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 int pg_jobc; /* (m) Job control process count. */ 115 struct mtx pg_mtx; /* Mutex to protect members */ 116 }; 117 118 /* 119 * pargs, used to hold a copy of the command line, if it had a sane length. 120 */ 121 struct pargs { 122 u_int ar_ref; /* Reference count. */ 123 u_int ar_length; /* Length. */ 124 u_char ar_args[1]; /* Arguments. */ 125 }; 126 127 /*- 128 * Description of a process. 129 * 130 * This structure contains the information needed to manage a thread of 131 * control, known in UN*X as a process; it has references to substructures 132 * containing descriptions of things that the process uses, but may share 133 * with related processes. The process structure and the substructures 134 * are always addressable except for those marked "(CPU)" below, 135 * which might be addressable only on a processor on which the process 136 * is running. 137 * 138 * Below is a key of locks used to protect each member of struct proc. The 139 * lock is indicated by a reference to a specific character in parens in the 140 * associated comment. 141 * * - not yet protected 142 * a - only touched by curproc or parent during fork/wait 143 * b - created at fork, never changes 144 * (exception aiods switch vmspaces, but they are also 145 * marked 'P_SYSTEM' so hopefully it will be left alone) 146 * c - locked by proc mtx 147 * d - locked by allproc_lock lock 148 * e - locked by proctree_lock lock 149 * f - session mtx 150 * g - process group mtx 151 * h - callout_lock mtx 152 * i - by curproc or the master session mtx 153 * j - locked by proc slock 154 * k - only accessed by curthread 155 * k*- only accessed by curthread and from an interrupt 156 * kx- only accessed by curthread and by debugger 157 * l - the attaching proc or attaching proc parent 158 * m - Giant 159 * n - not locked, lazy 160 * o - ktrace lock 161 * q - td_contested lock 162 * r - p_peers lock 163 * s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9) 164 * t - thread lock 165 * u - process stat lock 166 * w - process timer lock 167 * x - created at fork, only changes during single threading in exec 168 * y - created at first aio, doesn't change until exit or exec at which 169 * point we are single-threaded and only curthread changes it 170 * z - zombie threads lock 171 * 172 * If the locking key specifies two identifiers (for example, p_pptr) then 173 * either lock is sufficient for read access, but both locks must be held 174 * for write access. 175 */ 176 struct cpuset; 177 struct filecaps; 178 struct filemon; 179 struct kaioinfo; 180 struct kaudit_record; 181 struct kcov_info; 182 struct kdtrace_proc; 183 struct kdtrace_thread; 184 struct mqueue_notifier; 185 struct p_sched; 186 struct proc; 187 struct procdesc; 188 struct racct; 189 struct sbuf; 190 struct sleepqueue; 191 struct socket; 192 struct syscall_args; 193 struct td_sched; 194 struct thread; 195 struct trapframe; 196 struct turnstile; 197 struct vm_map; 198 struct vm_map_entry; 199 struct epoch_tracker; 200 201 /* 202 * XXX: Does this belong in resource.h or resourcevar.h instead? 203 * Resource usage extension. The times in rusage structs in the kernel are 204 * never up to date. The actual times are kept as runtimes and tick counts 205 * (with control info in the "previous" times), and are converted when 206 * userland asks for rusage info. Backwards compatibility prevents putting 207 * this directly in the user-visible rusage struct. 208 * 209 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux. 210 * Locking for td_rux: (t) for all fields. 211 */ 212 struct rusage_ext { 213 uint64_t rux_runtime; /* (cu) Real time. */ 214 uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */ 215 uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */ 216 uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */ 217 uint64_t rux_uu; /* (c) Previous user time in usec. */ 218 uint64_t rux_su; /* (c) Previous sys time in usec. */ 219 uint64_t rux_tu; /* (c) Previous total time in usec. */ 220 }; 221 222 /* 223 * Kernel runnable context (thread). 224 * This is what is put to sleep and reactivated. 225 * Thread context. Processes may have multiple threads. 226 */ 227 struct thread { 228 struct mtx *volatile td_lock; /* replaces sched lock */ 229 struct proc *td_proc; /* (*) Associated process. */ 230 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */ 231 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */ 232 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */ 233 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */ 234 LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */ 235 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */ 236 struct domainset_ref td_domain; /* (a) NUMA policy */ 237 struct seltd *td_sel; /* Select queue/channel. */ 238 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */ 239 struct turnstile *td_turnstile; /* (k) Associated turnstile. */ 240 struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */ 241 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */ 242 lwpid_t td_tid; /* (b) Thread ID. */ 243 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */ 244 #define td_siglist td_sigqueue.sq_signals 245 u_char td_lend_user_pri; /* (t) Lend user pri. */ 246 247 /* Cleared during fork1() */ 248 #define td_startzero td_flags 249 int td_flags; /* (t) TDF_* flags. */ 250 int td_inhibitors; /* (t) Why can not run. */ 251 int td_pflags; /* (k) Private thread (TDP_*) flags. */ 252 int td_pflags2; /* (k) Private thread (TDP2_*) flags. */ 253 int td_dupfd; /* (k) Ret value from fdopen. XXX */ 254 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */ 255 const void *td_wchan; /* (t) Sleep address. */ 256 const char *td_wmesg; /* (t) Reason for sleep. */ 257 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */ 258 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */ 259 short td_locks; /* (k) Debug: count of non-spin locks */ 260 short td_rw_rlocks; /* (k) Count of rwlock read locks. */ 261 short td_sx_slocks; /* (k) Count of sx shared locks. */ 262 short td_lk_slocks; /* (k) Count of lockmgr shared locks. */ 263 short td_stopsched; /* (k) Scheduler stopped. */ 264 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */ 265 const char *td_lockname; /* (t) Name of lock blocked on. */ 266 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */ 267 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */ 268 int td_intr_nesting_level; /* (k) Interrupt recursion. */ 269 int td_pinned; /* (k) Temporary cpu pin count. */ 270 struct ucred *td_realucred; /* (k) Reference to credentials. */ 271 struct ucred *td_ucred; /* (k) Used credentials, temporarily switchable. */ 272 struct plimit *td_limit; /* (k) Resource limits. */ 273 int td_slptick; /* (t) Time at sleep. */ 274 int td_blktick; /* (t) Time spent blocked. */ 275 int td_swvoltick; /* (t) Time at last SW_VOL switch. */ 276 int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */ 277 u_int td_cow; /* (*) Number of copy-on-write faults */ 278 struct rusage td_ru; /* (t) rusage information. */ 279 struct rusage_ext td_rux; /* (t) Internal rusage information. */ 280 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */ 281 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */ 282 u_int td_pticks; /* (t) Statclock hits for profiling */ 283 u_int td_sticks; /* (t) Statclock hits in system mode. */ 284 u_int td_iticks; /* (t) Statclock hits in intr mode. */ 285 u_int td_uticks; /* (t) Statclock hits in user mode. */ 286 int td_intrval; /* (t) Return value for sleepq. */ 287 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */ 288 volatile u_int td_generation; /* (k) For detection of preemption */ 289 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */ 290 int td_xsig; /* (c) Signal for ptrace */ 291 u_long td_profil_addr; /* (k) Temporary addr until AST. */ 292 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */ 293 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */ 294 struct file *td_fpop; /* (k) file referencing cdev under op */ 295 int td_dbgflags; /* (c) Userland debugger flags */ 296 siginfo_t td_si; /* (c) For debugger or core file */ 297 int td_ng_outbound; /* (k) Thread entered ng from above. */ 298 struct osd td_osd; /* (k) Object specific data. */ 299 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */ 300 pid_t td_dbg_forked; /* (c) Child pid for debugger. */ 301 struct vnode *td_vp_reserved;/* (k) Prealloated vnode. */ 302 u_int td_no_sleeping; /* (k) Sleeping disabled count. */ 303 void *td_su; /* (k) FFS SU private */ 304 sbintime_t td_sleeptimo; /* (t) Sleep timeout. */ 305 int td_rtcgen; /* (s) rtc_generation of abs. sleep */ 306 int td_errno; /* (k) Error from last syscall. */ 307 size_t td_vslock_sz; /* (k) amount of vslock-ed space */ 308 struct kcov_info *td_kcov_info; /* (*) Kernel code coverage data */ 309 u_int td_ucredref; /* (k) references on td_realucred */ 310 #define td_endzero td_sigmask 311 312 /* Copied during fork1() or create_thread(). */ 313 #define td_startcopy td_endzero 314 sigset_t td_sigmask; /* (c) Current signal mask. */ 315 u_char td_rqindex; /* (t) Run queue index. */ 316 u_char td_base_pri; /* (t) Thread base kernel priority. */ 317 u_char td_priority; /* (t) Thread active priority. */ 318 u_char td_pri_class; /* (t) Scheduling class. */ 319 u_char td_user_pri; /* (t) User pri from estcpu and nice. */ 320 u_char td_base_user_pri; /* (t) Base user pri */ 321 u_char td_pre_epoch_prio; /* (k) User pri on entry to epoch */ 322 uintptr_t td_rb_list; /* (k) Robust list head. */ 323 uintptr_t td_rbp_list; /* (k) Robust priv list head. */ 324 uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */ 325 struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on 326 fork for child tracing. */ 327 void *td_sigblock_ptr; /* (k) uptr for fast sigblock. */ 328 uint32_t td_sigblock_val; /* (k) fast sigblock value read at 329 td_sigblock_ptr on kern entry */ 330 #define td_endcopy td_pcb 331 332 /* 333 * Fields that must be manually set in fork1() or create_thread() 334 * or already have been set in the allocator, constructor, etc. 335 */ 336 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */ 337 enum td_states { 338 TDS_INACTIVE = 0x0, 339 TDS_INHIBITED, 340 TDS_CAN_RUN, 341 TDS_RUNQ, 342 TDS_RUNNING 343 } td_state; /* (t) thread state */ 344 union { 345 register_t tdu_retval[2]; 346 off_t tdu_off; 347 } td_uretoff; /* (k) Syscall aux returns. */ 348 #define td_retval td_uretoff.tdu_retval 349 u_int td_cowgen; /* (k) Generation of COW pointers. */ 350 /* LP64 hole */ 351 struct callout td_slpcallout; /* (h) Callout for sleep. */ 352 struct trapframe *td_frame; /* (k) */ 353 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */ 354 int td_kstack_pages; /* (a) Size of the kstack. */ 355 volatile u_int td_critnest; /* (k*) Critical section nest level. */ 356 struct mdthread td_md; /* (k) Any machine-dependent fields. */ 357 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */ 358 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */ 359 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */ 360 struct vnet *td_vnet; /* (k) Effective vnet. */ 361 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */ 362 struct trapframe *td_intr_frame;/* (k) Frame of the current irq */ 363 struct proc *td_rfppwait_p; /* (k) The vforked child */ 364 struct vm_page **td_ma; /* (k) uio pages held */ 365 int td_ma_cnt; /* (k) size of *td_ma */ 366 /* LP64 hole */ 367 void *td_emuldata; /* Emulator state data */ 368 int td_lastcpu; /* (t) Last cpu we were on. */ 369 int td_oncpu; /* (t) Which cpu we are on. */ 370 void *td_lkpi_task; /* LinuxKPI task struct pointer */ 371 int td_pmcpend; 372 #ifdef EPOCH_TRACE 373 SLIST_HEAD(, epoch_tracker) td_epochs; 374 #endif 375 }; 376 377 struct thread0_storage { 378 struct thread t0st_thread; 379 uint64_t t0st_sched[10]; 380 }; 381 382 struct mtx *thread_lock_block(struct thread *); 383 void thread_lock_block_wait(struct thread *); 384 void thread_lock_set(struct thread *, struct mtx *); 385 void thread_lock_unblock(struct thread *, struct mtx *); 386 #define THREAD_LOCK_ASSERT(td, type) \ 387 mtx_assert((td)->td_lock, (type)) 388 389 #define THREAD_LOCK_BLOCKED_ASSERT(td, type) \ 390 do { \ 391 struct mtx *__m = (td)->td_lock; \ 392 if (__m != &blocked_lock) \ 393 mtx_assert(__m, (type)); \ 394 } while (0) 395 396 #ifdef INVARIANTS 397 #define THREAD_LOCKPTR_ASSERT(td, lock) \ 398 do { \ 399 struct mtx *__m; \ 400 __m = (td)->td_lock; \ 401 KASSERT(__m == (lock), \ 402 ("Thread %p lock %p does not match %p", td, __m, (lock))); \ 403 } while (0) 404 405 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) \ 406 do { \ 407 struct mtx *__m; \ 408 __m = (td)->td_lock; \ 409 KASSERT(__m == (lock) || __m == &blocked_lock, \ 410 ("Thread %p lock %p does not match %p", td, __m, (lock))); \ 411 } while (0) 412 413 #define TD_LOCKS_INC(td) ((td)->td_locks++) 414 #define TD_LOCKS_DEC(td) do { \ 415 KASSERT(SCHEDULER_STOPPED_TD(td) || (td)->td_locks > 0, \ 416 ("thread %p owns no locks", (td))); \ 417 (td)->td_locks--; \ 418 } while (0) 419 #else 420 #define THREAD_LOCKPTR_ASSERT(td, lock) 421 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) 422 423 #define TD_LOCKS_INC(td) 424 #define TD_LOCKS_DEC(td) 425 #endif 426 427 /* 428 * Flags kept in td_flags: 429 * To change these you MUST have the scheduler lock. 430 */ 431 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */ 432 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */ 433 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */ 434 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */ 435 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */ 436 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */ 437 #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */ 438 #define TDF_UNUSED80 0x00000080 /* unused. */ 439 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */ 440 #define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */ 441 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */ 442 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */ 443 #define TDF_UNUSED12 0x00001000 /* --available-- */ 444 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */ 445 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */ 446 #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */ 447 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */ 448 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */ 449 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */ 450 #define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */ 451 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */ 452 #define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */ 453 #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */ 454 #define TDF_UNUSED23 0x00800000 /* --available-- */ 455 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */ 456 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */ 457 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */ 458 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */ 459 #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */ 460 #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */ 461 #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */ 462 463 /* Userland debug flags */ 464 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */ 465 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */ 466 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */ 467 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */ 468 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */ 469 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */ 470 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new 471 process */ 472 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child 473 only) */ 474 #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */ 475 #define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */ 476 #define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */ 477 #define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */ 478 #define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */ 479 #define TDB_STEP 0x00002000 /* (x86) PSL_T set for PT_STEP */ 480 481 /* 482 * "Private" flags kept in td_pflags: 483 * These are only written by curthread and thus need no locking. 484 */ 485 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */ 486 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */ 487 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */ 488 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */ 489 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */ 490 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */ 491 #define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */ 492 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */ 493 #define TDP_SIGFASTBLOCK 0x00000100 /* Fast sigblock active */ 494 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */ 495 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */ 496 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */ 497 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */ 498 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */ 499 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */ 500 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */ 501 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */ 502 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */ 503 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */ 504 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */ 505 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */ 506 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */ 507 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */ 508 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */ 509 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */ 510 #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */ 511 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */ 512 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */ 513 #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */ 514 #define TDP_FORKING 0x20000000 /* Thread is being created through fork() */ 515 #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */ 516 #define TDP_SIGFASTPENDING 0x80000000 /* Pending signal due to sigfastblock */ 517 518 #define TDP2_SBPAGES 0x00000001 /* Owns sbusy on some pages */ 519 520 /* 521 * Reasons that the current thread can not be run yet. 522 * More than one may apply. 523 */ 524 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */ 525 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */ 526 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */ 527 #define TDI_LOCK 0x0008 /* Stopped on a lock. */ 528 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */ 529 530 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING) 531 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL) 532 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED) 533 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED) 534 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK) 535 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT) 536 #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING) 537 #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ) 538 #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN) 539 #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED) 540 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED) 541 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD) 542 543 #define TD_CAN_ABORT(td) (TD_ON_SLEEPQ((td)) && \ 544 ((td)->td_flags & TDF_SINTR) != 0) 545 546 #define KTDSTATE(td) \ 547 (((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \ 548 ((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \ 549 ((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" : \ 550 ((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \ 551 ((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding") 552 553 #define TD_SET_INHIB(td, inhib) do { \ 554 (td)->td_state = TDS_INHIBITED; \ 555 (td)->td_inhibitors |= (inhib); \ 556 } while (0) 557 558 #define TD_CLR_INHIB(td, inhib) do { \ 559 if (((td)->td_inhibitors & (inhib)) && \ 560 (((td)->td_inhibitors &= ~(inhib)) == 0)) \ 561 (td)->td_state = TDS_CAN_RUN; \ 562 } while (0) 563 564 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING) 565 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED) 566 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK) 567 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED) 568 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT) 569 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING) 570 571 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING) 572 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED) 573 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK) 574 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED) 575 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT) 576 577 #define TD_SET_RUNNING(td) (td)->td_state = TDS_RUNNING 578 #define TD_SET_RUNQ(td) (td)->td_state = TDS_RUNQ 579 #define TD_SET_CAN_RUN(td) (td)->td_state = TDS_CAN_RUN 580 581 #define TD_SBDRY_INTR(td) \ 582 (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0) 583 #define TD_SBDRY_ERRNO(td) \ 584 (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART) 585 586 /* 587 * Process structure. 588 */ 589 struct proc { 590 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */ 591 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */ 592 struct mtx p_slock; /* process spin lock */ 593 struct ucred *p_ucred; /* (c) Process owner's identity. */ 594 struct filedesc *p_fd; /* (b) Open files. */ 595 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */ 596 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */ 597 struct plimit *p_limit; /* (c) Resource limits. */ 598 struct callout p_limco; /* (c) Limit callout handle */ 599 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */ 600 601 int p_flag; /* (c) P_* flags. */ 602 int p_flag2; /* (c) P2_* flags. */ 603 enum p_states { 604 PRS_NEW = 0, /* In creation */ 605 PRS_NORMAL, /* threads can be run. */ 606 PRS_ZOMBIE 607 } p_state; /* (j/c) Process status. */ 608 pid_t p_pid; /* (b) Process identifier. */ 609 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */ 610 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */ 611 struct proc *p_pptr; /* (c + e) Pointer to parent process. */ 612 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */ 613 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */ 614 struct proc *p_reaper; /* (e) My reaper. */ 615 LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants 616 (if I am reaper). */ 617 LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of 618 the same reaper. */ 619 struct mtx p_mtx; /* (n) Lock for this struct. */ 620 struct mtx p_statmtx; /* Lock for the stats */ 621 struct mtx p_itimmtx; /* Lock for the virt/prof timers */ 622 struct mtx p_profmtx; /* Lock for the profiling */ 623 struct ksiginfo *p_ksi; /* Locked by parent proc lock */ 624 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */ 625 #define p_siglist p_sigqueue.sq_signals 626 pid_t p_oppid; /* (c + e) Real parent pid. */ 627 628 /* The following fields are all zeroed upon creation in fork. */ 629 #define p_startzero p_vmspace 630 struct vmspace *p_vmspace; /* (b) Address space. */ 631 u_int p_swtick; /* (c) Tick when swapped in or out. */ 632 u_int p_cowgen; /* (c) Generation of COW pointers. */ 633 struct itimerval p_realtimer; /* (c) Alarm timer. */ 634 struct rusage p_ru; /* (a) Exit information. */ 635 struct rusage_ext p_rux; /* (cu) Internal resource usage. */ 636 struct rusage_ext p_crux; /* (c) Internal child resource usage. */ 637 int p_profthreads; /* (c) Num threads in addupc_task. */ 638 volatile int p_exitthreads; /* (j) Number of threads exiting */ 639 int p_traceflag; /* (o) Kernel trace points. */ 640 struct vnode *p_tracevp; /* (c + o) Trace to vnode. */ 641 struct ucred *p_tracecred; /* (o) Credentials to trace with. */ 642 struct vnode *p_textvp; /* (b) Vnode of executable. */ 643 u_int p_lock; /* (c) Proclock (prevent swap) count. */ 644 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */ 645 int p_sigparent; /* (c) Signal to parent on exit. */ 646 int p_sig; /* (n) For core dump/debugger XXX. */ 647 u_int p_ptevents; /* (c + e) ptrace() event mask. */ 648 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */ 649 struct thread *p_singlethread;/* (c + j) If single threading this is it */ 650 int p_suspcount; /* (j) Num threads in suspended mode. */ 651 struct thread *p_xthread; /* (c) Trap thread */ 652 int p_boundary_count;/* (j) Num threads at user boundary */ 653 int p_pendingcnt; /* how many signals are pending */ 654 struct itimers *p_itimers; /* (c) POSIX interval timers. */ 655 struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */ 656 u_int p_treeflag; /* (e) P_TREE flags */ 657 int p_pendingexits; /* (c) Count of pending thread exits. */ 658 struct filemon *p_filemon; /* (c) filemon-specific data. */ 659 int p_pdeathsig; /* (c) Signal from parent on exit. */ 660 /* End area that is zeroed on creation. */ 661 #define p_endzero p_magic 662 663 /* The following fields are all copied upon creation in fork. */ 664 #define p_startcopy p_endzero 665 u_int p_magic; /* (b) Magic number. */ 666 int p_osrel; /* (x) osreldate for the 667 binary (from ELF note, if any) */ 668 uint32_t p_fctl0; /* (x) ABI feature control, ELF note */ 669 char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */ 670 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */ 671 struct pargs *p_args; /* (c) Process arguments. */ 672 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */ 673 signed char p_nice; /* (c) Process "nice" value. */ 674 int p_fibnum; /* in this routing domain XXX MRT */ 675 pid_t p_reapsubtree; /* (e) Pid of the direct child of the 676 reaper which spawned 677 our subtree. */ 678 uint16_t p_elf_machine; /* (x) ELF machine type */ 679 uint64_t p_elf_flags; /* (x) ELF flags */ 680 /* End area that is copied on creation. */ 681 #define p_endcopy p_xexit 682 683 u_int p_xexit; /* (c) Exit code. */ 684 u_int p_xsig; /* (c) Stop/kill sig. */ 685 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */ 686 struct knlist *p_klist; /* (c) Knotes attached to this proc. */ 687 int p_numthreads; /* (c) Number of threads. */ 688 struct mdproc p_md; /* Any machine-dependent fields. */ 689 struct callout p_itcallout; /* (h + c) Interval timer callout. */ 690 u_short p_acflag; /* (c) Accounting flags. */ 691 struct proc *p_peers; /* (r) */ 692 struct proc *p_leader; /* (b) */ 693 void *p_emuldata; /* (c) Emulator state data. */ 694 struct label *p_label; /* (*) Proc (not subject) MAC label. */ 695 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */ 696 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/ 697 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */ 698 struct cv p_pwait; /* (*) wait cv for exit/exec. */ 699 uint64_t p_prev_runtime; /* (c) Resource usage accounting. */ 700 struct racct *p_racct; /* (b) Resource accounting. */ 701 int p_throttled; /* (c) Flag for racct pcpu throttling */ 702 /* 703 * An orphan is the child that has been re-parented to the 704 * debugger as a result of attaching to it. Need to keep 705 * track of them for parent to be able to collect the exit 706 * status of what used to be children. 707 */ 708 LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */ 709 LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */ 710 }; 711 712 #define p_session p_pgrp->pg_session 713 #define p_pgid p_pgrp->pg_id 714 715 #define NOCPU (-1) /* For when we aren't on a CPU. */ 716 #define NOCPU_OLD (255) 717 #define MAXCPU_OLD (254) 718 719 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock) 720 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock) 721 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type)) 722 723 #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx) 724 #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx) 725 #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type)) 726 727 #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx) 728 #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx) 729 #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type)) 730 731 #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx) 732 #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx) 733 #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type)) 734 735 /* These flags are kept in p_flag. */ 736 #define P_ADVLOCK 0x00001 /* Process may hold a POSIX advisory lock. */ 737 #define P_CONTROLT 0x00002 /* Has a controlling terminal. */ 738 #define P_KPROC 0x00004 /* Kernel process. */ 739 #define P_UNUSED3 0x00008 /* --available-- */ 740 #define P_PPWAIT 0x00010 /* Parent is waiting for child to exec/exit. */ 741 #define P_PROFIL 0x00020 /* Has started profiling. */ 742 #define P_STOPPROF 0x00040 /* Has thread requesting to stop profiling. */ 743 #define P_HADTHREADS 0x00080 /* Has had threads (no cleanup shortcuts) */ 744 #define P_SUGID 0x00100 /* Had set id privileges since last exec. */ 745 #define P_SYSTEM 0x00200 /* System proc: no sigs, stats or swapping. */ 746 #define P_SINGLE_EXIT 0x00400 /* Threads suspending should exit, not wait. */ 747 #define P_TRACED 0x00800 /* Debugged process being traced. */ 748 #define P_WAITED 0x01000 /* Someone is waiting for us. */ 749 #define P_WEXIT 0x02000 /* Working on exiting. */ 750 #define P_EXEC 0x04000 /* Process called exec. */ 751 #define P_WKILLED 0x08000 /* Killed, go to kernel/user boundary ASAP. */ 752 #define P_CONTINUED 0x10000 /* Proc has continued from a stopped state. */ 753 #define P_STOPPED_SIG 0x20000 /* Stopped due to SIGSTOP/SIGTSTP. */ 754 #define P_STOPPED_TRACE 0x40000 /* Stopped because of tracing. */ 755 #define P_STOPPED_SINGLE 0x80000 /* Only 1 thread can continue (not to user). */ 756 #define P_PROTECTED 0x100000 /* Do not kill on memory overcommit. */ 757 #define P_SIGEVENT 0x200000 /* Process pending signals changed. */ 758 #define P_SINGLE_BOUNDARY 0x400000 /* Threads should suspend at user boundary. */ 759 #define P_HWPMC 0x800000 /* Process is using HWPMCs */ 760 #define P_JAILED 0x1000000 /* Process is in jail. */ 761 #define P_TOTAL_STOP 0x2000000 /* Stopped in stop_all_proc. */ 762 #define P_INEXEC 0x4000000 /* Process is in execve(). */ 763 #define P_STATCHILD 0x8000000 /* Child process stopped or exited. */ 764 #define P_INMEM 0x10000000 /* Loaded into memory. */ 765 #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */ 766 #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */ 767 #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */ 768 769 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE) 770 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED) 771 #define P_KILLED(p) ((p)->p_flag & P_WKILLED) 772 773 /* These flags are kept in p_flag2. */ 774 #define P2_INHERIT_PROTECTED 0x00000001 /* New children get P_PROTECTED. */ 775 #define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or coredumps. */ 776 #define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on exec(2). */ 777 #define P2_AST_SU 0x00000008 /* Handles SU ast for kthreads. */ 778 #define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not yet handled. */ 779 #define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */ 780 #define P2_ASLR_ENABLE 0x00000040 /* Force enable ASLR. */ 781 #define P2_ASLR_DISABLE 0x00000080 /* Force disable ASLR. */ 782 #define P2_ASLR_IGNSTART 0x00000100 /* Enable ASLR to consume sbrk area. */ 783 #define P2_PROTMAX_ENABLE 0x00000200 /* Force enable implied PROT_MAX. */ 784 #define P2_PROTMAX_DISABLE 0x00000400 /* Force disable implied PROT_MAX. */ 785 #define P2_STKGAP_DISABLE 0x00000800 /* Disable stack gap for MAP_STACK */ 786 #define P2_STKGAP_DISABLE_EXEC 0x00001000 /* Stack gap disabled after exec */ 787 788 /* Flags protected by proctree_lock, kept in p_treeflags. */ 789 #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */ 790 #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan 791 list */ 792 #define P_TREE_REAPER 0x00000004 /* Reaper of subtree */ 793 794 /* 795 * These were process status values (p_stat), now they are only used in 796 * legacy conversion code. 797 */ 798 #define SIDL 1 /* Process being created by fork. */ 799 #define SRUN 2 /* Currently runnable. */ 800 #define SSLEEP 3 /* Sleeping on an address. */ 801 #define SSTOP 4 /* Process debugging or suspension. */ 802 #define SZOMB 5 /* Awaiting collection by parent. */ 803 #define SWAIT 6 /* Waiting for interrupt. */ 804 #define SLOCK 7 /* Blocked on a lock. */ 805 806 #define P_MAGIC 0xbeefface 807 808 #ifdef _KERNEL 809 810 /* Types and flags for mi_switch(). */ 811 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */ 812 #define SWT_NONE 0 /* Unspecified switch. */ 813 #define SWT_PREEMPT 1 /* Switching due to preemption. */ 814 #define SWT_OWEPREEMPT 2 /* Switching due to owepreempt. */ 815 #define SWT_TURNSTILE 3 /* Turnstile contention. */ 816 #define SWT_SLEEPQ 4 /* Sleepq wait. */ 817 #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */ 818 #define SWT_RELINQUISH 6 /* yield call. */ 819 #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */ 820 #define SWT_IDLE 8 /* Switching from the idle thread. */ 821 #define SWT_IWAIT 9 /* Waiting for interrupts. */ 822 #define SWT_SUSPEND 10 /* Thread suspended. */ 823 #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */ 824 #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */ 825 #define SWT_COUNT 13 /* Number of switch types. */ 826 /* Flags */ 827 #define SW_VOL 0x0100 /* Voluntary switch. */ 828 #define SW_INVOL 0x0200 /* Involuntary switch. */ 829 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */ 830 831 /* How values for thread_single(). */ 832 #define SINGLE_NO_EXIT 0 833 #define SINGLE_EXIT 1 834 #define SINGLE_BOUNDARY 2 835 #define SINGLE_ALLPROC 3 836 837 #ifdef MALLOC_DECLARE 838 MALLOC_DECLARE(M_PARGS); 839 MALLOC_DECLARE(M_PGRP); 840 MALLOC_DECLARE(M_SESSION); 841 MALLOC_DECLARE(M_SUBPROC); 842 #endif 843 844 #define FOREACH_PROC_IN_SYSTEM(p) \ 845 LIST_FOREACH((p), &allproc, p_list) 846 #define FOREACH_THREAD_IN_PROC(p, td) \ 847 TAILQ_FOREACH((td), &(p)->p_threads, td_plist) 848 849 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads) 850 851 /* 852 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit 853 * in a pid_t, as it is used to represent "no process group". 854 */ 855 #define PID_MAX 99999 856 #define NO_PID 100000 857 extern pid_t pid_max; 858 859 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p)) 860 861 862 /* Lock and unlock a process. */ 863 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx) 864 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx) 865 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx) 866 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx) 867 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type)) 868 869 /* Lock and unlock a process group. */ 870 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx) 871 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx) 872 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx) 873 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type)) 874 875 #define PGRP_LOCK_PGSIGNAL(pg) do { \ 876 if ((pg) != NULL) \ 877 PGRP_LOCK(pg); \ 878 } while (0) 879 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \ 880 if ((pg) != NULL) \ 881 PGRP_UNLOCK(pg); \ 882 } while (0) 883 884 /* Lock and unlock a session. */ 885 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx) 886 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx) 887 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx) 888 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type)) 889 890 /* 891 * Non-zero p_lock ensures that: 892 * - exit1() is not performed until p_lock reaches zero; 893 * - the process' threads stack are not swapped out if they are currently 894 * not (P_INMEM). 895 * 896 * PHOLD() asserts that the process (except the current process) is 897 * not exiting, increments p_lock and swaps threads stacks into memory, 898 * if needed. 899 * _PHOLD() is same as PHOLD(), it takes the process locked. 900 * _PHOLD_LITE() also takes the process locked, but comparing with 901 * _PHOLD(), it only guarantees that exit1() is not executed, 902 * faultin() is not called. 903 */ 904 #define PHOLD(p) do { \ 905 PROC_LOCK(p); \ 906 _PHOLD(p); \ 907 PROC_UNLOCK(p); \ 908 } while (0) 909 #define _PHOLD(p) do { \ 910 PROC_LOCK_ASSERT((p), MA_OWNED); \ 911 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 912 ("PHOLD of exiting process %p", p)); \ 913 (p)->p_lock++; \ 914 if (((p)->p_flag & P_INMEM) == 0) \ 915 faultin((p)); \ 916 } while (0) 917 #define _PHOLD_LITE(p) do { \ 918 PROC_LOCK_ASSERT((p), MA_OWNED); \ 919 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 920 ("PHOLD of exiting process %p", p)); \ 921 (p)->p_lock++; \ 922 } while (0) 923 #define PROC_ASSERT_HELD(p) do { \ 924 KASSERT((p)->p_lock > 0, ("process %p not held", p)); \ 925 } while (0) 926 927 #define PRELE(p) do { \ 928 PROC_LOCK((p)); \ 929 _PRELE((p)); \ 930 PROC_UNLOCK((p)); \ 931 } while (0) 932 #define _PRELE(p) do { \ 933 PROC_LOCK_ASSERT((p), MA_OWNED); \ 934 PROC_ASSERT_HELD(p); \ 935 (--(p)->p_lock); \ 936 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \ 937 wakeup(&(p)->p_lock); \ 938 } while (0) 939 #define PROC_ASSERT_NOT_HELD(p) do { \ 940 KASSERT((p)->p_lock == 0, ("process %p held", p)); \ 941 } while (0) 942 943 #define PROC_UPDATE_COW(p) do { \ 944 PROC_LOCK_ASSERT((p), MA_OWNED); \ 945 (p)->p_cowgen++; \ 946 } while (0) 947 948 /* Check whether a thread is safe to be swapped out. */ 949 #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP) 950 951 /* Control whether or not it is safe for curthread to sleep. */ 952 #define THREAD_NO_SLEEPING() do { \ 953 curthread->td_no_sleeping++; \ 954 MPASS(curthread->td_no_sleeping > 0); \ 955 } while (0) 956 957 #define THREAD_SLEEPING_OK() do { \ 958 MPASS(curthread->td_no_sleeping > 0); \ 959 curthread->td_no_sleeping--; \ 960 } while (0) 961 962 #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0) 963 964 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash]) 965 #define PIDHASHLOCK(pid) (&pidhashtbl_lock[((pid) & pidhashlock)]) 966 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl; 967 extern struct sx *pidhashtbl_lock; 968 extern u_long pidhash; 969 extern u_long pidhashlock; 970 #define TIDHASH(tid) (&tidhashtbl[(tid) & tidhash]) 971 extern LIST_HEAD(tidhashhead, thread) *tidhashtbl; 972 extern u_long tidhash; 973 extern struct rwlock tidhash_lock; 974 975 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash]) 976 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl; 977 extern u_long pgrphash; 978 979 extern struct sx allproc_lock; 980 extern int allproc_gen; 981 extern struct sx proctree_lock; 982 extern struct mtx ppeers_lock; 983 extern struct mtx procid_lock; 984 extern struct proc proc0; /* Process slot for swapper. */ 985 extern struct thread0_storage thread0_st; /* Primary thread in proc0. */ 986 #define thread0 (thread0_st.t0st_thread) 987 extern struct vmspace vmspace0; /* VM space for proc0. */ 988 extern int hogticks; /* Limit on kernel cpu hogs. */ 989 extern int lastpid; 990 extern int nprocs, maxproc; /* Current and max number of procs. */ 991 extern int maxprocperuid; /* Max procs per uid. */ 992 extern u_long ps_arg_cache_limit; 993 994 LIST_HEAD(proclist, proc); 995 TAILQ_HEAD(procqueue, proc); 996 TAILQ_HEAD(threadqueue, thread); 997 extern struct proclist allproc; /* List of all processes. */ 998 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */ 999 1000 extern struct uma_zone *proc_zone; 1001 1002 struct proc *pfind(pid_t); /* Find process by id. */ 1003 struct proc *pfind_any(pid_t); /* Find (zombie) process by id. */ 1004 struct proc *pfind_any_locked(pid_t pid); /* Find process by id, locked. */ 1005 struct pgrp *pgfind(pid_t); /* Find process group by id. */ 1006 void pidhash_slockall(void); /* Shared lock all pid hash lists. */ 1007 void pidhash_sunlockall(void); /* Shared unlock all pid hash lists. */ 1008 1009 struct fork_req { 1010 int fr_flags; 1011 int fr_pages; 1012 int *fr_pidp; 1013 struct proc **fr_procp; 1014 int *fr_pd_fd; 1015 int fr_pd_flags; 1016 struct filecaps *fr_pd_fcaps; 1017 int fr_flags2; 1018 #define FR2_DROPSIG_CAUGHT 0x00001 /* Drop caught non-DFL signals */ 1019 }; 1020 1021 /* 1022 * pget() flags. 1023 */ 1024 #define PGET_HOLD 0x00001 /* Hold the process. */ 1025 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */ 1026 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */ 1027 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */ 1028 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */ 1029 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */ 1030 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */ 1031 1032 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT) 1033 1034 int pget(pid_t pid, int flags, struct proc **pp); 1035 1036 void ast(struct trapframe *framep); 1037 struct thread *choosethread(void); 1038 int cr_cansee(struct ucred *u1, struct ucred *u2); 1039 int cr_canseesocket(struct ucred *cred, struct socket *so); 1040 int cr_canseeothergids(struct ucred *u1, struct ucred *u2); 1041 int cr_canseeotheruids(struct ucred *u1, struct ucred *u2); 1042 int cr_canseejailproc(struct ucred *u1, struct ucred *u2); 1043 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum); 1044 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp, 1045 struct session *sess); 1046 int enterthispgrp(struct proc *p, struct pgrp *pgrp); 1047 void faultin(struct proc *p); 1048 void fixjobc(struct proc *p, struct pgrp *pgrp, int entering); 1049 int fork1(struct thread *, struct fork_req *); 1050 void fork_rfppwait(struct thread *); 1051 void fork_exit(void (*)(void *, struct trapframe *), void *, 1052 struct trapframe *); 1053 void fork_return(struct thread *, struct trapframe *); 1054 int inferior(struct proc *p); 1055 void kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry, 1056 int *resident_count, bool *super); 1057 void kern_yield(int); 1058 void kick_proc0(void); 1059 void killjobc(void); 1060 int leavepgrp(struct proc *p); 1061 int maybe_preempt(struct thread *td); 1062 void maybe_yield(void); 1063 void mi_switch(int flags); 1064 int p_candebug(struct thread *td, struct proc *p); 1065 int p_cansee(struct thread *td, struct proc *p); 1066 int p_cansched(struct thread *td, struct proc *p); 1067 int p_cansignal(struct thread *td, struct proc *p, int signum); 1068 int p_canwait(struct thread *td, struct proc *p); 1069 struct pargs *pargs_alloc(int len); 1070 void pargs_drop(struct pargs *pa); 1071 void pargs_hold(struct pargs *pa); 1072 int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb); 1073 int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb); 1074 int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb); 1075 void procinit(void); 1076 int proc_iterate(int (*cb)(struct proc *, void *), void *cbarg); 1077 void proc_linkup0(struct proc *p, struct thread *td); 1078 void proc_linkup(struct proc *p, struct thread *td); 1079 struct proc *proc_realparent(struct proc *child); 1080 void proc_reap(struct thread *td, struct proc *p, int *status, int options); 1081 void proc_reparent(struct proc *child, struct proc *newparent, bool set_oppid); 1082 void proc_add_orphan(struct proc *child, struct proc *parent); 1083 void proc_set_traced(struct proc *p, bool stop); 1084 void proc_wkilled(struct proc *p); 1085 struct pstats *pstats_alloc(void); 1086 void pstats_fork(struct pstats *src, struct pstats *dst); 1087 void pstats_free(struct pstats *ps); 1088 void proc_clear_orphan(struct proc *p); 1089 void reaper_abandon_children(struct proc *p, bool exiting); 1090 int securelevel_ge(struct ucred *cr, int level); 1091 int securelevel_gt(struct ucred *cr, int level); 1092 void sess_hold(struct session *); 1093 void sess_release(struct session *); 1094 int setrunnable(struct thread *, int); 1095 void setsugid(struct proc *p); 1096 int should_yield(void); 1097 int sigonstack(size_t sp); 1098 void stopevent(struct proc *, u_int, u_int); 1099 struct thread *tdfind(lwpid_t, pid_t); 1100 void threadinit(void); 1101 void tidhash_add(struct thread *); 1102 void tidhash_remove(struct thread *); 1103 void cpu_idle(int); 1104 int cpu_idle_wakeup(int); 1105 extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */ 1106 void cpu_switch(struct thread *, struct thread *, struct mtx *); 1107 void cpu_throw(struct thread *, struct thread *) __dead2; 1108 void unsleep(struct thread *); 1109 void userret(struct thread *, struct trapframe *); 1110 1111 void cpu_exit(struct thread *); 1112 void exit1(struct thread *, int, int) __dead2; 1113 void cpu_copy_thread(struct thread *td, struct thread *td0); 1114 bool cpu_exec_vmspace_reuse(struct proc *p, struct vm_map *map); 1115 int cpu_fetch_syscall_args(struct thread *td); 1116 void cpu_fork(struct thread *, struct proc *, struct thread *, int); 1117 void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *); 1118 int cpu_procctl(struct thread *td, int idtype, id_t id, int com, 1119 void *data); 1120 void cpu_set_syscall_retval(struct thread *, int); 1121 void cpu_set_upcall(struct thread *, void (*)(void *), void *, 1122 stack_t *); 1123 int cpu_set_user_tls(struct thread *, void *tls_base); 1124 void cpu_thread_alloc(struct thread *); 1125 void cpu_thread_clean(struct thread *); 1126 void cpu_thread_exit(struct thread *); 1127 void cpu_thread_free(struct thread *); 1128 void cpu_thread_swapin(struct thread *); 1129 void cpu_thread_swapout(struct thread *); 1130 struct thread *thread_alloc(int pages); 1131 int thread_alloc_stack(struct thread *, int pages); 1132 int thread_check_susp(struct thread *td, bool sleep); 1133 void thread_cow_get_proc(struct thread *newtd, struct proc *p); 1134 void thread_cow_get(struct thread *newtd, struct thread *td); 1135 void thread_cow_free(struct thread *td); 1136 void thread_cow_update(struct thread *td); 1137 int thread_create(struct thread *td, struct rtprio *rtp, 1138 int (*initialize_thread)(struct thread *, void *), void *thunk); 1139 void thread_exit(void) __dead2; 1140 void thread_free(struct thread *td); 1141 void thread_link(struct thread *td, struct proc *p); 1142 void thread_reap(void); 1143 int thread_single(struct proc *p, int how); 1144 void thread_single_end(struct proc *p, int how); 1145 void thread_stash(struct thread *td); 1146 void thread_stopped(struct proc *p); 1147 void childproc_stopped(struct proc *child, int reason); 1148 void childproc_continued(struct proc *child); 1149 void childproc_exited(struct proc *child); 1150 int thread_suspend_check(int how); 1151 bool thread_suspend_check_needed(void); 1152 void thread_suspend_switch(struct thread *, struct proc *p); 1153 void thread_suspend_one(struct thread *td); 1154 void thread_unlink(struct thread *td); 1155 void thread_unsuspend(struct proc *p); 1156 void thread_wait(struct proc *p); 1157 struct thread *thread_find(struct proc *p, lwpid_t tid); 1158 1159 void stop_all_proc(void); 1160 void resume_all_proc(void); 1161 1162 static __inline int 1163 curthread_pflags_set(int flags) 1164 { 1165 struct thread *td; 1166 int save; 1167 1168 td = curthread; 1169 save = ~flags | (td->td_pflags & flags); 1170 td->td_pflags |= flags; 1171 return (save); 1172 } 1173 1174 static __inline void 1175 curthread_pflags_restore(int save) 1176 { 1177 1178 curthread->td_pflags &= save; 1179 } 1180 1181 static __inline int 1182 curthread_pflags2_set(int flags) 1183 { 1184 struct thread *td; 1185 int save; 1186 1187 td = curthread; 1188 save = ~flags | (td->td_pflags2 & flags); 1189 td->td_pflags2 |= flags; 1190 return (save); 1191 } 1192 1193 static __inline void 1194 curthread_pflags2_restore(int save) 1195 { 1196 1197 curthread->td_pflags2 &= save; 1198 } 1199 1200 static __inline __pure2 struct td_sched * 1201 td_get_sched(struct thread *td) 1202 { 1203 1204 return ((struct td_sched *)&td[1]); 1205 } 1206 1207 extern void (*softdep_ast_cleanup)(struct thread *); 1208 static __inline void 1209 td_softdep_cleanup(struct thread *td) 1210 { 1211 1212 if (td->td_su != NULL && softdep_ast_cleanup != NULL) 1213 softdep_ast_cleanup(td); 1214 } 1215 1216 #define PROC_ID_PID 0 1217 #define PROC_ID_GROUP 1 1218 #define PROC_ID_SESSION 2 1219 #define PROC_ID_REAP 3 1220 1221 void proc_id_set(int type, pid_t id); 1222 void proc_id_set_cond(int type, pid_t id); 1223 void proc_id_clear(int type, pid_t id); 1224 1225 EVENTHANDLER_LIST_DECLARE(process_ctor); 1226 EVENTHANDLER_LIST_DECLARE(process_dtor); 1227 EVENTHANDLER_LIST_DECLARE(process_init); 1228 EVENTHANDLER_LIST_DECLARE(process_fini); 1229 EVENTHANDLER_LIST_DECLARE(process_exit); 1230 EVENTHANDLER_LIST_DECLARE(process_fork); 1231 EVENTHANDLER_LIST_DECLARE(process_exec); 1232 1233 EVENTHANDLER_LIST_DECLARE(thread_ctor); 1234 EVENTHANDLER_LIST_DECLARE(thread_dtor); 1235 EVENTHANDLER_LIST_DECLARE(thread_init); 1236 1237 #endif /* _KERNEL */ 1238 1239 #endif /* !_SYS_PROC_H_ */ 1240