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 #include <sys/condvar.h> 46 #ifndef _KERNEL 47 #include <sys/filedesc.h> 48 #endif 49 #include <sys/queue.h> 50 #include <sys/_lock.h> 51 #include <sys/lock_profile.h> 52 #include <sys/_mutex.h> 53 #include <sys/osd.h> 54 #include <sys/priority.h> 55 #include <sys/rtprio.h> /* XXX. */ 56 #include <sys/runq.h> 57 #include <sys/resource.h> 58 #include <sys/sigio.h> 59 #include <sys/signal.h> 60 #include <sys/signalvar.h> 61 #ifndef _KERNEL 62 #include <sys/time.h> /* For structs itimerval, timeval. */ 63 #else 64 #include <sys/pcpu.h> 65 #include <sys/systm.h> 66 #endif 67 #include <sys/ucontext.h> 68 #include <sys/ucred.h> 69 #include <sys/types.h> 70 #include <sys/domainset.h> 71 72 #include <machine/proc.h> /* Machine-dependent proc substruct. */ 73 #ifdef _KERNEL 74 #include <machine/cpu.h> 75 #endif 76 77 78 /* 79 * One structure allocated per session. 80 * 81 * List of locks 82 * (m) locked by s_mtx mtx 83 * (e) locked by proctree_lock sx 84 * (c) const until freeing 85 */ 86 struct session { 87 u_int s_count; /* Ref cnt; pgrps in session - atomic. */ 88 struct proc *s_leader; /* (m + e) Session leader. */ 89 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */ 90 struct cdev_priv *s_ttydp; /* (m) Device of controlling tty. */ 91 struct tty *s_ttyp; /* (e) Controlling tty. */ 92 pid_t s_sid; /* (c) Session ID. */ 93 /* (m) Setlogin() name: */ 94 char s_login[roundup(MAXLOGNAME, sizeof(long))]; 95 struct mtx s_mtx; /* Mutex to protect members. */ 96 }; 97 98 /* 99 * One structure allocated per process group. 100 * 101 * List of locks 102 * (m) locked by pg_mtx mtx 103 * (e) locked by proctree_lock sx 104 * (c) const until freeing 105 */ 106 struct pgrp { 107 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */ 108 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */ 109 struct session *pg_session; /* (c) Pointer to session. */ 110 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */ 111 pid_t pg_id; /* (c) Process group id. */ 112 int pg_jobc; /* (m) Job control process count. */ 113 struct mtx pg_mtx; /* Mutex to protect members */ 114 }; 115 116 /* 117 * pargs, used to hold a copy of the command line, if it had a sane length. 118 */ 119 struct pargs { 120 u_int ar_ref; /* Reference count. */ 121 u_int ar_length; /* Length. */ 122 u_char ar_args[1]; /* Arguments. */ 123 }; 124 125 /*- 126 * Description of a process. 127 * 128 * This structure contains the information needed to manage a thread of 129 * control, known in UN*X as a process; it has references to substructures 130 * containing descriptions of things that the process uses, but may share 131 * with related processes. The process structure and the substructures 132 * are always addressable except for those marked "(CPU)" below, 133 * which might be addressable only on a processor on which the process 134 * is running. 135 * 136 * Below is a key of locks used to protect each member of struct proc. The 137 * lock is indicated by a reference to a specific character in parens in the 138 * associated comment. 139 * * - not yet protected 140 * a - only touched by curproc or parent during fork/wait 141 * b - created at fork, never changes 142 * (exception aiods switch vmspaces, but they are also 143 * marked 'P_SYSTEM' so hopefully it will be left alone) 144 * c - locked by proc mtx 145 * d - locked by allproc_lock lock 146 * e - locked by proctree_lock lock 147 * f - session mtx 148 * g - process group mtx 149 * h - callout_lock mtx 150 * i - by curproc or the master session mtx 151 * j - locked by proc slock 152 * k - only accessed by curthread 153 * k*- only accessed by curthread and from an interrupt 154 * kx- only accessed by curthread and by debugger 155 * l - the attaching proc or attaching proc parent 156 * m - Giant 157 * n - not locked, lazy 158 * o - ktrace lock 159 * q - td_contested lock 160 * r - p_peers lock 161 * s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9) 162 * t - thread lock 163 * u - process stat lock 164 * w - process timer lock 165 * x - created at fork, only changes during single threading in exec 166 * y - created at first aio, doesn't change until exit or exec at which 167 * point we are single-threaded and only curthread changes it 168 * z - zombie threads lock 169 * 170 * If the locking key specifies two identifiers (for example, p_pptr) then 171 * either lock is sufficient for read access, but both locks must be held 172 * for write access. 173 */ 174 struct cpuset; 175 struct filecaps; 176 struct filemon; 177 struct kaioinfo; 178 struct kaudit_record; 179 struct kdtrace_proc; 180 struct kdtrace_thread; 181 struct mqueue_notifier; 182 struct nlminfo; 183 struct p_sched; 184 struct proc; 185 struct procdesc; 186 struct racct; 187 struct sbuf; 188 struct sleepqueue; 189 struct socket; 190 struct syscall_args; 191 struct td_sched; 192 struct thread; 193 struct trapframe; 194 struct turnstile; 195 struct vm_map; 196 struct vm_map_entry; 197 198 /* 199 * XXX: Does this belong in resource.h or resourcevar.h instead? 200 * Resource usage extension. The times in rusage structs in the kernel are 201 * never up to date. The actual times are kept as runtimes and tick counts 202 * (with control info in the "previous" times), and are converted when 203 * userland asks for rusage info. Backwards compatibility prevents putting 204 * this directly in the user-visible rusage struct. 205 * 206 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux. 207 * Locking for td_rux: (t) for all fields. 208 */ 209 struct rusage_ext { 210 uint64_t rux_runtime; /* (cu) Real time. */ 211 uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */ 212 uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */ 213 uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */ 214 uint64_t rux_uu; /* (c) Previous user time in usec. */ 215 uint64_t rux_su; /* (c) Previous sys time in usec. */ 216 uint64_t rux_tu; /* (c) Previous total time in usec. */ 217 }; 218 219 /* 220 * Kernel runnable context (thread). 221 * This is what is put to sleep and reactivated. 222 * Thread context. Processes may have multiple threads. 223 */ 224 struct thread { 225 struct mtx *volatile td_lock; /* replaces sched lock */ 226 struct proc *td_proc; /* (*) Associated process. */ 227 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */ 228 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */ 229 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */ 230 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */ 231 LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */ 232 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */ 233 struct domainset_ref td_domain; /* (a) NUMA policy */ 234 struct seltd *td_sel; /* Select queue/channel. */ 235 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */ 236 struct turnstile *td_turnstile; /* (k) Associated turnstile. */ 237 struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */ 238 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */ 239 lwpid_t td_tid; /* (b) Thread ID. */ 240 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */ 241 #define td_siglist td_sigqueue.sq_signals 242 u_char td_lend_user_pri; /* (t) Lend user pri. */ 243 244 /* Cleared during fork1() */ 245 #define td_startzero td_flags 246 int td_flags; /* (t) TDF_* flags. */ 247 int td_inhibitors; /* (t) Why can not run. */ 248 int td_pflags; /* (k) Private thread (TDP_*) flags. */ 249 int td_dupfd; /* (k) Ret value from fdopen. XXX */ 250 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */ 251 void *td_wchan; /* (t) Sleep address. */ 252 const char *td_wmesg; /* (t) Reason for sleep. */ 253 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */ 254 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */ 255 short td_locks; /* (k) Debug: count of non-spin locks */ 256 short td_rw_rlocks; /* (k) Count of rwlock read locks. */ 257 short td_lk_slocks; /* (k) Count of lockmgr shared locks. */ 258 short td_stopsched; /* (k) Scheduler stopped. */ 259 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */ 260 const char *td_lockname; /* (t) Name of lock blocked on. */ 261 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */ 262 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */ 263 int td_intr_nesting_level; /* (k) Interrupt recursion. */ 264 int td_pinned; /* (k) Temporary cpu pin count. */ 265 struct ucred *td_ucred; /* (k) Reference to credentials. */ 266 struct plimit *td_limit; /* (k) Resource limits. */ 267 int td_slptick; /* (t) Time at sleep. */ 268 int td_blktick; /* (t) Time spent blocked. */ 269 int td_swvoltick; /* (t) Time at last SW_VOL switch. */ 270 int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */ 271 u_int td_cow; /* (*) Number of copy-on-write faults */ 272 struct rusage td_ru; /* (t) rusage information. */ 273 struct rusage_ext td_rux; /* (t) Internal rusage information. */ 274 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */ 275 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */ 276 u_int td_pticks; /* (t) Statclock hits for profiling */ 277 u_int td_sticks; /* (t) Statclock hits in system mode. */ 278 u_int td_iticks; /* (t) Statclock hits in intr mode. */ 279 u_int td_uticks; /* (t) Statclock hits in user mode. */ 280 int td_intrval; /* (t) Return value for sleepq. */ 281 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */ 282 volatile u_int td_generation; /* (k) For detection of preemption */ 283 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */ 284 int td_xsig; /* (c) Signal for ptrace */ 285 u_long td_profil_addr; /* (k) Temporary addr until AST. */ 286 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */ 287 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */ 288 struct file *td_fpop; /* (k) file referencing cdev under op */ 289 int td_dbgflags; /* (c) Userland debugger flags */ 290 siginfo_t td_si; /* (c) For debugger or core file */ 291 int td_ng_outbound; /* (k) Thread entered ng from above. */ 292 struct osd td_osd; /* (k) Object specific data. */ 293 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */ 294 pid_t td_dbg_forked; /* (c) Child pid for debugger. */ 295 u_int td_vp_reserv; /* (k) Count of reserved vnodes. */ 296 int td_no_sleeping; /* (k) Sleeping disabled count. */ 297 void *td_su; /* (k) FFS SU private */ 298 sbintime_t td_sleeptimo; /* (t) Sleep timeout. */ 299 int td_rtcgen; /* (s) rtc_generation of abs. sleep */ 300 size_t td_vslock_sz; /* (k) amount of vslock-ed space */ 301 #define td_endzero td_sigmask 302 303 /* Copied during fork1() or create_thread(). */ 304 #define td_startcopy td_endzero 305 sigset_t td_sigmask; /* (c) Current signal mask. */ 306 u_char td_rqindex; /* (t) Run queue index. */ 307 u_char td_base_pri; /* (t) Thread base kernel priority. */ 308 u_char td_priority; /* (t) Thread active priority. */ 309 u_char td_pri_class; /* (t) Scheduling class. */ 310 u_char td_user_pri; /* (t) User pri from estcpu and nice. */ 311 u_char td_base_user_pri; /* (t) Base user pri */ 312 uintptr_t td_rb_list; /* (k) Robust list head. */ 313 uintptr_t td_rbp_list; /* (k) Robust priv list head. */ 314 uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */ 315 struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on 316 fork for child tracing. */ 317 #define td_endcopy td_pcb 318 319 /* 320 * Fields that must be manually set in fork1() or create_thread() 321 * or already have been set in the allocator, constructor, etc. 322 */ 323 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */ 324 enum { 325 TDS_INACTIVE = 0x0, 326 TDS_INHIBITED, 327 TDS_CAN_RUN, 328 TDS_RUNQ, 329 TDS_RUNNING 330 } td_state; /* (t) thread state */ 331 union { 332 register_t tdu_retval[2]; 333 off_t tdu_off; 334 } td_uretoff; /* (k) Syscall aux returns. */ 335 #define td_retval td_uretoff.tdu_retval 336 u_int td_cowgen; /* (k) Generation of COW pointers. */ 337 struct callout td_slpcallout; /* (h) Callout for sleep. */ 338 struct trapframe *td_frame; /* (k) */ 339 struct vm_object *td_kstack_obj;/* (a) Kstack object. */ 340 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */ 341 int td_kstack_pages; /* (a) Size of the kstack. */ 342 volatile u_int td_critnest; /* (k*) Critical section nest level. */ 343 struct mdthread td_md; /* (k) Any machine-dependent fields. */ 344 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */ 345 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */ 346 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */ 347 int td_errno; /* Error returned by last syscall. */ 348 struct vnet *td_vnet; /* (k) Effective vnet. */ 349 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */ 350 struct trapframe *td_intr_frame;/* (k) Frame of the current irq */ 351 struct proc *td_rfppwait_p; /* (k) The vforked child */ 352 struct vm_page **td_ma; /* (k) uio pages held */ 353 int td_ma_cnt; /* (k) size of *td_ma */ 354 void *td_emuldata; /* Emulator state data */ 355 int td_lastcpu; /* (t) Last cpu we were on. */ 356 int td_oncpu; /* (t) Which cpu we are on. */ 357 void *td_lkpi_task; /* LinuxKPI task struct pointer */ 358 }; 359 360 struct thread0_storage { 361 struct thread t0st_thread; 362 uint64_t t0st_sched[10]; 363 }; 364 365 struct mtx *thread_lock_block(struct thread *); 366 void thread_lock_unblock(struct thread *, struct mtx *); 367 void thread_lock_set(struct thread *, struct mtx *); 368 #define THREAD_LOCK_ASSERT(td, type) \ 369 do { \ 370 struct mtx *__m = (td)->td_lock; \ 371 if (__m != &blocked_lock) \ 372 mtx_assert(__m, (type)); \ 373 } while (0) 374 375 #ifdef INVARIANTS 376 #define THREAD_LOCKPTR_ASSERT(td, lock) \ 377 do { \ 378 struct mtx *__m = (td)->td_lock; \ 379 KASSERT((__m == &blocked_lock || __m == (lock)), \ 380 ("Thread %p lock %p does not match %p", td, __m, (lock))); \ 381 } while (0) 382 383 #define TD_LOCKS_INC(td) ((td)->td_locks++) 384 #define TD_LOCKS_DEC(td) ((td)->td_locks--) 385 #else 386 #define THREAD_LOCKPTR_ASSERT(td, lock) 387 388 #define TD_LOCKS_INC(td) 389 #define TD_LOCKS_DEC(td) 390 #endif 391 392 /* 393 * Flags kept in td_flags: 394 * To change these you MUST have the scheduler lock. 395 */ 396 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */ 397 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */ 398 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */ 399 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */ 400 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */ 401 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */ 402 #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */ 403 #define TDF_SLEEPABORT 0x00000080 /* sleepq_abort was called. */ 404 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */ 405 #define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */ 406 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */ 407 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */ 408 #define TDF_UNUSED12 0x00001000 /* --available-- */ 409 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */ 410 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */ 411 #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */ 412 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */ 413 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */ 414 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */ 415 #define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */ 416 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */ 417 #define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */ 418 #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */ 419 #define TDF_UNUSED23 0x00800000 /* --available-- */ 420 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */ 421 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */ 422 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */ 423 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */ 424 #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */ 425 #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */ 426 #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */ 427 428 /* Userland debug flags */ 429 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */ 430 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */ 431 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */ 432 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */ 433 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */ 434 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */ 435 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new 436 process */ 437 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child 438 only) */ 439 #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */ 440 #define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */ 441 #define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */ 442 #define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */ 443 #define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */ 444 445 /* 446 * "Private" flags kept in td_pflags: 447 * These are only written by curthread and thus need no locking. 448 */ 449 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */ 450 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */ 451 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */ 452 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */ 453 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */ 454 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */ 455 #define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */ 456 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */ 457 #define TDP_UNUSED9 0x00000100 /* --available-- */ 458 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */ 459 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */ 460 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */ 461 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */ 462 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */ 463 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */ 464 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */ 465 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */ 466 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */ 467 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */ 468 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */ 469 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */ 470 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */ 471 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */ 472 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */ 473 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */ 474 #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */ 475 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */ 476 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */ 477 #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */ 478 #define TDP_FORKING 0x20000000 /* Thread is being created through fork() */ 479 #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */ 480 481 /* 482 * Reasons that the current thread can not be run yet. 483 * More than one may apply. 484 */ 485 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */ 486 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */ 487 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */ 488 #define TDI_LOCK 0x0008 /* Stopped on a lock. */ 489 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */ 490 491 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING) 492 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL) 493 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED) 494 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED) 495 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK) 496 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT) 497 #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING) 498 #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ) 499 #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN) 500 #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED) 501 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED) 502 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD) 503 504 #define KTDSTATE(td) \ 505 (((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \ 506 ((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \ 507 ((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" : \ 508 ((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \ 509 ((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding") 510 511 #define TD_SET_INHIB(td, inhib) do { \ 512 (td)->td_state = TDS_INHIBITED; \ 513 (td)->td_inhibitors |= (inhib); \ 514 } while (0) 515 516 #define TD_CLR_INHIB(td, inhib) do { \ 517 if (((td)->td_inhibitors & (inhib)) && \ 518 (((td)->td_inhibitors &= ~(inhib)) == 0)) \ 519 (td)->td_state = TDS_CAN_RUN; \ 520 } while (0) 521 522 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING) 523 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED) 524 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK) 525 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED) 526 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT) 527 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING) 528 529 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING) 530 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED) 531 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK) 532 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED) 533 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT) 534 535 #define TD_SET_RUNNING(td) (td)->td_state = TDS_RUNNING 536 #define TD_SET_RUNQ(td) (td)->td_state = TDS_RUNQ 537 #define TD_SET_CAN_RUN(td) (td)->td_state = TDS_CAN_RUN 538 539 #define TD_SBDRY_INTR(td) \ 540 (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0) 541 #define TD_SBDRY_ERRNO(td) \ 542 (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART) 543 544 /* 545 * Process structure. 546 */ 547 struct proc { 548 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */ 549 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */ 550 struct mtx p_slock; /* process spin lock */ 551 struct ucred *p_ucred; /* (c) Process owner's identity. */ 552 struct filedesc *p_fd; /* (b) Open files. */ 553 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */ 554 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */ 555 struct plimit *p_limit; /* (c) Resource limits. */ 556 struct callout p_limco; /* (c) Limit callout handle */ 557 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */ 558 559 int p_flag; /* (c) P_* flags. */ 560 int p_flag2; /* (c) P2_* flags. */ 561 enum { 562 PRS_NEW = 0, /* In creation */ 563 PRS_NORMAL, /* threads can be run. */ 564 PRS_ZOMBIE 565 } p_state; /* (j/c) Process status. */ 566 pid_t p_pid; /* (b) Process identifier. */ 567 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */ 568 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */ 569 struct proc *p_pptr; /* (c + e) Pointer to parent process. */ 570 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */ 571 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */ 572 struct proc *p_reaper; /* (e) My reaper. */ 573 LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants 574 (if I am reaper). */ 575 LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of 576 the same reaper. */ 577 struct mtx p_mtx; /* (n) Lock for this struct. */ 578 struct mtx p_statmtx; /* Lock for the stats */ 579 struct mtx p_itimmtx; /* Lock for the virt/prof timers */ 580 struct mtx p_profmtx; /* Lock for the profiling */ 581 struct ksiginfo *p_ksi; /* Locked by parent proc lock */ 582 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */ 583 #define p_siglist p_sigqueue.sq_signals 584 585 /* The following fields are all zeroed upon creation in fork. */ 586 #define p_startzero p_oppid 587 pid_t p_oppid; /* (c + e) Save ppid in ptrace. XXX */ 588 struct vmspace *p_vmspace; /* (b) Address space. */ 589 u_int p_swtick; /* (c) Tick when swapped in or out. */ 590 u_int p_cowgen; /* (c) Generation of COW pointers. */ 591 struct itimerval p_realtimer; /* (c) Alarm timer. */ 592 struct rusage p_ru; /* (a) Exit information. */ 593 struct rusage_ext p_rux; /* (cu) Internal resource usage. */ 594 struct rusage_ext p_crux; /* (c) Internal child resource usage. */ 595 int p_profthreads; /* (c) Num threads in addupc_task. */ 596 volatile int p_exitthreads; /* (j) Number of threads exiting */ 597 int p_traceflag; /* (o) Kernel trace points. */ 598 struct vnode *p_tracevp; /* (c + o) Trace to vnode. */ 599 struct ucred *p_tracecred; /* (o) Credentials to trace with. */ 600 struct vnode *p_textvp; /* (b) Vnode of executable. */ 601 u_int p_lock; /* (c) Proclock (prevent swap) count. */ 602 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */ 603 int p_sigparent; /* (c) Signal to parent on exit. */ 604 int p_sig; /* (n) For core dump/debugger XXX. */ 605 u_long p_code; /* (n) For core dump/debugger XXX. */ 606 u_int p_stops; /* (c) Stop event bitmask. */ 607 u_int p_stype; /* (c) Stop event type. */ 608 char p_step; /* (c) Process is stopped. */ 609 u_char p_pfsflags; /* (c) Procfs flags. */ 610 u_int p_ptevents; /* (c) ptrace() event mask. */ 611 struct nlminfo *p_nlminfo; /* (?) Only used by/for lockd. */ 612 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */ 613 struct thread *p_singlethread;/* (c + j) If single threading this is it */ 614 int p_suspcount; /* (j) Num threads in suspended mode. */ 615 struct thread *p_xthread; /* (c) Trap thread */ 616 int p_boundary_count;/* (j) Num threads at user boundary */ 617 int p_pendingcnt; /* how many signals are pending */ 618 struct itimers *p_itimers; /* (c) POSIX interval timers. */ 619 struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */ 620 u_int p_treeflag; /* (e) P_TREE flags */ 621 int p_pendingexits; /* (c) Count of pending thread exits. */ 622 struct filemon *p_filemon; /* (c) filemon-specific data. */ 623 /* End area that is zeroed on creation. */ 624 #define p_endzero p_magic 625 626 /* The following fields are all copied upon creation in fork. */ 627 #define p_startcopy p_endzero 628 u_int p_magic; /* (b) Magic number. */ 629 int p_osrel; /* (x) osreldate for the 630 binary (from ELF note, if any) */ 631 char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */ 632 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */ 633 struct pargs *p_args; /* (c) Process arguments. */ 634 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */ 635 signed char p_nice; /* (c) Process "nice" value. */ 636 int p_fibnum; /* in this routing domain XXX MRT */ 637 pid_t p_reapsubtree; /* (e) Pid of the direct child of the 638 reaper which spawned 639 our subtree. */ 640 uint16_t p_elf_machine; /* (x) ELF machine type */ 641 uint64_t p_elf_flags; /* (x) ELF flags */ 642 /* End area that is copied on creation. */ 643 #define p_endcopy p_xexit 644 645 u_int p_xexit; /* (c) Exit code. */ 646 u_int p_xsig; /* (c) Stop/kill sig. */ 647 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */ 648 struct knlist *p_klist; /* (c) Knotes attached to this proc. */ 649 int p_numthreads; /* (c) Number of threads. */ 650 struct mdproc p_md; /* Any machine-dependent fields. */ 651 struct callout p_itcallout; /* (h + c) Interval timer callout. */ 652 u_short p_acflag; /* (c) Accounting flags. */ 653 struct proc *p_peers; /* (r) */ 654 struct proc *p_leader; /* (b) */ 655 void *p_emuldata; /* (c) Emulator state data. */ 656 struct label *p_label; /* (*) Proc (not subject) MAC label. */ 657 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */ 658 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/ 659 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */ 660 struct cv p_pwait; /* (*) wait cv for exit/exec. */ 661 struct cv p_dbgwait; /* (*) wait cv for debugger attach 662 after fork. */ 663 uint64_t p_prev_runtime; /* (c) Resource usage accounting. */ 664 struct racct *p_racct; /* (b) Resource accounting. */ 665 int p_throttled; /* (c) Flag for racct pcpu throttling */ 666 /* 667 * An orphan is the child that has beed re-parented to the 668 * debugger as a result of attaching to it. Need to keep 669 * track of them for parent to be able to collect the exit 670 * status of what used to be children. 671 */ 672 LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */ 673 LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */ 674 }; 675 676 #define p_session p_pgrp->pg_session 677 #define p_pgid p_pgrp->pg_id 678 679 #define NOCPU (-1) /* For when we aren't on a CPU. */ 680 #define NOCPU_OLD (255) 681 #define MAXCPU_OLD (254) 682 683 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock) 684 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock) 685 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type)) 686 687 #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx) 688 #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx) 689 #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type)) 690 691 #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx) 692 #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx) 693 #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type)) 694 695 #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx) 696 #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx) 697 #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type)) 698 699 /* These flags are kept in p_flag. */ 700 #define P_ADVLOCK 0x00001 /* Process may hold a POSIX advisory lock. */ 701 #define P_CONTROLT 0x00002 /* Has a controlling terminal. */ 702 #define P_KPROC 0x00004 /* Kernel process. */ 703 #define P_UNUSED3 0x00008 /* --available-- */ 704 #define P_PPWAIT 0x00010 /* Parent is waiting for child to exec/exit. */ 705 #define P_PROFIL 0x00020 /* Has started profiling. */ 706 #define P_STOPPROF 0x00040 /* Has thread requesting to stop profiling. */ 707 #define P_HADTHREADS 0x00080 /* Has had threads (no cleanup shortcuts) */ 708 #define P_SUGID 0x00100 /* Had set id privileges since last exec. */ 709 #define P_SYSTEM 0x00200 /* System proc: no sigs, stats or swapping. */ 710 #define P_SINGLE_EXIT 0x00400 /* Threads suspending should exit, not wait. */ 711 #define P_TRACED 0x00800 /* Debugged process being traced. */ 712 #define P_WAITED 0x01000 /* Someone is waiting for us. */ 713 #define P_WEXIT 0x02000 /* Working on exiting. */ 714 #define P_EXEC 0x04000 /* Process called exec. */ 715 #define P_WKILLED 0x08000 /* Killed, go to kernel/user boundary ASAP. */ 716 #define P_CONTINUED 0x10000 /* Proc has continued from a stopped state. */ 717 #define P_STOPPED_SIG 0x20000 /* Stopped due to SIGSTOP/SIGTSTP. */ 718 #define P_STOPPED_TRACE 0x40000 /* Stopped because of tracing. */ 719 #define P_STOPPED_SINGLE 0x80000 /* Only 1 thread can continue (not to user). */ 720 #define P_PROTECTED 0x100000 /* Do not kill on memory overcommit. */ 721 #define P_SIGEVENT 0x200000 /* Process pending signals changed. */ 722 #define P_SINGLE_BOUNDARY 0x400000 /* Threads should suspend at user boundary. */ 723 #define P_HWPMC 0x800000 /* Process is using HWPMCs */ 724 #define P_JAILED 0x1000000 /* Process is in jail. */ 725 #define P_TOTAL_STOP 0x2000000 /* Stopped in stop_all_proc. */ 726 #define P_INEXEC 0x4000000 /* Process is in execve(). */ 727 #define P_STATCHILD 0x8000000 /* Child process stopped or exited. */ 728 #define P_INMEM 0x10000000 /* Loaded into memory. */ 729 #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */ 730 #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */ 731 #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */ 732 733 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE) 734 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED) 735 #define P_KILLED(p) ((p)->p_flag & P_WKILLED) 736 737 /* These flags are kept in p_flag2. */ 738 #define P2_INHERIT_PROTECTED 0x00000001 /* New children get P_PROTECTED. */ 739 #define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or coredumps. */ 740 #define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on exec(2). */ 741 #define P2_AST_SU 0x00000008 /* Handles SU ast for kthreads. */ 742 #define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not yet handled. */ 743 #define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */ 744 745 /* Flags protected by proctree_lock, kept in p_treeflags. */ 746 #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */ 747 #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan 748 list */ 749 #define P_TREE_REAPER 0x00000004 /* Reaper of subtree */ 750 751 /* 752 * These were process status values (p_stat), now they are only used in 753 * legacy conversion code. 754 */ 755 #define SIDL 1 /* Process being created by fork. */ 756 #define SRUN 2 /* Currently runnable. */ 757 #define SSLEEP 3 /* Sleeping on an address. */ 758 #define SSTOP 4 /* Process debugging or suspension. */ 759 #define SZOMB 5 /* Awaiting collection by parent. */ 760 #define SWAIT 6 /* Waiting for interrupt. */ 761 #define SLOCK 7 /* Blocked on a lock. */ 762 763 #define P_MAGIC 0xbeefface 764 765 #ifdef _KERNEL 766 767 /* Types and flags for mi_switch(). */ 768 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */ 769 #define SWT_NONE 0 /* Unspecified switch. */ 770 #define SWT_PREEMPT 1 /* Switching due to preemption. */ 771 #define SWT_OWEPREEMPT 2 /* Switching due to owepreempt. */ 772 #define SWT_TURNSTILE 3 /* Turnstile contention. */ 773 #define SWT_SLEEPQ 4 /* Sleepq wait. */ 774 #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */ 775 #define SWT_RELINQUISH 6 /* yield call. */ 776 #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */ 777 #define SWT_IDLE 8 /* Switching from the idle thread. */ 778 #define SWT_IWAIT 9 /* Waiting for interrupts. */ 779 #define SWT_SUSPEND 10 /* Thread suspended. */ 780 #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */ 781 #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */ 782 #define SWT_COUNT 13 /* Number of switch types. */ 783 /* Flags */ 784 #define SW_VOL 0x0100 /* Voluntary switch. */ 785 #define SW_INVOL 0x0200 /* Involuntary switch. */ 786 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */ 787 788 /* How values for thread_single(). */ 789 #define SINGLE_NO_EXIT 0 790 #define SINGLE_EXIT 1 791 #define SINGLE_BOUNDARY 2 792 #define SINGLE_ALLPROC 3 793 794 #ifdef MALLOC_DECLARE 795 MALLOC_DECLARE(M_PARGS); 796 MALLOC_DECLARE(M_PGRP); 797 MALLOC_DECLARE(M_SESSION); 798 MALLOC_DECLARE(M_SUBPROC); 799 #endif 800 801 #define FOREACH_PROC_IN_SYSTEM(p) \ 802 LIST_FOREACH((p), &allproc, p_list) 803 #define FOREACH_THREAD_IN_PROC(p, td) \ 804 TAILQ_FOREACH((td), &(p)->p_threads, td_plist) 805 806 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads) 807 808 /* 809 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit 810 * in a pid_t, as it is used to represent "no process group". 811 */ 812 #define PID_MAX 99999 813 #define NO_PID 100000 814 extern pid_t pid_max; 815 816 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p)) 817 818 819 #define STOPEVENT(p, e, v) do { \ 820 WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL, \ 821 "checking stopevent %d", (e)); \ 822 if ((p)->p_stops & (e)) { \ 823 PROC_LOCK(p); \ 824 stopevent((p), (e), (v)); \ 825 PROC_UNLOCK(p); \ 826 } \ 827 } while (0) 828 #define _STOPEVENT(p, e, v) do { \ 829 PROC_LOCK_ASSERT(p, MA_OWNED); \ 830 WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, &p->p_mtx.lock_object, \ 831 "checking stopevent %d", (e)); \ 832 if ((p)->p_stops & (e)) \ 833 stopevent((p), (e), (v)); \ 834 } while (0) 835 836 /* Lock and unlock a process. */ 837 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx) 838 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx) 839 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx) 840 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx) 841 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type)) 842 843 /* Lock and unlock a process group. */ 844 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx) 845 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx) 846 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx) 847 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type)) 848 849 #define PGRP_LOCK_PGSIGNAL(pg) do { \ 850 if ((pg) != NULL) \ 851 PGRP_LOCK(pg); \ 852 } while (0) 853 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \ 854 if ((pg) != NULL) \ 855 PGRP_UNLOCK(pg); \ 856 } while (0) 857 858 /* Lock and unlock a session. */ 859 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx) 860 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx) 861 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx) 862 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type)) 863 864 /* 865 * Non-zero p_lock ensures that: 866 * - exit1() is not performed until p_lock reaches zero; 867 * - the process' threads stack are not swapped out if they are currently 868 * not (P_INMEM). 869 * 870 * PHOLD() asserts that the process (except the current process) is 871 * not exiting, increments p_lock and swaps threads stacks into memory, 872 * if needed. 873 * _PHOLD() is same as PHOLD(), it takes the process locked. 874 * _PHOLD_LITE() also takes the process locked, but comparing with 875 * _PHOLD(), it only guarantees that exit1() is not executed, 876 * faultin() is not called. 877 */ 878 #define PHOLD(p) do { \ 879 PROC_LOCK(p); \ 880 _PHOLD(p); \ 881 PROC_UNLOCK(p); \ 882 } while (0) 883 #define _PHOLD(p) do { \ 884 PROC_LOCK_ASSERT((p), MA_OWNED); \ 885 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 886 ("PHOLD of exiting process %p", p)); \ 887 (p)->p_lock++; \ 888 if (((p)->p_flag & P_INMEM) == 0) \ 889 faultin((p)); \ 890 } while (0) 891 #define _PHOLD_LITE(p) do { \ 892 PROC_LOCK_ASSERT((p), MA_OWNED); \ 893 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 894 ("PHOLD of exiting process %p", p)); \ 895 (p)->p_lock++; \ 896 } while (0) 897 #define PROC_ASSERT_HELD(p) do { \ 898 KASSERT((p)->p_lock > 0, ("process %p not held", p)); \ 899 } while (0) 900 901 #define PRELE(p) do { \ 902 PROC_LOCK((p)); \ 903 _PRELE((p)); \ 904 PROC_UNLOCK((p)); \ 905 } while (0) 906 #define _PRELE(p) do { \ 907 PROC_LOCK_ASSERT((p), MA_OWNED); \ 908 PROC_ASSERT_HELD(p); \ 909 (--(p)->p_lock); \ 910 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \ 911 wakeup(&(p)->p_lock); \ 912 } while (0) 913 #define PROC_ASSERT_NOT_HELD(p) do { \ 914 KASSERT((p)->p_lock == 0, ("process %p held", p)); \ 915 } while (0) 916 917 #define PROC_UPDATE_COW(p) do { \ 918 PROC_LOCK_ASSERT((p), MA_OWNED); \ 919 (p)->p_cowgen++; \ 920 } while (0) 921 922 /* Check whether a thread is safe to be swapped out. */ 923 #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP) 924 925 /* Control whether or not it is safe for curthread to sleep. */ 926 #define THREAD_NO_SLEEPING() ((curthread)->td_no_sleeping++) 927 928 #define THREAD_SLEEPING_OK() ((curthread)->td_no_sleeping--) 929 930 #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0) 931 932 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash]) 933 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl; 934 extern u_long pidhash; 935 #define TIDHASH(tid) (&tidhashtbl[(tid) & tidhash]) 936 extern LIST_HEAD(tidhashhead, thread) *tidhashtbl; 937 extern u_long tidhash; 938 extern struct rwlock tidhash_lock; 939 940 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash]) 941 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl; 942 extern u_long pgrphash; 943 944 extern struct sx allproc_lock; 945 extern int allproc_gen; 946 extern struct sx proctree_lock; 947 extern struct mtx ppeers_lock; 948 extern struct proc proc0; /* Process slot for swapper. */ 949 extern struct thread0_storage thread0_st; /* Primary thread in proc0. */ 950 #define thread0 (thread0_st.t0st_thread) 951 extern struct vmspace vmspace0; /* VM space for proc0. */ 952 extern int hogticks; /* Limit on kernel cpu hogs. */ 953 extern int lastpid; 954 extern int nprocs, maxproc; /* Current and max number of procs. */ 955 extern int maxprocperuid; /* Max procs per uid. */ 956 extern u_long ps_arg_cache_limit; 957 958 LIST_HEAD(proclist, proc); 959 TAILQ_HEAD(procqueue, proc); 960 TAILQ_HEAD(threadqueue, thread); 961 extern struct proclist allproc; /* List of all processes. */ 962 extern struct proclist zombproc; /* List of zombie processes. */ 963 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */ 964 965 extern struct uma_zone *proc_zone; 966 967 struct proc *pfind(pid_t); /* Find process by id. */ 968 struct proc *pfind_any(pid_t); /* Find (zombie) process by id. */ 969 struct proc *pfind_locked(pid_t pid); 970 struct pgrp *pgfind(pid_t); /* Find process group by id. */ 971 struct proc *zpfind(pid_t); /* Find zombie process by id. */ 972 973 struct fork_req { 974 int fr_flags; 975 int fr_pages; 976 int *fr_pidp; 977 struct proc **fr_procp; 978 int *fr_pd_fd; 979 int fr_pd_flags; 980 struct filecaps *fr_pd_fcaps; 981 }; 982 983 /* 984 * pget() flags. 985 */ 986 #define PGET_HOLD 0x00001 /* Hold the process. */ 987 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */ 988 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */ 989 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */ 990 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */ 991 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */ 992 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */ 993 994 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT) 995 996 int pget(pid_t pid, int flags, struct proc **pp); 997 998 void ast(struct trapframe *framep); 999 struct thread *choosethread(void); 1000 int cr_cansee(struct ucred *u1, struct ucred *u2); 1001 int cr_canseesocket(struct ucred *cred, struct socket *so); 1002 int cr_canseeothergids(struct ucred *u1, struct ucred *u2); 1003 int cr_canseeotheruids(struct ucred *u1, struct ucred *u2); 1004 int cr_canseejailproc(struct ucred *u1, struct ucred *u2); 1005 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum); 1006 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp, 1007 struct session *sess); 1008 int enterthispgrp(struct proc *p, struct pgrp *pgrp); 1009 void faultin(struct proc *p); 1010 void fixjobc(struct proc *p, struct pgrp *pgrp, int entering); 1011 int fork1(struct thread *, struct fork_req *); 1012 void fork_exit(void (*)(void *, struct trapframe *), void *, 1013 struct trapframe *); 1014 void fork_return(struct thread *, struct trapframe *); 1015 int inferior(struct proc *p); 1016 void kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry, 1017 int *resident_count, bool *super); 1018 void kern_yield(int); 1019 void kick_proc0(void); 1020 void killjobc(void); 1021 int leavepgrp(struct proc *p); 1022 int maybe_preempt(struct thread *td); 1023 void maybe_yield(void); 1024 void mi_switch(int flags, struct thread *newtd); 1025 int p_candebug(struct thread *td, struct proc *p); 1026 int p_cansee(struct thread *td, struct proc *p); 1027 int p_cansched(struct thread *td, struct proc *p); 1028 int p_cansignal(struct thread *td, struct proc *p, int signum); 1029 int p_canwait(struct thread *td, struct proc *p); 1030 struct pargs *pargs_alloc(int len); 1031 void pargs_drop(struct pargs *pa); 1032 void pargs_hold(struct pargs *pa); 1033 int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb); 1034 int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb); 1035 int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb); 1036 void procinit(void); 1037 void proc_linkup0(struct proc *p, struct thread *td); 1038 void proc_linkup(struct proc *p, struct thread *td); 1039 struct proc *proc_realparent(struct proc *child); 1040 void proc_reap(struct thread *td, struct proc *p, int *status, int options); 1041 void proc_reparent(struct proc *child, struct proc *newparent); 1042 void proc_set_traced(struct proc *p, bool stop); 1043 struct pstats *pstats_alloc(void); 1044 void pstats_fork(struct pstats *src, struct pstats *dst); 1045 void pstats_free(struct pstats *ps); 1046 void reaper_abandon_children(struct proc *p, bool exiting); 1047 int securelevel_ge(struct ucred *cr, int level); 1048 int securelevel_gt(struct ucred *cr, int level); 1049 void sess_hold(struct session *); 1050 void sess_release(struct session *); 1051 int setrunnable(struct thread *); 1052 void setsugid(struct proc *p); 1053 int should_yield(void); 1054 int sigonstack(size_t sp); 1055 void stopevent(struct proc *, u_int, u_int); 1056 struct thread *tdfind(lwpid_t, pid_t); 1057 void threadinit(void); 1058 void tidhash_add(struct thread *); 1059 void tidhash_remove(struct thread *); 1060 void cpu_idle(int); 1061 int cpu_idle_wakeup(int); 1062 extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */ 1063 void cpu_switch(struct thread *, struct thread *, struct mtx *); 1064 void cpu_throw(struct thread *, struct thread *) __dead2; 1065 void unsleep(struct thread *); 1066 void userret(struct thread *, struct trapframe *); 1067 1068 void cpu_exit(struct thread *); 1069 void exit1(struct thread *, int, int) __dead2; 1070 void cpu_copy_thread(struct thread *td, struct thread *td0); 1071 int cpu_fetch_syscall_args(struct thread *td); 1072 void cpu_fork(struct thread *, struct proc *, struct thread *, int); 1073 void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *); 1074 void cpu_set_syscall_retval(struct thread *, int); 1075 void cpu_set_upcall(struct thread *, void (*)(void *), void *, 1076 stack_t *); 1077 int cpu_set_user_tls(struct thread *, void *tls_base); 1078 void cpu_thread_alloc(struct thread *); 1079 void cpu_thread_clean(struct thread *); 1080 void cpu_thread_exit(struct thread *); 1081 void cpu_thread_free(struct thread *); 1082 void cpu_thread_swapin(struct thread *); 1083 void cpu_thread_swapout(struct thread *); 1084 struct thread *thread_alloc(int pages); 1085 int thread_alloc_stack(struct thread *, int pages); 1086 void thread_cow_get_proc(struct thread *newtd, struct proc *p); 1087 void thread_cow_get(struct thread *newtd, struct thread *td); 1088 void thread_cow_free(struct thread *td); 1089 void thread_cow_update(struct thread *td); 1090 int thread_create(struct thread *td, struct rtprio *rtp, 1091 int (*initialize_thread)(struct thread *, void *), void *thunk); 1092 void thread_exit(void) __dead2; 1093 void thread_free(struct thread *td); 1094 void thread_link(struct thread *td, struct proc *p); 1095 void thread_reap(void); 1096 int thread_single(struct proc *p, int how); 1097 void thread_single_end(struct proc *p, int how); 1098 void thread_stash(struct thread *td); 1099 void thread_stopped(struct proc *p); 1100 void childproc_stopped(struct proc *child, int reason); 1101 void childproc_continued(struct proc *child); 1102 void childproc_exited(struct proc *child); 1103 int thread_suspend_check(int how); 1104 bool thread_suspend_check_needed(void); 1105 void thread_suspend_switch(struct thread *, struct proc *p); 1106 void thread_suspend_one(struct thread *td); 1107 void thread_unlink(struct thread *td); 1108 void thread_unsuspend(struct proc *p); 1109 void thread_wait(struct proc *p); 1110 struct thread *thread_find(struct proc *p, lwpid_t tid); 1111 1112 void stop_all_proc(void); 1113 void resume_all_proc(void); 1114 1115 static __inline int 1116 curthread_pflags_set(int flags) 1117 { 1118 struct thread *td; 1119 int save; 1120 1121 td = curthread; 1122 save = ~flags | (td->td_pflags & flags); 1123 td->td_pflags |= flags; 1124 return (save); 1125 } 1126 1127 static __inline void 1128 curthread_pflags_restore(int save) 1129 { 1130 1131 curthread->td_pflags &= save; 1132 } 1133 1134 static __inline __pure2 struct td_sched * 1135 td_get_sched(struct thread *td) 1136 { 1137 1138 return ((struct td_sched *)&td[1]); 1139 } 1140 1141 extern void (*softdep_ast_cleanup)(struct thread *); 1142 static __inline void 1143 td_softdep_cleanup(struct thread *td) 1144 { 1145 1146 if (td->td_su != NULL && softdep_ast_cleanup != NULL) 1147 softdep_ast_cleanup(td); 1148 } 1149 1150 #endif /* _KERNEL */ 1151 1152 #endif /* !_SYS_PROC_H_ */ 1153