CLONE(2) Linux Programmer's Manual CLONE(2)
NAME
clone - create a child process
SYNOPSIS
#include <sched.h>
int clone(int (*fn)(void *), void *child_stack, int flags, void *arg);
_syscall2(int, clone, int, flags, void *, child_stack)
_syscall5(int, clone, int, flags, void *, child_stack,
int *, parent_tidptr, struct user_desc *, newtls,
int *, child_tidptr)
DESCRIPTION
clone creates a new process, just like fork(2). clone is a library
function layered on top of the underlying clone system call, here-
inafter referred to as sys_clone. A description of sys_clone is given
towards the end of this page.
Unlike fork(2), these calls allow the child process to share parts of
its execution context with the calling process, such as the memory
space, the table of file(1,n) descriptors, and the table of signal(2,7) handlers.
(Note that on this manual page, "calling process" normally corresponds
to "parent process". But see the description of CLONE_PARENT below.)
The main use of clone is to implement threads: multiple threads of con-
trol in(1,8) a program that run concurrently in(1,8) a shared memory space.
When the child process is created with clone, it executes the function
application fn(arg). (This differs from fork(2), where execution con-
tinues in(1,8) the child from the point of the fork(2) call.) The fn argu-
ment is a pointer to a function that is called by the child process at
the beginning of its execution. The arg argument is passed to the fn
function.
When the fn(arg) function application returns, the child process termi-
nates. The integer returned by fn is the exit(3,n,1 builtins) code for the child
process. The child process may also terminate explicitly by calling
exit(3,n,1 builtins)(2) or after receiving a fatal signal.
The child_stack argument specifies the location of the stack used by
the child process. Since the child and calling process may share mem-
ory, it is not possible for the child process to execute in(1,8) the same
stack as the calling process. The calling process must therefore set(7,n,1 builtins)
up memory space for the child stack and pass a pointer to this space to
clone. Stacks grow downwards on all processors that run Linux (except
the HP PA processors), so child_stack usually points to the topmost
address of the memory space set(7,n,1 builtins) up for the child stack.
The low byte of flags contains the number of the signal(2,7) sent to the
parent when the child dies. If this signal(2,7) is specified as anything
other than SIGCHLD, then the parent process must specify the __WALL or
__WCLONE options when waiting for the child with wait(2). If no signal(2,7)
is specified, then the parent process is not signaled when the child
terminates.
flags may also be bitwise-or'ed with one or several of the following
constants, in(1,8) order to specify what is shared between the calling
process and the child process:
CLONE_PARENT (since Linux 2.3.12)
If CLONE_PARENT is set(7,n,1 builtins), then the parent of the new child (as
returned by getppid(2)) will be the same as that of the calling
process.
If CLONE_PARENT is not set(7,n,1 builtins), then (as with fork(2)) the child's
parent is the calling process.
Note that it is the parent process, as returned by getppid(2),
which is signaled when the child terminates, so that if(3,n)
CLONE_PARENT is set(7,n,1 builtins), then the parent of the calling process,
rather than the calling process itself, will be signaled.
CLONE_FS
If CLONE_FS is set(7,n,1 builtins), the caller and the child processes share the
same file(1,n) system information. This includes the root of the
file(1,n) system, the current working directory, and the umask. Any
call to chroot(1,2)(2), chdir(2), or umask(2) performed by the call-
ing process or the child process also takes effect in(1,8) the other
process.
If CLONE_FS is not set(7,n,1 builtins), the child process works on a copy of the
file(1,n) system information of the calling process at the time(1,2,n) of
the clone call. Calls to chroot(1,2)(2), chdir(2), umask(2) per-
formed later by one of the processes do not affect the other
process.
CLONE_FILES
If CLONE_FILES is set(7,n,1 builtins), the calling process and the child pro-
cesses share the same file(1,n) descriptor table. File descriptors
always refer to the same files in(1,8) the calling process and in(1,8) the
child process. Any file(1,n) descriptor created by the calling
process or by the child process is also valid in(1,8) the other
process. Similarly, if(3,n) one of the processes closes a file(1,n)
descriptor, or changes its associated flags, the other process
is also affected.
If CLONE_FILES is not set(7,n,1 builtins), the child process inherits a copy of
all file(1,n) descriptors opened in(1,8) the calling process at the time(1,2,n)
of clone. Operations on file(1,n) descriptors performed later by
either the calling process or the child process do not affect
the other process.
CLONE_NEWNS (since Linux 2.4.19)
Start the child in(1,8) a new namespace.
Every process lives in(1,8) a namespace. The namespace of a process
is the data (the set(7,n,1 builtins) of mounts) describing the file(1,n) hierarchy as
seen by that process. After a fork(2) or clone(2) where the
CLONE_NEWNS flag is not set(7,n,1 builtins), the child lives in(1,8) the same names-
pace as the parent. The system calls mount(2,8)(2) and umount(2)
change the namespace of the calling process, and hence affect
all processes that live in(1,8) the same namespace, but do not affect
processes in(1,8) a different namespace.
After a clone(2) where the CLONE_NEWNS flag is set(7,n,1 builtins), the cloned
child is started in(1,8) a new namespace, initialized with a copy of
the namespace of the parent.
Only a privileged process (one having the CAP_SYS_ADMIN capabil-
ity) may specify the CLONE_NEWNS flag. It is not permitted to
specify both CLONE_NEWNS and CLONE_FS in(1,8) the same clone call.
CLONE_SIGHAND
If CLONE_SIGHAND is set(7,n,1 builtins), the calling process and the child pro-
cesses share the same table of signal(2,7) handlers. If the calling
process or child process calls sigaction(2) to change the behav-
ior associated with a signal(2,7), the behavior is changed in(1,8) the
other process as well. However, the calling process and child
processes still have distinct signal(2,7) masks and sets of pending
signals. So, one of them may block or unblock some signals
using sigprocmask(2) without affecting the other process.
If CLONE_SIGHAND is not set(7,n,1 builtins), the child process inherits a copy
of the signal(2,7) handlers of the calling process at the time(1,2,n) clone
is called. Calls to sigaction(2) performed later by one of the
processes have no effect on the other process.
CLONE_PTRACE
If CLONE_PTRACE is specified, and the calling process is being
traced, then trace(3x,n,3x _nc_tracebits) the child also (see ptrace(2)).
CLONE_VFORK
If CLONE_VFORK is set(7,n,1 builtins), the execution of the calling process is
suspended until the child releases its virtual(5,8) memory resources
via a call to execve(2) or _exit(2) (as with vfork(2)).
If CLONE_VFORK is not set(7,n,1 builtins) then both the calling process and the
child are schedulable after the call, and an application should
not rely on execution occurring in(1,8) any particular order.
CLONE_VM
If CLONE_VM is set(7,n,1 builtins), the calling process and the child processes
run in(1,8) the same memory space. In particular, memory writes per-
formed by the calling process or by the child process are also
visible in(1,8) the other process. Moreover, any memory mapping or
unmapping performed with mmap(2) or munmap(2) by the child or
calling process also affects the other process.
If CLONE_VM is not set(7,n,1 builtins), the child process runs in(1,8) a separate
copy of the memory space of the calling process at the time(1,2,n) of
clone. Memory writes or file(1,n) mappings/unmappings performed by
one of the processes do not affect the other, as with fork(2).
CLONE_PID (obsolete)
If CLONE_PID is set(7,n,1 builtins), the child process is created with the same
process ID as the calling process. This is good for hacking the
system, but otherwise of not much use. Since 2.3.21 this flag
can be specified only by the system boot process (PID 0). It
disappeared in(1,8) Linux 2.5.16.
CLONE_THREAD (since Linux 2.4.0-test8)
If CLONE_THREAD is set(7,n,1 builtins), the child is placed in(1,8) the same thread
group as the calling process.
If CLONE_THREAD is not set(7,n,1 builtins), then the child is placed in(1,8) its own
(new) thread group, whose ID is the same as the process ID.
(Thread groups are feature added in(1,8) Linux 2.4 to support the
POSIX threads notion of a set(7,n,1 builtins) of threads sharing a single PID.
In Linux since 2.4, calls to getpid(2) return the thread group
ID of the caller.)
CLONE_SETTLS (since Linux 2.5.32)
The newtls parameter is the new TLS (Thread Local Storage)
descriptor. (See set_thread_area(2).)
CLONE_PARENT_SETTID (since Linux 2.5.49)
Store child thread ID at location parent_tidptr in(1,8) parent and
child memory. (In Linux 2.5.32-2.5.48 there was a flag
CLONE_SETTID that did this.)
CLONE_CHILD_SETTID (since Linux 2.5.49)
Store child thread ID at location child_tidptr in(1,8) child memory.
CLONE_CHILD_CLEARTID (since Linux 2.5.49)
Erase child thread ID at location child_tidptr in(1,8) child memory
when the child exits, and do a wakeup on the futex(2,4) at that
address. The address involved may be changed by the
set_tid_address(2) system call. This is used by threading
libraries.
sys_clone
The sys_clone system call corresponds more closely to fork(2) in(1,8) that
execution in(1,8) the child continues from the point of the call. Thus,
sys_clone only requires the flags and child_stack arguments, which have
the same meaning as for clone. (Note that the order of these arguments
differs from clone.)
Another difference for sys_clone is that the child_stack argument may
be zero, in(1,8) which case copy-on-write semantics ensure that the child
gets(3,n) separate copies of stack pages when either process modifies the
stack. In this case, for correct operation, the CLONE_VM option should
not be specified.
Since Linux 2.5.49 the system call has five parameters. The two new
parameters are parent_tidptr which points to the location (in(1,8) parent
and child memory) where the parent thread ID will be written in(1,8) case
CLONE_PARENT_SETTID was specified, and child_tidptr which points to the
location (in(1,8) child memory) where the child thread ID will be written in(1,8)
case CLONE_CHILD_SETTID was specified.
RETURN VALUE
On success, the thread ID of the child process is returned in(1,8) the
caller's thread of execution. On failure, a -1 will be returned in(1,8) the
caller's context, no child process will be created, and errno will be
set(7,n,1 builtins) appropriately.
ERRORS
EAGAIN Too many processes are already running.
EINVAL CLONE_SIGHAND was specified, but CLONE_VM was not. (Since Linux
2.6.0-test6.)
EINVAL CLONE_THREAD was specified, but CLONE_SIGHAND was not. (Since
Linux 2.5.35.)
EINVAL Precisely one of CLONE_DETACHED and CLONE_THREAD was specified.
(Since Linux 2.6.0-test6.)
EINVAL Both CLONE_FS and CLONE_NEWNS were specified in(1,8) flags.
EINVAL Returned by clone when a zero value is specified for
child_stack.
ENOMEM Cannot allocate sufficient memory to allocate a task structure
for the child, or to copy those parts of the caller's context
that need to be copied.
EPERM CLONE_NEWNS was specified by a non-root process (process without
CAP_SYS_ADMIN).
EPERM CLONE_PID was specified by a process other than process 0.
AVAILABILITY
There is no entry for clone in(1,8) libc5. glibc2 provides clone as
described in(1,8) this manual page.
NOTES
For kernel versions 2.4.7-2.4.18 the CLONE_THREAD flag implied the
CLONE_PARENT flag.
CONFORMING TO
The clone and sys_clone calls are Linux-specific and should not be used
in(1,8) programs intended to be portable.
SEE ALSO
fork(2), getpid(2), gettid(2), wait(2), capabilities(7)
Linux 2.6 2004-09-10 CLONE(2)