delete_module(2)

1DELETE_MODULE(2)           Linux Programmer's Manual          DELETE_MODULE(2)
2
3
4

NAME

6       delete_module - unload a kernel module
7

SYNOPSIS

9       #include <fcntl.h>            /* Definition of O_* constants */
10       #include <sys/syscall.h>      /* Definition of SYS_* constants */
11       #include <unistd.h>
12
13       int syscall(SYS_delete_module, const char *name, unsigned int flags);
14
15       Note:  glibc provides no wrapper for delete_module(), necessitating the
16       use of syscall(2).
17

DESCRIPTION

19       The delete_module() system call attempts to remove the unused  loadable
20       module  entry  identified by name.  If the module has an exit function,
21       then that function is executed before unloading the module.  The  flags
22       argument  is  used  to  modify  the behavior of the system call, as de‐
23       scribed below.  This system call requires privilege.
24
25       Module removal is attempted according to the following rules:
26
27       1.  If there are other loaded modules that depend on  (i.e.,  refer  to
28           symbols defined in) this module, then the call fails.
29
30       2.  Otherwise,  if the reference count for the module (i.e., the number
31           of processes currently using the module) is zero, then  the  module
32           is immediately unloaded.
33
34       3.  If  a  module  has a nonzero reference count, then the behavior de‐
35           pends on the bits set in flags.  In normal usage (see  NOTES),  the
36           O_NONBLOCK flag is always specified, and the O_TRUNC flag may addi‐
37           tionally be specified.
38
39           The various combinations for flags have the following effect:
40
41           flags == O_NONBLOCK
42                  The call returns immediately, with an error.
43
44           flags == (O_NONBLOCK | O_TRUNC)
45                  The module is unloaded immediately, regardless of whether it
46                  has a nonzero reference count.
47
48           (flags & O_NONBLOCK) == 0
49                  If  flags  does  not specify O_NONBLOCK, the following steps
50                  occur:
51
52                  *  The module is marked so that no new references  are  per‐
53                     mitted.
54
55                  *  If the module's reference count is nonzero, the caller is
56                     placed in an uninterruptible sleep  state  (TASK_UNINTER‐
57                     RUPTIBLE)  until  the  reference  count is zero, at which
58                     point the call unblocks.
59
60                  *  The module is unloaded in the usual way.
61
62       The O_TRUNC flag has one further effect on the rules  described  above.
63       By default, if a module has an init function but no exit function, then
64       an attempt to remove the module fails.  However, if O_TRUNC was  speci‐
65       fied, this requirement is bypassed.
66
67       Using  the O_TRUNC flag is dangerous!  If the kernel was not built with
68       CONFIG_MODULE_FORCE_UNLOAD, this flag is silently ignored.   (Normally,
69       CONFIG_MODULE_FORCE_UNLOAD  is  enabled.)   Using  this flag taints the
70       kernel (TAINT_FORCED_RMMOD).
71

RETURN VALUE

73       On success, zero is returned.  On error, -1 is returned  and  errno  is
74       set to indicate the error.
75

ERRORS

77       EBUSY  The module is not "live" (i.e., it is still being initialized or
78              is already marked for removal); or, the module has an init func‐
79              tion  but has no exit function, and O_TRUNC was not specified in
80              flags.
81
82       EFAULT name refers to a location outside the process's  accessible  ad‐
83              dress space.
84
85       ENOENT No module by that name exists.
86
87       EPERM  The  caller  was not privileged (did not have the CAP_SYS_MODULE
88              capability), or module unloading is disabled (see /proc/sys/ker‐
89              nel/modules_disabled in proc(5)).
90
91       EWOULDBLOCK
92              Other  modules  depend on this module; or, O_NONBLOCK was speci‐
93              fied in flags, but the reference count of this module is nonzero
94              and O_TRUNC was not specified in flags.
95

CONFORMING TO

97       delete_module() is Linux-specific.
98

NOTES

100       The delete_module() system call is not supported by glibc.  No declara‐
101       tion is provided in glibc headers, but, through  a  quirk  of  history,
102       glibc  versions  before  2.23  did  export an ABI for this system call.
103       Therefore, in order to employ this system call,  it  is  (before  glibc
104       2.23) sufficient to manually declare the interface in your code; alter‐
105       natively, you can invoke the system call using syscall(2).
106
107       The uninterruptible sleep that may occur if O_NONBLOCK is omitted  from
108       flags  is  considered undesirable, because the sleeping process is left
109       in an unkillable state.  As at Linux 3.7, specifying O_NONBLOCK is  op‐
110       tional, but in future kernels it is likely to become mandatory.
111
112   Linux 2.4 and earlier
113       In Linux 2.4 and earlier, the system call took only one argument:
114
115          int delete_module(const char *name);
116
117       If name is NULL, all unused modules marked auto-clean are removed.
118
119       Some  further details of differences in the behavior of delete_module()
120       in Linux 2.4 and earlier are not currently  explained  in  this  manual
121       page.
122

COLOPHON

128       This page is part of release 5.13 of the Linux  man-pages  project.   A
129       description  of  the project, information about reporting bugs, and the
130       latest    version    of    this    page,    can     be     found     at
131       https://www.kernel.org/doc/man-pages/.
132
133
134
135Linux                             2021-03-22                  DELETE_MODULE(2)