of the cap_user_*_t types) is subject
28 to extension with each kernel revision, but old programs will keep
29 working.
30 The portable interfaces are cap_set_proc(3) and cap_get_proc(3); if
32 possible you should use those interfaces in applications. If you wish
33 to use the Linux extensions in applications, you should use the easier-
34 to-use interfaces capsetp(3) and capgetp(3).
35 Current details
37 Now that you have been warned, some current kernel details. The struc‐
38 tures are defined as follows.
39 #define _LINUX_CAPABILITY_VERSION_1 0x19980330
41 #define _LINUX_CAPABILITY_U32S_1 1
42 #define _LINUX_CAPABILITY_VERSION_2 0x20071026
44 #define _LINUX_CAPABILITY_U32S_2 2
45 typedef struct __user_cap_header_struct {
47 __u32 version;
48 int pid;
49 } *cap_user_header_t;
50 typedef struct __user_cap_data_struct {
52 __u32 effective;
53 __u32 permitted;
54 __u32 inheritable;
55 } *cap_user_data_t;
56 effective, permitted, inheritable are bitmasks of the capabilities
58 defined in capability(7). Note the CAP_* values are bit indexes and
59 need to be bit-shifted before ORing into the bit fields. To define the
60 structures for passing to the system call you have to use the struct
61 __user_cap_header_struct and struct __user_cap_data_struct names
62 because the typedefs are only pointers.
63 Kernels prior to 2.6.25 prefer 32-bit capabilities with version
65 _LINUX_CAPABILITY_VERSION_1, and kernels 2.6.25+ prefer 64-bit capabil‐
66 ities with version _LINUX_CAPABILITY_VERSION_2. Note, 64-bit capabili‐
67 ties use datap[0] and datap[1], whereas 32-bit capabilities only use
68 datap[0].
69 Another change affecting the behavior of these system calls is kernel
71 support for file capabilities (VFS capability support). This support
72 is currently a compile time option (added in kernel 2.6.24).
73 For capget() calls, one can probe the capabilities of any process by
75 specifying its process ID with the hdrp->pid field value.
76 With VFS Capability Support
78 VFS Capability support creates a file-attribute method for adding capa‐
79 bilities to privileged executables. This privilege model obsoletes
80 kernel support for one process asynchronously setting the capabilities
81 of another. That is, with VFS support, for capset() calls the only
82 permitted values for hdrp->pid are 0 or getpid(2), which are equiva‐
83 lent.
84 Without VFS Capability Support
86 When the kernel does not support VFS capabilities, capset() calls can
87 operate on the capabilities of the thread specified by the pid field of
88 hdrp when that is nonzero, or on the capabilities of the calling thread
89 if pid is 0. If pid refers to a single-threaded process, then pid can
90 be specified as a traditional process ID; operating on a thread of a
91 multithreaded process requires a thread ID of the type returned by get‐
92 tid(2). For capset(), pid can also be: -1, meaning perform the change
93 on all threads except the caller and init(8); or a value less than -1,
94 in which case the change is applied to all members of the process group
95 whose ID is -pid.
96 For details on the data, see capabilities(7).
98
100 On success, zero is returned. On error, -1 is returned, and errno is
101 set appropriately.
102 The calls will fail with the error EINVAL, and set the version field of
104 hdrp to the kernel preferred value of _LINUX_CAPABILITY_VERSION_? when
105 an unsupported version value is specified. In this way, one can probe
106 what the current preferred capability revision is.
107
109 EFAULT Bad memory address. hdrp must not be NULL. datap may only be
110 NULL when the user is trying to determine the preferred capabil‐
111 ity version format supported by the kernel.
112 EINVAL One of the arguments was invalid.
114 EPERM An attempt was made to add a capability to the Permitted set, or
116 to set a capability in the Effective or Inheritable sets that is
117 not in the Permitted set.
118 EPERM The caller attempted to use capset() to modify the capabilities
120 of a thread other than itself, but lacked sufficient privilege.
121 For kernels supporting VFS capabilities, this is never permit‐
122 ted. For kernels lacking VFS support, the CAP_SETPCAP capabil‐
123 ity is required. (A bug in kernels before 2.6.11 meant that
124 this error could also occur if a thread without this capability
125 tried to change its own capabilities by specifying the pid field
126 as a nonzero value (i.e., the value returned by getpid(2))
127 instead of 0.)
128 ESRCH No such thread.
130
132 These system calls are Linux-specific.
133
135 The portable interface to the capability querying and setting functions
136 is provided by the libcap library and is available here:
137 http://www.kernel.org/pub/linux/libs/security/linux-privs
138
140 clone(2), gettid(2), capabilities(7)
141
143 This page is part of release 3.25 of the Linux man-pages project. A
144 description of the project, and information about reporting bugs, can
145 be found at http://www.kernel.org/doc/man-pages/.
146Linux 2009-01-26 CAPGET(2)