1 /* $NetBSD: kern_fork.c,v 1.196 2016/11/04 18:14:04 christos Exp $ */ 2 3 /*- 4 * Copyright (c) 1999, 2001, 2004, 2006, 2007, 2008 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility, 9 * NASA Ames Research Center, by Charles M. Hannum, and by Andrew Doran. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * Copyright (c) 1982, 1986, 1989, 1991, 1993 35 * The Regents of the University of California. All rights reserved. 36 * (c) UNIX System Laboratories, Inc. 37 * All or some portions of this file are derived from material licensed 38 * to the University of California by American Telephone and Telegraph 39 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 40 * the permission of UNIX System Laboratories, Inc. 41 * 42 * Redistribution and use in source and binary forms, with or without 43 * modification, are permitted provided that the following conditions 44 * are met: 45 * 1. Redistributions of source code must retain the above copyright 46 * notice, this list of conditions and the following disclaimer. 47 * 2. Redistributions in binary form must reproduce the above copyright 48 * notice, this list of conditions and the following disclaimer in the 49 * documentation and/or other materials provided with the distribution. 50 * 3. Neither the name of the University nor the names of its contributors 51 * may be used to endorse or promote products derived from this software 52 * without specific prior written permission. 53 * 54 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 64 * SUCH DAMAGE. 65 * 66 * @(#)kern_fork.c 8.8 (Berkeley) 2/14/95 67 */ 68 69 #include <sys/cdefs.h> 70 __KERNEL_RCSID(0, "$NetBSD: kern_fork.c,v 1.196 2016/11/04 18:14:04 christos Exp $"); 71 72 #include "opt_ktrace.h" 73 #include "opt_dtrace.h" 74 75 #include <sys/param.h> 76 #include <sys/systm.h> 77 #include <sys/filedesc.h> 78 #include <sys/kernel.h> 79 #include <sys/pool.h> 80 #include <sys/mount.h> 81 #include <sys/proc.h> 82 #include <sys/ras.h> 83 #include <sys/resourcevar.h> 84 #include <sys/vnode.h> 85 #include <sys/file.h> 86 #include <sys/acct.h> 87 #include <sys/ktrace.h> 88 #include <sys/sched.h> 89 #include <sys/signalvar.h> 90 #include <sys/kauth.h> 91 #include <sys/atomic.h> 92 #include <sys/syscallargs.h> 93 #include <sys/uidinfo.h> 94 #include <sys/sdt.h> 95 #include <sys/ptrace.h> 96 97 #include <uvm/uvm_extern.h> 98 99 /* 100 * DTrace SDT provider definitions 101 */ 102 SDT_PROVIDER_DECLARE(proc); 103 SDT_PROBE_DEFINE3(proc, kernel, , create, 104 "struct proc *", /* new process */ 105 "struct proc *", /* parent process */ 106 "int" /* flags */); 107 108 u_int nprocs __cacheline_aligned = 1; /* process 0 */ 109 110 /* 111 * Number of ticks to sleep if fork() would fail due to process hitting 112 * limits. Exported in miliseconds to userland via sysctl. 113 */ 114 int forkfsleep = 0; 115 116 int 117 sys_fork(struct lwp *l, const void *v, register_t *retval) 118 { 119 120 return fork1(l, 0, SIGCHLD, NULL, 0, NULL, NULL, retval, NULL); 121 } 122 123 /* 124 * vfork(2) system call compatible with 4.4BSD (i.e. BSD with Mach VM). 125 * Address space is not shared, but parent is blocked until child exit. 126 */ 127 int 128 sys_vfork(struct lwp *l, const void *v, register_t *retval) 129 { 130 131 return fork1(l, FORK_PPWAIT, SIGCHLD, NULL, 0, NULL, NULL, 132 retval, NULL); 133 } 134 135 /* 136 * New vfork(2) system call for NetBSD, which implements original 3BSD vfork(2) 137 * semantics. Address space is shared, and parent is blocked until child exit. 138 */ 139 int 140 sys___vfork14(struct lwp *l, const void *v, register_t *retval) 141 { 142 143 return fork1(l, FORK_PPWAIT|FORK_SHAREVM, SIGCHLD, NULL, 0, 144 NULL, NULL, retval, NULL); 145 } 146 147 /* 148 * Linux-compatible __clone(2) system call. 149 */ 150 int 151 sys___clone(struct lwp *l, const struct sys___clone_args *uap, 152 register_t *retval) 153 { 154 /* { 155 syscallarg(int) flags; 156 syscallarg(void *) stack; 157 } */ 158 int flags, sig; 159 160 /* 161 * We don't support the CLONE_PID or CLONE_PTRACE flags. 162 */ 163 if (SCARG(uap, flags) & (CLONE_PID|CLONE_PTRACE)) 164 return EINVAL; 165 166 /* 167 * Linux enforces CLONE_VM with CLONE_SIGHAND, do same. 168 */ 169 if (SCARG(uap, flags) & CLONE_SIGHAND 170 && (SCARG(uap, flags) & CLONE_VM) == 0) 171 return EINVAL; 172 173 flags = 0; 174 175 if (SCARG(uap, flags) & CLONE_VM) 176 flags |= FORK_SHAREVM; 177 if (SCARG(uap, flags) & CLONE_FS) 178 flags |= FORK_SHARECWD; 179 if (SCARG(uap, flags) & CLONE_FILES) 180 flags |= FORK_SHAREFILES; 181 if (SCARG(uap, flags) & CLONE_SIGHAND) 182 flags |= FORK_SHARESIGS; 183 if (SCARG(uap, flags) & CLONE_VFORK) 184 flags |= FORK_PPWAIT; 185 186 sig = SCARG(uap, flags) & CLONE_CSIGNAL; 187 if (sig < 0 || sig >= _NSIG) 188 return EINVAL; 189 190 /* 191 * Note that the Linux API does not provide a portable way of 192 * specifying the stack area; the caller must know if the stack 193 * grows up or down. So, we pass a stack size of 0, so that the 194 * code that makes this adjustment is a noop. 195 */ 196 return fork1(l, flags, sig, SCARG(uap, stack), 0, 197 NULL, NULL, retval, NULL); 198 } 199 200 /* 201 * Print the 'table full' message once per 10 seconds. 202 */ 203 static struct timeval fork_tfmrate = { 10, 0 }; 204 205 /* 206 * General fork call. Note that another LWP in the process may call exec() 207 * or exit() while we are forking. It's safe to continue here, because 208 * neither operation will complete until all LWPs have exited the process. 209 */ 210 int 211 fork1(struct lwp *l1, int flags, int exitsig, void *stack, size_t stacksize, 212 void (*func)(void *), void *arg, register_t *retval, 213 struct proc **rnewprocp) 214 { 215 struct proc *p1, *p2, *parent; 216 struct plimit *p1_lim; 217 uid_t uid; 218 struct lwp *l2; 219 int count; 220 vaddr_t uaddr; 221 int tnprocs; 222 int tracefork; 223 int error = 0; 224 225 p1 = l1->l_proc; 226 uid = kauth_cred_getuid(l1->l_cred); 227 tnprocs = atomic_inc_uint_nv(&nprocs); 228 229 /* 230 * Although process entries are dynamically created, we still keep 231 * a global limit on the maximum number we will create. 232 */ 233 if (__predict_false(tnprocs >= maxproc)) 234 error = -1; 235 else 236 error = kauth_authorize_process(l1->l_cred, 237 KAUTH_PROCESS_FORK, p1, KAUTH_ARG(tnprocs), NULL, NULL); 238 239 if (error) { 240 static struct timeval lasttfm; 241 atomic_dec_uint(&nprocs); 242 if (ratecheck(&lasttfm, &fork_tfmrate)) 243 tablefull("proc", "increase kern.maxproc or NPROC"); 244 if (forkfsleep) 245 kpause("forkmx", false, forkfsleep, NULL); 246 return EAGAIN; 247 } 248 249 /* 250 * Enforce limits. 251 */ 252 count = chgproccnt(uid, 1); 253 if (__predict_false(count > p1->p_rlimit[RLIMIT_NPROC].rlim_cur)) { 254 if (kauth_authorize_process(l1->l_cred, KAUTH_PROCESS_RLIMIT, 255 p1, KAUTH_ARG(KAUTH_REQ_PROCESS_RLIMIT_BYPASS), 256 &p1->p_rlimit[RLIMIT_NPROC], KAUTH_ARG(RLIMIT_NPROC)) != 0) { 257 (void)chgproccnt(uid, -1); 258 atomic_dec_uint(&nprocs); 259 if (forkfsleep) 260 kpause("forkulim", false, forkfsleep, NULL); 261 return EAGAIN; 262 } 263 } 264 265 /* 266 * Allocate virtual address space for the U-area now, while it 267 * is still easy to abort the fork operation if we're out of 268 * kernel virtual address space. 269 */ 270 uaddr = uvm_uarea_alloc(); 271 if (__predict_false(uaddr == 0)) { 272 (void)chgproccnt(uid, -1); 273 atomic_dec_uint(&nprocs); 274 return ENOMEM; 275 } 276 277 /* 278 * We are now committed to the fork. From here on, we may 279 * block on resources, but resource allocation may NOT fail. 280 */ 281 282 /* Allocate new proc. */ 283 p2 = proc_alloc(); 284 285 /* 286 * Make a proc table entry for the new process. 287 * Start by zeroing the section of proc that is zero-initialized, 288 * then copy the section that is copied directly from the parent. 289 */ 290 memset(&p2->p_startzero, 0, 291 (unsigned) ((char *)&p2->p_endzero - (char *)&p2->p_startzero)); 292 memcpy(&p2->p_startcopy, &p1->p_startcopy, 293 (unsigned) ((char *)&p2->p_endcopy - (char *)&p2->p_startcopy)); 294 295 TAILQ_INIT(&p2->p_sigpend.sp_info); 296 297 LIST_INIT(&p2->p_lwps); 298 LIST_INIT(&p2->p_sigwaiters); 299 300 /* 301 * Duplicate sub-structures as needed. 302 * Increase reference counts on shared objects. 303 * Inherit flags we want to keep. The flags related to SIGCHLD 304 * handling are important in order to keep a consistent behaviour 305 * for the child after the fork. If we are a 32-bit process, the 306 * child will be too. 307 */ 308 p2->p_flag = 309 p1->p_flag & (PK_SUGID | PK_NOCLDWAIT | PK_CLDSIGIGN | PK_32); 310 p2->p_emul = p1->p_emul; 311 p2->p_execsw = p1->p_execsw; 312 313 if (flags & FORK_SYSTEM) { 314 /* 315 * Mark it as a system process. Set P_NOCLDWAIT so that 316 * children are reparented to init(8) when they exit. 317 * init(8) can easily wait them out for us. 318 */ 319 p2->p_flag |= (PK_SYSTEM | PK_NOCLDWAIT); 320 } 321 322 mutex_init(&p2->p_stmutex, MUTEX_DEFAULT, IPL_HIGH); 323 mutex_init(&p2->p_auxlock, MUTEX_DEFAULT, IPL_NONE); 324 rw_init(&p2->p_reflock); 325 cv_init(&p2->p_waitcv, "wait"); 326 cv_init(&p2->p_lwpcv, "lwpwait"); 327 328 /* 329 * Share a lock between the processes if they are to share signal 330 * state: we must synchronize access to it. 331 */ 332 if (flags & FORK_SHARESIGS) { 333 p2->p_lock = p1->p_lock; 334 mutex_obj_hold(p1->p_lock); 335 } else 336 p2->p_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 337 338 kauth_proc_fork(p1, p2); 339 340 p2->p_raslist = NULL; 341 #if defined(__HAVE_RAS) 342 ras_fork(p1, p2); 343 #endif 344 345 /* bump references to the text vnode (for procfs) */ 346 p2->p_textvp = p1->p_textvp; 347 if (p2->p_textvp) 348 vref(p2->p_textvp); 349 350 if (flags & FORK_SHAREFILES) 351 fd_share(p2); 352 else if (flags & FORK_CLEANFILES) 353 p2->p_fd = fd_init(NULL); 354 else 355 p2->p_fd = fd_copy(); 356 357 /* XXX racy */ 358 p2->p_mqueue_cnt = p1->p_mqueue_cnt; 359 360 if (flags & FORK_SHARECWD) 361 cwdshare(p2); 362 else 363 p2->p_cwdi = cwdinit(); 364 365 /* 366 * Note: p_limit (rlimit stuff) is copy-on-write, so normally 367 * we just need increase pl_refcnt. 368 */ 369 p1_lim = p1->p_limit; 370 if (!p1_lim->pl_writeable) { 371 lim_addref(p1_lim); 372 p2->p_limit = p1_lim; 373 } else { 374 p2->p_limit = lim_copy(p1_lim); 375 } 376 377 if (flags & FORK_PPWAIT) { 378 /* Mark ourselves as waiting for a child. */ 379 l1->l_pflag |= LP_VFORKWAIT; 380 p2->p_lflag = PL_PPWAIT; 381 p2->p_vforklwp = l1; 382 } else { 383 p2->p_lflag = 0; 384 } 385 p2->p_sflag = 0; 386 p2->p_slflag = 0; 387 parent = (flags & FORK_NOWAIT) ? initproc : p1; 388 p2->p_pptr = parent; 389 p2->p_ppid = parent->p_pid; 390 LIST_INIT(&p2->p_children); 391 392 p2->p_aio = NULL; 393 394 #ifdef KTRACE 395 /* 396 * Copy traceflag and tracefile if enabled. 397 * If not inherited, these were zeroed above. 398 */ 399 if (p1->p_traceflag & KTRFAC_INHERIT) { 400 mutex_enter(&ktrace_lock); 401 p2->p_traceflag = p1->p_traceflag; 402 if ((p2->p_tracep = p1->p_tracep) != NULL) 403 ktradref(p2); 404 mutex_exit(&ktrace_lock); 405 } 406 #endif 407 408 /* 409 * Create signal actions for the child process. 410 */ 411 p2->p_sigacts = sigactsinit(p1, flags & FORK_SHARESIGS); 412 mutex_enter(p1->p_lock); 413 p2->p_sflag |= 414 (p1->p_sflag & (PS_STOPFORK | PS_STOPEXEC | PS_NOCLDSTOP)); 415 sched_proc_fork(p1, p2); 416 mutex_exit(p1->p_lock); 417 418 p2->p_stflag = p1->p_stflag; 419 420 /* 421 * p_stats. 422 * Copy parts of p_stats, and zero out the rest. 423 */ 424 p2->p_stats = pstatscopy(p1->p_stats); 425 426 /* 427 * Set up the new process address space. 428 */ 429 uvm_proc_fork(p1, p2, (flags & FORK_SHAREVM) ? true : false); 430 431 /* 432 * Finish creating the child process. 433 * It will return through a different path later. 434 */ 435 lwp_create(l1, p2, uaddr, (flags & FORK_PPWAIT) ? LWP_VFORK : 0, 436 stack, stacksize, (func != NULL) ? func : child_return, arg, &l2, 437 l1->l_class); 438 439 /* 440 * Inherit l_private from the parent. 441 * Note that we cannot use lwp_setprivate() here since that 442 * also sets the CPU TLS register, which is incorrect if the 443 * process has changed that without letting the kernel know. 444 */ 445 l2->l_private = l1->l_private; 446 447 /* 448 * If emulation has a process fork hook, call it now. 449 */ 450 if (p2->p_emul->e_proc_fork) 451 (*p2->p_emul->e_proc_fork)(p2, l1, flags); 452 453 /* 454 * ...and finally, any other random fork hooks that subsystems 455 * might have registered. 456 */ 457 doforkhooks(p2, p1); 458 459 SDT_PROBE(proc, kernel, , create, p2, p1, flags, 0, 0); 460 461 /* 462 * It's now safe for the scheduler and other processes to see the 463 * child process. 464 */ 465 mutex_enter(proc_lock); 466 467 if (p1->p_session->s_ttyvp != NULL && p1->p_lflag & PL_CONTROLT) 468 p2->p_lflag |= PL_CONTROLT; 469 470 LIST_INSERT_HEAD(&parent->p_children, p2, p_sibling); 471 p2->p_exitsig = exitsig; /* signal for parent on exit */ 472 473 /* 474 * We don't want to tracefork vfork()ed processes because they 475 * will not receive the SIGTRAP until it is too late. 476 */ 477 tracefork = (p1->p_slflag & (PSL_TRACEFORK|PSL_TRACED)) == 478 (PSL_TRACEFORK|PSL_TRACED) && (flags && FORK_PPWAIT) == 0; 479 if (tracefork) { 480 proc_changeparent(p2, p1->p_pptr); 481 /* 482 * Set ptrace status. 483 */ 484 p1->p_fpid = p2->p_pid; 485 p2->p_fpid = p1->p_pid; 486 } 487 488 LIST_INSERT_AFTER(p1, p2, p_pglist); 489 LIST_INSERT_HEAD(&allproc, p2, p_list); 490 491 p2->p_trace_enabled = trace_is_enabled(p2); 492 #ifdef __HAVE_SYSCALL_INTERN 493 (*p2->p_emul->e_syscall_intern)(p2); 494 #endif 495 496 /* 497 * Update stats now that we know the fork was successful. 498 */ 499 uvmexp.forks++; 500 if (flags & FORK_PPWAIT) 501 uvmexp.forks_ppwait++; 502 if (flags & FORK_SHAREVM) 503 uvmexp.forks_sharevm++; 504 505 /* 506 * Pass a pointer to the new process to the caller. 507 */ 508 if (rnewprocp != NULL) 509 *rnewprocp = p2; 510 511 if (ktrpoint(KTR_EMUL)) 512 p2->p_traceflag |= KTRFAC_TRC_EMUL; 513 514 /* 515 * Notify any interested parties about the new process. 516 */ 517 if (!SLIST_EMPTY(&p1->p_klist)) { 518 mutex_exit(proc_lock); 519 KNOTE(&p1->p_klist, NOTE_FORK | p2->p_pid); 520 mutex_enter(proc_lock); 521 } 522 523 /* 524 * Make child runnable, set start time, and add to run queue except 525 * if the parent requested the child to start in SSTOP state. 526 */ 527 mutex_enter(p2->p_lock); 528 529 /* 530 * Start profiling. 531 */ 532 if ((p2->p_stflag & PST_PROFIL) != 0) { 533 mutex_spin_enter(&p2->p_stmutex); 534 startprofclock(p2); 535 mutex_spin_exit(&p2->p_stmutex); 536 } 537 538 getmicrotime(&p2->p_stats->p_start); 539 p2->p_acflag = AFORK; 540 lwp_lock(l2); 541 KASSERT(p2->p_nrlwps == 1); 542 if (p2->p_sflag & PS_STOPFORK) { 543 struct schedstate_percpu *spc = &l2->l_cpu->ci_schedstate; 544 p2->p_nrlwps = 0; 545 p2->p_stat = SSTOP; 546 p2->p_waited = 0; 547 p1->p_nstopchild++; 548 l2->l_stat = LSSTOP; 549 KASSERT(l2->l_wchan == NULL); 550 lwp_unlock_to(l2, spc->spc_lwplock); 551 } else { 552 p2->p_nrlwps = 1; 553 p2->p_stat = SACTIVE; 554 l2->l_stat = LSRUN; 555 sched_enqueue(l2, false); 556 lwp_unlock(l2); 557 } 558 559 /* 560 * Return child pid to parent process, 561 * marking us as parent via retval[1]. 562 */ 563 if (retval != NULL) { 564 retval[0] = p2->p_pid; 565 retval[1] = 0; 566 } 567 mutex_exit(p2->p_lock); 568 569 /* 570 * Preserve synchronization semantics of vfork. If waiting for 571 * child to exec or exit, sleep until it clears LP_VFORKWAIT. 572 */ 573 #if 0 574 while (l1->l_pflag & LP_VFORKWAIT) { 575 cv_wait(&l1->l_waitcv, proc_lock); 576 } 577 #else 578 while (p2->p_lflag & PL_PPWAIT) 579 cv_wait(&p1->p_waitcv, proc_lock); 580 #endif 581 582 /* 583 * Let the parent know that we are tracing its child. 584 */ 585 if (tracefork) { 586 ksiginfo_t ksi; 587 588 KSI_INIT_EMPTY(&ksi); 589 ksi.ksi_signo = SIGTRAP; 590 ksi.ksi_lid = l1->l_lid; 591 kpsignal(p1, &ksi, NULL); 592 } 593 mutex_exit(proc_lock); 594 595 return 0; 596 } 597