freebsd-skq/sys/compat/linux/linux_fork.c
dchagin fc55d94b46 Rework signal code to allow using it by other modules, like linprocfs:
1. Linux sigset always 64 bit on all platforms. In order to move Linux
sigset code to the linux_common module define it as 64 bit int. Move
Linux sigset manipulation routines to the MI path.

2. Move Linux signal number definitions to the MI path. In general, they
are the same on all platforms except for a few signals.

3. Map Linux RT signals to the FreeBSD RT signals and hide signal conversion
tables to avoid conversion errors.

4. Emulate Linux SIGPWR signal via FreeBSD SIGRTMIN signal which is outside
of allowed on Linux signal numbers.

PR:		197216
2015-05-24 17:47:20 +00:00

464 lines
11 KiB
C

/*-
* Copyright (c) 2004 Tim J. Robbins
* Copyright (c) 2002 Doug Rabson
* Copyright (c) 2000 Marcel Moolenaar
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer
* in this position and unchanged.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_compat.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/imgact.h>
#include <sys/ktr.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/racct.h>
#include <sys/sched.h>
#include <sys/syscallsubr.h>
#include <sys/sx.h>
#include <sys/unistd.h>
#include <sys/wait.h>
#include <vm/vm.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#ifdef COMPAT_LINUX32
#include <machine/../linux32/linux.h>
#include <machine/../linux32/linux32_proto.h>
#else
#include <machine/../linux/linux.h>
#include <machine/../linux/linux_proto.h>
#endif
#include <compat/linux/linux_emul.h>
#include <compat/linux/linux_futex.h>
#include <compat/linux/linux_misc.h>
#include <compat/linux/linux_util.h>
int
linux_fork(struct thread *td, struct linux_fork_args *args)
{
int error;
struct proc *p2;
struct thread *td2;
#ifdef DEBUG
if (ldebug(fork))
printf(ARGS(fork, ""));
#endif
if ((error = fork1(td, RFFDG | RFPROC | RFSTOPPED, 0, &p2, NULL, 0))
!= 0)
return (error);
td2 = FIRST_THREAD_IN_PROC(p2);
linux_proc_init(td, td2, 0);
td->td_retval[0] = p2->p_pid;
/*
* Make this runnable after we are finished with it.
*/
thread_lock(td2);
TD_SET_CAN_RUN(td2);
sched_add(td2, SRQ_BORING);
thread_unlock(td2);
return (0);
}
int
linux_vfork(struct thread *td, struct linux_vfork_args *args)
{
int error;
struct proc *p2;
struct thread *td2;
#ifdef DEBUG
if (ldebug(vfork))
printf(ARGS(vfork, ""));
#endif
/* Exclude RFPPWAIT */
if ((error = fork1(td, RFFDG | RFPROC | RFMEM | RFSTOPPED, 0, &p2,
NULL, 0)) != 0)
return (error);
td2 = FIRST_THREAD_IN_PROC(p2);
linux_proc_init(td, td2, 0);
PROC_LOCK(p2);
p2->p_flag |= P_PPWAIT;
PROC_UNLOCK(p2);
td->td_retval[0] = p2->p_pid;
/*
* Make this runnable after we are finished with it.
*/
thread_lock(td2);
TD_SET_CAN_RUN(td2);
sched_add(td2, SRQ_BORING);
thread_unlock(td2);
/* wait for the children to exit, ie. emulate vfork */
PROC_LOCK(p2);
while (p2->p_flag & P_PPWAIT)
cv_wait(&p2->p_pwait, &p2->p_mtx);
PROC_UNLOCK(p2);
return (0);
}
static int
linux_clone_proc(struct thread *td, struct linux_clone_args *args)
{
int error, ff = RFPROC | RFSTOPPED;
struct proc *p2;
struct thread *td2;
int exit_signal;
struct linux_emuldata *em;
#ifdef DEBUG
if (ldebug(clone)) {
printf(ARGS(clone, "flags %x, stack %p, parent tid: %p, "
"child tid: %p"), (unsigned)args->flags,
args->stack, args->parent_tidptr, args->child_tidptr);
}
#endif
exit_signal = args->flags & 0x000000ff;
if (LINUX_SIG_VALID(exit_signal)) {
exit_signal = linux_to_bsd_signal(exit_signal);
} else if (exit_signal != 0)
return (EINVAL);
if (args->flags & LINUX_CLONE_VM)
ff |= RFMEM;
if (args->flags & LINUX_CLONE_SIGHAND)
ff |= RFSIGSHARE;
/*
* XXX: In Linux, sharing of fs info (chroot/cwd/umask)
* and open files is independant. In FreeBSD, its in one
* structure but in reality it does not cause any problems
* because both of these flags are usually set together.
*/
if (!(args->flags & (LINUX_CLONE_FILES | LINUX_CLONE_FS)))
ff |= RFFDG;
if (args->flags & LINUX_CLONE_PARENT_SETTID)
if (args->parent_tidptr == NULL)
return (EINVAL);
error = fork1(td, ff, 0, &p2, NULL, 0);
if (error)
return (error);
td2 = FIRST_THREAD_IN_PROC(p2);
/* create the emuldata */
linux_proc_init(td, td2, args->flags);
em = em_find(td2);
KASSERT(em != NULL, ("clone_proc: emuldata not found.\n"));
if (args->flags & LINUX_CLONE_CHILD_SETTID)
em->child_set_tid = args->child_tidptr;
else
em->child_set_tid = NULL;
if (args->flags & LINUX_CLONE_CHILD_CLEARTID)
em->child_clear_tid = args->child_tidptr;
else
em->child_clear_tid = NULL;
if (args->flags & LINUX_CLONE_PARENT_SETTID) {
error = copyout(&p2->p_pid, args->parent_tidptr,
sizeof(p2->p_pid));
if (error)
printf(LMSG("copyout failed!"));
}
PROC_LOCK(p2);
p2->p_sigparent = exit_signal;
PROC_UNLOCK(p2);
/*
* In a case of stack = NULL, we are supposed to COW calling process
* stack. This is what normal fork() does, so we just keep tf_rsp arg
* intact.
*/
linux_set_upcall_kse(td2, PTROUT(args->stack));
if (args->flags & LINUX_CLONE_SETTLS)
linux_set_cloned_tls(td2, args->tls);
#ifdef DEBUG
if (ldebug(clone))
printf(LMSG("clone: successful rfork to %d, "
"stack %p sig = %d"), (int)p2->p_pid, args->stack,
exit_signal);
#endif
if (args->flags & LINUX_CLONE_VFORK) {
PROC_LOCK(p2);
p2->p_flag |= P_PPWAIT;
PROC_UNLOCK(p2);
}
/*
* Make this runnable after we are finished with it.
*/
thread_lock(td2);
TD_SET_CAN_RUN(td2);
sched_add(td2, SRQ_BORING);
thread_unlock(td2);
td->td_retval[0] = p2->p_pid;
if (args->flags & LINUX_CLONE_VFORK) {
/* wait for the children to exit, ie. emulate vfork */
PROC_LOCK(p2);
while (p2->p_flag & P_PPWAIT)
cv_wait(&p2->p_pwait, &p2->p_mtx);
PROC_UNLOCK(p2);
}
return (0);
}
static int
linux_clone_thread(struct thread *td, struct linux_clone_args *args)
{
struct linux_emuldata *em;
struct thread *newtd;
struct proc *p;
int error;
#ifdef DEBUG
if (ldebug(clone)) {
printf(ARGS(clone, "thread: flags %x, stack %p, parent tid: %p, "
"child tid: %p"), (unsigned)args->flags,
args->stack, args->parent_tidptr, args->child_tidptr);
}
#endif
LINUX_CTR4(clone_thread, "thread(%d) flags %x ptid %p ctid %p",
td->td_tid, (unsigned)args->flags,
args->parent_tidptr, args->child_tidptr);
if (args->flags & LINUX_CLONE_PARENT_SETTID)
if (args->parent_tidptr == NULL)
return (EINVAL);
/* Threads should be created with own stack */
if (args->stack == NULL)
return (EINVAL);
p = td->td_proc;
/* Initialize our td */
error = kern_thr_alloc(p, 0, &newtd);
if (error)
return (error);
cpu_set_upcall(newtd, td);
bzero(&newtd->td_startzero,
__rangeof(struct thread, td_startzero, td_endzero));
bcopy(&td->td_startcopy, &newtd->td_startcopy,
__rangeof(struct thread, td_startcopy, td_endcopy));
newtd->td_proc = p;
newtd->td_ucred = crhold(td->td_ucred);
/* create the emuldata */
linux_proc_init(td, newtd, args->flags);
em = em_find(newtd);
KASSERT(em != NULL, ("clone_thread: emuldata not found.\n"));
if (args->flags & LINUX_CLONE_SETTLS)
linux_set_cloned_tls(newtd, args->tls);
if (args->flags & LINUX_CLONE_CHILD_SETTID)
em->child_set_tid = args->child_tidptr;
else
em->child_set_tid = NULL;
if (args->flags & LINUX_CLONE_CHILD_CLEARTID)
em->child_clear_tid = args->child_tidptr;
else
em->child_clear_tid = NULL;
cpu_thread_clean(newtd);
linux_set_upcall_kse(newtd, PTROUT(args->stack));
PROC_LOCK(p);
p->p_flag |= P_HADTHREADS;
bcopy(p->p_comm, newtd->td_name, sizeof(newtd->td_name));
if (args->flags & LINUX_CLONE_PARENT)
thread_link(newtd, p->p_pptr);
else
thread_link(newtd, p);
thread_lock(td);
/* let the scheduler know about these things. */
sched_fork_thread(td, newtd);
thread_unlock(td);
if (P_SHOULDSTOP(p))
newtd->td_flags |= TDF_ASTPENDING | TDF_NEEDSUSPCHK;
PROC_UNLOCK(p);
tidhash_add(newtd);
#ifdef DEBUG
if (ldebug(clone))
printf(ARGS(clone, "successful clone to %d, stack %p"),
(int)newtd->td_tid, args->stack);
#endif
LINUX_CTR2(clone_thread, "thread(%d) successful clone to %d",
td->td_tid, newtd->td_tid);
if (args->flags & LINUX_CLONE_PARENT_SETTID) {
error = copyout(&newtd->td_tid, args->parent_tidptr,
sizeof(newtd->td_tid));
if (error)
printf(LMSG("clone_thread: copyout failed!"));
}
/*
* Make this runnable after we are finished with it.
*/
thread_lock(newtd);
TD_SET_CAN_RUN(newtd);
sched_add(newtd, SRQ_BORING);
thread_unlock(newtd);
td->td_retval[0] = newtd->td_tid;
return (0);
}
int
linux_clone(struct thread *td, struct linux_clone_args *args)
{
if (args->flags & LINUX_CLONE_THREAD)
return (linux_clone_thread(td, args));
else
return (linux_clone_proc(td, args));
}
int
linux_exit(struct thread *td, struct linux_exit_args *args)
{
struct linux_emuldata *em;
em = em_find(td);
KASSERT(em != NULL, ("exit: emuldata not found.\n"));
LINUX_CTR2(exit, "thread(%d) (%d)", em->em_tid, args->rval);
linux_thread_detach(td);
/*
* XXX. When the last two threads of a process
* exit via pthread_exit() try thr_exit() first.
*/
kern_thr_exit(td);
exit1(td, W_EXITCODE(args->rval, 0));
/* NOTREACHED */
}
int
linux_set_tid_address(struct thread *td, struct linux_set_tid_address_args *args)
{
struct linux_emuldata *em;
em = em_find(td);
KASSERT(em != NULL, ("set_tid_address: emuldata not found.\n"));
em->child_clear_tid = args->tidptr;
td->td_retval[0] = em->em_tid;
LINUX_CTR3(set_tid_address, "tidptr(%d) %p, returns %d",
em->em_tid, args->tidptr, td->td_retval[0]);
return (0);
}
void
linux_thread_detach(struct thread *td)
{
struct linux_sys_futex_args cup;
struct linux_emuldata *em;
int *child_clear_tid;
int error;
em = em_find(td);
KASSERT(em != NULL, ("thread_detach: emuldata not found.\n"));
LINUX_CTR1(thread_detach, "thread(%d)", em->em_tid);
release_futexes(td, em);
child_clear_tid = em->child_clear_tid;
if (child_clear_tid != NULL) {
LINUX_CTR2(thread_detach, "thread(%d) %p",
em->em_tid, child_clear_tid);
error = suword32(child_clear_tid, 0);
if (error != 0)
return;
cup.uaddr = child_clear_tid;
cup.op = LINUX_FUTEX_WAKE;
cup.val = 1; /* wake one */
cup.timeout = NULL;
cup.uaddr2 = NULL;
cup.val3 = 0;
error = linux_sys_futex(td, &cup);
/*
* this cannot happen at the moment and if this happens it
* probably means there is a user space bug
*/
if (error != 0)
linux_msg(td, "futex stuff in thread_detach failed.");
}
}