1 /*- 2 * SPDX-License-Identifier: BSD-3-Clause 3 * 4 * Copyright (c) 1986, 1989, 1991, 1993 5 * The Regents of the University of California. All rights reserved. 6 * (c) UNIX System Laboratories, Inc. 7 * All or some portions of this file are derived from material licensed 8 * to the University of California by American Telephone and Telegraph 9 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 10 * the permission of UNIX System Laboratories, Inc. 11 * 12 * Redistribution and use in source and binary forms, with or without 13 * modification, are permitted provided that the following conditions 14 * are met: 15 * 1. Redistributions of source code must retain the above copyright 16 * notice, this list of conditions and the following disclaimer. 17 * 2. Redistributions in binary form must reproduce the above copyright 18 * notice, this list of conditions and the following disclaimer in the 19 * documentation and/or other materials provided with the distribution. 20 * 3. Neither the name of the University nor the names of its contributors 21 * may be used to endorse or promote products derived from this software 22 * without specific prior written permission. 23 * 24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 27 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 34 * SUCH DAMAGE. 35 * 36 * @(#)proc.h 8.15 (Berkeley) 5/19/95 37 * $FreeBSD$ 38 */ 39 40 #ifndef _SYS_PROC_H_ 41 #define _SYS_PROC_H_ 42 43 #include <sys/callout.h> /* For struct callout. */ 44 #include <sys/event.h> /* For struct klist. */ 45 #ifdef _KERNEL 46 #include <sys/_eventhandler.h> 47 #endif 48 #include <sys/condvar.h> 49 #ifndef _KERNEL 50 #include <sys/filedesc.h> 51 #endif 52 #include <sys/queue.h> 53 #include <sys/_lock.h> 54 #include <sys/lock_profile.h> 55 #include <sys/_mutex.h> 56 #include <sys/osd.h> 57 #include <sys/priority.h> 58 #include <sys/rtprio.h> /* XXX. */ 59 #include <sys/runq.h> 60 #include <sys/resource.h> 61 #include <sys/sigio.h> 62 #include <sys/signal.h> 63 #include <sys/signalvar.h> 64 #ifndef _KERNEL 65 #include <sys/time.h> /* For structs itimerval, timeval. */ 66 #else 67 #include <sys/pcpu.h> 68 #include <sys/systm.h> 69 #endif 70 #include <sys/ucontext.h> 71 #include <sys/ucred.h> 72 #include <sys/types.h> 73 #include <sys/_domainset.h> 74 75 #include <machine/proc.h> /* Machine-dependent proc substruct. */ 76 #ifdef _KERNEL 77 #include <machine/cpu.h> 78 #endif 79 80 /* 81 * One structure allocated per session. 82 * 83 * List of locks 84 * (m) locked by s_mtx mtx 85 * (e) locked by proctree_lock sx 86 * (c) const until freeing 87 */ 88 struct session { 89 u_int s_count; /* Ref cnt; pgrps in session - atomic. */ 90 struct proc *s_leader; /* (m + e) Session leader. */ 91 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */ 92 struct cdev_priv *s_ttydp; /* (m) Device of controlling tty. */ 93 struct tty *s_ttyp; /* (e) Controlling tty. */ 94 pid_t s_sid; /* (c) Session ID. */ 95 /* (m) Setlogin() name: */ 96 char s_login[roundup(MAXLOGNAME, sizeof(long))]; 97 struct mtx s_mtx; /* Mutex to protect members. */ 98 }; 99 100 /* 101 * One structure allocated per process group. 102 * 103 * List of locks 104 * (m) locked by pg_mtx mtx 105 * (e) locked by proctree_lock sx 106 * (c) const until freeing 107 */ 108 struct pgrp { 109 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */ 110 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */ 111 struct session *pg_session; /* (c) Pointer to session. */ 112 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */ 113 pid_t pg_id; /* (c) Process group id. */ 114 struct mtx pg_mtx; /* Mutex to protect members */ 115 int pg_flags; /* (m) PGRP_ flags */ 116 }; 117 118 #define PGRP_ORPHANED 0x00000001 /* Group is orphaned */ 119 120 /* 121 * pargs, used to hold a copy of the command line, if it had a sane length. 122 */ 123 struct pargs { 124 u_int ar_ref; /* Reference count. */ 125 u_int ar_length; /* Length. */ 126 u_char ar_args[1]; /* Arguments. */ 127 }; 128 129 /*- 130 * Description of a process. 131 * 132 * This structure contains the information needed to manage a thread of 133 * control, known in UN*X as a process; it has references to substructures 134 * containing descriptions of things that the process uses, but may share 135 * with related processes. The process structure and the substructures 136 * are always addressable except for those marked "(CPU)" below, 137 * which might be addressable only on a processor on which the process 138 * is running. 139 * 140 * Below is a key of locks used to protect each member of struct proc. The 141 * lock is indicated by a reference to a specific character in parens in the 142 * associated comment. 143 * * - not yet protected 144 * a - only touched by curproc or parent during fork/wait 145 * b - created at fork, never changes 146 * (exception aiods switch vmspaces, but they are also 147 * marked 'P_SYSTEM' so hopefully it will be left alone) 148 * c - locked by proc mtx 149 * d - locked by allproc_lock lock 150 * e - locked by proctree_lock lock 151 * f - session mtx 152 * g - process group mtx 153 * h - callout_lock mtx 154 * i - by curproc or the master session mtx 155 * j - locked by proc slock 156 * k - only accessed by curthread 157 * k*- only accessed by curthread and from an interrupt 158 * kx- only accessed by curthread and by debugger 159 * l - the attaching proc or attaching proc parent 160 * m - Giant 161 * n - not locked, lazy 162 * o - ktrace lock 163 * q - td_contested lock 164 * r - p_peers lock 165 * s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9) 166 * t - thread lock 167 * u - process stat lock 168 * w - process timer lock 169 * x - created at fork, only changes during single threading in exec 170 * y - created at first aio, doesn't change until exit or exec at which 171 * point we are single-threaded and only curthread changes it 172 * z - zombie threads lock 173 * 174 * If the locking key specifies two identifiers (for example, p_pptr) then 175 * either lock is sufficient for read access, but both locks must be held 176 * for write access. 177 */ 178 struct cpuset; 179 struct filecaps; 180 struct filemon; 181 struct kaioinfo; 182 struct kaudit_record; 183 struct kcov_info; 184 struct kdtrace_proc; 185 struct kdtrace_thread; 186 struct kmsan_td; 187 struct kq_timer_cb_data; 188 struct mqueue_notifier; 189 struct p_sched; 190 struct proc; 191 struct procdesc; 192 struct racct; 193 struct sbuf; 194 struct sleepqueue; 195 struct socket; 196 struct syscall_args; 197 struct td_sched; 198 struct thread; 199 struct trapframe; 200 struct turnstile; 201 struct vm_map; 202 struct vm_map_entry; 203 struct epoch_tracker; 204 205 /* 206 * XXX: Does this belong in resource.h or resourcevar.h instead? 207 * Resource usage extension. The times in rusage structs in the kernel are 208 * never up to date. The actual times are kept as runtimes and tick counts 209 * (with control info in the "previous" times), and are converted when 210 * userland asks for rusage info. Backwards compatibility prevents putting 211 * this directly in the user-visible rusage struct. 212 * 213 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux. 214 * Locking for td_rux: (t) for all fields. 215 */ 216 struct rusage_ext { 217 uint64_t rux_runtime; /* (cu) Real time. */ 218 uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */ 219 uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */ 220 uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */ 221 uint64_t rux_uu; /* (c) Previous user time in usec. */ 222 uint64_t rux_su; /* (c) Previous sys time in usec. */ 223 uint64_t rux_tu; /* (c) Previous total time in usec. */ 224 }; 225 226 /* 227 * Kernel runnable context (thread). 228 * This is what is put to sleep and reactivated. 229 * Thread context. Processes may have multiple threads. 230 */ 231 struct thread { 232 struct mtx *volatile td_lock; /* replaces sched lock */ 233 struct proc *td_proc; /* (*) Associated process. */ 234 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */ 235 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */ 236 union { 237 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */ 238 struct thread *td_zombie; /* Zombie list linkage */ 239 }; 240 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */ 241 LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */ 242 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */ 243 struct domainset_ref td_domain; /* (a) NUMA policy */ 244 struct seltd *td_sel; /* Select queue/channel. */ 245 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */ 246 struct turnstile *td_turnstile; /* (k) Associated turnstile. */ 247 struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */ 248 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */ 249 lwpid_t td_tid; /* (b) Thread ID. */ 250 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */ 251 #define td_siglist td_sigqueue.sq_signals 252 u_char td_lend_user_pri; /* (t) Lend user pri. */ 253 u_char td_allocdomain; /* (b) NUMA domain backing this struct thread. */ 254 struct kmsan_td *td_kmsan; /* (k) KMSAN state */ 255 256 /* Cleared during fork1() */ 257 #define td_startzero td_flags 258 int td_flags; /* (t) TDF_* flags. */ 259 int td_inhibitors; /* (t) Why can not run. */ 260 int td_pflags; /* (k) Private thread (TDP_*) flags. */ 261 int td_pflags2; /* (k) Private thread (TDP2_*) flags. */ 262 int td_dupfd; /* (k) Ret value from fdopen. XXX */ 263 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */ 264 const void *td_wchan; /* (t) Sleep address. */ 265 const char *td_wmesg; /* (t) Reason for sleep. */ 266 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */ 267 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */ 268 short td_locks; /* (k) Debug: count of non-spin locks */ 269 short td_rw_rlocks; /* (k) Count of rwlock read locks. */ 270 short td_sx_slocks; /* (k) Count of sx shared locks. */ 271 short td_lk_slocks; /* (k) Count of lockmgr shared locks. */ 272 short td_stopsched; /* (k) Scheduler stopped. */ 273 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */ 274 const char *td_lockname; /* (t) Name of lock blocked on. */ 275 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */ 276 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */ 277 int td_intr_nesting_level; /* (k) Interrupt recursion. */ 278 int td_pinned; /* (k) Temporary cpu pin count. */ 279 struct ucred *td_realucred; /* (k) Reference to credentials. */ 280 struct ucred *td_ucred; /* (k) Used credentials, temporarily switchable. */ 281 struct plimit *td_limit; /* (k) Resource limits. */ 282 int td_slptick; /* (t) Time at sleep. */ 283 int td_blktick; /* (t) Time spent blocked. */ 284 int td_swvoltick; /* (t) Time at last SW_VOL switch. */ 285 int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */ 286 u_int td_cow; /* (*) Number of copy-on-write faults */ 287 struct rusage td_ru; /* (t) rusage information. */ 288 struct rusage_ext td_rux; /* (t) Internal rusage information. */ 289 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */ 290 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */ 291 u_int td_pticks; /* (t) Statclock hits for profiling */ 292 u_int td_sticks; /* (t) Statclock hits in system mode. */ 293 u_int td_iticks; /* (t) Statclock hits in intr mode. */ 294 u_int td_uticks; /* (t) Statclock hits in user mode. */ 295 int td_intrval; /* (t) Return value for sleepq. */ 296 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */ 297 volatile u_int td_generation; /* (k) For detection of preemption */ 298 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */ 299 int td_xsig; /* (c) Signal for ptrace */ 300 u_long td_profil_addr; /* (k) Temporary addr until AST. */ 301 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */ 302 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */ 303 struct file *td_fpop; /* (k) file referencing cdev under op */ 304 int td_dbgflags; /* (c) Userland debugger flags */ 305 siginfo_t td_si; /* (c) For debugger or core file */ 306 int td_ng_outbound; /* (k) Thread entered ng from above. */ 307 struct osd td_osd; /* (k) Object specific data. */ 308 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */ 309 pid_t td_dbg_forked; /* (c) Child pid for debugger. */ 310 struct vnode *td_vp_reserved;/* (k) Preallocated vnode. */ 311 u_int td_no_sleeping; /* (k) Sleeping disabled count. */ 312 void *td_su; /* (k) FFS SU private */ 313 sbintime_t td_sleeptimo; /* (t) Sleep timeout. */ 314 int td_rtcgen; /* (s) rtc_generation of abs. sleep */ 315 int td_errno; /* (k) Error from last syscall. */ 316 size_t td_vslock_sz; /* (k) amount of vslock-ed space */ 317 struct kcov_info *td_kcov_info; /* (*) Kernel code coverage data */ 318 u_int td_ucredref; /* (k) references on td_realucred */ 319 #define td_endzero td_sigmask 320 321 /* Copied during fork1() or create_thread(). */ 322 #define td_startcopy td_endzero 323 sigset_t td_sigmask; /* (c) Current signal mask. */ 324 u_char td_rqindex; /* (t) Run queue index. */ 325 u_char td_base_pri; /* (t) Thread base kernel priority. */ 326 u_char td_priority; /* (t) Thread active priority. */ 327 u_char td_pri_class; /* (t) Scheduling class. */ 328 u_char td_user_pri; /* (t) User pri from estcpu and nice. */ 329 u_char td_base_user_pri; /* (t) Base user pri */ 330 uintptr_t td_rb_list; /* (k) Robust list head. */ 331 uintptr_t td_rbp_list; /* (k) Robust priv list head. */ 332 uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */ 333 struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on 334 fork for child tracing. */ 335 void *td_sigblock_ptr; /* (k) uptr for fast sigblock. */ 336 uint32_t td_sigblock_val; /* (k) fast sigblock value read at 337 td_sigblock_ptr on kern entry */ 338 #define td_endcopy td_pcb 339 340 /* 341 * Fields that must be manually set in fork1() or create_thread() 342 * or already have been set in the allocator, constructor, etc. 343 */ 344 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */ 345 enum td_states { 346 TDS_INACTIVE = 0x0, 347 TDS_INHIBITED, 348 TDS_CAN_RUN, 349 TDS_RUNQ, 350 TDS_RUNNING 351 } td_state; /* (t) thread state */ 352 /* Note: td_state must be accessed using TD_{GET,SET}_STATE(). */ 353 union { 354 register_t tdu_retval[2]; 355 off_t tdu_off; 356 } td_uretoff; /* (k) Syscall aux returns. */ 357 #define td_retval td_uretoff.tdu_retval 358 u_int td_cowgen; /* (k) Generation of COW pointers. */ 359 /* LP64 hole */ 360 struct callout td_slpcallout; /* (h) Callout for sleep. */ 361 struct trapframe *td_frame; /* (k) */ 362 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */ 363 int td_kstack_pages; /* (a) Size of the kstack. */ 364 volatile u_int td_critnest; /* (k*) Critical section nest level. */ 365 struct mdthread td_md; /* (k) Any machine-dependent fields. */ 366 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */ 367 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */ 368 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */ 369 struct vnet *td_vnet; /* (k) Effective vnet. */ 370 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */ 371 struct trapframe *td_intr_frame;/* (k) Frame of the current irq */ 372 struct proc *td_rfppwait_p; /* (k) The vforked child */ 373 struct vm_page **td_ma; /* (k) uio pages held */ 374 int td_ma_cnt; /* (k) size of *td_ma */ 375 /* LP64 hole */ 376 void *td_emuldata; /* Emulator state data */ 377 int td_lastcpu; /* (t) Last cpu we were on. */ 378 int td_oncpu; /* (t) Which cpu we are on. */ 379 void *td_lkpi_task; /* LinuxKPI task struct pointer */ 380 int td_pmcpend; 381 void *td_coredump; /* (c) coredump request. */ 382 off_t td_ktr_io_lim; /* (k) limit for ktrace file size */ 383 #ifdef EPOCH_TRACE 384 SLIST_HEAD(, epoch_tracker) td_epochs; 385 #endif 386 }; 387 388 struct thread0_storage { 389 struct thread t0st_thread; 390 uint64_t t0st_sched[10]; 391 }; 392 393 struct mtx *thread_lock_block(struct thread *); 394 void thread_lock_block_wait(struct thread *); 395 void thread_lock_set(struct thread *, struct mtx *); 396 void thread_lock_unblock(struct thread *, struct mtx *); 397 #define THREAD_LOCK_ASSERT(td, type) \ 398 mtx_assert((td)->td_lock, (type)) 399 400 #define THREAD_LOCK_BLOCKED_ASSERT(td, type) \ 401 do { \ 402 struct mtx *__m = (td)->td_lock; \ 403 if (__m != &blocked_lock) \ 404 mtx_assert(__m, (type)); \ 405 } while (0) 406 407 #ifdef INVARIANTS 408 #define THREAD_LOCKPTR_ASSERT(td, lock) \ 409 do { \ 410 struct mtx *__m; \ 411 __m = (td)->td_lock; \ 412 KASSERT(__m == (lock), \ 413 ("Thread %p lock %p does not match %p", td, __m, (lock))); \ 414 } while (0) 415 416 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) \ 417 do { \ 418 struct mtx *__m; \ 419 __m = (td)->td_lock; \ 420 KASSERT(__m == (lock) || __m == &blocked_lock, \ 421 ("Thread %p lock %p does not match %p", td, __m, (lock))); \ 422 } while (0) 423 424 #define TD_LOCKS_INC(td) ((td)->td_locks++) 425 #define TD_LOCKS_DEC(td) do { \ 426 KASSERT(SCHEDULER_STOPPED_TD(td) || (td)->td_locks > 0, \ 427 ("thread %p owns no locks", (td))); \ 428 (td)->td_locks--; \ 429 } while (0) 430 #else 431 #define THREAD_LOCKPTR_ASSERT(td, lock) 432 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) 433 434 #define TD_LOCKS_INC(td) 435 #define TD_LOCKS_DEC(td) 436 #endif 437 438 /* 439 * Flags kept in td_flags: 440 * To change these you MUST have the scheduler lock. 441 */ 442 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */ 443 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */ 444 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */ 445 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */ 446 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */ 447 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */ 448 #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */ 449 #define TDF_SIGWAIT 0x00000080 /* Ignore ignored signals */ 450 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */ 451 #define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */ 452 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */ 453 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */ 454 #define TDF_KQTICKLED 0x00001000 /* AST drain kqueue taskqueue */ 455 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */ 456 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */ 457 #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */ 458 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */ 459 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */ 460 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */ 461 #define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */ 462 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */ 463 #define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */ 464 #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */ 465 #define TDF_UNUSED23 0x00800000 /* --available-- */ 466 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */ 467 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */ 468 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */ 469 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */ 470 #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */ 471 #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */ 472 #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */ 473 474 /* Userland debug flags */ 475 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */ 476 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */ 477 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */ 478 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */ 479 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */ 480 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */ 481 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new 482 process */ 483 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child 484 only) */ 485 #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */ 486 #define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */ 487 #define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */ 488 #define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */ 489 #define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */ 490 #define TDB_STEP 0x00002000 /* (x86) PSL_T set for PT_STEP */ 491 #define TDB_SSWITCH 0x00004000 /* Suspended in ptracestop */ 492 #define TDB_COREDUMPRQ 0x00008000 /* Coredump request */ 493 494 /* 495 * "Private" flags kept in td_pflags: 496 * These are only written by curthread and thus need no locking. 497 */ 498 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */ 499 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */ 500 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */ 501 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */ 502 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */ 503 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */ 504 #define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */ 505 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */ 506 #define TDP_SIGFASTBLOCK 0x00000100 /* Fast sigblock active */ 507 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */ 508 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */ 509 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */ 510 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */ 511 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */ 512 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */ 513 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */ 514 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */ 515 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */ 516 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */ 517 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */ 518 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */ 519 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */ 520 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */ 521 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */ 522 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */ 523 #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */ 524 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */ 525 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */ 526 #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */ 527 #define TDP_UNUSED0 0x20000000 /* UNUSED */ 528 #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */ 529 #define TDP_SIGFASTPENDING 0x80000000 /* Pending signal due to sigfastblock */ 530 531 #define TDP2_SBPAGES 0x00000001 /* Owns sbusy on some pages */ 532 #define TDP2_COMPAT32RB 0x00000002 /* compat32 ABI for robust lists */ 533 #define TDP2_ACCT 0x00000004 /* Doing accounting */ 534 535 /* 536 * Reasons that the current thread can not be run yet. 537 * More than one may apply. 538 */ 539 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */ 540 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */ 541 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */ 542 #define TDI_LOCK 0x0008 /* Stopped on a lock. */ 543 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */ 544 545 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING) 546 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL) 547 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED) 548 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED) 549 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK) 550 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT) 551 #ifdef _KERNEL 552 #define TD_GET_STATE(td) atomic_load_int(&(td)->td_state) 553 #else 554 #define TD_GET_STATE(td) ((td)->td_state) 555 #endif 556 #define TD_IS_RUNNING(td) (TD_GET_STATE(td) == TDS_RUNNING) 557 #define TD_ON_RUNQ(td) (TD_GET_STATE(td) == TDS_RUNQ) 558 #define TD_CAN_RUN(td) (TD_GET_STATE(td) == TDS_CAN_RUN) 559 #define TD_IS_INHIBITED(td) (TD_GET_STATE(td) == TDS_INHIBITED) 560 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED) 561 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD) 562 563 #define TD_CAN_ABORT(td) (TD_ON_SLEEPQ((td)) && \ 564 ((td)->td_flags & TDF_SINTR) != 0) 565 566 #define KTDSTATE(td) \ 567 (((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \ 568 ((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \ 569 ((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" : \ 570 ((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \ 571 ((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding") 572 573 #define TD_SET_INHIB(td, inhib) do { \ 574 TD_SET_STATE(td, TDS_INHIBITED); \ 575 (td)->td_inhibitors |= (inhib); \ 576 } while (0) 577 578 #define TD_CLR_INHIB(td, inhib) do { \ 579 if (((td)->td_inhibitors & (inhib)) && \ 580 (((td)->td_inhibitors &= ~(inhib)) == 0)) \ 581 TD_SET_STATE(td, TDS_CAN_RUN); \ 582 } while (0) 583 584 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING) 585 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED) 586 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK) 587 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED) 588 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT) 589 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING) 590 591 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING) 592 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED) 593 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK) 594 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED) 595 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT) 596 597 #ifdef _KERNEL 598 #define TD_SET_STATE(td, state) atomic_store_int(&(td)->td_state, state) 599 #else 600 #define TD_SET_STATE(td, state) (td)->td_state = state 601 #endif 602 #define TD_SET_RUNNING(td) TD_SET_STATE(td, TDS_RUNNING) 603 #define TD_SET_RUNQ(td) TD_SET_STATE(td, TDS_RUNQ) 604 #define TD_SET_CAN_RUN(td) TD_SET_STATE(td, TDS_CAN_RUN) 605 606 607 #define TD_SBDRY_INTR(td) \ 608 (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0) 609 #define TD_SBDRY_ERRNO(td) \ 610 (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART) 611 612 /* 613 * Process structure. 614 */ 615 struct proc { 616 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */ 617 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */ 618 struct mtx p_slock; /* process spin lock */ 619 struct ucred *p_ucred; /* (c) Process owner's identity. */ 620 struct filedesc *p_fd; /* (b) Open files. */ 621 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */ 622 struct pwddesc *p_pd; /* (b) Cwd, chroot, jail, umask */ 623 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */ 624 struct plimit *p_limit; /* (c) Resource limits. */ 625 struct callout p_limco; /* (c) Limit callout handle */ 626 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */ 627 628 int p_flag; /* (c) P_* flags. */ 629 int p_flag2; /* (c) P2_* flags. */ 630 enum p_states { 631 PRS_NEW = 0, /* In creation */ 632 PRS_NORMAL, /* threads can be run. */ 633 PRS_ZOMBIE 634 } p_state; /* (j/c) Process status. */ 635 pid_t p_pid; /* (b) Process identifier. */ 636 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */ 637 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */ 638 struct proc *p_pptr; /* (c + e) Pointer to parent process. */ 639 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */ 640 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */ 641 struct proc *p_reaper; /* (e) My reaper. */ 642 LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants 643 (if I am reaper). */ 644 LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of 645 the same reaper. */ 646 struct mtx p_mtx; /* (n) Lock for this struct. */ 647 struct mtx p_statmtx; /* Lock for the stats */ 648 struct mtx p_itimmtx; /* Lock for the virt/prof timers */ 649 struct mtx p_profmtx; /* Lock for the profiling */ 650 struct ksiginfo *p_ksi; /* Locked by parent proc lock */ 651 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */ 652 #define p_siglist p_sigqueue.sq_signals 653 pid_t p_oppid; /* (c + e) Real parent pid. */ 654 655 /* The following fields are all zeroed upon creation in fork. */ 656 #define p_startzero p_vmspace 657 struct vmspace *p_vmspace; /* (b) Address space. */ 658 u_int p_swtick; /* (c) Tick when swapped in or out. */ 659 u_int p_cowgen; /* (c) Generation of COW pointers. */ 660 struct itimerval p_realtimer; /* (c) Alarm timer. */ 661 struct rusage p_ru; /* (a) Exit information. */ 662 struct rusage_ext p_rux; /* (cu) Internal resource usage. */ 663 struct rusage_ext p_crux; /* (c) Internal child resource usage. */ 664 int p_profthreads; /* (c) Num threads in addupc_task. */ 665 volatile int p_exitthreads; /* (j) Number of threads exiting */ 666 int p_traceflag; /* (o) Kernel trace points. */ 667 struct ktr_io_params *p_ktrioparms; /* (c + o) Params for ktrace. */ 668 struct vnode *p_textvp; /* (b) Vnode of executable. */ 669 struct vnode *p_textdvp; /* (b) Dir containing textvp. */ 670 char *p_binname; /* (b) Binary hardlink name. */ 671 u_int p_lock; /* (c) Proclock (prevent swap) count. */ 672 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */ 673 int p_sigparent; /* (c) Signal to parent on exit. */ 674 int p_sig; /* (n) For core dump/debugger XXX. */ 675 u_int p_ptevents; /* (c + e) ptrace() event mask. */ 676 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */ 677 struct thread *p_singlethread;/* (c + j) If single threading this is it */ 678 int p_suspcount; /* (j) Num threads in suspended mode. */ 679 struct thread *p_xthread; /* (c) Trap thread */ 680 int p_boundary_count;/* (j) Num threads at user boundary */ 681 int p_pendingcnt; /* how many signals are pending */ 682 struct itimers *p_itimers; /* (c) POSIX interval timers. */ 683 struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */ 684 u_int p_treeflag; /* (e) P_TREE flags */ 685 int p_pendingexits; /* (c) Count of pending thread exits. */ 686 struct filemon *p_filemon; /* (c) filemon-specific data. */ 687 int p_pdeathsig; /* (c) Signal from parent on exit. */ 688 /* End area that is zeroed on creation. */ 689 #define p_endzero p_magic 690 691 /* The following fields are all copied upon creation in fork. */ 692 #define p_startcopy p_endzero 693 u_int p_magic; /* (b) Magic number. */ 694 int p_osrel; /* (x) osreldate for the 695 binary (from ELF note, if any) */ 696 uint32_t p_fctl0; /* (x) ABI feature control, ELF note */ 697 char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */ 698 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */ 699 struct pargs *p_args; /* (c) Process arguments. */ 700 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */ 701 signed char p_nice; /* (c) Process "nice" value. */ 702 int p_fibnum; /* in this routing domain XXX MRT */ 703 pid_t p_reapsubtree; /* (e) Pid of the direct child of the 704 reaper which spawned 705 our subtree. */ 706 uint64_t p_elf_flags; /* (x) ELF flags */ 707 void *p_elf_brandinfo; /* (x) Elf_Brandinfo, NULL for 708 non ELF binaries. */ 709 /* End area that is copied on creation. */ 710 #define p_endcopy p_xexit 711 712 u_int p_xexit; /* (c) Exit code. */ 713 u_int p_xsig; /* (c) Stop/kill sig. */ 714 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */ 715 struct knlist *p_klist; /* (c) Knotes attached to this proc. */ 716 int p_numthreads; /* (c) Number of threads. */ 717 struct mdproc p_md; /* Any machine-dependent fields. */ 718 struct callout p_itcallout; /* (h + c) Interval timer callout. */ 719 u_short p_acflag; /* (c) Accounting flags. */ 720 struct proc *p_peers; /* (r) */ 721 struct proc *p_leader; /* (b) */ 722 void *p_emuldata; /* (c) Emulator state data. */ 723 struct label *p_label; /* (*) Proc (not subject) MAC label. */ 724 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */ 725 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/ 726 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */ 727 struct cv p_pwait; /* (*) wait cv for exit/exec. */ 728 uint64_t p_prev_runtime; /* (c) Resource usage accounting. */ 729 struct racct *p_racct; /* (b) Resource accounting. */ 730 int p_throttled; /* (c) Flag for racct pcpu throttling */ 731 /* 732 * An orphan is the child that has been re-parented to the 733 * debugger as a result of attaching to it. Need to keep 734 * track of them for parent to be able to collect the exit 735 * status of what used to be children. 736 */ 737 LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */ 738 LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */ 739 740 TAILQ_HEAD(, kq_timer_cb_data) p_kqtim_stop; /* (c) */ 741 }; 742 743 #define p_session p_pgrp->pg_session 744 #define p_pgid p_pgrp->pg_id 745 746 #define NOCPU (-1) /* For when we aren't on a CPU. */ 747 #define NOCPU_OLD (255) 748 #define MAXCPU_OLD (254) 749 750 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock) 751 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock) 752 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type)) 753 754 #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx) 755 #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx) 756 #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type)) 757 758 #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx) 759 #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx) 760 #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type)) 761 762 #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx) 763 #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx) 764 #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type)) 765 766 /* These flags are kept in p_flag. */ 767 #define P_ADVLOCK 0x00000001 /* Process may hold a POSIX advisory 768 lock. */ 769 #define P_CONTROLT 0x00000002 /* Has a controlling terminal. */ 770 #define P_KPROC 0x00000004 /* Kernel process. */ 771 #define P_UNUSED3 0x00000008 /* --available-- */ 772 #define P_PPWAIT 0x00000010 /* Parent is waiting for child to 773 exec/exit. */ 774 #define P_PROFIL 0x00000020 /* Has started profiling. */ 775 #define P_STOPPROF 0x00000040 /* Has thread requesting to stop 776 profiling. */ 777 #define P_HADTHREADS 0x00000080 /* Has had threads (no cleanup 778 shortcuts) */ 779 #define P_SUGID 0x00000100 /* Had set id privileges since last 780 exec. */ 781 #define P_SYSTEM 0x00000200 /* System proc: no sigs, stats or 782 swapping. */ 783 #define P_SINGLE_EXIT 0x00000400 /* Threads suspending should exit, 784 not wait. */ 785 #define P_TRACED 0x00000800 /* Debugged process being traced. */ 786 #define P_WAITED 0x00001000 /* Someone is waiting for us. */ 787 #define P_WEXIT 0x00002000 /* Working on exiting. */ 788 #define P_EXEC 0x00004000 /* Process called exec. */ 789 #define P_WKILLED 0x00008000 /* Killed, go to kernel/user boundary 790 ASAP. */ 791 #define P_CONTINUED 0x00010000 /* Proc has continued from a stopped 792 state. */ 793 #define P_STOPPED_SIG 0x00020000 /* Stopped due to SIGSTOP/SIGTSTP. */ 794 #define P_STOPPED_TRACE 0x00040000 /* Stopped because of tracing. */ 795 #define P_STOPPED_SINGLE 0x00080000 /* Only 1 thread can continue (not to 796 user). */ 797 #define P_PROTECTED 0x00100000 /* Do not kill on memory overcommit. */ 798 #define P_SIGEVENT 0x00200000 /* Process pending signals changed. */ 799 #define P_SINGLE_BOUNDARY 0x00400000 /* Threads should suspend at user 800 boundary. */ 801 #define P_HWPMC 0x00800000 /* Process is using HWPMCs */ 802 #define P_JAILED 0x01000000 /* Process is in jail. */ 803 #define P_TOTAL_STOP 0x02000000 /* Stopped in stop_all_proc. */ 804 #define P_INEXEC 0x04000000 /* Process is in execve(). */ 805 #define P_STATCHILD 0x08000000 /* Child process stopped or exited. */ 806 #define P_INMEM 0x10000000 /* Loaded into memory. */ 807 #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */ 808 #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */ 809 #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */ 810 811 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE) 812 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED) 813 #define P_KILLED(p) ((p)->p_flag & P_WKILLED) 814 815 /* These flags are kept in p_flag2. */ 816 #define P2_INHERIT_PROTECTED 0x00000001 /* New children get 817 P_PROTECTED. */ 818 #define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or 819 coredumps. */ 820 #define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on 821 exec(2). */ 822 #define P2_AST_SU 0x00000008 /* Handles SU ast for 823 kthreads. */ 824 #define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not 825 yet handled. */ 826 #define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */ 827 #define P2_ASLR_ENABLE 0x00000040 /* Force enable ASLR. */ 828 #define P2_ASLR_DISABLE 0x00000080 /* Force disable ASLR. */ 829 #define P2_ASLR_IGNSTART 0x00000100 /* Enable ASLR to consume sbrk 830 area. */ 831 #define P2_PROTMAX_ENABLE 0x00000200 /* Force enable implied 832 PROT_MAX. */ 833 #define P2_PROTMAX_DISABLE 0x00000400 /* Force disable implied 834 PROT_MAX. */ 835 #define P2_STKGAP_DISABLE 0x00000800 /* Disable stack gap for 836 MAP_STACK */ 837 #define P2_STKGAP_DISABLE_EXEC 0x00001000 /* Stack gap disabled 838 after exec */ 839 #define P2_ITSTOPPED 0x00002000 840 #define P2_PTRACEREQ 0x00004000 /* Active ptrace req */ 841 #define P2_NO_NEW_PRIVS 0x00008000 /* Ignore setuid */ 842 #define P2_WXORX_DISABLE 0x00010000 /* WX mappings enabled */ 843 #define P2_WXORX_ENABLE_EXEC 0x00020000 /* WXORX enabled after exec */ 844 845 /* Flags protected by proctree_lock, kept in p_treeflags. */ 846 #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */ 847 #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan 848 list */ 849 #define P_TREE_REAPER 0x00000004 /* Reaper of subtree */ 850 #define P_TREE_GRPEXITED 0x00000008 /* exit1() done with job ctl */ 851 852 /* 853 * These were process status values (p_stat), now they are only used in 854 * legacy conversion code. 855 */ 856 #define SIDL 1 /* Process being created by fork. */ 857 #define SRUN 2 /* Currently runnable. */ 858 #define SSLEEP 3 /* Sleeping on an address. */ 859 #define SSTOP 4 /* Process debugging or suspension. */ 860 #define SZOMB 5 /* Awaiting collection by parent. */ 861 #define SWAIT 6 /* Waiting for interrupt. */ 862 #define SLOCK 7 /* Blocked on a lock. */ 863 864 #define P_MAGIC 0xbeefface 865 866 #ifdef _KERNEL 867 868 /* Types and flags for mi_switch(). */ 869 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */ 870 #define SWT_NONE 0 /* Unspecified switch. */ 871 #define SWT_PREEMPT 1 /* Switching due to preemption. */ 872 #define SWT_OWEPREEMPT 2 /* Switching due to owepreempt. */ 873 #define SWT_TURNSTILE 3 /* Turnstile contention. */ 874 #define SWT_SLEEPQ 4 /* Sleepq wait. */ 875 #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */ 876 #define SWT_RELINQUISH 6 /* yield call. */ 877 #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */ 878 #define SWT_IDLE 8 /* Switching from the idle thread. */ 879 #define SWT_IWAIT 9 /* Waiting for interrupts. */ 880 #define SWT_SUSPEND 10 /* Thread suspended. */ 881 #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */ 882 #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */ 883 #define SWT_COUNT 13 /* Number of switch types. */ 884 /* Flags */ 885 #define SW_VOL 0x0100 /* Voluntary switch. */ 886 #define SW_INVOL 0x0200 /* Involuntary switch. */ 887 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */ 888 889 /* How values for thread_single(). */ 890 #define SINGLE_NO_EXIT 0 891 #define SINGLE_EXIT 1 892 #define SINGLE_BOUNDARY 2 893 #define SINGLE_ALLPROC 3 894 895 #ifdef MALLOC_DECLARE 896 MALLOC_DECLARE(M_PARGS); 897 MALLOC_DECLARE(M_SESSION); 898 MALLOC_DECLARE(M_SUBPROC); 899 #endif 900 901 #define FOREACH_PROC_IN_SYSTEM(p) \ 902 LIST_FOREACH((p), &allproc, p_list) 903 #define FOREACH_THREAD_IN_PROC(p, td) \ 904 TAILQ_FOREACH((td), &(p)->p_threads, td_plist) 905 906 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads) 907 908 /* 909 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit 910 * in a pid_t, as it is used to represent "no process group". 911 */ 912 #define PID_MAX 99999 913 #define NO_PID 100000 914 #define THREAD0_TID NO_PID 915 extern pid_t pid_max; 916 917 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p)) 918 919 /* Lock and unlock a process. */ 920 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx) 921 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx) 922 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx) 923 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx) 924 #define PROC_WAIT_UNLOCKED(p) mtx_wait_unlocked(&(p)->p_mtx) 925 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type)) 926 927 /* Lock and unlock a process group. */ 928 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx) 929 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx) 930 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx) 931 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type)) 932 933 #define PGRP_LOCK_PGSIGNAL(pg) do { \ 934 if ((pg) != NULL) \ 935 PGRP_LOCK(pg); \ 936 } while (0) 937 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \ 938 if ((pg) != NULL) \ 939 PGRP_UNLOCK(pg); \ 940 } while (0) 941 942 /* Lock and unlock a session. */ 943 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx) 944 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx) 945 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx) 946 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type)) 947 948 /* 949 * Non-zero p_lock ensures that: 950 * - exit1() is not performed until p_lock reaches zero; 951 * - the process' threads stack are not swapped out if they are currently 952 * not (P_INMEM). 953 * 954 * PHOLD() asserts that the process (except the current process) is 955 * not exiting, increments p_lock and swaps threads stacks into memory, 956 * if needed. 957 * _PHOLD() is same as PHOLD(), it takes the process locked. 958 * _PHOLD_LITE() also takes the process locked, but comparing with 959 * _PHOLD(), it only guarantees that exit1() is not executed, 960 * faultin() is not called. 961 */ 962 #define PHOLD(p) do { \ 963 PROC_LOCK(p); \ 964 _PHOLD(p); \ 965 PROC_UNLOCK(p); \ 966 } while (0) 967 #define _PHOLD(p) do { \ 968 PROC_LOCK_ASSERT((p), MA_OWNED); \ 969 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 970 ("PHOLD of exiting process %p", p)); \ 971 (p)->p_lock++; \ 972 if (((p)->p_flag & P_INMEM) == 0) \ 973 faultin((p)); \ 974 } while (0) 975 #define _PHOLD_LITE(p) do { \ 976 PROC_LOCK_ASSERT((p), MA_OWNED); \ 977 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 978 ("PHOLD of exiting process %p", p)); \ 979 (p)->p_lock++; \ 980 } while (0) 981 #define PROC_ASSERT_HELD(p) do { \ 982 KASSERT((p)->p_lock > 0, ("process %p not held", p)); \ 983 } while (0) 984 985 #define PRELE(p) do { \ 986 PROC_LOCK((p)); \ 987 _PRELE((p)); \ 988 PROC_UNLOCK((p)); \ 989 } while (0) 990 #define _PRELE(p) do { \ 991 PROC_LOCK_ASSERT((p), MA_OWNED); \ 992 PROC_ASSERT_HELD(p); \ 993 (--(p)->p_lock); \ 994 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \ 995 wakeup(&(p)->p_lock); \ 996 } while (0) 997 #define PROC_ASSERT_NOT_HELD(p) do { \ 998 KASSERT((p)->p_lock == 0, ("process %p held", p)); \ 999 } while (0) 1000 1001 #define PROC_UPDATE_COW(p) do { \ 1002 PROC_LOCK_ASSERT((p), MA_OWNED); \ 1003 (p)->p_cowgen++; \ 1004 } while (0) 1005 1006 /* Check whether a thread is safe to be swapped out. */ 1007 #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP) 1008 1009 /* Control whether or not it is safe for curthread to sleep. */ 1010 #define THREAD_NO_SLEEPING() do { \ 1011 curthread->td_no_sleeping++; \ 1012 MPASS(curthread->td_no_sleeping > 0); \ 1013 } while (0) 1014 1015 #define THREAD_SLEEPING_OK() do { \ 1016 MPASS(curthread->td_no_sleeping > 0); \ 1017 curthread->td_no_sleeping--; \ 1018 } while (0) 1019 1020 #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0) 1021 1022 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash]) 1023 #define PIDHASHLOCK(pid) (&pidhashtbl_lock[((pid) & pidhashlock)]) 1024 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl; 1025 extern struct sx *pidhashtbl_lock; 1026 extern u_long pidhash; 1027 extern u_long pidhashlock; 1028 1029 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash]) 1030 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl; 1031 extern u_long pgrphash; 1032 1033 extern struct sx allproc_lock; 1034 extern int allproc_gen; 1035 extern struct sx proctree_lock; 1036 extern struct mtx ppeers_lock; 1037 extern struct mtx procid_lock; 1038 extern struct proc proc0; /* Process slot for swapper. */ 1039 extern struct thread0_storage thread0_st; /* Primary thread in proc0. */ 1040 #define thread0 (thread0_st.t0st_thread) 1041 extern struct vmspace vmspace0; /* VM space for proc0. */ 1042 extern int hogticks; /* Limit on kernel cpu hogs. */ 1043 extern int lastpid; 1044 extern int nprocs, maxproc; /* Current and max number of procs. */ 1045 extern int maxprocperuid; /* Max procs per uid. */ 1046 extern u_long ps_arg_cache_limit; 1047 1048 LIST_HEAD(proclist, proc); 1049 TAILQ_HEAD(procqueue, proc); 1050 TAILQ_HEAD(threadqueue, thread); 1051 extern struct proclist allproc; /* List of all processes. */ 1052 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */ 1053 1054 extern struct uma_zone *proc_zone; 1055 extern struct uma_zone *pgrp_zone; 1056 1057 struct proc *pfind(pid_t); /* Find process by id. */ 1058 struct proc *pfind_any(pid_t); /* Find (zombie) process by id. */ 1059 struct proc *pfind_any_locked(pid_t pid); /* Find process by id, locked. */ 1060 struct pgrp *pgfind(pid_t); /* Find process group by id. */ 1061 void pidhash_slockall(void); /* Shared lock all pid hash lists. */ 1062 void pidhash_sunlockall(void); /* Shared unlock all pid hash lists. */ 1063 1064 struct fork_req { 1065 int fr_flags; 1066 int fr_pages; 1067 int *fr_pidp; 1068 struct proc **fr_procp; 1069 int *fr_pd_fd; 1070 int fr_pd_flags; 1071 struct filecaps *fr_pd_fcaps; 1072 int fr_flags2; 1073 #define FR2_DROPSIG_CAUGHT 0x00000001 /* Drop caught non-DFL signals */ 1074 #define FR2_SHARE_PATHS 0x00000002 /* Invert sense of RFFDG for paths */ 1075 #define FR2_KPROC 0x00000004 /* Create a kernel process */ 1076 }; 1077 1078 /* 1079 * pget() flags. 1080 */ 1081 #define PGET_HOLD 0x00001 /* Hold the process. */ 1082 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */ 1083 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */ 1084 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */ 1085 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */ 1086 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */ 1087 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */ 1088 1089 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT) 1090 1091 int pget(pid_t pid, int flags, struct proc **pp); 1092 1093 void ast(struct trapframe *framep); 1094 struct thread *choosethread(void); 1095 int cr_cansee(struct ucred *u1, struct ucred *u2); 1096 int cr_canseesocket(struct ucred *cred, struct socket *so); 1097 int cr_canseeothergids(struct ucred *u1, struct ucred *u2); 1098 int cr_canseeotheruids(struct ucred *u1, struct ucred *u2); 1099 int cr_canseejailproc(struct ucred *u1, struct ucred *u2); 1100 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum); 1101 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp, 1102 struct session *sess); 1103 int enterthispgrp(struct proc *p, struct pgrp *pgrp); 1104 void faultin(struct proc *p); 1105 int fork1(struct thread *, struct fork_req *); 1106 void fork_rfppwait(struct thread *); 1107 void fork_exit(void (*)(void *, struct trapframe *), void *, 1108 struct trapframe *); 1109 void fork_return(struct thread *, struct trapframe *); 1110 int inferior(struct proc *p); 1111 void itimer_proc_continue(struct proc *p); 1112 void kqtimer_proc_continue(struct proc *p); 1113 void kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry, 1114 int *resident_count, bool *super); 1115 void kern_yield(int); 1116 void kick_proc0(void); 1117 void killjobc(void); 1118 int leavepgrp(struct proc *p); 1119 int maybe_preempt(struct thread *td); 1120 void maybe_yield(void); 1121 void mi_switch(int flags); 1122 int p_candebug(struct thread *td, struct proc *p); 1123 int p_cansee(struct thread *td, struct proc *p); 1124 int p_cansched(struct thread *td, struct proc *p); 1125 int p_cansignal(struct thread *td, struct proc *p, int signum); 1126 int p_canwait(struct thread *td, struct proc *p); 1127 struct pargs *pargs_alloc(int len); 1128 void pargs_drop(struct pargs *pa); 1129 void pargs_hold(struct pargs *pa); 1130 void proc_add_orphan(struct proc *child, struct proc *parent); 1131 int proc_get_binpath(struct proc *p, char *binname, char **fullpath, 1132 char **freepath); 1133 int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb); 1134 int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb); 1135 int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb); 1136 void procinit(void); 1137 int proc_iterate(int (*cb)(struct proc *, void *), void *cbarg); 1138 void proc_linkup0(struct proc *p, struct thread *td); 1139 void proc_linkup(struct proc *p, struct thread *td); 1140 struct proc *proc_realparent(struct proc *child); 1141 void proc_reap(struct thread *td, struct proc *p, int *status, int options); 1142 void proc_reparent(struct proc *child, struct proc *newparent, bool set_oppid); 1143 void proc_set_traced(struct proc *p, bool stop); 1144 void proc_wkilled(struct proc *p); 1145 struct pstats *pstats_alloc(void); 1146 void pstats_fork(struct pstats *src, struct pstats *dst); 1147 void pstats_free(struct pstats *ps); 1148 void proc_clear_orphan(struct proc *p); 1149 void reaper_abandon_children(struct proc *p, bool exiting); 1150 int securelevel_ge(struct ucred *cr, int level); 1151 int securelevel_gt(struct ucred *cr, int level); 1152 void sess_hold(struct session *); 1153 void sess_release(struct session *); 1154 int setrunnable(struct thread *, int); 1155 void setsugid(struct proc *p); 1156 int should_yield(void); 1157 int sigonstack(size_t sp); 1158 void stopevent(struct proc *, u_int, u_int); 1159 struct thread *tdfind(lwpid_t, pid_t); 1160 void threadinit(void); 1161 void tidhash_add(struct thread *); 1162 void tidhash_remove(struct thread *); 1163 void cpu_idle(int); 1164 int cpu_idle_wakeup(int); 1165 extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */ 1166 void cpu_switch(struct thread *, struct thread *, struct mtx *); 1167 void cpu_throw(struct thread *, struct thread *) __dead2; 1168 bool curproc_sigkilled(void); 1169 void userret(struct thread *, struct trapframe *); 1170 1171 void cpu_exit(struct thread *); 1172 void exit1(struct thread *, int, int) __dead2; 1173 void cpu_copy_thread(struct thread *td, struct thread *td0); 1174 bool cpu_exec_vmspace_reuse(struct proc *p, struct vm_map *map); 1175 int cpu_fetch_syscall_args(struct thread *td); 1176 void cpu_fork(struct thread *, struct proc *, struct thread *, int); 1177 void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *); 1178 int cpu_procctl(struct thread *td, int idtype, id_t id, int com, 1179 void *data); 1180 void cpu_set_syscall_retval(struct thread *, int); 1181 void cpu_set_upcall(struct thread *, void (*)(void *), void *, 1182 stack_t *); 1183 int cpu_set_user_tls(struct thread *, void *tls_base); 1184 void cpu_thread_alloc(struct thread *); 1185 void cpu_thread_clean(struct thread *); 1186 void cpu_thread_exit(struct thread *); 1187 void cpu_thread_free(struct thread *); 1188 void cpu_thread_swapin(struct thread *); 1189 void cpu_thread_swapout(struct thread *); 1190 struct thread *thread_alloc(int pages); 1191 int thread_alloc_stack(struct thread *, int pages); 1192 int thread_check_susp(struct thread *td, bool sleep); 1193 void thread_cow_get_proc(struct thread *newtd, struct proc *p); 1194 void thread_cow_get(struct thread *newtd, struct thread *td); 1195 void thread_cow_free(struct thread *td); 1196 void thread_cow_update(struct thread *td); 1197 int thread_create(struct thread *td, struct rtprio *rtp, 1198 int (*initialize_thread)(struct thread *, void *), void *thunk); 1199 void thread_exit(void) __dead2; 1200 void thread_free(struct thread *td); 1201 void thread_link(struct thread *td, struct proc *p); 1202 void thread_reap_barrier(void); 1203 int thread_single(struct proc *p, int how); 1204 void thread_single_end(struct proc *p, int how); 1205 void thread_stash(struct thread *td); 1206 void thread_stopped(struct proc *p); 1207 void childproc_stopped(struct proc *child, int reason); 1208 void childproc_continued(struct proc *child); 1209 void childproc_exited(struct proc *child); 1210 void thread_run_flash(struct thread *td); 1211 int thread_suspend_check(int how); 1212 bool thread_suspend_check_needed(void); 1213 void thread_suspend_switch(struct thread *, struct proc *p); 1214 void thread_suspend_one(struct thread *td); 1215 void thread_unlink(struct thread *td); 1216 void thread_unsuspend(struct proc *p); 1217 void thread_wait(struct proc *p); 1218 1219 void stop_all_proc(void); 1220 void resume_all_proc(void); 1221 1222 static __inline int 1223 curthread_pflags_set(int flags) 1224 { 1225 struct thread *td; 1226 int save; 1227 1228 td = curthread; 1229 save = ~flags | (td->td_pflags & flags); 1230 td->td_pflags |= flags; 1231 return (save); 1232 } 1233 1234 static __inline void 1235 curthread_pflags_restore(int save) 1236 { 1237 1238 curthread->td_pflags &= save; 1239 } 1240 1241 static __inline int 1242 curthread_pflags2_set(int flags) 1243 { 1244 struct thread *td; 1245 int save; 1246 1247 td = curthread; 1248 save = ~flags | (td->td_pflags2 & flags); 1249 td->td_pflags2 |= flags; 1250 return (save); 1251 } 1252 1253 static __inline void 1254 curthread_pflags2_restore(int save) 1255 { 1256 1257 curthread->td_pflags2 &= save; 1258 } 1259 1260 static __inline bool 1261 kstack_contains(struct thread *td, vm_offset_t va, size_t len) 1262 { 1263 return (va >= td->td_kstack && va + len >= va && 1264 va + len <= td->td_kstack + td->td_kstack_pages * PAGE_SIZE); 1265 } 1266 1267 static __inline __pure2 struct td_sched * 1268 td_get_sched(struct thread *td) 1269 { 1270 1271 return ((struct td_sched *)&td[1]); 1272 } 1273 1274 extern void (*softdep_ast_cleanup)(struct thread *); 1275 static __inline void 1276 td_softdep_cleanup(struct thread *td) 1277 { 1278 1279 if (td->td_su != NULL && softdep_ast_cleanup != NULL) 1280 softdep_ast_cleanup(td); 1281 } 1282 1283 #define PROC_ID_PID 0 1284 #define PROC_ID_GROUP 1 1285 #define PROC_ID_SESSION 2 1286 #define PROC_ID_REAP 3 1287 1288 void proc_id_set(int type, pid_t id); 1289 void proc_id_set_cond(int type, pid_t id); 1290 void proc_id_clear(int type, pid_t id); 1291 1292 EVENTHANDLER_LIST_DECLARE(process_ctor); 1293 EVENTHANDLER_LIST_DECLARE(process_dtor); 1294 EVENTHANDLER_LIST_DECLARE(process_init); 1295 EVENTHANDLER_LIST_DECLARE(process_fini); 1296 EVENTHANDLER_LIST_DECLARE(process_exit); 1297 EVENTHANDLER_LIST_DECLARE(process_fork); 1298 EVENTHANDLER_LIST_DECLARE(process_exec); 1299 1300 EVENTHANDLER_LIST_DECLARE(thread_ctor); 1301 EVENTHANDLER_LIST_DECLARE(thread_dtor); 1302 EVENTHANDLER_LIST_DECLARE(thread_init); 1303 1304 #endif /* _KERNEL */ 1305 1306 #endif /* !_SYS_PROC_H_ */ 1307