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 td_sched; 197 struct thread; 198 struct trapframe; 199 struct turnstile; 200 struct vm_map; 201 struct vm_map_entry; 202 struct epoch_tracker; 203 204 struct syscall_args { 205 u_int code; 206 u_int original_code; 207 struct sysent *callp; 208 register_t args[8]; 209 }; 210 211 /* 212 * XXX: Does this belong in resource.h or resourcevar.h instead? 213 * Resource usage extension. The times in rusage structs in the kernel are 214 * never up to date. The actual times are kept as runtimes and tick counts 215 * (with control info in the "previous" times), and are converted when 216 * userland asks for rusage info. Backwards compatibility prevents putting 217 * this directly in the user-visible rusage struct. 218 * 219 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux. 220 * Locking for td_rux: (t) for all fields. 221 */ 222 struct rusage_ext { 223 uint64_t rux_runtime; /* (cu) Real time. */ 224 uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */ 225 uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */ 226 uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */ 227 uint64_t rux_uu; /* (c) Previous user time in usec. */ 228 uint64_t rux_su; /* (c) Previous sys time in usec. */ 229 uint64_t rux_tu; /* (c) Previous total time in usec. */ 230 }; 231 232 /* 233 * Kernel runnable context (thread). 234 * This is what is put to sleep and reactivated. 235 * Thread context. Processes may have multiple threads. 236 */ 237 struct thread { 238 struct mtx *volatile td_lock; /* replaces sched lock */ 239 struct proc *td_proc; /* (*) Associated process. */ 240 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */ 241 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */ 242 union { 243 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */ 244 struct thread *td_zombie; /* Zombie list linkage */ 245 }; 246 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */ 247 LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */ 248 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */ 249 struct domainset_ref td_domain; /* (a) NUMA policy */ 250 struct seltd *td_sel; /* Select queue/channel. */ 251 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */ 252 struct turnstile *td_turnstile; /* (k) Associated turnstile. */ 253 struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */ 254 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */ 255 lwpid_t td_tid; /* (b) Thread ID. */ 256 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */ 257 #define td_siglist td_sigqueue.sq_signals 258 u_char td_lend_user_pri; /* (t) Lend user pri. */ 259 u_char td_allocdomain; /* (b) NUMA domain backing this struct thread. */ 260 struct kmsan_td *td_kmsan; /* (k) KMSAN state */ 261 262 /* Cleared during fork1() */ 263 #define td_startzero td_flags 264 int td_flags; /* (t) TDF_* flags. */ 265 int td_inhibitors; /* (t) Why can not run. */ 266 int td_pflags; /* (k) Private thread (TDP_*) flags. */ 267 int td_pflags2; /* (k) Private thread (TDP2_*) flags. */ 268 int td_dupfd; /* (k) Ret value from fdopen. XXX */ 269 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */ 270 const void *td_wchan; /* (t) Sleep address. */ 271 const char *td_wmesg; /* (t) Reason for sleep. */ 272 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */ 273 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */ 274 short td_locks; /* (k) Debug: count of non-spin locks */ 275 short td_rw_rlocks; /* (k) Count of rwlock read locks. */ 276 short td_sx_slocks; /* (k) Count of sx shared locks. */ 277 short td_lk_slocks; /* (k) Count of lockmgr shared locks. */ 278 short td_stopsched; /* (k) Scheduler stopped. */ 279 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */ 280 const char *td_lockname; /* (t) Name of lock blocked on. */ 281 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */ 282 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */ 283 int td_intr_nesting_level; /* (k) Interrupt recursion. */ 284 int td_pinned; /* (k) Temporary cpu pin count. */ 285 struct ucred *td_realucred; /* (k) Reference to credentials. */ 286 struct ucred *td_ucred; /* (k) Used credentials, temporarily switchable. */ 287 struct plimit *td_limit; /* (k) Resource limits. */ 288 int td_slptick; /* (t) Time at sleep. */ 289 int td_blktick; /* (t) Time spent blocked. */ 290 int td_swvoltick; /* (t) Time at last SW_VOL switch. */ 291 int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */ 292 u_int td_cow; /* (*) Number of copy-on-write faults */ 293 struct rusage td_ru; /* (t) rusage information. */ 294 struct rusage_ext td_rux; /* (t) Internal rusage information. */ 295 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */ 296 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */ 297 u_int td_pticks; /* (t) Statclock hits for profiling */ 298 u_int td_sticks; /* (t) Statclock hits in system mode. */ 299 u_int td_iticks; /* (t) Statclock hits in intr mode. */ 300 u_int td_uticks; /* (t) Statclock hits in user mode. */ 301 int td_intrval; /* (t) Return value for sleepq. */ 302 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */ 303 volatile u_int td_generation; /* (k) For detection of preemption */ 304 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */ 305 int td_xsig; /* (c) Signal for ptrace */ 306 u_long td_profil_addr; /* (k) Temporary addr until AST. */ 307 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */ 308 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */ 309 struct file *td_fpop; /* (k) file referencing cdev under op */ 310 int td_dbgflags; /* (c) Userland debugger flags */ 311 siginfo_t td_si; /* (c) For debugger or core file */ 312 int td_ng_outbound; /* (k) Thread entered ng from above. */ 313 struct osd td_osd; /* (k) Object specific data. */ 314 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */ 315 pid_t td_dbg_forked; /* (c) Child pid for debugger. */ 316 struct vnode *td_vp_reserved;/* (k) Preallocated vnode. */ 317 u_int td_no_sleeping; /* (k) Sleeping disabled count. */ 318 void *td_su; /* (k) FFS SU private */ 319 sbintime_t td_sleeptimo; /* (t) Sleep timeout. */ 320 int td_rtcgen; /* (s) rtc_generation of abs. sleep */ 321 int td_errno; /* (k) Error from last syscall. */ 322 size_t td_vslock_sz; /* (k) amount of vslock-ed space */ 323 struct kcov_info *td_kcov_info; /* (*) Kernel code coverage data */ 324 u_int td_ucredref; /* (k) references on td_realucred */ 325 #define td_endzero td_sigmask 326 327 /* Copied during fork1() or create_thread(). */ 328 #define td_startcopy td_endzero 329 sigset_t td_sigmask; /* (c) Current signal mask. */ 330 u_char td_rqindex; /* (t) Run queue index. */ 331 u_char td_base_pri; /* (t) Thread base kernel priority. */ 332 u_char td_priority; /* (t) Thread active priority. */ 333 u_char td_pri_class; /* (t) Scheduling class. */ 334 u_char td_user_pri; /* (t) User pri from estcpu and nice. */ 335 u_char td_base_user_pri; /* (t) Base user pri */ 336 uintptr_t td_rb_list; /* (k) Robust list head. */ 337 uintptr_t td_rbp_list; /* (k) Robust priv list head. */ 338 uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */ 339 struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on 340 fork for child tracing. */ 341 void *td_sigblock_ptr; /* (k) uptr for fast sigblock. */ 342 uint32_t td_sigblock_val; /* (k) fast sigblock value read at 343 td_sigblock_ptr on kern entry */ 344 #define td_endcopy td_pcb 345 346 /* 347 * Fields that must be manually set in fork1() or create_thread() 348 * or already have been set in the allocator, constructor, etc. 349 */ 350 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */ 351 enum td_states { 352 TDS_INACTIVE = 0x0, 353 TDS_INHIBITED, 354 TDS_CAN_RUN, 355 TDS_RUNQ, 356 TDS_RUNNING 357 } td_state; /* (t) thread state */ 358 /* Note: td_state must be accessed using TD_{GET,SET}_STATE(). */ 359 union { 360 syscallarg_t tdu_retval[2]; 361 off_t tdu_off; 362 } td_uretoff; /* (k) Syscall aux returns. */ 363 #define td_retval td_uretoff.tdu_retval 364 u_int td_cowgen; /* (k) Generation of COW pointers. */ 365 /* LP64 hole */ 366 struct callout td_slpcallout; /* (h) Callout for sleep. */ 367 struct trapframe *td_frame; /* (k) */ 368 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */ 369 int td_kstack_pages; /* (a) Size of the kstack. */ 370 volatile u_int td_critnest; /* (k*) Critical section nest level. */ 371 struct mdthread td_md; /* (k) Any machine-dependent fields. */ 372 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */ 373 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */ 374 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */ 375 struct vnet *td_vnet; /* (k) Effective vnet. */ 376 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */ 377 struct trapframe *td_intr_frame;/* (k) Frame of the current irq */ 378 struct proc *td_rfppwait_p; /* (k) The vforked child */ 379 struct vm_page **td_ma; /* (k) uio pages held */ 380 int td_ma_cnt; /* (k) size of *td_ma */ 381 /* LP64 hole */ 382 void *td_emuldata; /* Emulator state data */ 383 int td_lastcpu; /* (t) Last cpu we were on. */ 384 int td_oncpu; /* (t) Which cpu we are on. */ 385 void *td_lkpi_task; /* LinuxKPI task struct pointer */ 386 int td_pmcpend; 387 void *td_coredump; /* (c) coredump request. */ 388 off_t td_ktr_io_lim; /* (k) limit for ktrace file size */ 389 #ifdef EPOCH_TRACE 390 SLIST_HEAD(, epoch_tracker) td_epochs; 391 #endif 392 }; 393 394 struct thread0_storage { 395 struct thread t0st_thread; 396 uint64_t t0st_sched[10]; 397 }; 398 399 struct mtx *thread_lock_block(struct thread *); 400 void thread_lock_block_wait(struct thread *); 401 void thread_lock_set(struct thread *, struct mtx *); 402 void thread_lock_unblock(struct thread *, struct mtx *); 403 #define THREAD_LOCK_ASSERT(td, type) \ 404 mtx_assert((td)->td_lock, (type)) 405 406 #define THREAD_LOCK_BLOCKED_ASSERT(td, type) \ 407 do { \ 408 struct mtx *__m = (td)->td_lock; \ 409 if (__m != &blocked_lock) \ 410 mtx_assert(__m, (type)); \ 411 } while (0) 412 413 #ifdef INVARIANTS 414 #define THREAD_LOCKPTR_ASSERT(td, lock) \ 415 do { \ 416 struct mtx *__m; \ 417 __m = (td)->td_lock; \ 418 KASSERT(__m == (lock), \ 419 ("Thread %p lock %p does not match %p", td, __m, (lock))); \ 420 } while (0) 421 422 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) \ 423 do { \ 424 struct mtx *__m; \ 425 __m = (td)->td_lock; \ 426 KASSERT(__m == (lock) || __m == &blocked_lock, \ 427 ("Thread %p lock %p does not match %p", td, __m, (lock))); \ 428 } while (0) 429 430 #define TD_LOCKS_INC(td) ((td)->td_locks++) 431 #define TD_LOCKS_DEC(td) do { \ 432 KASSERT(SCHEDULER_STOPPED_TD(td) || (td)->td_locks > 0, \ 433 ("thread %p owns no locks", (td))); \ 434 (td)->td_locks--; \ 435 } while (0) 436 #else 437 #define THREAD_LOCKPTR_ASSERT(td, lock) 438 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) 439 440 #define TD_LOCKS_INC(td) 441 #define TD_LOCKS_DEC(td) 442 #endif 443 444 /* 445 * Flags kept in td_flags: 446 * To change these you MUST have the scheduler lock. 447 */ 448 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */ 449 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */ 450 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */ 451 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */ 452 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */ 453 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */ 454 #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */ 455 #define TDF_SIGWAIT 0x00000080 /* Ignore ignored signals */ 456 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */ 457 #define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */ 458 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */ 459 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */ 460 #define TDF_KQTICKLED 0x00001000 /* AST drain kqueue taskqueue */ 461 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */ 462 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */ 463 #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */ 464 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */ 465 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */ 466 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */ 467 #define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */ 468 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */ 469 #define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */ 470 #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */ 471 #define TDF_DOING_SA 0x00800000 /* Doing SINGLE_ALLPROC, do not unsuspend me */ 472 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */ 473 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */ 474 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */ 475 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */ 476 #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */ 477 #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */ 478 #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */ 479 480 /* Userland debug flags */ 481 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */ 482 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */ 483 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */ 484 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */ 485 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */ 486 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */ 487 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new 488 process */ 489 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child 490 only) */ 491 #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */ 492 #define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */ 493 #define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */ 494 #define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */ 495 #define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */ 496 #define TDB_STEP 0x00002000 /* (x86) PSL_T set for PT_STEP */ 497 #define TDB_SSWITCH 0x00004000 /* Suspended in ptracestop */ 498 #define TDB_COREDUMPRQ 0x00008000 /* Coredump request */ 499 500 /* 501 * "Private" flags kept in td_pflags: 502 * These are only written by curthread and thus need no locking. 503 */ 504 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */ 505 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */ 506 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */ 507 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */ 508 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */ 509 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */ 510 #define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */ 511 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */ 512 #define TDP_SIGFASTBLOCK 0x00000100 /* Fast sigblock active */ 513 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */ 514 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */ 515 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */ 516 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */ 517 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */ 518 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */ 519 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */ 520 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */ 521 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */ 522 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */ 523 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */ 524 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */ 525 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */ 526 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */ 527 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */ 528 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */ 529 #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */ 530 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */ 531 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */ 532 #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */ 533 #define TDP_UNUSED0 0x20000000 /* UNUSED */ 534 #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */ 535 #define TDP_SIGFASTPENDING 0x80000000 /* Pending signal due to sigfastblock */ 536 537 #define TDP2_SBPAGES 0x00000001 /* Owns sbusy on some pages */ 538 #define TDP2_COMPAT32RB 0x00000002 /* compat32 ABI for robust lists */ 539 #define TDP2_ACCT 0x00000004 /* Doing accounting */ 540 541 /* 542 * Reasons that the current thread can not be run yet. 543 * More than one may apply. 544 */ 545 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */ 546 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */ 547 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */ 548 #define TDI_LOCK 0x0008 /* Stopped on a lock. */ 549 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */ 550 551 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING) 552 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL) 553 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED) 554 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED) 555 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK) 556 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT) 557 #ifdef _KERNEL 558 #define TD_GET_STATE(td) atomic_load_int(&(td)->td_state) 559 #else 560 #define TD_GET_STATE(td) ((td)->td_state) 561 #endif 562 #define TD_IS_RUNNING(td) (TD_GET_STATE(td) == TDS_RUNNING) 563 #define TD_ON_RUNQ(td) (TD_GET_STATE(td) == TDS_RUNQ) 564 #define TD_CAN_RUN(td) (TD_GET_STATE(td) == TDS_CAN_RUN) 565 #define TD_IS_INHIBITED(td) (TD_GET_STATE(td) == TDS_INHIBITED) 566 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED) 567 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD) 568 569 #define TD_CAN_ABORT(td) (TD_ON_SLEEPQ((td)) && \ 570 ((td)->td_flags & TDF_SINTR) != 0) 571 572 #define KTDSTATE(td) \ 573 (((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \ 574 ((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \ 575 ((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" : \ 576 ((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \ 577 ((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding") 578 579 #define TD_SET_INHIB(td, inhib) do { \ 580 TD_SET_STATE(td, TDS_INHIBITED); \ 581 (td)->td_inhibitors |= (inhib); \ 582 } while (0) 583 584 #define TD_CLR_INHIB(td, inhib) do { \ 585 if (((td)->td_inhibitors & (inhib)) && \ 586 (((td)->td_inhibitors &= ~(inhib)) == 0)) \ 587 TD_SET_STATE(td, TDS_CAN_RUN); \ 588 } while (0) 589 590 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING) 591 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED) 592 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK) 593 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED) 594 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT) 595 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING) 596 597 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING) 598 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED) 599 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK) 600 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED) 601 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT) 602 603 #ifdef _KERNEL 604 #define TD_SET_STATE(td, state) atomic_store_int(&(td)->td_state, state) 605 #else 606 #define TD_SET_STATE(td, state) (td)->td_state = state 607 #endif 608 #define TD_SET_RUNNING(td) TD_SET_STATE(td, TDS_RUNNING) 609 #define TD_SET_RUNQ(td) TD_SET_STATE(td, TDS_RUNQ) 610 #define TD_SET_CAN_RUN(td) TD_SET_STATE(td, TDS_CAN_RUN) 611 612 613 #define TD_SBDRY_INTR(td) \ 614 (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0) 615 #define TD_SBDRY_ERRNO(td) \ 616 (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART) 617 618 /* 619 * Process structure. 620 */ 621 struct proc { 622 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */ 623 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */ 624 struct mtx p_slock; /* process spin lock */ 625 struct ucred *p_ucred; /* (c) Process owner's identity. */ 626 struct filedesc *p_fd; /* (b) Open files. */ 627 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */ 628 struct pwddesc *p_pd; /* (b) Cwd, chroot, jail, umask */ 629 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */ 630 struct plimit *p_limit; /* (c) Resource limits. */ 631 struct callout p_limco; /* (c) Limit callout handle */ 632 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */ 633 634 int p_flag; /* (c) P_* flags. */ 635 int p_flag2; /* (c) P2_* flags. */ 636 enum p_states { 637 PRS_NEW = 0, /* In creation */ 638 PRS_NORMAL, /* threads can be run. */ 639 PRS_ZOMBIE 640 } p_state; /* (j/c) Process status. */ 641 pid_t p_pid; /* (b) Process identifier. */ 642 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */ 643 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */ 644 struct proc *p_pptr; /* (c + e) Pointer to parent process. */ 645 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */ 646 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */ 647 struct proc *p_reaper; /* (e) My reaper. */ 648 LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants 649 (if I am reaper). */ 650 LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of 651 the same reaper. */ 652 struct mtx p_mtx; /* (n) Lock for this struct. */ 653 struct mtx p_statmtx; /* Lock for the stats */ 654 struct mtx p_itimmtx; /* Lock for the virt/prof timers */ 655 struct mtx p_profmtx; /* Lock for the profiling */ 656 struct ksiginfo *p_ksi; /* Locked by parent proc lock */ 657 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */ 658 #define p_siglist p_sigqueue.sq_signals 659 pid_t p_oppid; /* (c + e) Real parent pid. */ 660 661 /* The following fields are all zeroed upon creation in fork. */ 662 #define p_startzero p_vmspace 663 struct vmspace *p_vmspace; /* (b) Address space. */ 664 u_int p_swtick; /* (c) Tick when swapped in or out. */ 665 u_int p_cowgen; /* (c) Generation of COW pointers. */ 666 struct itimerval p_realtimer; /* (c) Alarm timer. */ 667 struct rusage p_ru; /* (a) Exit information. */ 668 struct rusage_ext p_rux; /* (cu) Internal resource usage. */ 669 struct rusage_ext p_crux; /* (c) Internal child resource usage. */ 670 int p_profthreads; /* (c) Num threads in addupc_task. */ 671 volatile int p_exitthreads; /* (j) Number of threads exiting */ 672 int p_traceflag; /* (o) Kernel trace points. */ 673 struct ktr_io_params *p_ktrioparms; /* (c + o) Params for ktrace. */ 674 struct vnode *p_textvp; /* (b) Vnode of executable. */ 675 struct vnode *p_textdvp; /* (b) Dir containing textvp. */ 676 char *p_binname; /* (b) Binary hardlink name. */ 677 u_int p_lock; /* (c) Proclock (prevent swap) count. */ 678 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */ 679 int p_sigparent; /* (c) Signal to parent on exit. */ 680 int p_sig; /* (n) For core dump/debugger XXX. */ 681 u_int p_ptevents; /* (c + e) ptrace() event mask. */ 682 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */ 683 struct thread *p_singlethread;/* (c + j) If single threading this is it */ 684 int p_suspcount; /* (j) Num threads in suspended mode. */ 685 struct thread *p_xthread; /* (c) Trap thread */ 686 int p_boundary_count;/* (j) Num threads at user boundary */ 687 int p_pendingcnt; /* how many signals are pending */ 688 struct itimers *p_itimers; /* (c) POSIX interval timers. */ 689 struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */ 690 u_int p_treeflag; /* (e) P_TREE flags */ 691 int p_pendingexits; /* (c) Count of pending thread exits. */ 692 struct filemon *p_filemon; /* (c) filemon-specific data. */ 693 int p_pdeathsig; /* (c) Signal from parent on exit. */ 694 int p_singlethr; /* (c) Count of threads doing 695 external thread_single() */ 696 /* End area that is zeroed on creation. */ 697 #define p_endzero p_magic 698 699 /* The following fields are all copied upon creation in fork. */ 700 #define p_startcopy p_endzero 701 u_int p_magic; /* (b) Magic number. */ 702 int p_osrel; /* (x) osreldate for the 703 binary (from ELF note, if any) */ 704 uint32_t p_fctl0; /* (x) ABI feature control, ELF note */ 705 char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */ 706 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */ 707 struct pargs *p_args; /* (c) Process arguments. */ 708 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */ 709 signed char p_nice; /* (c) Process "nice" value. */ 710 int p_fibnum; /* in this routing domain XXX MRT */ 711 pid_t p_reapsubtree; /* (e) Pid of the direct child of the 712 reaper which spawned 713 our subtree. */ 714 uint64_t p_elf_flags; /* (x) ELF flags */ 715 void *p_elf_brandinfo; /* (x) Elf_Brandinfo, NULL for 716 non ELF binaries. */ 717 /* End area that is copied on creation. */ 718 #define p_endcopy p_xexit 719 720 u_int p_xexit; /* (c) Exit code. */ 721 u_int p_xsig; /* (c) Stop/kill sig. */ 722 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */ 723 struct knlist *p_klist; /* (c) Knotes attached to this proc. */ 724 int p_numthreads; /* (c) Number of threads. */ 725 struct mdproc p_md; /* Any machine-dependent fields. */ 726 struct callout p_itcallout; /* (h + c) Interval timer callout. */ 727 u_short p_acflag; /* (c) Accounting flags. */ 728 struct proc *p_peers; /* (r) */ 729 struct proc *p_leader; /* (b) */ 730 void *p_emuldata; /* (c) Emulator state data. */ 731 struct label *p_label; /* (*) Proc (not subject) MAC label. */ 732 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */ 733 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/ 734 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */ 735 struct cv p_pwait; /* (*) wait cv for exit/exec. */ 736 uint64_t p_prev_runtime; /* (c) Resource usage accounting. */ 737 struct racct *p_racct; /* (b) Resource accounting. */ 738 int p_throttled; /* (c) Flag for racct pcpu throttling */ 739 /* 740 * An orphan is the child that has been re-parented to the 741 * debugger as a result of attaching to it. Need to keep 742 * track of them for parent to be able to collect the exit 743 * status of what used to be children. 744 */ 745 LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */ 746 LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */ 747 748 TAILQ_HEAD(, kq_timer_cb_data) p_kqtim_stop; /* (c) */ 749 }; 750 751 #define p_session p_pgrp->pg_session 752 #define p_pgid p_pgrp->pg_id 753 754 #define NOCPU (-1) /* For when we aren't on a CPU. */ 755 #define NOCPU_OLD (255) 756 #define MAXCPU_OLD (254) 757 758 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock) 759 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock) 760 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type)) 761 762 #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx) 763 #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx) 764 #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type)) 765 766 #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx) 767 #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx) 768 #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type)) 769 770 #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx) 771 #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx) 772 #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type)) 773 774 /* These flags are kept in p_flag. */ 775 #define P_ADVLOCK 0x00000001 /* Process may hold a POSIX advisory 776 lock. */ 777 #define P_CONTROLT 0x00000002 /* Has a controlling terminal. */ 778 #define P_KPROC 0x00000004 /* Kernel process. */ 779 #define P_UNUSED3 0x00000008 /* --available-- */ 780 #define P_PPWAIT 0x00000010 /* Parent is waiting for child to 781 exec/exit. */ 782 #define P_PROFIL 0x00000020 /* Has started profiling. */ 783 #define P_STOPPROF 0x00000040 /* Has thread requesting to stop 784 profiling. */ 785 #define P_HADTHREADS 0x00000080 /* Has had threads (no cleanup 786 shortcuts) */ 787 #define P_SUGID 0x00000100 /* Had set id privileges since last 788 exec. */ 789 #define P_SYSTEM 0x00000200 /* System proc: no sigs, stats or 790 swapping. */ 791 #define P_SINGLE_EXIT 0x00000400 /* Threads suspending should exit, 792 not wait. */ 793 #define P_TRACED 0x00000800 /* Debugged process being traced. */ 794 #define P_WAITED 0x00001000 /* Someone is waiting for us. */ 795 #define P_WEXIT 0x00002000 /* Working on exiting. */ 796 #define P_EXEC 0x00004000 /* Process called exec. */ 797 #define P_WKILLED 0x00008000 /* Killed, go to kernel/user boundary 798 ASAP. */ 799 #define P_CONTINUED 0x00010000 /* Proc has continued from a stopped 800 state. */ 801 #define P_STOPPED_SIG 0x00020000 /* Stopped due to SIGSTOP/SIGTSTP. */ 802 #define P_STOPPED_TRACE 0x00040000 /* Stopped because of tracing. */ 803 #define P_STOPPED_SINGLE 0x00080000 /* Only 1 thread can continue (not to 804 user). */ 805 #define P_PROTECTED 0x00100000 /* Do not kill on memory overcommit. */ 806 #define P_SIGEVENT 0x00200000 /* Process pending signals changed. */ 807 #define P_SINGLE_BOUNDARY 0x00400000 /* Threads should suspend at user 808 boundary. */ 809 #define P_HWPMC 0x00800000 /* Process is using HWPMCs */ 810 #define P_JAILED 0x01000000 /* Process is in jail. */ 811 #define P_TOTAL_STOP 0x02000000 /* Stopped in stop_all_proc. */ 812 #define P_INEXEC 0x04000000 /* Process is in execve(). */ 813 #define P_STATCHILD 0x08000000 /* Child process stopped or exited. */ 814 #define P_INMEM 0x10000000 /* Loaded into memory. */ 815 #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */ 816 #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */ 817 #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */ 818 819 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE) 820 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED) 821 #define P_KILLED(p) ((p)->p_flag & P_WKILLED) 822 823 /* These flags are kept in p_flag2. */ 824 #define P2_INHERIT_PROTECTED 0x00000001 /* New children get 825 P_PROTECTED. */ 826 #define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or 827 coredumps. */ 828 #define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on 829 exec(2). */ 830 #define P2_AST_SU 0x00000008 /* Handles SU ast for 831 kthreads. */ 832 #define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not 833 yet handled. */ 834 #define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */ 835 #define P2_ASLR_ENABLE 0x00000040 /* Force enable ASLR. */ 836 #define P2_ASLR_DISABLE 0x00000080 /* Force disable ASLR. */ 837 #define P2_ASLR_IGNSTART 0x00000100 /* Enable ASLR to consume sbrk 838 area. */ 839 #define P2_PROTMAX_ENABLE 0x00000200 /* Force enable implied 840 PROT_MAX. */ 841 #define P2_PROTMAX_DISABLE 0x00000400 /* Force disable implied 842 PROT_MAX. */ 843 #define P2_STKGAP_DISABLE 0x00000800 /* Disable stack gap for 844 MAP_STACK */ 845 #define P2_STKGAP_DISABLE_EXEC 0x00001000 /* Stack gap disabled 846 after exec */ 847 #define P2_ITSTOPPED 0x00002000 848 #define P2_PTRACEREQ 0x00004000 /* Active ptrace req */ 849 #define P2_NO_NEW_PRIVS 0x00008000 /* Ignore setuid */ 850 #define P2_WXORX_DISABLE 0x00010000 /* WX mappings enabled */ 851 #define P2_WXORX_ENABLE_EXEC 0x00020000 /* WXORX enabled after exec */ 852 #define P2_WEXIT 0x00040000 /* exit just started, no 853 external thread_single() is 854 permitted */ 855 856 /* Flags protected by proctree_lock, kept in p_treeflags. */ 857 #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */ 858 #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan 859 list */ 860 #define P_TREE_REAPER 0x00000004 /* Reaper of subtree */ 861 #define P_TREE_GRPEXITED 0x00000008 /* exit1() done with job ctl */ 862 863 /* 864 * These were process status values (p_stat), now they are only used in 865 * legacy conversion code. 866 */ 867 #define SIDL 1 /* Process being created by fork. */ 868 #define SRUN 2 /* Currently runnable. */ 869 #define SSLEEP 3 /* Sleeping on an address. */ 870 #define SSTOP 4 /* Process debugging or suspension. */ 871 #define SZOMB 5 /* Awaiting collection by parent. */ 872 #define SWAIT 6 /* Waiting for interrupt. */ 873 #define SLOCK 7 /* Blocked on a lock. */ 874 875 #define P_MAGIC 0xbeefface 876 877 #ifdef _KERNEL 878 879 /* Types and flags for mi_switch(). */ 880 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */ 881 #define SWT_NONE 0 /* Unspecified switch. */ 882 #define SWT_PREEMPT 1 /* Switching due to preemption. */ 883 #define SWT_OWEPREEMPT 2 /* Switching due to owepreempt. */ 884 #define SWT_TURNSTILE 3 /* Turnstile contention. */ 885 #define SWT_SLEEPQ 4 /* Sleepq wait. */ 886 #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */ 887 #define SWT_RELINQUISH 6 /* yield call. */ 888 #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */ 889 #define SWT_IDLE 8 /* Switching from the idle thread. */ 890 #define SWT_IWAIT 9 /* Waiting for interrupts. */ 891 #define SWT_SUSPEND 10 /* Thread suspended. */ 892 #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */ 893 #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */ 894 #define SWT_COUNT 13 /* Number of switch types. */ 895 /* Flags */ 896 #define SW_VOL 0x0100 /* Voluntary switch. */ 897 #define SW_INVOL 0x0200 /* Involuntary switch. */ 898 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */ 899 900 /* How values for thread_single(). */ 901 #define SINGLE_NO_EXIT 0 902 #define SINGLE_EXIT 1 903 #define SINGLE_BOUNDARY 2 904 #define SINGLE_ALLPROC 3 905 906 #ifdef MALLOC_DECLARE 907 MALLOC_DECLARE(M_PARGS); 908 MALLOC_DECLARE(M_SESSION); 909 MALLOC_DECLARE(M_SUBPROC); 910 #endif 911 912 #define FOREACH_PROC_IN_SYSTEM(p) \ 913 LIST_FOREACH((p), &allproc, p_list) 914 #define FOREACH_THREAD_IN_PROC(p, td) \ 915 TAILQ_FOREACH((td), &(p)->p_threads, td_plist) 916 917 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads) 918 919 /* 920 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit 921 * in a pid_t, as it is used to represent "no process group". 922 */ 923 #define PID_MAX 99999 924 #define NO_PID 100000 925 #define THREAD0_TID NO_PID 926 extern pid_t pid_max; 927 928 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p)) 929 930 /* Lock and unlock a process. */ 931 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx) 932 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx) 933 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx) 934 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx) 935 #define PROC_WAIT_UNLOCKED(p) mtx_wait_unlocked(&(p)->p_mtx) 936 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type)) 937 938 /* Lock and unlock a process group. */ 939 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx) 940 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx) 941 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx) 942 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type)) 943 944 #define PGRP_LOCK_PGSIGNAL(pg) do { \ 945 if ((pg) != NULL) \ 946 PGRP_LOCK(pg); \ 947 } while (0) 948 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \ 949 if ((pg) != NULL) \ 950 PGRP_UNLOCK(pg); \ 951 } while (0) 952 953 /* Lock and unlock a session. */ 954 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx) 955 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx) 956 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx) 957 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type)) 958 959 /* 960 * Non-zero p_lock ensures that: 961 * - exit1() is not performed until p_lock reaches zero; 962 * - the process' threads stack are not swapped out if they are currently 963 * not (P_INMEM). 964 * 965 * PHOLD() asserts that the process (except the current process) is 966 * not exiting, increments p_lock and swaps threads stacks into memory, 967 * if needed. 968 * _PHOLD() is same as PHOLD(), it takes the process locked. 969 * _PHOLD_LITE() also takes the process locked, but comparing with 970 * _PHOLD(), it only guarantees that exit1() is not executed, 971 * faultin() is not called. 972 */ 973 #define PHOLD(p) do { \ 974 PROC_LOCK(p); \ 975 _PHOLD(p); \ 976 PROC_UNLOCK(p); \ 977 } while (0) 978 #define _PHOLD(p) do { \ 979 PROC_LOCK_ASSERT((p), MA_OWNED); \ 980 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 981 ("PHOLD of exiting process %p", p)); \ 982 (p)->p_lock++; \ 983 if (((p)->p_flag & P_INMEM) == 0) \ 984 faultin((p)); \ 985 } while (0) 986 #define _PHOLD_LITE(p) do { \ 987 PROC_LOCK_ASSERT((p), MA_OWNED); \ 988 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 989 ("PHOLD of exiting process %p", p)); \ 990 (p)->p_lock++; \ 991 } while (0) 992 #define PROC_ASSERT_HELD(p) do { \ 993 KASSERT((p)->p_lock > 0, ("process %p not held", p)); \ 994 } while (0) 995 996 #define PRELE(p) do { \ 997 PROC_LOCK((p)); \ 998 _PRELE((p)); \ 999 PROC_UNLOCK((p)); \ 1000 } while (0) 1001 #define _PRELE(p) do { \ 1002 PROC_LOCK_ASSERT((p), MA_OWNED); \ 1003 PROC_ASSERT_HELD(p); \ 1004 (--(p)->p_lock); \ 1005 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \ 1006 wakeup(&(p)->p_lock); \ 1007 } while (0) 1008 #define PROC_ASSERT_NOT_HELD(p) do { \ 1009 KASSERT((p)->p_lock == 0, ("process %p held", p)); \ 1010 } while (0) 1011 1012 #define PROC_UPDATE_COW(p) do { \ 1013 struct proc *_p = (p); \ 1014 PROC_LOCK_ASSERT((_p), MA_OWNED); \ 1015 atomic_store_int(&_p->p_cowgen, _p->p_cowgen + 1); \ 1016 } while (0) 1017 1018 #define PROC_COW_CHANGECOUNT(td, p) ({ \ 1019 struct thread *_td = (td); \ 1020 struct proc *_p = (p); \ 1021 MPASS(_td == curthread); \ 1022 PROC_LOCK_ASSERT(_p, MA_OWNED); \ 1023 _p->p_cowgen - _td->td_cowgen; \ 1024 }) 1025 1026 /* Check whether a thread is safe to be swapped out. */ 1027 #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP) 1028 1029 /* Control whether or not it is safe for curthread to sleep. */ 1030 #define THREAD_NO_SLEEPING() do { \ 1031 curthread->td_no_sleeping++; \ 1032 MPASS(curthread->td_no_sleeping > 0); \ 1033 } while (0) 1034 1035 #define THREAD_SLEEPING_OK() do { \ 1036 MPASS(curthread->td_no_sleeping > 0); \ 1037 curthread->td_no_sleeping--; \ 1038 } while (0) 1039 1040 #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0) 1041 1042 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash]) 1043 #define PIDHASHLOCK(pid) (&pidhashtbl_lock[((pid) & pidhashlock)]) 1044 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl; 1045 extern struct sx *pidhashtbl_lock; 1046 extern u_long pidhash; 1047 extern u_long pidhashlock; 1048 1049 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash]) 1050 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl; 1051 extern u_long pgrphash; 1052 1053 extern struct sx allproc_lock; 1054 extern int allproc_gen; 1055 extern struct sx proctree_lock; 1056 extern struct mtx ppeers_lock; 1057 extern struct mtx procid_lock; 1058 extern struct proc proc0; /* Process slot for swapper. */ 1059 extern struct thread0_storage thread0_st; /* Primary thread in proc0. */ 1060 #define thread0 (thread0_st.t0st_thread) 1061 extern struct vmspace vmspace0; /* VM space for proc0. */ 1062 extern int hogticks; /* Limit on kernel cpu hogs. */ 1063 extern int lastpid; 1064 extern int nprocs, maxproc; /* Current and max number of procs. */ 1065 extern int maxprocperuid; /* Max procs per uid. */ 1066 extern u_long ps_arg_cache_limit; 1067 1068 LIST_HEAD(proclist, proc); 1069 TAILQ_HEAD(procqueue, proc); 1070 TAILQ_HEAD(threadqueue, thread); 1071 extern struct proclist allproc; /* List of all processes. */ 1072 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */ 1073 1074 extern struct uma_zone *proc_zone; 1075 extern struct uma_zone *pgrp_zone; 1076 1077 struct proc *pfind(pid_t); /* Find process by id. */ 1078 struct proc *pfind_any(pid_t); /* Find (zombie) process by id. */ 1079 struct proc *pfind_any_locked(pid_t pid); /* Find process by id, locked. */ 1080 struct pgrp *pgfind(pid_t); /* Find process group by id. */ 1081 void pidhash_slockall(void); /* Shared lock all pid hash lists. */ 1082 void pidhash_sunlockall(void); /* Shared unlock all pid hash lists. */ 1083 1084 struct fork_req { 1085 int fr_flags; 1086 int fr_pages; 1087 int *fr_pidp; 1088 struct proc **fr_procp; 1089 int *fr_pd_fd; 1090 int fr_pd_flags; 1091 struct filecaps *fr_pd_fcaps; 1092 int fr_flags2; 1093 #define FR2_DROPSIG_CAUGHT 0x00000001 /* Drop caught non-DFL signals */ 1094 #define FR2_SHARE_PATHS 0x00000002 /* Invert sense of RFFDG for paths */ 1095 #define FR2_KPROC 0x00000004 /* Create a kernel process */ 1096 }; 1097 1098 /* 1099 * pget() flags. 1100 */ 1101 #define PGET_HOLD 0x00001 /* Hold the process. */ 1102 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */ 1103 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */ 1104 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */ 1105 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */ 1106 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */ 1107 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */ 1108 1109 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT) 1110 1111 int pget(pid_t pid, int flags, struct proc **pp); 1112 1113 void ast(struct trapframe *framep); 1114 struct thread *choosethread(void); 1115 int cr_cansee(struct ucred *u1, struct ucred *u2); 1116 int cr_canseesocket(struct ucred *cred, struct socket *so); 1117 int cr_canseeothergids(struct ucred *u1, struct ucred *u2); 1118 int cr_canseeotheruids(struct ucred *u1, struct ucred *u2); 1119 int cr_canseejailproc(struct ucred *u1, struct ucred *u2); 1120 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum); 1121 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp, 1122 struct session *sess); 1123 int enterthispgrp(struct proc *p, struct pgrp *pgrp); 1124 void faultin(struct proc *p); 1125 int fork1(struct thread *, struct fork_req *); 1126 void fork_rfppwait(struct thread *); 1127 void fork_exit(void (*)(void *, struct trapframe *), void *, 1128 struct trapframe *); 1129 void fork_return(struct thread *, struct trapframe *); 1130 int inferior(struct proc *p); 1131 void itimer_proc_continue(struct proc *p); 1132 void kqtimer_proc_continue(struct proc *p); 1133 void kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry, 1134 int *resident_count, bool *super); 1135 void kern_yield(int); 1136 void kick_proc0(void); 1137 void killjobc(void); 1138 int leavepgrp(struct proc *p); 1139 int maybe_preempt(struct thread *td); 1140 void maybe_yield(void); 1141 void mi_switch(int flags); 1142 int p_candebug(struct thread *td, struct proc *p); 1143 int p_cansee(struct thread *td, struct proc *p); 1144 int p_cansched(struct thread *td, struct proc *p); 1145 int p_cansignal(struct thread *td, struct proc *p, int signum); 1146 int p_canwait(struct thread *td, struct proc *p); 1147 struct pargs *pargs_alloc(int len); 1148 void pargs_drop(struct pargs *pa); 1149 void pargs_hold(struct pargs *pa); 1150 void proc_add_orphan(struct proc *child, struct proc *parent); 1151 int proc_get_binpath(struct proc *p, char *binname, char **fullpath, 1152 char **freepath); 1153 int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb); 1154 int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb); 1155 int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb); 1156 void procinit(void); 1157 int proc_iterate(int (*cb)(struct proc *, void *), void *cbarg); 1158 void proc_linkup0(struct proc *p, struct thread *td); 1159 void proc_linkup(struct proc *p, struct thread *td); 1160 struct proc *proc_realparent(struct proc *child); 1161 void proc_reap(struct thread *td, struct proc *p, int *status, int options); 1162 void proc_reparent(struct proc *child, struct proc *newparent, bool set_oppid); 1163 void proc_set_p2_wexit(struct proc *p); 1164 void proc_set_traced(struct proc *p, bool stop); 1165 void proc_wkilled(struct proc *p); 1166 struct pstats *pstats_alloc(void); 1167 void pstats_fork(struct pstats *src, struct pstats *dst); 1168 void pstats_free(struct pstats *ps); 1169 void proc_clear_orphan(struct proc *p); 1170 void reaper_abandon_children(struct proc *p, bool exiting); 1171 int securelevel_ge(struct ucred *cr, int level); 1172 int securelevel_gt(struct ucred *cr, int level); 1173 void sess_hold(struct session *); 1174 void sess_release(struct session *); 1175 int setrunnable(struct thread *, int); 1176 void setsugid(struct proc *p); 1177 int should_yield(void); 1178 int sigonstack(size_t sp); 1179 void stopevent(struct proc *, u_int, u_int); 1180 struct thread *tdfind(lwpid_t, pid_t); 1181 void threadinit(void); 1182 void tidhash_add(struct thread *); 1183 void tidhash_remove(struct thread *); 1184 void cpu_idle(int); 1185 int cpu_idle_wakeup(int); 1186 extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */ 1187 void cpu_switch(struct thread *, struct thread *, struct mtx *); 1188 void cpu_throw(struct thread *, struct thread *) __dead2; 1189 bool curproc_sigkilled(void); 1190 void userret(struct thread *, struct trapframe *); 1191 1192 void cpu_exit(struct thread *); 1193 void exit1(struct thread *, int, int) __dead2; 1194 void cpu_copy_thread(struct thread *td, struct thread *td0); 1195 bool cpu_exec_vmspace_reuse(struct proc *p, struct vm_map *map); 1196 int cpu_fetch_syscall_args(struct thread *td); 1197 void cpu_fork(struct thread *, struct proc *, struct thread *, int); 1198 void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *); 1199 int cpu_procctl(struct thread *td, int idtype, id_t id, int com, 1200 void *data); 1201 void cpu_set_syscall_retval(struct thread *, int); 1202 void cpu_set_upcall(struct thread *, void (*)(void *), void *, 1203 stack_t *); 1204 int cpu_set_user_tls(struct thread *, void *tls_base); 1205 void cpu_thread_alloc(struct thread *); 1206 void cpu_thread_clean(struct thread *); 1207 void cpu_thread_exit(struct thread *); 1208 void cpu_thread_free(struct thread *); 1209 void cpu_thread_swapin(struct thread *); 1210 void cpu_thread_swapout(struct thread *); 1211 struct thread *thread_alloc(int pages); 1212 int thread_alloc_stack(struct thread *, int pages); 1213 int thread_check_susp(struct thread *td, bool sleep); 1214 void thread_cow_get_proc(struct thread *newtd, struct proc *p); 1215 void thread_cow_get(struct thread *newtd, struct thread *td); 1216 void thread_cow_free(struct thread *td); 1217 void thread_cow_update(struct thread *td); 1218 void thread_cow_synced(struct thread *td); 1219 int thread_create(struct thread *td, struct rtprio *rtp, 1220 int (*initialize_thread)(struct thread *, void *), void *thunk); 1221 void thread_exit(void) __dead2; 1222 void thread_free(struct thread *td); 1223 void thread_link(struct thread *td, struct proc *p); 1224 void thread_reap_barrier(void); 1225 int thread_single(struct proc *p, int how); 1226 void thread_single_end(struct proc *p, int how); 1227 void thread_stash(struct thread *td); 1228 void thread_stopped(struct proc *p); 1229 void childproc_stopped(struct proc *child, int reason); 1230 void childproc_continued(struct proc *child); 1231 void childproc_exited(struct proc *child); 1232 void thread_run_flash(struct thread *td); 1233 int thread_suspend_check(int how); 1234 bool thread_suspend_check_needed(void); 1235 void thread_suspend_switch(struct thread *, struct proc *p); 1236 void thread_suspend_one(struct thread *td); 1237 void thread_unlink(struct thread *td); 1238 void thread_unsuspend(struct proc *p); 1239 void thread_wait(struct proc *p); 1240 1241 bool stop_all_proc_block(void); 1242 void stop_all_proc_unblock(void); 1243 void stop_all_proc(void); 1244 void resume_all_proc(void); 1245 1246 static __inline int 1247 curthread_pflags_set(int flags) 1248 { 1249 struct thread *td; 1250 int save; 1251 1252 td = curthread; 1253 save = ~flags | (td->td_pflags & flags); 1254 td->td_pflags |= flags; 1255 return (save); 1256 } 1257 1258 static __inline void 1259 curthread_pflags_restore(int save) 1260 { 1261 1262 curthread->td_pflags &= save; 1263 } 1264 1265 static __inline int 1266 curthread_pflags2_set(int flags) 1267 { 1268 struct thread *td; 1269 int save; 1270 1271 td = curthread; 1272 save = ~flags | (td->td_pflags2 & flags); 1273 td->td_pflags2 |= flags; 1274 return (save); 1275 } 1276 1277 static __inline void 1278 curthread_pflags2_restore(int save) 1279 { 1280 1281 curthread->td_pflags2 &= save; 1282 } 1283 1284 static __inline bool 1285 kstack_contains(struct thread *td, vm_offset_t va, size_t len) 1286 { 1287 return (va >= td->td_kstack && va + len >= va && 1288 va + len <= td->td_kstack + td->td_kstack_pages * PAGE_SIZE); 1289 } 1290 1291 static __inline __pure2 struct td_sched * 1292 td_get_sched(struct thread *td) 1293 { 1294 1295 return ((struct td_sched *)&td[1]); 1296 } 1297 1298 extern void (*softdep_ast_cleanup)(struct thread *); 1299 static __inline void 1300 td_softdep_cleanup(struct thread *td) 1301 { 1302 1303 if (td->td_su != NULL && softdep_ast_cleanup != NULL) 1304 softdep_ast_cleanup(td); 1305 } 1306 1307 #define PROC_ID_PID 0 1308 #define PROC_ID_GROUP 1 1309 #define PROC_ID_SESSION 2 1310 #define PROC_ID_REAP 3 1311 1312 void proc_id_set(int type, pid_t id); 1313 void proc_id_set_cond(int type, pid_t id); 1314 void proc_id_clear(int type, pid_t id); 1315 1316 EVENTHANDLER_LIST_DECLARE(process_ctor); 1317 EVENTHANDLER_LIST_DECLARE(process_dtor); 1318 EVENTHANDLER_LIST_DECLARE(process_init); 1319 EVENTHANDLER_LIST_DECLARE(process_fini); 1320 EVENTHANDLER_LIST_DECLARE(process_exit); 1321 EVENTHANDLER_LIST_DECLARE(process_fork); 1322 EVENTHANDLER_LIST_DECLARE(process_exec); 1323 1324 EVENTHANDLER_LIST_DECLARE(thread_ctor); 1325 EVENTHANDLER_LIST_DECLARE(thread_dtor); 1326 EVENTHANDLER_LIST_DECLARE(thread_init); 1327 1328 #endif /* _KERNEL */ 1329 1330 #endif /* !_SYS_PROC_H_ */ 1331