kern_proc.c revision 1.175
1 /* $NetBSD: kern_proc.c,v 1.175 2011/04/18 00:26:11 rmind Exp $ */ 2 3 /*- 4 * Copyright (c) 1999, 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, 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 * 37 * Redistribution and use in source and binary forms, with or without 38 * modification, are permitted provided that the following conditions 39 * are met: 40 * 1. Redistributions of source code must retain the above copyright 41 * notice, this list of conditions and the following disclaimer. 42 * 2. Redistributions in binary form must reproduce the above copyright 43 * notice, this list of conditions and the following disclaimer in the 44 * documentation and/or other materials provided with the distribution. 45 * 3. Neither the name of the University nor the names of its contributors 46 * may be used to endorse or promote products derived from this software 47 * without specific prior written permission. 48 * 49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 59 * SUCH DAMAGE. 60 * 61 * @(#)kern_proc.c 8.7 (Berkeley) 2/14/95 62 */ 63 64 #include <sys/cdefs.h> 65 __KERNEL_RCSID(0, "$NetBSD: kern_proc.c,v 1.175 2011/04/18 00:26:11 rmind Exp $"); 66 67 #ifdef _KERNEL_OPT 68 #include "opt_kstack.h" 69 #include "opt_maxuprc.h" 70 #include "opt_dtrace.h" 71 #include "opt_compat_netbsd32.h" 72 #endif 73 74 #include <sys/param.h> 75 #include <sys/systm.h> 76 #include <sys/kernel.h> 77 #include <sys/proc.h> 78 #include <sys/resourcevar.h> 79 #include <sys/buf.h> 80 #include <sys/acct.h> 81 #include <sys/wait.h> 82 #include <sys/file.h> 83 #include <ufs/ufs/quota.h> 84 #include <sys/uio.h> 85 #include <sys/pool.h> 86 #include <sys/pset.h> 87 #include <sys/mbuf.h> 88 #include <sys/ioctl.h> 89 #include <sys/tty.h> 90 #include <sys/signalvar.h> 91 #include <sys/ras.h> 92 #include <sys/sa.h> 93 #include <sys/savar.h> 94 #include <sys/filedesc.h> 95 #include "sys/syscall_stats.h" 96 #include <sys/kauth.h> 97 #include <sys/sleepq.h> 98 #include <sys/atomic.h> 99 #include <sys/kmem.h> 100 #include <sys/dtrace_bsd.h> 101 #include <sys/sysctl.h> 102 #include <sys/exec.h> 103 #include <sys/cpu.h> 104 105 #include <uvm/uvm_extern.h> 106 #include <uvm/uvm_extern.h> 107 108 #ifdef COMPAT_NETBSD32 109 #include <compat/netbsd32/netbsd32.h> 110 #endif 111 112 /* 113 * Other process lists 114 */ 115 116 struct proclist allproc; 117 struct proclist zombproc; /* resources have been freed */ 118 119 kmutex_t *proc_lock; 120 121 /* 122 * pid to proc lookup is done by indexing the pid_table array. 123 * Since pid numbers are only allocated when an empty slot 124 * has been found, there is no need to search any lists ever. 125 * (an orphaned pgrp will lock the slot, a session will lock 126 * the pgrp with the same number.) 127 * If the table is too small it is reallocated with twice the 128 * previous size and the entries 'unzipped' into the two halves. 129 * A linked list of free entries is passed through the pt_proc 130 * field of 'free' items - set odd to be an invalid ptr. 131 */ 132 133 struct pid_table { 134 struct proc *pt_proc; 135 struct pgrp *pt_pgrp; 136 pid_t pt_pid; 137 }; 138 #if 1 /* strongly typed cast - should be a noop */ 139 static inline uint p2u(struct proc *p) { return (uint)(uintptr_t)p; } 140 #else 141 #define p2u(p) ((uint)p) 142 #endif 143 #define P_VALID(p) (!(p2u(p) & 1)) 144 #define P_NEXT(p) (p2u(p) >> 1) 145 #define P_FREE(pid) ((struct proc *)(uintptr_t)((pid) << 1 | 1)) 146 147 #define INITIAL_PID_TABLE_SIZE (1 << 5) 148 static struct pid_table *pid_table; 149 static uint pid_tbl_mask = INITIAL_PID_TABLE_SIZE - 1; 150 static uint pid_alloc_lim; /* max we allocate before growing table */ 151 static uint pid_alloc_cnt; /* number of allocated pids */ 152 153 /* links through free slots - never empty! */ 154 static uint next_free_pt, last_free_pt; 155 static pid_t pid_max = PID_MAX; /* largest value we allocate */ 156 157 /* Components of the first process -- never freed. */ 158 159 extern struct emul emul_netbsd; /* defined in kern_exec.c */ 160 161 struct session session0 = { 162 .s_count = 1, 163 .s_sid = 0, 164 }; 165 struct pgrp pgrp0 = { 166 .pg_members = LIST_HEAD_INITIALIZER(&pgrp0.pg_members), 167 .pg_session = &session0, 168 }; 169 filedesc_t filedesc0; 170 struct cwdinfo cwdi0 = { 171 .cwdi_cmask = CMASK, /* see cmask below */ 172 .cwdi_refcnt = 1, 173 }; 174 struct plimit limit0; 175 struct pstats pstat0; 176 struct vmspace vmspace0; 177 struct sigacts sigacts0; 178 struct proc proc0 = { 179 .p_lwps = LIST_HEAD_INITIALIZER(&proc0.p_lwps), 180 .p_sigwaiters = LIST_HEAD_INITIALIZER(&proc0.p_sigwaiters), 181 .p_nlwps = 1, 182 .p_nrlwps = 1, 183 .p_nlwpid = 1, /* must match lwp0.l_lid */ 184 .p_pgrp = &pgrp0, 185 .p_comm = "system", 186 /* 187 * Set P_NOCLDWAIT so that kernel threads are reparented to init(8) 188 * when they exit. init(8) can easily wait them out for us. 189 */ 190 .p_flag = PK_SYSTEM | PK_NOCLDWAIT, 191 .p_stat = SACTIVE, 192 .p_nice = NZERO, 193 .p_emul = &emul_netbsd, 194 .p_cwdi = &cwdi0, 195 .p_limit = &limit0, 196 .p_fd = &filedesc0, 197 .p_vmspace = &vmspace0, 198 .p_stats = &pstat0, 199 .p_sigacts = &sigacts0, 200 }; 201 kauth_cred_t cred0; 202 203 int nofile = NOFILE; 204 int maxuprc = MAXUPRC; 205 int cmask = CMASK; 206 207 MALLOC_DEFINE(M_EMULDATA, "emuldata", "Per-process emulation data"); 208 209 static int sysctl_doeproc(SYSCTLFN_PROTO); 210 static int sysctl_kern_proc_args(SYSCTLFN_PROTO); 211 static void fill_kproc2(struct proc *, struct kinfo_proc2 *, bool); 212 213 /* 214 * The process list descriptors, used during pid allocation and 215 * by sysctl. No locking on this data structure is needed since 216 * it is completely static. 217 */ 218 const struct proclist_desc proclists[] = { 219 { &allproc }, 220 { &zombproc }, 221 { NULL }, 222 }; 223 224 static struct pgrp * pg_remove(pid_t); 225 static void pg_delete(pid_t); 226 static void orphanpg(struct pgrp *); 227 228 static specificdata_domain_t proc_specificdata_domain; 229 230 static pool_cache_t proc_cache; 231 232 static kauth_listener_t proc_listener; 233 234 static int 235 proc_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie, 236 void *arg0, void *arg1, void *arg2, void *arg3) 237 { 238 struct proc *p; 239 int result; 240 241 result = KAUTH_RESULT_DEFER; 242 p = arg0; 243 244 switch (action) { 245 case KAUTH_PROCESS_CANSEE: { 246 enum kauth_process_req req; 247 248 req = (enum kauth_process_req)arg1; 249 250 switch (req) { 251 case KAUTH_REQ_PROCESS_CANSEE_ARGS: 252 case KAUTH_REQ_PROCESS_CANSEE_ENTRY: 253 case KAUTH_REQ_PROCESS_CANSEE_OPENFILES: 254 result = KAUTH_RESULT_ALLOW; 255 256 break; 257 258 case KAUTH_REQ_PROCESS_CANSEE_ENV: 259 if (kauth_cred_getuid(cred) != 260 kauth_cred_getuid(p->p_cred) || 261 kauth_cred_getuid(cred) != 262 kauth_cred_getsvuid(p->p_cred)) 263 break; 264 265 result = KAUTH_RESULT_ALLOW; 266 267 break; 268 269 default: 270 break; 271 } 272 273 break; 274 } 275 276 case KAUTH_PROCESS_FORK: { 277 int lnprocs = (int)(unsigned long)arg2; 278 279 /* 280 * Don't allow a nonprivileged user to use the last few 281 * processes. The variable lnprocs is the current number of 282 * processes, maxproc is the limit. 283 */ 284 if (__predict_false((lnprocs >= maxproc - 5))) 285 break; 286 287 result = KAUTH_RESULT_ALLOW; 288 289 break; 290 } 291 292 case KAUTH_PROCESS_CORENAME: 293 case KAUTH_PROCESS_STOPFLAG: 294 if (proc_uidmatch(cred, p->p_cred) == 0) 295 result = KAUTH_RESULT_ALLOW; 296 297 break; 298 299 default: 300 break; 301 } 302 303 return result; 304 } 305 306 /* 307 * Initialize global process hashing structures. 308 */ 309 void 310 procinit(void) 311 { 312 const struct proclist_desc *pd; 313 u_int i; 314 #define LINK_EMPTY ((PID_MAX + INITIAL_PID_TABLE_SIZE) & ~(INITIAL_PID_TABLE_SIZE - 1)) 315 316 for (pd = proclists; pd->pd_list != NULL; pd++) 317 LIST_INIT(pd->pd_list); 318 319 proc_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 320 pid_table = kmem_alloc(INITIAL_PID_TABLE_SIZE 321 * sizeof(struct pid_table), KM_SLEEP); 322 323 /* Set free list running through table... 324 Preset 'use count' above PID_MAX so we allocate pid 1 next. */ 325 for (i = 0; i <= pid_tbl_mask; i++) { 326 pid_table[i].pt_proc = P_FREE(LINK_EMPTY + i + 1); 327 pid_table[i].pt_pgrp = 0; 328 pid_table[i].pt_pid = 0; 329 } 330 /* slot 0 is just grabbed */ 331 next_free_pt = 1; 332 /* Need to fix last entry. */ 333 last_free_pt = pid_tbl_mask; 334 pid_table[last_free_pt].pt_proc = P_FREE(LINK_EMPTY); 335 /* point at which we grow table - to avoid reusing pids too often */ 336 pid_alloc_lim = pid_tbl_mask - 1; 337 #undef LINK_EMPTY 338 339 proc_specificdata_domain = specificdata_domain_create(); 340 KASSERT(proc_specificdata_domain != NULL); 341 342 proc_cache = pool_cache_init(sizeof(struct proc), 0, 0, 0, 343 "procpl", NULL, IPL_NONE, NULL, NULL, NULL); 344 345 proc_listener = kauth_listen_scope(KAUTH_SCOPE_PROCESS, 346 proc_listener_cb, NULL); 347 } 348 349 void 350 procinit_sysctl(void) 351 { 352 static struct sysctllog *clog; 353 354 sysctl_createv(&clog, 0, NULL, NULL, 355 CTLFLAG_PERMANENT, 356 CTLTYPE_NODE, "kern", NULL, 357 NULL, 0, NULL, 0, 358 CTL_KERN, CTL_EOL); 359 360 sysctl_createv(&clog, 0, NULL, NULL, 361 CTLFLAG_PERMANENT, 362 CTLTYPE_NODE, "proc", 363 SYSCTL_DESCR("System-wide process information"), 364 sysctl_doeproc, 0, NULL, 0, 365 CTL_KERN, KERN_PROC, CTL_EOL); 366 sysctl_createv(&clog, 0, NULL, NULL, 367 CTLFLAG_PERMANENT, 368 CTLTYPE_NODE, "proc2", 369 SYSCTL_DESCR("Machine-independent process information"), 370 sysctl_doeproc, 0, NULL, 0, 371 CTL_KERN, KERN_PROC2, CTL_EOL); 372 sysctl_createv(&clog, 0, NULL, NULL, 373 CTLFLAG_PERMANENT, 374 CTLTYPE_NODE, "proc_args", 375 SYSCTL_DESCR("Process argument information"), 376 sysctl_kern_proc_args, 0, NULL, 0, 377 CTL_KERN, KERN_PROC_ARGS, CTL_EOL); 378 379 /* 380 "nodes" under these: 381 382 KERN_PROC_ALL 383 KERN_PROC_PID pid 384 KERN_PROC_PGRP pgrp 385 KERN_PROC_SESSION sess 386 KERN_PROC_TTY tty 387 KERN_PROC_UID uid 388 KERN_PROC_RUID uid 389 KERN_PROC_GID gid 390 KERN_PROC_RGID gid 391 392 all in all, probably not worth the effort... 393 */ 394 } 395 396 /* 397 * Initialize process 0. 398 */ 399 void 400 proc0_init(void) 401 { 402 struct proc *p; 403 struct pgrp *pg; 404 rlim_t lim; 405 int i; 406 407 p = &proc0; 408 pg = &pgrp0; 409 410 mutex_init(&p->p_stmutex, MUTEX_DEFAULT, IPL_HIGH); 411 mutex_init(&p->p_auxlock, MUTEX_DEFAULT, IPL_NONE); 412 p->p_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 413 414 rw_init(&p->p_reflock); 415 cv_init(&p->p_waitcv, "wait"); 416 cv_init(&p->p_lwpcv, "lwpwait"); 417 418 LIST_INSERT_HEAD(&p->p_lwps, &lwp0, l_sibling); 419 420 pid_table[0].pt_proc = p; 421 LIST_INSERT_HEAD(&allproc, p, p_list); 422 423 pid_table[0].pt_pgrp = pg; 424 LIST_INSERT_HEAD(&pg->pg_members, p, p_pglist); 425 426 #ifdef __HAVE_SYSCALL_INTERN 427 (*p->p_emul->e_syscall_intern)(p); 428 #endif 429 430 /* Create credentials. */ 431 cred0 = kauth_cred_alloc(); 432 p->p_cred = cred0; 433 434 /* Create the CWD info. */ 435 rw_init(&cwdi0.cwdi_lock); 436 437 /* Create the limits structures. */ 438 mutex_init(&limit0.pl_lock, MUTEX_DEFAULT, IPL_NONE); 439 for (i = 0; i < __arraycount(limit0.pl_rlimit); i++) 440 limit0.pl_rlimit[i].rlim_cur = 441 limit0.pl_rlimit[i].rlim_max = RLIM_INFINITY; 442 443 limit0.pl_rlimit[RLIMIT_NOFILE].rlim_max = maxfiles; 444 limit0.pl_rlimit[RLIMIT_NOFILE].rlim_cur = 445 maxfiles < nofile ? maxfiles : nofile; 446 447 limit0.pl_rlimit[RLIMIT_NPROC].rlim_max = maxproc; 448 limit0.pl_rlimit[RLIMIT_NPROC].rlim_cur = 449 maxproc < maxuprc ? maxproc : maxuprc; 450 451 lim = MIN(VM_MAXUSER_ADDRESS, ctob((rlim_t)uvmexp.free)); 452 limit0.pl_rlimit[RLIMIT_RSS].rlim_max = lim; 453 limit0.pl_rlimit[RLIMIT_MEMLOCK].rlim_max = lim; 454 limit0.pl_rlimit[RLIMIT_MEMLOCK].rlim_cur = lim / 3; 455 limit0.pl_corename = defcorename; 456 limit0.pl_refcnt = 1; 457 limit0.pl_sv_limit = NULL; 458 459 /* Configure virtual memory system, set vm rlimits. */ 460 uvm_init_limits(p); 461 462 /* Initialize file descriptor table for proc0. */ 463 fd_init(&filedesc0); 464 465 /* 466 * Initialize proc0's vmspace, which uses the kernel pmap. 467 * All kernel processes (which never have user space mappings) 468 * share proc0's vmspace, and thus, the kernel pmap. 469 */ 470 uvmspace_init(&vmspace0, pmap_kernel(), round_page(VM_MIN_ADDRESS), 471 trunc_page(VM_MAX_ADDRESS)); 472 473 /* Initialize signal state for proc0. XXX IPL_SCHED */ 474 mutex_init(&p->p_sigacts->sa_mutex, MUTEX_DEFAULT, IPL_SCHED); 475 siginit(p); 476 477 proc_initspecific(p); 478 kdtrace_proc_ctor(NULL, p); 479 } 480 481 /* 482 * Session reference counting. 483 */ 484 485 void 486 proc_sesshold(struct session *ss) 487 { 488 489 KASSERT(mutex_owned(proc_lock)); 490 ss->s_count++; 491 } 492 493 void 494 proc_sessrele(struct session *ss) 495 { 496 497 KASSERT(mutex_owned(proc_lock)); 498 /* 499 * We keep the pgrp with the same id as the session in order to 500 * stop a process being given the same pid. Since the pgrp holds 501 * a reference to the session, it must be a 'zombie' pgrp by now. 502 */ 503 if (--ss->s_count == 0) { 504 struct pgrp *pg; 505 506 pg = pg_remove(ss->s_sid); 507 mutex_exit(proc_lock); 508 509 kmem_free(pg, sizeof(struct pgrp)); 510 kmem_free(ss, sizeof(struct session)); 511 } else { 512 mutex_exit(proc_lock); 513 } 514 } 515 516 /* 517 * Check that the specified process group is in the session of the 518 * specified process. 519 * Treats -ve ids as process ids. 520 * Used to validate TIOCSPGRP requests. 521 */ 522 int 523 pgid_in_session(struct proc *p, pid_t pg_id) 524 { 525 struct pgrp *pgrp; 526 struct session *session; 527 int error; 528 529 mutex_enter(proc_lock); 530 if (pg_id < 0) { 531 struct proc *p1 = proc_find(-pg_id); 532 if (p1 == NULL) { 533 error = EINVAL; 534 goto fail; 535 } 536 pgrp = p1->p_pgrp; 537 } else { 538 pgrp = pgrp_find(pg_id); 539 if (pgrp == NULL) { 540 error = EINVAL; 541 goto fail; 542 } 543 } 544 session = pgrp->pg_session; 545 error = (session != p->p_pgrp->pg_session) ? EPERM : 0; 546 fail: 547 mutex_exit(proc_lock); 548 return error; 549 } 550 551 /* 552 * p_inferior: is p an inferior of q? 553 */ 554 static inline bool 555 p_inferior(struct proc *p, struct proc *q) 556 { 557 558 KASSERT(mutex_owned(proc_lock)); 559 560 for (; p != q; p = p->p_pptr) 561 if (p->p_pid == 0) 562 return false; 563 return true; 564 } 565 566 /* 567 * proc_find: locate a process by the ID. 568 * 569 * => Must be called with proc_lock held. 570 */ 571 proc_t * 572 proc_find_raw(pid_t pid) 573 { 574 struct pid_table *pt; 575 proc_t *p; 576 577 KASSERT(mutex_owned(proc_lock)); 578 pt = &pid_table[pid & pid_tbl_mask]; 579 p = pt->pt_proc; 580 if (__predict_false(!P_VALID(p) || pt->pt_pid != pid)) { 581 return NULL; 582 } 583 return p; 584 } 585 586 proc_t * 587 proc_find(pid_t pid) 588 { 589 proc_t *p; 590 591 p = proc_find_raw(pid); 592 if (__predict_false(p == NULL)) { 593 return NULL; 594 } 595 596 /* 597 * Only allow live processes to be found by PID. 598 * XXX: p_stat might change, since unlocked. 599 */ 600 if (__predict_true(p->p_stat == SACTIVE || p->p_stat == SSTOP)) { 601 return p; 602 } 603 return NULL; 604 } 605 606 /* 607 * pgrp_find: locate a process group by the ID. 608 * 609 * => Must be called with proc_lock held. 610 */ 611 struct pgrp * 612 pgrp_find(pid_t pgid) 613 { 614 struct pgrp *pg; 615 616 KASSERT(mutex_owned(proc_lock)); 617 618 pg = pid_table[pgid & pid_tbl_mask].pt_pgrp; 619 620 /* 621 * Cannot look up a process group that only exists because the 622 * session has not died yet (traditional). 623 */ 624 if (pg == NULL || pg->pg_id != pgid || LIST_EMPTY(&pg->pg_members)) { 625 return NULL; 626 } 627 return pg; 628 } 629 630 static void 631 expand_pid_table(void) 632 { 633 size_t pt_size, tsz; 634 struct pid_table *n_pt, *new_pt; 635 struct proc *proc; 636 struct pgrp *pgrp; 637 pid_t pid, rpid; 638 u_int i; 639 uint new_pt_mask; 640 641 pt_size = pid_tbl_mask + 1; 642 tsz = pt_size * 2 * sizeof(struct pid_table); 643 new_pt = kmem_alloc(tsz, KM_SLEEP); 644 new_pt_mask = pt_size * 2 - 1; 645 646 mutex_enter(proc_lock); 647 if (pt_size != pid_tbl_mask + 1) { 648 /* Another process beat us to it... */ 649 mutex_exit(proc_lock); 650 kmem_free(new_pt, tsz); 651 return; 652 } 653 654 /* 655 * Copy entries from old table into new one. 656 * If 'pid' is 'odd' we need to place in the upper half, 657 * even pid's to the lower half. 658 * Free items stay in the low half so we don't have to 659 * fixup the reference to them. 660 * We stuff free items on the front of the freelist 661 * because we can't write to unmodified entries. 662 * Processing the table backwards maintains a semblance 663 * of issuing pid numbers that increase with time. 664 */ 665 i = pt_size - 1; 666 n_pt = new_pt + i; 667 for (; ; i--, n_pt--) { 668 proc = pid_table[i].pt_proc; 669 pgrp = pid_table[i].pt_pgrp; 670 if (!P_VALID(proc)) { 671 /* Up 'use count' so that link is valid */ 672 pid = (P_NEXT(proc) + pt_size) & ~pt_size; 673 rpid = 0; 674 proc = P_FREE(pid); 675 if (pgrp) 676 pid = pgrp->pg_id; 677 } else { 678 pid = pid_table[i].pt_pid; 679 rpid = pid; 680 } 681 682 /* Save entry in appropriate half of table */ 683 n_pt[pid & pt_size].pt_proc = proc; 684 n_pt[pid & pt_size].pt_pgrp = pgrp; 685 n_pt[pid & pt_size].pt_pid = rpid; 686 687 /* Put other piece on start of free list */ 688 pid = (pid ^ pt_size) & ~pid_tbl_mask; 689 n_pt[pid & pt_size].pt_proc = 690 P_FREE((pid & ~pt_size) | next_free_pt); 691 n_pt[pid & pt_size].pt_pgrp = 0; 692 n_pt[pid & pt_size].pt_pid = 0; 693 694 next_free_pt = i | (pid & pt_size); 695 if (i == 0) 696 break; 697 } 698 699 /* Save old table size and switch tables */ 700 tsz = pt_size * sizeof(struct pid_table); 701 n_pt = pid_table; 702 pid_table = new_pt; 703 pid_tbl_mask = new_pt_mask; 704 705 /* 706 * pid_max starts as PID_MAX (= 30000), once we have 16384 707 * allocated pids we need it to be larger! 708 */ 709 if (pid_tbl_mask > PID_MAX) { 710 pid_max = pid_tbl_mask * 2 + 1; 711 pid_alloc_lim |= pid_alloc_lim << 1; 712 } else 713 pid_alloc_lim <<= 1; /* doubles number of free slots... */ 714 715 mutex_exit(proc_lock); 716 kmem_free(n_pt, tsz); 717 } 718 719 struct proc * 720 proc_alloc(void) 721 { 722 struct proc *p; 723 724 p = pool_cache_get(proc_cache, PR_WAITOK); 725 p->p_stat = SIDL; /* protect against others */ 726 proc_initspecific(p); 727 kdtrace_proc_ctor(NULL, p); 728 p->p_pid = -1; 729 proc_alloc_pid(p); 730 return p; 731 } 732 733 pid_t 734 proc_alloc_pid(struct proc *p) 735 { 736 struct pid_table *pt; 737 pid_t pid; 738 int nxt; 739 740 for (;;expand_pid_table()) { 741 if (__predict_false(pid_alloc_cnt >= pid_alloc_lim)) 742 /* ensure pids cycle through 2000+ values */ 743 continue; 744 mutex_enter(proc_lock); 745 pt = &pid_table[next_free_pt]; 746 #ifdef DIAGNOSTIC 747 if (__predict_false(P_VALID(pt->pt_proc) || pt->pt_pgrp)) 748 panic("proc_alloc: slot busy"); 749 #endif 750 nxt = P_NEXT(pt->pt_proc); 751 if (nxt & pid_tbl_mask) 752 break; 753 /* Table full - expand (NB last entry not used....) */ 754 mutex_exit(proc_lock); 755 } 756 757 /* pid is 'saved use count' + 'size' + entry */ 758 pid = (nxt & ~pid_tbl_mask) + pid_tbl_mask + 1 + next_free_pt; 759 if ((uint)pid > (uint)pid_max) 760 pid &= pid_tbl_mask; 761 next_free_pt = nxt & pid_tbl_mask; 762 763 /* Grab table slot */ 764 pt->pt_proc = p; 765 766 KASSERT(pt->pt_pid == 0); 767 pt->pt_pid = pid; 768 if (p->p_pid == -1) { 769 p->p_pid = pid; 770 } 771 pid_alloc_cnt++; 772 mutex_exit(proc_lock); 773 774 return pid; 775 } 776 777 /* 778 * Free a process id - called from proc_free (in kern_exit.c) 779 * 780 * Called with the proc_lock held. 781 */ 782 void 783 proc_free_pid(pid_t pid) 784 { 785 struct pid_table *pt; 786 787 KASSERT(mutex_owned(proc_lock)); 788 789 pt = &pid_table[pid & pid_tbl_mask]; 790 791 /* save pid use count in slot */ 792 pt->pt_proc = P_FREE(pid & ~pid_tbl_mask); 793 KASSERT(pt->pt_pid == pid); 794 pt->pt_pid = 0; 795 796 if (pt->pt_pgrp == NULL) { 797 /* link last freed entry onto ours */ 798 pid &= pid_tbl_mask; 799 pt = &pid_table[last_free_pt]; 800 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pid); 801 pt->pt_pid = 0; 802 last_free_pt = pid; 803 pid_alloc_cnt--; 804 } 805 806 atomic_dec_uint(&nprocs); 807 } 808 809 void 810 proc_free_mem(struct proc *p) 811 { 812 813 kdtrace_proc_dtor(NULL, p); 814 pool_cache_put(proc_cache, p); 815 } 816 817 /* 818 * proc_enterpgrp: move p to a new or existing process group (and session). 819 * 820 * If we are creating a new pgrp, the pgid should equal 821 * the calling process' pid. 822 * If is only valid to enter a process group that is in the session 823 * of the process. 824 * Also mksess should only be set if we are creating a process group 825 * 826 * Only called from sys_setsid and sys_setpgid. 827 */ 828 int 829 proc_enterpgrp(struct proc *curp, pid_t pid, pid_t pgid, bool mksess) 830 { 831 struct pgrp *new_pgrp, *pgrp; 832 struct session *sess; 833 struct proc *p; 834 int rval; 835 pid_t pg_id = NO_PGID; 836 837 sess = mksess ? kmem_alloc(sizeof(*sess), KM_SLEEP) : NULL; 838 839 /* Allocate data areas we might need before doing any validity checks */ 840 mutex_enter(proc_lock); /* Because pid_table might change */ 841 if (pid_table[pgid & pid_tbl_mask].pt_pgrp == 0) { 842 mutex_exit(proc_lock); 843 new_pgrp = kmem_alloc(sizeof(*new_pgrp), KM_SLEEP); 844 mutex_enter(proc_lock); 845 } else 846 new_pgrp = NULL; 847 rval = EPERM; /* most common error (to save typing) */ 848 849 /* Check pgrp exists or can be created */ 850 pgrp = pid_table[pgid & pid_tbl_mask].pt_pgrp; 851 if (pgrp != NULL && pgrp->pg_id != pgid) 852 goto done; 853 854 /* Can only set another process under restricted circumstances. */ 855 if (pid != curp->p_pid) { 856 /* Must exist and be one of our children... */ 857 p = proc_find(pid); 858 if (p == NULL || !p_inferior(p, curp)) { 859 rval = ESRCH; 860 goto done; 861 } 862 /* ... in the same session... */ 863 if (sess != NULL || p->p_session != curp->p_session) 864 goto done; 865 /* ... existing pgid must be in same session ... */ 866 if (pgrp != NULL && pgrp->pg_session != p->p_session) 867 goto done; 868 /* ... and not done an exec. */ 869 if (p->p_flag & PK_EXEC) { 870 rval = EACCES; 871 goto done; 872 } 873 } else { 874 /* ... setsid() cannot re-enter a pgrp */ 875 if (mksess && (curp->p_pgid == curp->p_pid || 876 pgrp_find(curp->p_pid))) 877 goto done; 878 p = curp; 879 } 880 881 /* Changing the process group/session of a session 882 leader is definitely off limits. */ 883 if (SESS_LEADER(p)) { 884 if (sess == NULL && p->p_pgrp == pgrp) 885 /* unless it's a definite noop */ 886 rval = 0; 887 goto done; 888 } 889 890 /* Can only create a process group with id of process */ 891 if (pgrp == NULL && pgid != pid) 892 goto done; 893 894 /* Can only create a session if creating pgrp */ 895 if (sess != NULL && pgrp != NULL) 896 goto done; 897 898 /* Check we allocated memory for a pgrp... */ 899 if (pgrp == NULL && new_pgrp == NULL) 900 goto done; 901 902 /* Don't attach to 'zombie' pgrp */ 903 if (pgrp != NULL && LIST_EMPTY(&pgrp->pg_members)) 904 goto done; 905 906 /* Expect to succeed now */ 907 rval = 0; 908 909 if (pgrp == p->p_pgrp) 910 /* nothing to do */ 911 goto done; 912 913 /* Ok all setup, link up required structures */ 914 915 if (pgrp == NULL) { 916 pgrp = new_pgrp; 917 new_pgrp = NULL; 918 if (sess != NULL) { 919 sess->s_sid = p->p_pid; 920 sess->s_leader = p; 921 sess->s_count = 1; 922 sess->s_ttyvp = NULL; 923 sess->s_ttyp = NULL; 924 sess->s_flags = p->p_session->s_flags & ~S_LOGIN_SET; 925 memcpy(sess->s_login, p->p_session->s_login, 926 sizeof(sess->s_login)); 927 p->p_lflag &= ~PL_CONTROLT; 928 } else { 929 sess = p->p_pgrp->pg_session; 930 proc_sesshold(sess); 931 } 932 pgrp->pg_session = sess; 933 sess = NULL; 934 935 pgrp->pg_id = pgid; 936 LIST_INIT(&pgrp->pg_members); 937 #ifdef DIAGNOSTIC 938 if (__predict_false(pid_table[pgid & pid_tbl_mask].pt_pgrp)) 939 panic("enterpgrp: pgrp table slot in use"); 940 if (__predict_false(mksess && p != curp)) 941 panic("enterpgrp: mksession and p != curproc"); 942 #endif 943 pid_table[pgid & pid_tbl_mask].pt_pgrp = pgrp; 944 pgrp->pg_jobc = 0; 945 } 946 947 /* 948 * Adjust eligibility of affected pgrps to participate in job control. 949 * Increment eligibility counts before decrementing, otherwise we 950 * could reach 0 spuriously during the first call. 951 */ 952 fixjobc(p, pgrp, 1); 953 fixjobc(p, p->p_pgrp, 0); 954 955 /* Interlock with ttread(). */ 956 mutex_spin_enter(&tty_lock); 957 958 /* Move process to requested group. */ 959 LIST_REMOVE(p, p_pglist); 960 if (LIST_EMPTY(&p->p_pgrp->pg_members)) 961 /* defer delete until we've dumped the lock */ 962 pg_id = p->p_pgrp->pg_id; 963 p->p_pgrp = pgrp; 964 LIST_INSERT_HEAD(&pgrp->pg_members, p, p_pglist); 965 966 /* Done with the swap; we can release the tty mutex. */ 967 mutex_spin_exit(&tty_lock); 968 969 done: 970 if (pg_id != NO_PGID) { 971 /* Releases proc_lock. */ 972 pg_delete(pg_id); 973 } else { 974 mutex_exit(proc_lock); 975 } 976 if (sess != NULL) 977 kmem_free(sess, sizeof(*sess)); 978 if (new_pgrp != NULL) 979 kmem_free(new_pgrp, sizeof(*new_pgrp)); 980 #ifdef DEBUG_PGRP 981 if (__predict_false(rval)) 982 printf("enterpgrp(%d,%d,%d), curproc %d, rval %d\n", 983 pid, pgid, mksess, curp->p_pid, rval); 984 #endif 985 return rval; 986 } 987 988 /* 989 * proc_leavepgrp: remove a process from its process group. 990 * => must be called with the proc_lock held, which will be released; 991 */ 992 void 993 proc_leavepgrp(struct proc *p) 994 { 995 struct pgrp *pgrp; 996 997 KASSERT(mutex_owned(proc_lock)); 998 999 /* Interlock with ttread() */ 1000 mutex_spin_enter(&tty_lock); 1001 pgrp = p->p_pgrp; 1002 LIST_REMOVE(p, p_pglist); 1003 p->p_pgrp = NULL; 1004 mutex_spin_exit(&tty_lock); 1005 1006 if (LIST_EMPTY(&pgrp->pg_members)) { 1007 /* Releases proc_lock. */ 1008 pg_delete(pgrp->pg_id); 1009 } else { 1010 mutex_exit(proc_lock); 1011 } 1012 } 1013 1014 /* 1015 * pg_remove: remove a process group from the table. 1016 * => must be called with the proc_lock held; 1017 * => returns process group to free; 1018 */ 1019 static struct pgrp * 1020 pg_remove(pid_t pg_id) 1021 { 1022 struct pgrp *pgrp; 1023 struct pid_table *pt; 1024 1025 KASSERT(mutex_owned(proc_lock)); 1026 1027 pt = &pid_table[pg_id & pid_tbl_mask]; 1028 pgrp = pt->pt_pgrp; 1029 1030 KASSERT(pgrp != NULL); 1031 KASSERT(pgrp->pg_id == pg_id); 1032 KASSERT(LIST_EMPTY(&pgrp->pg_members)); 1033 1034 pt->pt_pgrp = NULL; 1035 1036 if (!P_VALID(pt->pt_proc)) { 1037 /* Orphaned pgrp, put slot onto free list. */ 1038 KASSERT((P_NEXT(pt->pt_proc) & pid_tbl_mask) == 0); 1039 pg_id &= pid_tbl_mask; 1040 pt = &pid_table[last_free_pt]; 1041 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pg_id); 1042 KASSERT(pt->pt_pid == 0); 1043 last_free_pt = pg_id; 1044 pid_alloc_cnt--; 1045 } 1046 return pgrp; 1047 } 1048 1049 /* 1050 * pg_delete: delete and free a process group. 1051 * => must be called with the proc_lock held, which will be released. 1052 */ 1053 static void 1054 pg_delete(pid_t pg_id) 1055 { 1056 struct pgrp *pg; 1057 struct tty *ttyp; 1058 struct session *ss; 1059 1060 KASSERT(mutex_owned(proc_lock)); 1061 1062 pg = pid_table[pg_id & pid_tbl_mask].pt_pgrp; 1063 if (pg == NULL || pg->pg_id != pg_id || !LIST_EMPTY(&pg->pg_members)) { 1064 mutex_exit(proc_lock); 1065 return; 1066 } 1067 1068 ss = pg->pg_session; 1069 1070 /* Remove reference (if any) from tty to this process group */ 1071 mutex_spin_enter(&tty_lock); 1072 ttyp = ss->s_ttyp; 1073 if (ttyp != NULL && ttyp->t_pgrp == pg) { 1074 ttyp->t_pgrp = NULL; 1075 KASSERT(ttyp->t_session == ss); 1076 } 1077 mutex_spin_exit(&tty_lock); 1078 1079 /* 1080 * The leading process group in a session is freed by proc_sessrele(), 1081 * if last reference. Note: proc_sessrele() releases proc_lock. 1082 */ 1083 pg = (ss->s_sid != pg->pg_id) ? pg_remove(pg_id) : NULL; 1084 proc_sessrele(ss); 1085 1086 if (pg != NULL) { 1087 /* Free it, if was not done by proc_sessrele(). */ 1088 kmem_free(pg, sizeof(struct pgrp)); 1089 } 1090 } 1091 1092 /* 1093 * Adjust pgrp jobc counters when specified process changes process group. 1094 * We count the number of processes in each process group that "qualify" 1095 * the group for terminal job control (those with a parent in a different 1096 * process group of the same session). If that count reaches zero, the 1097 * process group becomes orphaned. Check both the specified process' 1098 * process group and that of its children. 1099 * entering == 0 => p is leaving specified group. 1100 * entering == 1 => p is entering specified group. 1101 * 1102 * Call with proc_lock held. 1103 */ 1104 void 1105 fixjobc(struct proc *p, struct pgrp *pgrp, int entering) 1106 { 1107 struct pgrp *hispgrp; 1108 struct session *mysession = pgrp->pg_session; 1109 struct proc *child; 1110 1111 KASSERT(mutex_owned(proc_lock)); 1112 1113 /* 1114 * Check p's parent to see whether p qualifies its own process 1115 * group; if so, adjust count for p's process group. 1116 */ 1117 hispgrp = p->p_pptr->p_pgrp; 1118 if (hispgrp != pgrp && hispgrp->pg_session == mysession) { 1119 if (entering) { 1120 pgrp->pg_jobc++; 1121 p->p_lflag &= ~PL_ORPHANPG; 1122 } else if (--pgrp->pg_jobc == 0) 1123 orphanpg(pgrp); 1124 } 1125 1126 /* 1127 * Check this process' children to see whether they qualify 1128 * their process groups; if so, adjust counts for children's 1129 * process groups. 1130 */ 1131 LIST_FOREACH(child, &p->p_children, p_sibling) { 1132 hispgrp = child->p_pgrp; 1133 if (hispgrp != pgrp && hispgrp->pg_session == mysession && 1134 !P_ZOMBIE(child)) { 1135 if (entering) { 1136 child->p_lflag &= ~PL_ORPHANPG; 1137 hispgrp->pg_jobc++; 1138 } else if (--hispgrp->pg_jobc == 0) 1139 orphanpg(hispgrp); 1140 } 1141 } 1142 } 1143 1144 /* 1145 * A process group has become orphaned; 1146 * if there are any stopped processes in the group, 1147 * hang-up all process in that group. 1148 * 1149 * Call with proc_lock held. 1150 */ 1151 static void 1152 orphanpg(struct pgrp *pg) 1153 { 1154 struct proc *p; 1155 1156 KASSERT(mutex_owned(proc_lock)); 1157 1158 LIST_FOREACH(p, &pg->pg_members, p_pglist) { 1159 if (p->p_stat == SSTOP) { 1160 p->p_lflag |= PL_ORPHANPG; 1161 psignal(p, SIGHUP); 1162 psignal(p, SIGCONT); 1163 } 1164 } 1165 } 1166 1167 #ifdef DDB 1168 #include <ddb/db_output.h> 1169 void pidtbl_dump(void); 1170 void 1171 pidtbl_dump(void) 1172 { 1173 struct pid_table *pt; 1174 struct proc *p; 1175 struct pgrp *pgrp; 1176 int id; 1177 1178 db_printf("pid table %p size %x, next %x, last %x\n", 1179 pid_table, pid_tbl_mask+1, 1180 next_free_pt, last_free_pt); 1181 for (pt = pid_table, id = 0; id <= pid_tbl_mask; id++, pt++) { 1182 p = pt->pt_proc; 1183 if (!P_VALID(p) && !pt->pt_pgrp) 1184 continue; 1185 db_printf(" id %x: ", id); 1186 if (P_VALID(p)) 1187 db_printf("slotpid %d proc %p id %d (0x%x) %s\n", 1188 pt->pt_pid, p, p->p_pid, p->p_pid, p->p_comm); 1189 else 1190 db_printf("next %x use %x\n", 1191 P_NEXT(p) & pid_tbl_mask, 1192 P_NEXT(p) & ~pid_tbl_mask); 1193 if ((pgrp = pt->pt_pgrp)) { 1194 db_printf("\tsession %p, sid %d, count %d, login %s\n", 1195 pgrp->pg_session, pgrp->pg_session->s_sid, 1196 pgrp->pg_session->s_count, 1197 pgrp->pg_session->s_login); 1198 db_printf("\tpgrp %p, pg_id %d, pg_jobc %d, members %p\n", 1199 pgrp, pgrp->pg_id, pgrp->pg_jobc, 1200 LIST_FIRST(&pgrp->pg_members)); 1201 LIST_FOREACH(p, &pgrp->pg_members, p_pglist) { 1202 db_printf("\t\tpid %d addr %p pgrp %p %s\n", 1203 p->p_pid, p, p->p_pgrp, p->p_comm); 1204 } 1205 } 1206 } 1207 } 1208 #endif /* DDB */ 1209 1210 #ifdef KSTACK_CHECK_MAGIC 1211 1212 #define KSTACK_MAGIC 0xdeadbeaf 1213 1214 /* XXX should be per process basis? */ 1215 static int kstackleftmin = KSTACK_SIZE; 1216 static int kstackleftthres = KSTACK_SIZE / 8; 1217 1218 void 1219 kstack_setup_magic(const struct lwp *l) 1220 { 1221 uint32_t *ip; 1222 uint32_t const *end; 1223 1224 KASSERT(l != NULL); 1225 KASSERT(l != &lwp0); 1226 1227 /* 1228 * fill all the stack with magic number 1229 * so that later modification on it can be detected. 1230 */ 1231 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1232 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1233 for (; ip < end; ip++) { 1234 *ip = KSTACK_MAGIC; 1235 } 1236 } 1237 1238 void 1239 kstack_check_magic(const struct lwp *l) 1240 { 1241 uint32_t const *ip, *end; 1242 int stackleft; 1243 1244 KASSERT(l != NULL); 1245 1246 /* don't check proc0 */ /*XXX*/ 1247 if (l == &lwp0) 1248 return; 1249 1250 #ifdef __MACHINE_STACK_GROWS_UP 1251 /* stack grows upwards (eg. hppa) */ 1252 ip = (uint32_t *)((void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1253 end = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1254 for (ip--; ip >= end; ip--) 1255 if (*ip != KSTACK_MAGIC) 1256 break; 1257 1258 stackleft = (void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE - (void *)ip; 1259 #else /* __MACHINE_STACK_GROWS_UP */ 1260 /* stack grows downwards (eg. i386) */ 1261 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1262 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1263 for (; ip < end; ip++) 1264 if (*ip != KSTACK_MAGIC) 1265 break; 1266 1267 stackleft = ((const char *)ip) - (const char *)KSTACK_LOWEST_ADDR(l); 1268 #endif /* __MACHINE_STACK_GROWS_UP */ 1269 1270 if (kstackleftmin > stackleft) { 1271 kstackleftmin = stackleft; 1272 if (stackleft < kstackleftthres) 1273 printf("warning: kernel stack left %d bytes" 1274 "(pid %u:lid %u)\n", stackleft, 1275 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1276 } 1277 1278 if (stackleft <= 0) { 1279 panic("magic on the top of kernel stack changed for " 1280 "pid %u, lid %u: maybe kernel stack overflow", 1281 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1282 } 1283 } 1284 #endif /* KSTACK_CHECK_MAGIC */ 1285 1286 int 1287 proclist_foreach_call(struct proclist *list, 1288 int (*callback)(struct proc *, void *arg), void *arg) 1289 { 1290 struct proc marker; 1291 struct proc *p; 1292 int ret = 0; 1293 1294 marker.p_flag = PK_MARKER; 1295 mutex_enter(proc_lock); 1296 for (p = LIST_FIRST(list); ret == 0 && p != NULL;) { 1297 if (p->p_flag & PK_MARKER) { 1298 p = LIST_NEXT(p, p_list); 1299 continue; 1300 } 1301 LIST_INSERT_AFTER(p, &marker, p_list); 1302 ret = (*callback)(p, arg); 1303 KASSERT(mutex_owned(proc_lock)); 1304 p = LIST_NEXT(&marker, p_list); 1305 LIST_REMOVE(&marker, p_list); 1306 } 1307 mutex_exit(proc_lock); 1308 1309 return ret; 1310 } 1311 1312 int 1313 proc_vmspace_getref(struct proc *p, struct vmspace **vm) 1314 { 1315 1316 /* XXXCDC: how should locking work here? */ 1317 1318 /* curproc exception is for coredump. */ 1319 1320 if ((p != curproc && (p->p_sflag & PS_WEXIT) != 0) || 1321 (p->p_vmspace->vm_refcnt < 1)) { /* XXX */ 1322 return EFAULT; 1323 } 1324 1325 uvmspace_addref(p->p_vmspace); 1326 *vm = p->p_vmspace; 1327 1328 return 0; 1329 } 1330 1331 /* 1332 * Acquire a write lock on the process credential. 1333 */ 1334 void 1335 proc_crmod_enter(void) 1336 { 1337 struct lwp *l = curlwp; 1338 struct proc *p = l->l_proc; 1339 struct plimit *lim; 1340 kauth_cred_t oc; 1341 char *cn; 1342 1343 /* Reset what needs to be reset in plimit. */ 1344 if (p->p_limit->pl_corename != defcorename) { 1345 lim_privatise(p, false); 1346 lim = p->p_limit; 1347 mutex_enter(&lim->pl_lock); 1348 cn = lim->pl_corename; 1349 lim->pl_corename = defcorename; 1350 mutex_exit(&lim->pl_lock); 1351 if (cn != defcorename) 1352 free(cn, M_TEMP); 1353 } 1354 1355 mutex_enter(p->p_lock); 1356 1357 /* Ensure the LWP cached credentials are up to date. */ 1358 if ((oc = l->l_cred) != p->p_cred) { 1359 kauth_cred_hold(p->p_cred); 1360 l->l_cred = p->p_cred; 1361 kauth_cred_free(oc); 1362 } 1363 1364 } 1365 1366 /* 1367 * Set in a new process credential, and drop the write lock. The credential 1368 * must have a reference already. Optionally, free a no-longer required 1369 * credential. The scheduler also needs to inspect p_cred, so we also 1370 * briefly acquire the sched state mutex. 1371 */ 1372 void 1373 proc_crmod_leave(kauth_cred_t scred, kauth_cred_t fcred, bool sugid) 1374 { 1375 struct lwp *l = curlwp, *l2; 1376 struct proc *p = l->l_proc; 1377 kauth_cred_t oc; 1378 1379 KASSERT(mutex_owned(p->p_lock)); 1380 1381 /* Is there a new credential to set in? */ 1382 if (scred != NULL) { 1383 p->p_cred = scred; 1384 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 1385 if (l2 != l) 1386 l2->l_prflag |= LPR_CRMOD; 1387 } 1388 1389 /* Ensure the LWP cached credentials are up to date. */ 1390 if ((oc = l->l_cred) != scred) { 1391 kauth_cred_hold(scred); 1392 l->l_cred = scred; 1393 } 1394 } else 1395 oc = NULL; /* XXXgcc */ 1396 1397 if (sugid) { 1398 /* 1399 * Mark process as having changed credentials, stops 1400 * tracing etc. 1401 */ 1402 p->p_flag |= PK_SUGID; 1403 } 1404 1405 mutex_exit(p->p_lock); 1406 1407 /* If there is a credential to be released, free it now. */ 1408 if (fcred != NULL) { 1409 KASSERT(scred != NULL); 1410 kauth_cred_free(fcred); 1411 if (oc != scred) 1412 kauth_cred_free(oc); 1413 } 1414 } 1415 1416 /* 1417 * proc_specific_key_create -- 1418 * Create a key for subsystem proc-specific data. 1419 */ 1420 int 1421 proc_specific_key_create(specificdata_key_t *keyp, specificdata_dtor_t dtor) 1422 { 1423 1424 return (specificdata_key_create(proc_specificdata_domain, keyp, dtor)); 1425 } 1426 1427 /* 1428 * proc_specific_key_delete -- 1429 * Delete a key for subsystem proc-specific data. 1430 */ 1431 void 1432 proc_specific_key_delete(specificdata_key_t key) 1433 { 1434 1435 specificdata_key_delete(proc_specificdata_domain, key); 1436 } 1437 1438 /* 1439 * proc_initspecific -- 1440 * Initialize a proc's specificdata container. 1441 */ 1442 void 1443 proc_initspecific(struct proc *p) 1444 { 1445 int error; 1446 1447 error = specificdata_init(proc_specificdata_domain, &p->p_specdataref); 1448 KASSERT(error == 0); 1449 } 1450 1451 /* 1452 * proc_finispecific -- 1453 * Finalize a proc's specificdata container. 1454 */ 1455 void 1456 proc_finispecific(struct proc *p) 1457 { 1458 1459 specificdata_fini(proc_specificdata_domain, &p->p_specdataref); 1460 } 1461 1462 /* 1463 * proc_getspecific -- 1464 * Return proc-specific data corresponding to the specified key. 1465 */ 1466 void * 1467 proc_getspecific(struct proc *p, specificdata_key_t key) 1468 { 1469 1470 return (specificdata_getspecific(proc_specificdata_domain, 1471 &p->p_specdataref, key)); 1472 } 1473 1474 /* 1475 * proc_setspecific -- 1476 * Set proc-specific data corresponding to the specified key. 1477 */ 1478 void 1479 proc_setspecific(struct proc *p, specificdata_key_t key, void *data) 1480 { 1481 1482 specificdata_setspecific(proc_specificdata_domain, 1483 &p->p_specdataref, key, data); 1484 } 1485 1486 int 1487 proc_uidmatch(kauth_cred_t cred, kauth_cred_t target) 1488 { 1489 int r = 0; 1490 1491 if (kauth_cred_getuid(cred) != kauth_cred_getuid(target) || 1492 kauth_cred_getuid(cred) != kauth_cred_getsvuid(target)) { 1493 /* 1494 * suid proc of ours or proc not ours 1495 */ 1496 r = EPERM; 1497 } else if (kauth_cred_getgid(target) != kauth_cred_getsvgid(target)) { 1498 /* 1499 * sgid proc has sgid back to us temporarily 1500 */ 1501 r = EPERM; 1502 } else { 1503 /* 1504 * our rgid must be in target's group list (ie, 1505 * sub-processes started by a sgid process) 1506 */ 1507 int ismember = 0; 1508 1509 if (kauth_cred_ismember_gid(cred, 1510 kauth_cred_getgid(target), &ismember) != 0 || 1511 !ismember) 1512 r = EPERM; 1513 } 1514 1515 return (r); 1516 } 1517 1518 /* 1519 * sysctl stuff 1520 */ 1521 1522 #define KERN_PROCSLOP (5 * sizeof(struct kinfo_proc)) 1523 1524 static const u_int sysctl_flagmap[] = { 1525 PK_ADVLOCK, P_ADVLOCK, 1526 PK_EXEC, P_EXEC, 1527 PK_NOCLDWAIT, P_NOCLDWAIT, 1528 PK_32, P_32, 1529 PK_CLDSIGIGN, P_CLDSIGIGN, 1530 PK_SUGID, P_SUGID, 1531 0 1532 }; 1533 1534 static const u_int sysctl_sflagmap[] = { 1535 PS_NOCLDSTOP, P_NOCLDSTOP, 1536 PS_WEXIT, P_WEXIT, 1537 PS_STOPFORK, P_STOPFORK, 1538 PS_STOPEXEC, P_STOPEXEC, 1539 PS_STOPEXIT, P_STOPEXIT, 1540 0 1541 }; 1542 1543 static const u_int sysctl_slflagmap[] = { 1544 PSL_TRACED, P_TRACED, 1545 PSL_FSTRACE, P_FSTRACE, 1546 PSL_CHTRACED, P_CHTRACED, 1547 PSL_SYSCALL, P_SYSCALL, 1548 0 1549 }; 1550 1551 static const u_int sysctl_lflagmap[] = { 1552 PL_CONTROLT, P_CONTROLT, 1553 PL_PPWAIT, P_PPWAIT, 1554 0 1555 }; 1556 1557 static const u_int sysctl_stflagmap[] = { 1558 PST_PROFIL, P_PROFIL, 1559 0 1560 1561 }; 1562 1563 /* used by kern_lwp also */ 1564 const u_int sysctl_lwpflagmap[] = { 1565 LW_SINTR, L_SINTR, 1566 LW_SYSTEM, L_SYSTEM, 1567 LW_SA, L_SA, /* WRS ??? */ 1568 0 1569 }; 1570 1571 /* 1572 * Find the most ``active'' lwp of a process and return it for ps display 1573 * purposes 1574 */ 1575 static struct lwp * 1576 proc_active_lwp(struct proc *p) 1577 { 1578 static const int ostat[] = { 1579 0, 1580 2, /* LSIDL */ 1581 6, /* LSRUN */ 1582 5, /* LSSLEEP */ 1583 4, /* LSSTOP */ 1584 0, /* LSZOMB */ 1585 1, /* LSDEAD */ 1586 7, /* LSONPROC */ 1587 3 /* LSSUSPENDED */ 1588 }; 1589 1590 struct lwp *l, *lp = NULL; 1591 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 1592 KASSERT(l->l_stat >= 0 && l->l_stat < __arraycount(ostat)); 1593 if (lp == NULL || 1594 ostat[l->l_stat] > ostat[lp->l_stat] || 1595 (ostat[l->l_stat] == ostat[lp->l_stat] && 1596 l->l_cpticks > lp->l_cpticks)) { 1597 lp = l; 1598 continue; 1599 } 1600 } 1601 return lp; 1602 } 1603 1604 static int 1605 sysctl_doeproc(SYSCTLFN_ARGS) 1606 { 1607 union { 1608 struct kinfo_proc kproc; 1609 struct kinfo_proc2 kproc2; 1610 } *kbuf; 1611 struct proc *p, *next, *marker; 1612 char *where, *dp; 1613 int type, op, arg, error; 1614 u_int elem_size, kelem_size, elem_count; 1615 size_t buflen, needed; 1616 bool match, zombie, mmmbrains; 1617 1618 if (namelen == 1 && name[0] == CTL_QUERY) 1619 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1620 1621 dp = where = oldp; 1622 buflen = where != NULL ? *oldlenp : 0; 1623 error = 0; 1624 needed = 0; 1625 type = rnode->sysctl_num; 1626 1627 if (type == KERN_PROC) { 1628 if (namelen != 2 && !(namelen == 1 && name[0] == KERN_PROC_ALL)) 1629 return (EINVAL); 1630 op = name[0]; 1631 if (op != KERN_PROC_ALL) 1632 arg = name[1]; 1633 else 1634 arg = 0; /* Quell compiler warning */ 1635 elem_count = 0; /* Ditto */ 1636 kelem_size = elem_size = sizeof(kbuf->kproc); 1637 } else { 1638 if (namelen != 4) 1639 return (EINVAL); 1640 op = name[0]; 1641 arg = name[1]; 1642 elem_size = name[2]; 1643 elem_count = name[3]; 1644 kelem_size = sizeof(kbuf->kproc2); 1645 } 1646 1647 sysctl_unlock(); 1648 1649 kbuf = kmem_alloc(sizeof(*kbuf), KM_SLEEP); 1650 marker = kmem_alloc(sizeof(*marker), KM_SLEEP); 1651 marker->p_flag = PK_MARKER; 1652 1653 mutex_enter(proc_lock); 1654 mmmbrains = false; 1655 for (p = LIST_FIRST(&allproc);; p = next) { 1656 if (p == NULL) { 1657 if (!mmmbrains) { 1658 p = LIST_FIRST(&zombproc); 1659 mmmbrains = true; 1660 } 1661 if (p == NULL) 1662 break; 1663 } 1664 next = LIST_NEXT(p, p_list); 1665 if ((p->p_flag & PK_MARKER) != 0) 1666 continue; 1667 1668 /* 1669 * Skip embryonic processes. 1670 */ 1671 if (p->p_stat == SIDL) 1672 continue; 1673 1674 mutex_enter(p->p_lock); 1675 error = kauth_authorize_process(l->l_cred, 1676 KAUTH_PROCESS_CANSEE, p, 1677 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENTRY), NULL, NULL); 1678 if (error != 0) { 1679 mutex_exit(p->p_lock); 1680 continue; 1681 } 1682 1683 /* 1684 * TODO - make more efficient (see notes below). 1685 * do by session. 1686 */ 1687 switch (op) { 1688 case KERN_PROC_PID: 1689 /* could do this with just a lookup */ 1690 match = (p->p_pid == (pid_t)arg); 1691 break; 1692 1693 case KERN_PROC_PGRP: 1694 /* could do this by traversing pgrp */ 1695 match = (p->p_pgrp->pg_id == (pid_t)arg); 1696 break; 1697 1698 case KERN_PROC_SESSION: 1699 match = (p->p_session->s_sid == (pid_t)arg); 1700 break; 1701 1702 case KERN_PROC_TTY: 1703 match = true; 1704 if (arg == (int) KERN_PROC_TTY_REVOKE) { 1705 if ((p->p_lflag & PL_CONTROLT) == 0 || 1706 p->p_session->s_ttyp == NULL || 1707 p->p_session->s_ttyvp != NULL) { 1708 match = false; 1709 } 1710 } else if ((p->p_lflag & PL_CONTROLT) == 0 || 1711 p->p_session->s_ttyp == NULL) { 1712 if ((dev_t)arg != KERN_PROC_TTY_NODEV) { 1713 match = false; 1714 } 1715 } else if (p->p_session->s_ttyp->t_dev != (dev_t)arg) { 1716 match = false; 1717 } 1718 break; 1719 1720 case KERN_PROC_UID: 1721 match = (kauth_cred_geteuid(p->p_cred) == (uid_t)arg); 1722 break; 1723 1724 case KERN_PROC_RUID: 1725 match = (kauth_cred_getuid(p->p_cred) == (uid_t)arg); 1726 break; 1727 1728 case KERN_PROC_GID: 1729 match = (kauth_cred_getegid(p->p_cred) == (uid_t)arg); 1730 break; 1731 1732 case KERN_PROC_RGID: 1733 match = (kauth_cred_getgid(p->p_cred) == (uid_t)arg); 1734 break; 1735 1736 case KERN_PROC_ALL: 1737 match = true; 1738 /* allow everything */ 1739 break; 1740 1741 default: 1742 error = EINVAL; 1743 mutex_exit(p->p_lock); 1744 goto cleanup; 1745 } 1746 if (!match) { 1747 mutex_exit(p->p_lock); 1748 continue; 1749 } 1750 1751 /* 1752 * Grab a hold on the process. 1753 */ 1754 if (mmmbrains) { 1755 zombie = true; 1756 } else { 1757 zombie = !rw_tryenter(&p->p_reflock, RW_READER); 1758 } 1759 if (zombie) { 1760 LIST_INSERT_AFTER(p, marker, p_list); 1761 } 1762 1763 if (buflen >= elem_size && 1764 (type == KERN_PROC || elem_count > 0)) { 1765 if (type == KERN_PROC) { 1766 kbuf->kproc.kp_proc = *p; 1767 fill_eproc(p, &kbuf->kproc.kp_eproc, zombie); 1768 } else { 1769 fill_kproc2(p, &kbuf->kproc2, zombie); 1770 elem_count--; 1771 } 1772 mutex_exit(p->p_lock); 1773 mutex_exit(proc_lock); 1774 /* 1775 * Copy out elem_size, but not larger than kelem_size 1776 */ 1777 error = sysctl_copyout(l, kbuf, dp, 1778 min(kelem_size, elem_size)); 1779 mutex_enter(proc_lock); 1780 if (error) { 1781 goto bah; 1782 } 1783 dp += elem_size; 1784 buflen -= elem_size; 1785 } else { 1786 mutex_exit(p->p_lock); 1787 } 1788 needed += elem_size; 1789 1790 /* 1791 * Release reference to process. 1792 */ 1793 if (zombie) { 1794 next = LIST_NEXT(marker, p_list); 1795 LIST_REMOVE(marker, p_list); 1796 } else { 1797 rw_exit(&p->p_reflock); 1798 next = LIST_NEXT(p, p_list); 1799 } 1800 } 1801 mutex_exit(proc_lock); 1802 1803 if (where != NULL) { 1804 *oldlenp = dp - where; 1805 if (needed > *oldlenp) { 1806 error = ENOMEM; 1807 goto out; 1808 } 1809 } else { 1810 needed += KERN_PROCSLOP; 1811 *oldlenp = needed; 1812 } 1813 if (kbuf) 1814 kmem_free(kbuf, sizeof(*kbuf)); 1815 if (marker) 1816 kmem_free(marker, sizeof(*marker)); 1817 sysctl_relock(); 1818 return 0; 1819 bah: 1820 if (zombie) 1821 LIST_REMOVE(marker, p_list); 1822 else 1823 rw_exit(&p->p_reflock); 1824 cleanup: 1825 mutex_exit(proc_lock); 1826 out: 1827 if (kbuf) 1828 kmem_free(kbuf, sizeof(*kbuf)); 1829 if (marker) 1830 kmem_free(marker, sizeof(*marker)); 1831 sysctl_relock(); 1832 return error; 1833 } 1834 1835 int 1836 copyin_psstrings(struct proc *p, struct ps_strings *arginfo) 1837 { 1838 1839 #ifdef COMPAT_NETBSD32 1840 if (p->p_flag & PK_32) { 1841 struct ps_strings32 arginfo32; 1842 1843 int error = copyin_proc(p, (void *)p->p_psstrp, &arginfo32, 1844 sizeof(arginfo32)); 1845 if (error) 1846 return error; 1847 arginfo->ps_argvstr = (void *)(uintptr_t)arginfo32.ps_argvstr; 1848 arginfo->ps_nargvstr = arginfo32.ps_nargvstr; 1849 arginfo->ps_envstr = (void *)(uintptr_t)arginfo32.ps_envstr; 1850 arginfo->ps_nenvstr = arginfo32.ps_nenvstr; 1851 return 0; 1852 } 1853 #endif 1854 return copyin_proc(p, (void *)p->p_psstrp, arginfo, sizeof(*arginfo)); 1855 } 1856 1857 static int 1858 copy_procargs_sysctl_cb(void *cookie_, const void *src, size_t off, size_t len) 1859 { 1860 void **cookie = cookie_; 1861 struct lwp *l = cookie[0]; 1862 char *dst = cookie[1]; 1863 1864 return sysctl_copyout(l, src, dst + off, len); 1865 } 1866 1867 /* 1868 * sysctl helper routine for kern.proc_args pseudo-subtree. 1869 */ 1870 static int 1871 sysctl_kern_proc_args(SYSCTLFN_ARGS) 1872 { 1873 struct ps_strings pss; 1874 struct proc *p; 1875 pid_t pid; 1876 int type, error; 1877 void *cookie[2]; 1878 1879 if (namelen == 1 && name[0] == CTL_QUERY) 1880 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1881 1882 if (newp != NULL || namelen != 2) 1883 return (EINVAL); 1884 pid = name[0]; 1885 type = name[1]; 1886 1887 switch (type) { 1888 case KERN_PROC_ARGV: 1889 case KERN_PROC_NARGV: 1890 case KERN_PROC_ENV: 1891 case KERN_PROC_NENV: 1892 /* ok */ 1893 break; 1894 default: 1895 return (EINVAL); 1896 } 1897 1898 sysctl_unlock(); 1899 1900 /* check pid */ 1901 mutex_enter(proc_lock); 1902 if ((p = proc_find(pid)) == NULL) { 1903 error = EINVAL; 1904 goto out_locked; 1905 } 1906 mutex_enter(p->p_lock); 1907 1908 /* Check permission. */ 1909 if (type == KERN_PROC_ARGV || type == KERN_PROC_NARGV) 1910 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1911 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ARGS), NULL, NULL); 1912 else if (type == KERN_PROC_ENV || type == KERN_PROC_NENV) 1913 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1914 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENV), NULL, NULL); 1915 else 1916 error = EINVAL; /* XXXGCC */ 1917 if (error) { 1918 mutex_exit(p->p_lock); 1919 goto out_locked; 1920 } 1921 1922 if (oldp == NULL) { 1923 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) 1924 *oldlenp = sizeof (int); 1925 else 1926 *oldlenp = ARG_MAX; /* XXX XXX XXX */ 1927 error = 0; 1928 mutex_exit(p->p_lock); 1929 goto out_locked; 1930 } 1931 1932 /* 1933 * Zombies don't have a stack, so we can't read their psstrings. 1934 * System processes also don't have a user stack. 1935 */ 1936 if (P_ZOMBIE(p) || (p->p_flag & PK_SYSTEM) != 0) { 1937 error = EINVAL; 1938 mutex_exit(p->p_lock); 1939 goto out_locked; 1940 } 1941 1942 error = rw_tryenter(&p->p_reflock, RW_READER) ? 0 : EBUSY; 1943 mutex_exit(p->p_lock); 1944 if (error) { 1945 goto out_locked; 1946 } 1947 mutex_exit(proc_lock); 1948 1949 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) { 1950 int value; 1951 if ((error = copyin_psstrings(p, &pss)) == 0) { 1952 if (type == KERN_PROC_NARGV) 1953 value = pss.ps_nargvstr; 1954 else 1955 value = pss.ps_nenvstr; 1956 error = sysctl_copyout(l, &value, oldp, sizeof(value)); 1957 *oldlenp = sizeof(value); 1958 } 1959 } else { 1960 cookie[0] = l; 1961 cookie[1] = oldp; 1962 error = copy_procargs(p, type, oldlenp, 1963 copy_procargs_sysctl_cb, cookie); 1964 } 1965 rw_exit(&p->p_reflock); 1966 sysctl_relock(); 1967 return error; 1968 1969 out_locked: 1970 mutex_exit(proc_lock); 1971 sysctl_relock(); 1972 return error; 1973 } 1974 1975 int 1976 copy_procargs(struct proc *p, int oid, size_t *limit, 1977 int (*cb)(void *, const void *, size_t, size_t), void *cookie) 1978 { 1979 struct ps_strings pss; 1980 size_t len, i, loaded, entry_len; 1981 struct uio auio; 1982 struct iovec aiov; 1983 int error, argvlen; 1984 char *arg; 1985 char **argv; 1986 vaddr_t user_argv; 1987 struct vmspace *vmspace; 1988 1989 /* 1990 * Allocate a temporary buffer to hold the argument vector and 1991 * the arguments themselve. 1992 */ 1993 arg = kmem_alloc(PAGE_SIZE, KM_SLEEP); 1994 argv = kmem_alloc(PAGE_SIZE, KM_SLEEP); 1995 1996 /* 1997 * Lock the process down in memory. 1998 */ 1999 vmspace = p->p_vmspace; 2000 uvmspace_addref(vmspace); 2001 2002 /* 2003 * Read in the ps_strings structure. 2004 */ 2005 if ((error = copyin_psstrings(p, &pss)) != 0) 2006 goto done; 2007 2008 /* 2009 * Now read the address of the argument vector. 2010 */ 2011 switch (oid) { 2012 case KERN_PROC_ARGV: 2013 user_argv = (uintptr_t)pss.ps_argvstr; 2014 argvlen = pss.ps_nargvstr; 2015 break; 2016 case KERN_PROC_ENV: 2017 user_argv = (uintptr_t)pss.ps_envstr; 2018 argvlen = pss.ps_nenvstr; 2019 break; 2020 default: 2021 error = EINVAL; 2022 goto done; 2023 } 2024 2025 if (argvlen < 0) { 2026 error = EIO; 2027 goto done; 2028 } 2029 2030 #ifdef COMPAT_NETBSD32 2031 if (p->p_flag & PK_32) 2032 entry_len = sizeof(netbsd32_charp); 2033 else 2034 #endif 2035 entry_len = sizeof(char *); 2036 2037 /* 2038 * Now copy each string. 2039 */ 2040 len = 0; /* bytes written to user buffer */ 2041 loaded = 0; /* bytes from argv already processed */ 2042 i = 0; /* To make compiler happy */ 2043 2044 for (; argvlen; --argvlen) { 2045 int finished = 0; 2046 vaddr_t base; 2047 size_t xlen; 2048 int j; 2049 2050 if (loaded == 0) { 2051 size_t rem = entry_len * argvlen; 2052 loaded = MIN(rem, PAGE_SIZE); 2053 error = copyin_vmspace(vmspace, 2054 (const void *)user_argv, argv, loaded); 2055 if (error) 2056 break; 2057 user_argv += loaded; 2058 i = 0; 2059 } 2060 2061 #ifdef COMPAT_NETBSD32 2062 if (p->p_flag & PK_32) { 2063 netbsd32_charp *argv32; 2064 2065 argv32 = (netbsd32_charp *)argv; 2066 base = (vaddr_t)NETBSD32PTR64(argv32[i++]); 2067 } else 2068 #endif 2069 base = (vaddr_t)argv[i++]; 2070 loaded -= entry_len; 2071 2072 /* 2073 * The program has messed around with its arguments, 2074 * possibly deleting some, and replacing them with 2075 * NULL's. Treat this as the last argument and not 2076 * a failure. 2077 */ 2078 if (base == 0) 2079 break; 2080 2081 while (!finished) { 2082 xlen = PAGE_SIZE - (base & PAGE_MASK); 2083 2084 aiov.iov_base = arg; 2085 aiov.iov_len = PAGE_SIZE; 2086 auio.uio_iov = &aiov; 2087 auio.uio_iovcnt = 1; 2088 auio.uio_offset = base; 2089 auio.uio_resid = xlen; 2090 auio.uio_rw = UIO_READ; 2091 UIO_SETUP_SYSSPACE(&auio); 2092 error = uvm_io(&vmspace->vm_map, &auio); 2093 if (error) 2094 goto done; 2095 2096 /* Look for the end of the string */ 2097 for (j = 0; j < xlen; j++) { 2098 if (arg[j] == '\0') { 2099 xlen = j + 1; 2100 finished = 1; 2101 break; 2102 } 2103 } 2104 2105 /* Check for user buffer overflow */ 2106 if (len + xlen > *limit) { 2107 finished = 1; 2108 if (len > *limit) 2109 xlen = 0; 2110 else 2111 xlen = *limit - len; 2112 } 2113 2114 /* Copyout the page */ 2115 error = (*cb)(cookie, arg, len, xlen); 2116 if (error) 2117 goto done; 2118 2119 len += xlen; 2120 base += xlen; 2121 } 2122 } 2123 *limit = len; 2124 2125 done: 2126 kmem_free(argv, PAGE_SIZE); 2127 kmem_free(arg, PAGE_SIZE); 2128 uvmspace_free(vmspace); 2129 return error; 2130 } 2131 2132 /* 2133 * Fill in an eproc structure for the specified process. 2134 */ 2135 void 2136 fill_eproc(struct proc *p, struct eproc *ep, bool zombie) 2137 { 2138 struct tty *tp; 2139 struct lwp *l; 2140 2141 KASSERT(mutex_owned(proc_lock)); 2142 KASSERT(mutex_owned(p->p_lock)); 2143 2144 memset(ep, 0, sizeof(*ep)); 2145 2146 ep->e_paddr = p; 2147 ep->e_sess = p->p_session; 2148 if (p->p_cred) { 2149 kauth_cred_topcred(p->p_cred, &ep->e_pcred); 2150 kauth_cred_toucred(p->p_cred, &ep->e_ucred); 2151 } 2152 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2153 struct vmspace *vm = p->p_vmspace; 2154 2155 ep->e_vm.vm_rssize = vm_resident_count(vm); 2156 ep->e_vm.vm_tsize = vm->vm_tsize; 2157 ep->e_vm.vm_dsize = vm->vm_dsize; 2158 ep->e_vm.vm_ssize = vm->vm_ssize; 2159 ep->e_vm.vm_map.size = vm->vm_map.size; 2160 2161 /* Pick the primary (first) LWP */ 2162 l = proc_active_lwp(p); 2163 KASSERT(l != NULL); 2164 lwp_lock(l); 2165 if (l->l_wchan) 2166 strncpy(ep->e_wmesg, l->l_wmesg, WMESGLEN); 2167 lwp_unlock(l); 2168 } 2169 if (p->p_pptr) 2170 ep->e_ppid = p->p_pptr->p_pid; 2171 if (p->p_pgrp && p->p_session) { 2172 ep->e_pgid = p->p_pgrp->pg_id; 2173 ep->e_jobc = p->p_pgrp->pg_jobc; 2174 ep->e_sid = p->p_session->s_sid; 2175 if ((p->p_lflag & PL_CONTROLT) && 2176 (tp = ep->e_sess->s_ttyp)) { 2177 ep->e_tdev = tp->t_dev; 2178 ep->e_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2179 ep->e_tsess = tp->t_session; 2180 } else 2181 ep->e_tdev = (uint32_t)NODEV; 2182 ep->e_flag = ep->e_sess->s_ttyvp ? EPROC_CTTY : 0; 2183 if (SESS_LEADER(p)) 2184 ep->e_flag |= EPROC_SLEADER; 2185 strncpy(ep->e_login, ep->e_sess->s_login, MAXLOGNAME); 2186 } 2187 ep->e_xsize = ep->e_xrssize = 0; 2188 ep->e_xccount = ep->e_xswrss = 0; 2189 } 2190 2191 /* 2192 * Fill in a kinfo_proc2 structure for the specified process. 2193 */ 2194 static void 2195 fill_kproc2(struct proc *p, struct kinfo_proc2 *ki, bool zombie) 2196 { 2197 struct tty *tp; 2198 struct lwp *l, *l2; 2199 struct timeval ut, st, rt; 2200 sigset_t ss1, ss2; 2201 struct rusage ru; 2202 struct vmspace *vm; 2203 2204 KASSERT(mutex_owned(proc_lock)); 2205 KASSERT(mutex_owned(p->p_lock)); 2206 2207 sigemptyset(&ss1); 2208 sigemptyset(&ss2); 2209 memset(ki, 0, sizeof(*ki)); 2210 2211 ki->p_paddr = PTRTOUINT64(p); 2212 ki->p_fd = PTRTOUINT64(p->p_fd); 2213 ki->p_cwdi = PTRTOUINT64(p->p_cwdi); 2214 ki->p_stats = PTRTOUINT64(p->p_stats); 2215 ki->p_limit = PTRTOUINT64(p->p_limit); 2216 ki->p_vmspace = PTRTOUINT64(p->p_vmspace); 2217 ki->p_sigacts = PTRTOUINT64(p->p_sigacts); 2218 ki->p_sess = PTRTOUINT64(p->p_session); 2219 ki->p_tsess = 0; /* may be changed if controlling tty below */ 2220 ki->p_ru = PTRTOUINT64(&p->p_stats->p_ru); 2221 ki->p_eflag = 0; 2222 ki->p_exitsig = p->p_exitsig; 2223 ki->p_flag = L_INMEM; /* Process never swapped out */ 2224 ki->p_flag |= sysctl_map_flags(sysctl_flagmap, p->p_flag); 2225 ki->p_flag |= sysctl_map_flags(sysctl_sflagmap, p->p_sflag); 2226 ki->p_flag |= sysctl_map_flags(sysctl_slflagmap, p->p_slflag); 2227 ki->p_flag |= sysctl_map_flags(sysctl_lflagmap, p->p_lflag); 2228 ki->p_flag |= sysctl_map_flags(sysctl_stflagmap, p->p_stflag); 2229 ki->p_pid = p->p_pid; 2230 if (p->p_pptr) 2231 ki->p_ppid = p->p_pptr->p_pid; 2232 else 2233 ki->p_ppid = 0; 2234 ki->p_uid = kauth_cred_geteuid(p->p_cred); 2235 ki->p_ruid = kauth_cred_getuid(p->p_cred); 2236 ki->p_gid = kauth_cred_getegid(p->p_cred); 2237 ki->p_rgid = kauth_cred_getgid(p->p_cred); 2238 ki->p_svuid = kauth_cred_getsvuid(p->p_cred); 2239 ki->p_svgid = kauth_cred_getsvgid(p->p_cred); 2240 ki->p_ngroups = kauth_cred_ngroups(p->p_cred); 2241 kauth_cred_getgroups(p->p_cred, ki->p_groups, 2242 min(ki->p_ngroups, sizeof(ki->p_groups) / sizeof(ki->p_groups[0])), 2243 UIO_SYSSPACE); 2244 2245 ki->p_uticks = p->p_uticks; 2246 ki->p_sticks = p->p_sticks; 2247 ki->p_iticks = p->p_iticks; 2248 ki->p_tpgid = NO_PGID; /* may be changed if controlling tty below */ 2249 ki->p_tracep = PTRTOUINT64(p->p_tracep); 2250 ki->p_traceflag = p->p_traceflag; 2251 2252 memcpy(&ki->p_sigignore, &p->p_sigctx.ps_sigignore,sizeof(ki_sigset_t)); 2253 memcpy(&ki->p_sigcatch, &p->p_sigctx.ps_sigcatch, sizeof(ki_sigset_t)); 2254 2255 ki->p_cpticks = 0; 2256 ki->p_pctcpu = p->p_pctcpu; 2257 ki->p_estcpu = 0; 2258 ki->p_stat = p->p_stat; /* Will likely be overridden by LWP status */ 2259 ki->p_realstat = p->p_stat; 2260 ki->p_nice = p->p_nice; 2261 ki->p_xstat = p->p_xstat; 2262 ki->p_acflag = p->p_acflag; 2263 2264 strncpy(ki->p_comm, p->p_comm, 2265 min(sizeof(ki->p_comm), sizeof(p->p_comm))); 2266 strncpy(ki->p_ename, p->p_emul->e_name, sizeof(ki->p_ename)); 2267 2268 ki->p_nlwps = p->p_nlwps; 2269 ki->p_realflag = ki->p_flag; 2270 2271 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2272 vm = p->p_vmspace; 2273 ki->p_vm_rssize = vm_resident_count(vm); 2274 ki->p_vm_tsize = vm->vm_tsize; 2275 ki->p_vm_dsize = vm->vm_dsize; 2276 ki->p_vm_ssize = vm->vm_ssize; 2277 ki->p_vm_vsize = vm->vm_map.size; 2278 /* 2279 * Since the stack is initially mapped mostly with 2280 * PROT_NONE and grown as needed, adjust the "mapped size" 2281 * to skip the unused stack portion. 2282 */ 2283 ki->p_vm_msize = 2284 atop(vm->vm_map.size) - vm->vm_issize + vm->vm_ssize; 2285 2286 /* Pick the primary (first) LWP */ 2287 l = proc_active_lwp(p); 2288 KASSERT(l != NULL); 2289 lwp_lock(l); 2290 ki->p_nrlwps = p->p_nrlwps; 2291 ki->p_forw = 0; 2292 ki->p_back = 0; 2293 ki->p_addr = PTRTOUINT64(l->l_addr); 2294 ki->p_stat = l->l_stat; 2295 ki->p_flag |= sysctl_map_flags(sysctl_lwpflagmap, l->l_flag); 2296 ki->p_swtime = l->l_swtime; 2297 ki->p_slptime = l->l_slptime; 2298 if (l->l_stat == LSONPROC) 2299 ki->p_schedflags = l->l_cpu->ci_schedstate.spc_flags; 2300 else 2301 ki->p_schedflags = 0; 2302 ki->p_priority = lwp_eprio(l); 2303 ki->p_usrpri = l->l_priority; 2304 if (l->l_wchan) 2305 strncpy(ki->p_wmesg, l->l_wmesg, sizeof(ki->p_wmesg)); 2306 ki->p_wchan = PTRTOUINT64(l->l_wchan); 2307 ki->p_cpuid = cpu_index(l->l_cpu); 2308 lwp_unlock(l); 2309 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 2310 /* This is hardly correct, but... */ 2311 sigplusset(&l->l_sigpend.sp_set, &ss1); 2312 sigplusset(&l->l_sigmask, &ss2); 2313 ki->p_cpticks += l->l_cpticks; 2314 ki->p_pctcpu += l->l_pctcpu; 2315 ki->p_estcpu += l->l_estcpu; 2316 } 2317 } 2318 sigplusset(&p->p_sigpend.sp_set, &ss2); 2319 memcpy(&ki->p_siglist, &ss1, sizeof(ki_sigset_t)); 2320 memcpy(&ki->p_sigmask, &ss2, sizeof(ki_sigset_t)); 2321 2322 if (p->p_session != NULL) { 2323 ki->p_sid = p->p_session->s_sid; 2324 ki->p__pgid = p->p_pgrp->pg_id; 2325 if (p->p_session->s_ttyvp) 2326 ki->p_eflag |= EPROC_CTTY; 2327 if (SESS_LEADER(p)) 2328 ki->p_eflag |= EPROC_SLEADER; 2329 strncpy(ki->p_login, p->p_session->s_login, 2330 min(sizeof ki->p_login - 1, sizeof p->p_session->s_login)); 2331 ki->p_jobc = p->p_pgrp->pg_jobc; 2332 if ((p->p_lflag & PL_CONTROLT) && (tp = p->p_session->s_ttyp)) { 2333 ki->p_tdev = tp->t_dev; 2334 ki->p_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2335 ki->p_tsess = PTRTOUINT64(tp->t_session); 2336 } else { 2337 ki->p_tdev = (int32_t)NODEV; 2338 } 2339 } 2340 2341 if (!P_ZOMBIE(p) && !zombie) { 2342 ki->p_uvalid = 1; 2343 ki->p_ustart_sec = p->p_stats->p_start.tv_sec; 2344 ki->p_ustart_usec = p->p_stats->p_start.tv_usec; 2345 2346 calcru(p, &ut, &st, NULL, &rt); 2347 ki->p_rtime_sec = rt.tv_sec; 2348 ki->p_rtime_usec = rt.tv_usec; 2349 ki->p_uutime_sec = ut.tv_sec; 2350 ki->p_uutime_usec = ut.tv_usec; 2351 ki->p_ustime_sec = st.tv_sec; 2352 ki->p_ustime_usec = st.tv_usec; 2353 2354 memcpy(&ru, &p->p_stats->p_ru, sizeof(ru)); 2355 ki->p_uru_nvcsw = 0; 2356 ki->p_uru_nivcsw = 0; 2357 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 2358 ki->p_uru_nvcsw += (l2->l_ncsw - l2->l_nivcsw); 2359 ki->p_uru_nivcsw += l2->l_nivcsw; 2360 ruadd(&ru, &l2->l_ru); 2361 } 2362 ki->p_uru_maxrss = ru.ru_maxrss; 2363 ki->p_uru_ixrss = ru.ru_ixrss; 2364 ki->p_uru_idrss = ru.ru_idrss; 2365 ki->p_uru_isrss = ru.ru_isrss; 2366 ki->p_uru_minflt = ru.ru_minflt; 2367 ki->p_uru_majflt = ru.ru_majflt; 2368 ki->p_uru_nswap = ru.ru_nswap; 2369 ki->p_uru_inblock = ru.ru_inblock; 2370 ki->p_uru_oublock = ru.ru_oublock; 2371 ki->p_uru_msgsnd = ru.ru_msgsnd; 2372 ki->p_uru_msgrcv = ru.ru_msgrcv; 2373 ki->p_uru_nsignals = ru.ru_nsignals; 2374 2375 timeradd(&p->p_stats->p_cru.ru_utime, 2376 &p->p_stats->p_cru.ru_stime, &ut); 2377 ki->p_uctime_sec = ut.tv_sec; 2378 ki->p_uctime_usec = ut.tv_usec; 2379 } 2380 } 2381
Indexes created Sun Sep 21 20:09:37 GMT 2025