1 /*- 2 * Copyright (c) 1986, 1989, 1991, 1993 3 * The Regents of the University of California. All rights reserved. 4 * (c) UNIX System Laboratories, Inc. 5 * All or some portions of this file are derived from material licensed 6 * to the University of California by American Telephone and Telegraph 7 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 8 * the permission of UNIX System Laboratories, Inc. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. Neither the name of the University nor the names of its contributors 19 * may be used to endorse or promote products derived from this software 20 * without specific prior written permission. 21 * 22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 32 * SUCH DAMAGE. 33 * 34 * @(#)proc.h 8.15 (Berkeley) 5/19/95 35 * $FreeBSD$ 36 */ 37 38 #ifndef _SYS_PROC_H_ 39 #define _SYS_PROC_H_ 40 41 #include <sys/callout.h> /* For struct callout. */ 42 #include <sys/event.h> /* For struct klist. */ 43 #include <sys/condvar.h> 44 #ifndef _KERNEL 45 #include <sys/filedesc.h> 46 #endif 47 #include <sys/queue.h> 48 #include <sys/_lock.h> 49 #include <sys/lock_profile.h> 50 #include <sys/_mutex.h> 51 #include <sys/osd.h> 52 #include <sys/priority.h> 53 #include <sys/rtprio.h> /* XXX. */ 54 #include <sys/runq.h> 55 #include <sys/resource.h> 56 #include <sys/sigio.h> 57 #include <sys/signal.h> 58 #include <sys/signalvar.h> 59 #ifndef _KERNEL 60 #include <sys/time.h> /* For structs itimerval, timeval. */ 61 #else 62 #include <sys/pcpu.h> 63 #endif 64 #include <sys/ucontext.h> 65 #include <sys/ucred.h> 66 #include <sys/_vm_domain.h> 67 #include <machine/proc.h> /* Machine-dependent proc substruct. */ 68 69 /* 70 * One structure allocated per session. 71 * 72 * List of locks 73 * (m) locked by s_mtx mtx 74 * (e) locked by proctree_lock sx 75 * (c) const until freeing 76 */ 77 struct session { 78 u_int s_count; /* Ref cnt; pgrps in session - atomic. */ 79 struct proc *s_leader; /* (m + e) Session leader. */ 80 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */ 81 struct cdev_priv *s_ttydp; /* (m) Device of controlling tty. */ 82 struct tty *s_ttyp; /* (e) Controlling tty. */ 83 pid_t s_sid; /* (c) Session ID. */ 84 /* (m) Setlogin() name: */ 85 char s_login[roundup(MAXLOGNAME, sizeof(long))]; 86 struct mtx s_mtx; /* Mutex to protect members. */ 87 }; 88 89 /* 90 * One structure allocated per process group. 91 * 92 * List of locks 93 * (m) locked by pg_mtx mtx 94 * (e) locked by proctree_lock sx 95 * (c) const until freeing 96 */ 97 struct pgrp { 98 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */ 99 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */ 100 struct session *pg_session; /* (c) Pointer to session. */ 101 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */ 102 pid_t pg_id; /* (c) Process group id. */ 103 int pg_jobc; /* (m) Job control process count. */ 104 struct mtx pg_mtx; /* Mutex to protect members */ 105 }; 106 107 /* 108 * pargs, used to hold a copy of the command line, if it had a sane length. 109 */ 110 struct pargs { 111 u_int ar_ref; /* Reference count. */ 112 u_int ar_length; /* Length. */ 113 u_char ar_args[1]; /* Arguments. */ 114 }; 115 116 /*- 117 * Description of a process. 118 * 119 * This structure contains the information needed to manage a thread of 120 * control, known in UN*X as a process; it has references to substructures 121 * containing descriptions of things that the process uses, but may share 122 * with related processes. The process structure and the substructures 123 * are always addressable except for those marked "(CPU)" below, 124 * which might be addressable only on a processor on which the process 125 * is running. 126 * 127 * Below is a key of locks used to protect each member of struct proc. The 128 * lock is indicated by a reference to a specific character in parens in the 129 * associated comment. 130 * * - not yet protected 131 * a - only touched by curproc or parent during fork/wait 132 * b - created at fork, never changes 133 * (exception aiods switch vmspaces, but they are also 134 * marked 'P_SYSTEM' so hopefully it will be left alone) 135 * c - locked by proc mtx 136 * d - locked by allproc_lock lock 137 * e - locked by proctree_lock lock 138 * f - session mtx 139 * g - process group mtx 140 * h - callout_lock mtx 141 * i - by curproc or the master session mtx 142 * j - locked by proc slock 143 * k - only accessed by curthread 144 * k*- only accessed by curthread and from an interrupt 145 * kx- only accessed by curthread and by debugger 146 * l - the attaching proc or attaching proc parent 147 * m - Giant 148 * n - not locked, lazy 149 * o - ktrace lock 150 * q - td_contested lock 151 * r - p_peers lock 152 * s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9) 153 * t - thread lock 154 * u - process stat lock 155 * w - process timer lock 156 * x - created at fork, only changes during single threading in exec 157 * y - created at first aio, doesn't change until exit or exec at which 158 * point we are single-threaded and only curthread changes it 159 * z - zombie threads lock 160 * 161 * If the locking key specifies two identifiers (for example, p_pptr) then 162 * either lock is sufficient for read access, but both locks must be held 163 * for write access. 164 */ 165 struct cpuset; 166 struct filecaps; 167 struct filemon; 168 struct kaioinfo; 169 struct kaudit_record; 170 struct kdtrace_proc; 171 struct kdtrace_thread; 172 struct mqueue_notifier; 173 struct nlminfo; 174 struct p_sched; 175 struct proc; 176 struct procdesc; 177 struct racct; 178 struct sbuf; 179 struct sleepqueue; 180 struct syscall_args; 181 struct td_sched; 182 struct thread; 183 struct trapframe; 184 struct turnstile; 185 186 /* 187 * XXX: Does this belong in resource.h or resourcevar.h instead? 188 * Resource usage extension. The times in rusage structs in the kernel are 189 * never up to date. The actual times are kept as runtimes and tick counts 190 * (with control info in the "previous" times), and are converted when 191 * userland asks for rusage info. Backwards compatibility prevents putting 192 * this directly in the user-visible rusage struct. 193 * 194 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux. 195 * Locking for td_rux: (t) for all fields. 196 */ 197 struct rusage_ext { 198 uint64_t rux_runtime; /* (cu) Real time. */ 199 uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */ 200 uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */ 201 uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */ 202 uint64_t rux_uu; /* (c) Previous user time in usec. */ 203 uint64_t rux_su; /* (c) Previous sys time in usec. */ 204 uint64_t rux_tu; /* (c) Previous total time in usec. */ 205 }; 206 207 /* 208 * Kernel runnable context (thread). 209 * This is what is put to sleep and reactivated. 210 * Thread context. Processes may have multiple threads. 211 */ 212 struct thread { 213 struct mtx *volatile td_lock; /* replaces sched lock */ 214 struct proc *td_proc; /* (*) Associated process. */ 215 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */ 216 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */ 217 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */ 218 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */ 219 LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */ 220 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */ 221 struct seltd *td_sel; /* Select queue/channel. */ 222 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */ 223 struct turnstile *td_turnstile; /* (k) Associated turnstile. */ 224 struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */ 225 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */ 226 struct vm_domain_policy td_vm_dom_policy; /* (c) current numa domain policy */ 227 lwpid_t td_tid; /* (b) Thread ID. */ 228 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */ 229 #define td_siglist td_sigqueue.sq_signals 230 u_char td_lend_user_pri; /* (t) Lend user pri. */ 231 232 /* Cleared during fork1() */ 233 #define td_startzero td_flags 234 int td_flags; /* (t) TDF_* flags. */ 235 int td_inhibitors; /* (t) Why can not run. */ 236 int td_pflags; /* (k) Private thread (TDP_*) flags. */ 237 int td_dupfd; /* (k) Ret value from fdopen. XXX */ 238 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */ 239 void *td_wchan; /* (t) Sleep address. */ 240 const char *td_wmesg; /* (t) Reason for sleep. */ 241 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */ 242 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */ 243 short td_locks; /* (k) Debug: count of non-spin locks */ 244 short td_rw_rlocks; /* (k) Count of rwlock read locks. */ 245 short td_lk_slocks; /* (k) Count of lockmgr shared locks. */ 246 short td_stopsched; /* (k) Scheduler stopped. */ 247 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */ 248 const char *td_lockname; /* (t) Name of lock blocked on. */ 249 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */ 250 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */ 251 int td_intr_nesting_level; /* (k) Interrupt recursion. */ 252 int td_pinned; /* (k) Temporary cpu pin count. */ 253 struct ucred *td_ucred; /* (k) Reference to credentials. */ 254 struct plimit *td_limit; /* (k) Resource limits. */ 255 int td_slptick; /* (t) Time at sleep. */ 256 int td_blktick; /* (t) Time spent blocked. */ 257 int td_swvoltick; /* (t) Time at last SW_VOL switch. */ 258 int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */ 259 u_int td_cow; /* (*) Number of copy-on-write faults */ 260 struct rusage td_ru; /* (t) rusage information. */ 261 struct rusage_ext td_rux; /* (t) Internal rusage information. */ 262 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */ 263 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */ 264 u_int td_pticks; /* (t) Statclock hits for profiling */ 265 u_int td_sticks; /* (t) Statclock hits in system mode. */ 266 u_int td_iticks; /* (t) Statclock hits in intr mode. */ 267 u_int td_uticks; /* (t) Statclock hits in user mode. */ 268 int td_intrval; /* (t) Return value for sleepq. */ 269 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */ 270 volatile u_int td_generation; /* (k) For detection of preemption */ 271 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */ 272 int td_xsig; /* (c) Signal for ptrace */ 273 u_long td_profil_addr; /* (k) Temporary addr until AST. */ 274 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */ 275 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */ 276 struct file *td_fpop; /* (k) file referencing cdev under op */ 277 int td_dbgflags; /* (c) Userland debugger flags */ 278 siginfo_t td_si; /* (c) For debugger or core file */ 279 int td_ng_outbound; /* (k) Thread entered ng from above. */ 280 struct osd td_osd; /* (k) Object specific data. */ 281 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */ 282 pid_t td_dbg_forked; /* (c) Child pid for debugger. */ 283 u_int td_vp_reserv; /* (k) Count of reserved vnodes. */ 284 int td_no_sleeping; /* (k) Sleeping disabled count. */ 285 int td_dom_rr_idx; /* (k) RR Numa domain selection. */ 286 void *td_su; /* (k) FFS SU private */ 287 sbintime_t td_sleeptimo; /* (t) Sleep timeout. */ 288 int td_rtcgen; /* (s) rtc_generation of abs. sleep */ 289 #define td_endzero td_sigmask 290 291 /* Copied during fork1() or create_thread(). */ 292 #define td_startcopy td_endzero 293 sigset_t td_sigmask; /* (c) Current signal mask. */ 294 u_char td_rqindex; /* (t) Run queue index. */ 295 u_char td_base_pri; /* (t) Thread base kernel priority. */ 296 u_char td_priority; /* (t) Thread active priority. */ 297 u_char td_pri_class; /* (t) Scheduling class. */ 298 u_char td_user_pri; /* (t) User pri from estcpu and nice. */ 299 u_char td_base_user_pri; /* (t) Base user pri */ 300 uintptr_t td_rb_list; /* (k) Robust list head. */ 301 uintptr_t td_rbp_list; /* (k) Robust priv list head. */ 302 uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */ 303 struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on 304 fork for child tracing. */ 305 #define td_endcopy td_pcb 306 307 /* 308 * Fields that must be manually set in fork1() or create_thread() 309 * or already have been set in the allocator, constructor, etc. 310 */ 311 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */ 312 enum { 313 TDS_INACTIVE = 0x0, 314 TDS_INHIBITED, 315 TDS_CAN_RUN, 316 TDS_RUNQ, 317 TDS_RUNNING 318 } td_state; /* (t) thread state */ 319 union { 320 register_t tdu_retval[2]; 321 off_t tdu_off; 322 } td_uretoff; /* (k) Syscall aux returns. */ 323 #define td_retval td_uretoff.tdu_retval 324 u_int td_cowgen; /* (k) Generation of COW pointers. */ 325 struct callout td_slpcallout; /* (h) Callout for sleep. */ 326 struct trapframe *td_frame; /* (k) */ 327 struct vm_object *td_kstack_obj;/* (a) Kstack object. */ 328 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */ 329 int td_kstack_pages; /* (a) Size of the kstack. */ 330 volatile u_int td_critnest; /* (k*) Critical section nest level. */ 331 struct mdthread td_md; /* (k) Any machine-dependent fields. */ 332 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */ 333 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */ 334 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */ 335 int td_errno; /* Error returned by last syscall. */ 336 struct vnet *td_vnet; /* (k) Effective vnet. */ 337 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */ 338 struct trapframe *td_intr_frame;/* (k) Frame of the current irq */ 339 struct proc *td_rfppwait_p; /* (k) The vforked child */ 340 struct vm_page **td_ma; /* (k) uio pages held */ 341 int td_ma_cnt; /* (k) size of *td_ma */ 342 void *td_emuldata; /* Emulator state data */ 343 int td_lastcpu; /* (t) Last cpu we were on. */ 344 int td_oncpu; /* (t) Which cpu we are on. */ 345 void *td_lkpi_task; /* LinuxKPI task struct pointer */ 346 }; 347 348 struct thread0_storage { 349 struct thread t0st_thread; 350 uint64_t t0st_sched[10]; 351 }; 352 353 struct mtx *thread_lock_block(struct thread *); 354 void thread_lock_unblock(struct thread *, struct mtx *); 355 void thread_lock_set(struct thread *, struct mtx *); 356 #define THREAD_LOCK_ASSERT(td, type) \ 357 do { \ 358 struct mtx *__m = (td)->td_lock; \ 359 if (__m != &blocked_lock) \ 360 mtx_assert(__m, (type)); \ 361 } while (0) 362 363 #ifdef INVARIANTS 364 #define THREAD_LOCKPTR_ASSERT(td, lock) \ 365 do { \ 366 struct mtx *__m = (td)->td_lock; \ 367 KASSERT((__m == &blocked_lock || __m == (lock)), \ 368 ("Thread %p lock %p does not match %p", td, __m, (lock))); \ 369 } while (0) 370 371 #define TD_LOCKS_INC(td) ((td)->td_locks++) 372 #define TD_LOCKS_DEC(td) ((td)->td_locks--) 373 #else 374 #define THREAD_LOCKPTR_ASSERT(td, lock) 375 376 #define TD_LOCKS_INC(td) 377 #define TD_LOCKS_DEC(td) 378 #endif 379 380 /* 381 * Flags kept in td_flags: 382 * To change these you MUST have the scheduler lock. 383 */ 384 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */ 385 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */ 386 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */ 387 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */ 388 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */ 389 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */ 390 #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */ 391 #define TDF_SLEEPABORT 0x00000080 /* sleepq_abort was called. */ 392 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */ 393 #define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */ 394 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */ 395 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */ 396 #define TDF_UNUSED12 0x00001000 /* --available-- */ 397 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */ 398 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */ 399 #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */ 400 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */ 401 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */ 402 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */ 403 #define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */ 404 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */ 405 #define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */ 406 #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */ 407 #define TDF_UNUSED23 0x00800000 /* --available-- */ 408 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */ 409 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */ 410 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */ 411 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */ 412 #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */ 413 #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */ 414 #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */ 415 416 /* Userland debug flags */ 417 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */ 418 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */ 419 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */ 420 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */ 421 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */ 422 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */ 423 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new 424 process */ 425 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child 426 only) */ 427 #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */ 428 #define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */ 429 #define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */ 430 #define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */ 431 #define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */ 432 433 /* 434 * "Private" flags kept in td_pflags: 435 * These are only written by curthread and thus need no locking. 436 */ 437 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */ 438 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */ 439 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */ 440 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */ 441 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */ 442 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */ 443 #define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */ 444 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */ 445 #define TDP_UNUSED9 0x00000100 /* --available-- */ 446 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */ 447 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */ 448 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */ 449 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */ 450 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */ 451 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */ 452 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */ 453 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */ 454 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */ 455 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */ 456 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */ 457 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */ 458 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */ 459 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */ 460 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */ 461 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */ 462 #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */ 463 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */ 464 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */ 465 #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */ 466 #define TDP_FORKING 0x20000000 /* Thread is being created through fork() */ 467 #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */ 468 469 /* 470 * Reasons that the current thread can not be run yet. 471 * More than one may apply. 472 */ 473 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */ 474 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */ 475 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */ 476 #define TDI_LOCK 0x0008 /* Stopped on a lock. */ 477 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */ 478 479 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING) 480 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL) 481 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED) 482 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED) 483 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK) 484 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT) 485 #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING) 486 #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ) 487 #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN) 488 #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED) 489 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED) 490 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD) 491 492 #define KTDSTATE(td) \ 493 (((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \ 494 ((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \ 495 ((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" : \ 496 ((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \ 497 ((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding") 498 499 #define TD_SET_INHIB(td, inhib) do { \ 500 (td)->td_state = TDS_INHIBITED; \ 501 (td)->td_inhibitors |= (inhib); \ 502 } while (0) 503 504 #define TD_CLR_INHIB(td, inhib) do { \ 505 if (((td)->td_inhibitors & (inhib)) && \ 506 (((td)->td_inhibitors &= ~(inhib)) == 0)) \ 507 (td)->td_state = TDS_CAN_RUN; \ 508 } while (0) 509 510 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING) 511 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED) 512 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK) 513 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED) 514 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT) 515 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING) 516 517 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING) 518 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED) 519 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK) 520 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED) 521 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT) 522 523 #define TD_SET_RUNNING(td) (td)->td_state = TDS_RUNNING 524 #define TD_SET_RUNQ(td) (td)->td_state = TDS_RUNQ 525 #define TD_SET_CAN_RUN(td) (td)->td_state = TDS_CAN_RUN 526 527 #define TD_SBDRY_INTR(td) \ 528 (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0) 529 #define TD_SBDRY_ERRNO(td) \ 530 (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART) 531 532 /* 533 * Process structure. 534 */ 535 struct proc { 536 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */ 537 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */ 538 struct mtx p_slock; /* process spin lock */ 539 struct ucred *p_ucred; /* (c) Process owner's identity. */ 540 struct filedesc *p_fd; /* (b) Open files. */ 541 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */ 542 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */ 543 struct plimit *p_limit; /* (c) Resource limits. */ 544 struct callout p_limco; /* (c) Limit callout handle */ 545 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */ 546 547 int p_flag; /* (c) P_* flags. */ 548 int p_flag2; /* (c) P2_* flags. */ 549 enum { 550 PRS_NEW = 0, /* In creation */ 551 PRS_NORMAL, /* threads can be run. */ 552 PRS_ZOMBIE 553 } p_state; /* (j/c) Process status. */ 554 pid_t p_pid; /* (b) Process identifier. */ 555 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */ 556 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */ 557 struct proc *p_pptr; /* (c + e) Pointer to parent process. */ 558 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */ 559 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */ 560 struct proc *p_reaper; /* (e) My reaper. */ 561 LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants 562 (if I am reaper). */ 563 LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of 564 the same reaper. */ 565 struct mtx p_mtx; /* (n) Lock for this struct. */ 566 struct mtx p_statmtx; /* Lock for the stats */ 567 struct mtx p_itimmtx; /* Lock for the virt/prof timers */ 568 struct mtx p_profmtx; /* Lock for the profiling */ 569 struct ksiginfo *p_ksi; /* Locked by parent proc lock */ 570 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */ 571 #define p_siglist p_sigqueue.sq_signals 572 573 /* The following fields are all zeroed upon creation in fork. */ 574 #define p_startzero p_oppid 575 pid_t p_oppid; /* (c + e) Save ppid in ptrace. XXX */ 576 struct vmspace *p_vmspace; /* (b) Address space. */ 577 u_int p_swtick; /* (c) Tick when swapped in or out. */ 578 u_int p_cowgen; /* (c) Generation of COW pointers. */ 579 struct itimerval p_realtimer; /* (c) Alarm timer. */ 580 struct rusage p_ru; /* (a) Exit information. */ 581 struct rusage_ext p_rux; /* (cu) Internal resource usage. */ 582 struct rusage_ext p_crux; /* (c) Internal child resource usage. */ 583 int p_profthreads; /* (c) Num threads in addupc_task. */ 584 volatile int p_exitthreads; /* (j) Number of threads exiting */ 585 int p_traceflag; /* (o) Kernel trace points. */ 586 struct vnode *p_tracevp; /* (c + o) Trace to vnode. */ 587 struct ucred *p_tracecred; /* (o) Credentials to trace with. */ 588 struct vnode *p_textvp; /* (b) Vnode of executable. */ 589 u_int p_lock; /* (c) Proclock (prevent swap) count. */ 590 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */ 591 int p_sigparent; /* (c) Signal to parent on exit. */ 592 int p_sig; /* (n) For core dump/debugger XXX. */ 593 u_long p_code; /* (n) For core dump/debugger XXX. */ 594 u_int p_stops; /* (c) Stop event bitmask. */ 595 u_int p_stype; /* (c) Stop event type. */ 596 char p_step; /* (c) Process is stopped. */ 597 u_char p_pfsflags; /* (c) Procfs flags. */ 598 u_int p_ptevents; /* (c) ptrace() event mask. */ 599 struct nlminfo *p_nlminfo; /* (?) Only used by/for lockd. */ 600 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */ 601 struct thread *p_singlethread;/* (c + j) If single threading this is it */ 602 int p_suspcount; /* (j) Num threads in suspended mode. */ 603 struct thread *p_xthread; /* (c) Trap thread */ 604 int p_boundary_count;/* (j) Num threads at user boundary */ 605 int p_pendingcnt; /* how many signals are pending */ 606 struct itimers *p_itimers; /* (c) POSIX interval timers. */ 607 struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */ 608 u_int p_treeflag; /* (e) P_TREE flags */ 609 int p_pendingexits; /* (c) Count of pending thread exits. */ 610 struct filemon *p_filemon; /* (c) filemon-specific data. */ 611 /* End area that is zeroed on creation. */ 612 #define p_endzero p_magic 613 614 /* The following fields are all copied upon creation in fork. */ 615 #define p_startcopy p_endzero 616 u_int p_magic; /* (b) Magic number. */ 617 int p_osrel; /* (x) osreldate for the 618 binary (from ELF note, if any) */ 619 char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */ 620 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */ 621 struct pargs *p_args; /* (c) Process arguments. */ 622 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */ 623 signed char p_nice; /* (c) Process "nice" value. */ 624 int p_fibnum; /* in this routing domain XXX MRT */ 625 pid_t p_reapsubtree; /* (e) Pid of the direct child of the 626 reaper which spawned 627 our subtree. */ 628 uint16_t p_elf_machine; /* (x) ELF machine type */ 629 uint64_t p_elf_flags; /* (x) ELF flags */ 630 /* End area that is copied on creation. */ 631 #define p_endcopy p_xexit 632 633 u_int p_xexit; /* (c) Exit code. */ 634 u_int p_xsig; /* (c) Stop/kill sig. */ 635 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */ 636 struct knlist *p_klist; /* (c) Knotes attached to this proc. */ 637 int p_numthreads; /* (c) Number of threads. */ 638 struct mdproc p_md; /* Any machine-dependent fields. */ 639 struct callout p_itcallout; /* (h + c) Interval timer callout. */ 640 u_short p_acflag; /* (c) Accounting flags. */ 641 struct proc *p_peers; /* (r) */ 642 struct proc *p_leader; /* (b) */ 643 void *p_emuldata; /* (c) Emulator state data. */ 644 struct label *p_label; /* (*) Proc (not subject) MAC label. */ 645 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */ 646 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/ 647 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */ 648 struct cv p_pwait; /* (*) wait cv for exit/exec. */ 649 struct cv p_dbgwait; /* (*) wait cv for debugger attach 650 after fork. */ 651 uint64_t p_prev_runtime; /* (c) Resource usage accounting. */ 652 struct racct *p_racct; /* (b) Resource accounting. */ 653 int p_throttled; /* (c) Flag for racct pcpu throttling */ 654 struct vm_domain_policy p_vm_dom_policy; /* (c) process default VM domain, or -1 */ 655 /* 656 * An orphan is the child that has beed re-parented to the 657 * debugger as a result of attaching to it. Need to keep 658 * track of them for parent to be able to collect the exit 659 * status of what used to be children. 660 */ 661 LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */ 662 LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */ 663 }; 664 665 #define p_session p_pgrp->pg_session 666 #define p_pgid p_pgrp->pg_id 667 668 #define NOCPU (-1) /* For when we aren't on a CPU. */ 669 #define NOCPU_OLD (255) 670 #define MAXCPU_OLD (254) 671 672 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock) 673 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock) 674 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type)) 675 676 #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx) 677 #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx) 678 #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type)) 679 680 #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx) 681 #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx) 682 #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type)) 683 684 #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx) 685 #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx) 686 #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type)) 687 688 /* These flags are kept in p_flag. */ 689 #define P_ADVLOCK 0x00001 /* Process may hold a POSIX advisory lock. */ 690 #define P_CONTROLT 0x00002 /* Has a controlling terminal. */ 691 #define P_KPROC 0x00004 /* Kernel process. */ 692 #define P_UNUSED3 0x00008 /* --available-- */ 693 #define P_PPWAIT 0x00010 /* Parent is waiting for child to exec/exit. */ 694 #define P_PROFIL 0x00020 /* Has started profiling. */ 695 #define P_STOPPROF 0x00040 /* Has thread requesting to stop profiling. */ 696 #define P_HADTHREADS 0x00080 /* Has had threads (no cleanup shortcuts) */ 697 #define P_SUGID 0x00100 /* Had set id privileges since last exec. */ 698 #define P_SYSTEM 0x00200 /* System proc: no sigs, stats or swapping. */ 699 #define P_SINGLE_EXIT 0x00400 /* Threads suspending should exit, not wait. */ 700 #define P_TRACED 0x00800 /* Debugged process being traced. */ 701 #define P_WAITED 0x01000 /* Someone is waiting for us. */ 702 #define P_WEXIT 0x02000 /* Working on exiting. */ 703 #define P_EXEC 0x04000 /* Process called exec. */ 704 #define P_WKILLED 0x08000 /* Killed, go to kernel/user boundary ASAP. */ 705 #define P_CONTINUED 0x10000 /* Proc has continued from a stopped state. */ 706 #define P_STOPPED_SIG 0x20000 /* Stopped due to SIGSTOP/SIGTSTP. */ 707 #define P_STOPPED_TRACE 0x40000 /* Stopped because of tracing. */ 708 #define P_STOPPED_SINGLE 0x80000 /* Only 1 thread can continue (not to user). */ 709 #define P_PROTECTED 0x100000 /* Do not kill on memory overcommit. */ 710 #define P_SIGEVENT 0x200000 /* Process pending signals changed. */ 711 #define P_SINGLE_BOUNDARY 0x400000 /* Threads should suspend at user boundary. */ 712 #define P_HWPMC 0x800000 /* Process is using HWPMCs */ 713 #define P_JAILED 0x1000000 /* Process is in jail. */ 714 #define P_TOTAL_STOP 0x2000000 /* Stopped in stop_all_proc. */ 715 #define P_INEXEC 0x4000000 /* Process is in execve(). */ 716 #define P_STATCHILD 0x8000000 /* Child process stopped or exited. */ 717 #define P_INMEM 0x10000000 /* Loaded into memory. */ 718 #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */ 719 #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */ 720 #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */ 721 722 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE) 723 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED) 724 #define P_KILLED(p) ((p)->p_flag & P_WKILLED) 725 726 /* These flags are kept in p_flag2. */ 727 #define P2_INHERIT_PROTECTED 0x00000001 /* New children get P_PROTECTED. */ 728 #define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or coredumps. */ 729 #define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on exec(2). */ 730 #define P2_AST_SU 0x00000008 /* Handles SU ast for kthreads. */ 731 #define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not yet handled. */ 732 #define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */ 733 734 /* Flags protected by proctree_lock, kept in p_treeflags. */ 735 #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */ 736 #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan 737 list */ 738 #define P_TREE_REAPER 0x00000004 /* Reaper of subtree */ 739 740 /* 741 * These were process status values (p_stat), now they are only used in 742 * legacy conversion code. 743 */ 744 #define SIDL 1 /* Process being created by fork. */ 745 #define SRUN 2 /* Currently runnable. */ 746 #define SSLEEP 3 /* Sleeping on an address. */ 747 #define SSTOP 4 /* Process debugging or suspension. */ 748 #define SZOMB 5 /* Awaiting collection by parent. */ 749 #define SWAIT 6 /* Waiting for interrupt. */ 750 #define SLOCK 7 /* Blocked on a lock. */ 751 752 #define P_MAGIC 0xbeefface 753 754 #ifdef _KERNEL 755 756 /* Types and flags for mi_switch(). */ 757 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */ 758 #define SWT_NONE 0 /* Unspecified switch. */ 759 #define SWT_PREEMPT 1 /* Switching due to preemption. */ 760 #define SWT_OWEPREEMPT 2 /* Switching due to owepreempt. */ 761 #define SWT_TURNSTILE 3 /* Turnstile contention. */ 762 #define SWT_SLEEPQ 4 /* Sleepq wait. */ 763 #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */ 764 #define SWT_RELINQUISH 6 /* yield call. */ 765 #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */ 766 #define SWT_IDLE 8 /* Switching from the idle thread. */ 767 #define SWT_IWAIT 9 /* Waiting for interrupts. */ 768 #define SWT_SUSPEND 10 /* Thread suspended. */ 769 #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */ 770 #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */ 771 #define SWT_COUNT 13 /* Number of switch types. */ 772 /* Flags */ 773 #define SW_VOL 0x0100 /* Voluntary switch. */ 774 #define SW_INVOL 0x0200 /* Involuntary switch. */ 775 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */ 776 777 /* How values for thread_single(). */ 778 #define SINGLE_NO_EXIT 0 779 #define SINGLE_EXIT 1 780 #define SINGLE_BOUNDARY 2 781 #define SINGLE_ALLPROC 3 782 783 #ifdef MALLOC_DECLARE 784 MALLOC_DECLARE(M_PARGS); 785 MALLOC_DECLARE(M_PGRP); 786 MALLOC_DECLARE(M_SESSION); 787 MALLOC_DECLARE(M_SUBPROC); 788 #endif 789 790 #define FOREACH_PROC_IN_SYSTEM(p) \ 791 LIST_FOREACH((p), &allproc, p_list) 792 #define FOREACH_THREAD_IN_PROC(p, td) \ 793 TAILQ_FOREACH((td), &(p)->p_threads, td_plist) 794 795 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads) 796 797 /* 798 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit 799 * in a pid_t, as it is used to represent "no process group". 800 */ 801 #define PID_MAX 99999 802 #define NO_PID 100000 803 extern pid_t pid_max; 804 805 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p)) 806 807 808 #define STOPEVENT(p, e, v) do { \ 809 WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL, \ 810 "checking stopevent %d", (e)); \ 811 if ((p)->p_stops & (e)) { \ 812 PROC_LOCK(p); \ 813 stopevent((p), (e), (v)); \ 814 PROC_UNLOCK(p); \ 815 } \ 816 } while (0) 817 #define _STOPEVENT(p, e, v) do { \ 818 PROC_LOCK_ASSERT(p, MA_OWNED); \ 819 WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, &p->p_mtx.lock_object, \ 820 "checking stopevent %d", (e)); \ 821 if ((p)->p_stops & (e)) \ 822 stopevent((p), (e), (v)); \ 823 } while (0) 824 825 /* Lock and unlock a process. */ 826 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx) 827 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx) 828 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx) 829 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx) 830 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type)) 831 832 /* Lock and unlock a process group. */ 833 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx) 834 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx) 835 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx) 836 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type)) 837 838 #define PGRP_LOCK_PGSIGNAL(pg) do { \ 839 if ((pg) != NULL) \ 840 PGRP_LOCK(pg); \ 841 } while (0) 842 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \ 843 if ((pg) != NULL) \ 844 PGRP_UNLOCK(pg); \ 845 } while (0) 846 847 /* Lock and unlock a session. */ 848 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx) 849 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx) 850 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx) 851 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type)) 852 853 /* 854 * Non-zero p_lock ensures that: 855 * - exit1() is not performed until p_lock reaches zero; 856 * - the process' threads stack are not swapped out if they are currently 857 * not (P_INMEM). 858 * 859 * PHOLD() asserts that the process (except the current process) is 860 * not exiting, increments p_lock and swaps threads stacks into memory, 861 * if needed. 862 * _PHOLD() is same as PHOLD(), it takes the process locked. 863 * _PHOLD_LITE() also takes the process locked, but comparing with 864 * _PHOLD(), it only guarantees that exit1() is not executed, 865 * faultin() is not called. 866 */ 867 #define PHOLD(p) do { \ 868 PROC_LOCK(p); \ 869 _PHOLD(p); \ 870 PROC_UNLOCK(p); \ 871 } while (0) 872 #define _PHOLD(p) do { \ 873 PROC_LOCK_ASSERT((p), MA_OWNED); \ 874 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 875 ("PHOLD of exiting process %p", p)); \ 876 (p)->p_lock++; \ 877 if (((p)->p_flag & P_INMEM) == 0) \ 878 faultin((p)); \ 879 } while (0) 880 #define _PHOLD_LITE(p) do { \ 881 PROC_LOCK_ASSERT((p), MA_OWNED); \ 882 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ 883 ("PHOLD of exiting process %p", p)); \ 884 (p)->p_lock++; \ 885 } while (0) 886 #define PROC_ASSERT_HELD(p) do { \ 887 KASSERT((p)->p_lock > 0, ("process %p not held", p)); \ 888 } while (0) 889 890 #define PRELE(p) do { \ 891 PROC_LOCK((p)); \ 892 _PRELE((p)); \ 893 PROC_UNLOCK((p)); \ 894 } while (0) 895 #define _PRELE(p) do { \ 896 PROC_LOCK_ASSERT((p), MA_OWNED); \ 897 PROC_ASSERT_HELD(p); \ 898 (--(p)->p_lock); \ 899 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \ 900 wakeup(&(p)->p_lock); \ 901 } while (0) 902 #define PROC_ASSERT_NOT_HELD(p) do { \ 903 KASSERT((p)->p_lock == 0, ("process %p held", p)); \ 904 } while (0) 905 906 #define PROC_UPDATE_COW(p) do { \ 907 PROC_LOCK_ASSERT((p), MA_OWNED); \ 908 (p)->p_cowgen++; \ 909 } while (0) 910 911 /* Check whether a thread is safe to be swapped out. */ 912 #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP) 913 914 /* Control whether or not it is safe for curthread to sleep. */ 915 #define THREAD_NO_SLEEPING() ((curthread)->td_no_sleeping++) 916 917 #define THREAD_SLEEPING_OK() ((curthread)->td_no_sleeping--) 918 919 #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0) 920 921 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash]) 922 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl; 923 extern u_long pidhash; 924 #define TIDHASH(tid) (&tidhashtbl[(tid) & tidhash]) 925 extern LIST_HEAD(tidhashhead, thread) *tidhashtbl; 926 extern u_long tidhash; 927 extern struct rwlock tidhash_lock; 928 929 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash]) 930 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl; 931 extern u_long pgrphash; 932 933 extern struct sx allproc_lock; 934 extern int allproc_gen; 935 extern struct sx proctree_lock; 936 extern struct mtx ppeers_lock; 937 extern struct proc proc0; /* Process slot for swapper. */ 938 extern struct thread0_storage thread0_st; /* Primary thread in proc0. */ 939 #define thread0 (thread0_st.t0st_thread) 940 extern struct vmspace vmspace0; /* VM space for proc0. */ 941 extern int hogticks; /* Limit on kernel cpu hogs. */ 942 extern int lastpid; 943 extern int nprocs, maxproc; /* Current and max number of procs. */ 944 extern int maxprocperuid; /* Max procs per uid. */ 945 extern u_long ps_arg_cache_limit; 946 947 LIST_HEAD(proclist, proc); 948 TAILQ_HEAD(procqueue, proc); 949 TAILQ_HEAD(threadqueue, thread); 950 extern struct proclist allproc; /* List of all processes. */ 951 extern struct proclist zombproc; /* List of zombie processes. */ 952 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */ 953 954 extern struct uma_zone *proc_zone; 955 956 struct proc *pfind(pid_t); /* Find process by id. */ 957 struct proc *pfind_locked(pid_t pid); 958 struct pgrp *pgfind(pid_t); /* Find process group by id. */ 959 struct proc *zpfind(pid_t); /* Find zombie process by id. */ 960 961 struct fork_req { 962 int fr_flags; 963 int fr_pages; 964 int *fr_pidp; 965 struct proc **fr_procp; 966 int *fr_pd_fd; 967 int fr_pd_flags; 968 struct filecaps *fr_pd_fcaps; 969 }; 970 971 /* 972 * pget() flags. 973 */ 974 #define PGET_HOLD 0x00001 /* Hold the process. */ 975 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */ 976 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */ 977 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */ 978 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */ 979 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */ 980 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */ 981 982 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT) 983 984 int pget(pid_t pid, int flags, struct proc **pp); 985 986 void ast(struct trapframe *framep); 987 struct thread *choosethread(void); 988 int cr_cansee(struct ucred *u1, struct ucred *u2); 989 int cr_canseesocket(struct ucred *cred, struct socket *so); 990 int cr_canseeothergids(struct ucred *u1, struct ucred *u2); 991 int cr_canseeotheruids(struct ucred *u1, struct ucred *u2); 992 int cr_canseejailproc(struct ucred *u1, struct ucred *u2); 993 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum); 994 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp, 995 struct session *sess); 996 int enterthispgrp(struct proc *p, struct pgrp *pgrp); 997 void faultin(struct proc *p); 998 void fixjobc(struct proc *p, struct pgrp *pgrp, int entering); 999 int fork1(struct thread *, struct fork_req *); 1000 void fork_exit(void (*)(void *, struct trapframe *), void *, 1001 struct trapframe *); 1002 void fork_return(struct thread *, struct trapframe *); 1003 int inferior(struct proc *p); 1004 void kern_yield(int); 1005 void kick_proc0(void); 1006 void killjobc(void); 1007 int leavepgrp(struct proc *p); 1008 int maybe_preempt(struct thread *td); 1009 void maybe_yield(void); 1010 void mi_switch(int flags, struct thread *newtd); 1011 int p_candebug(struct thread *td, struct proc *p); 1012 int p_cansee(struct thread *td, struct proc *p); 1013 int p_cansched(struct thread *td, struct proc *p); 1014 int p_cansignal(struct thread *td, struct proc *p, int signum); 1015 int p_canwait(struct thread *td, struct proc *p); 1016 struct pargs *pargs_alloc(int len); 1017 void pargs_drop(struct pargs *pa); 1018 void pargs_hold(struct pargs *pa); 1019 int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb); 1020 int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb); 1021 int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb); 1022 void procinit(void); 1023 void proc_linkup0(struct proc *p, struct thread *td); 1024 void proc_linkup(struct proc *p, struct thread *td); 1025 struct proc *proc_realparent(struct proc *child); 1026 void proc_reap(struct thread *td, struct proc *p, int *status, int options); 1027 void proc_reparent(struct proc *child, struct proc *newparent); 1028 void proc_set_traced(struct proc *p, bool stop); 1029 struct pstats *pstats_alloc(void); 1030 void pstats_fork(struct pstats *src, struct pstats *dst); 1031 void pstats_free(struct pstats *ps); 1032 void reaper_abandon_children(struct proc *p, bool exiting); 1033 int securelevel_ge(struct ucred *cr, int level); 1034 int securelevel_gt(struct ucred *cr, int level); 1035 void sess_hold(struct session *); 1036 void sess_release(struct session *); 1037 int setrunnable(struct thread *); 1038 void setsugid(struct proc *p); 1039 int should_yield(void); 1040 int sigonstack(size_t sp); 1041 void stopevent(struct proc *, u_int, u_int); 1042 struct thread *tdfind(lwpid_t, pid_t); 1043 void threadinit(void); 1044 void tidhash_add(struct thread *); 1045 void tidhash_remove(struct thread *); 1046 void cpu_idle(int); 1047 int cpu_idle_wakeup(int); 1048 extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */ 1049 void cpu_switch(struct thread *, struct thread *, struct mtx *); 1050 void cpu_throw(struct thread *, struct thread *) __dead2; 1051 void unsleep(struct thread *); 1052 void userret(struct thread *, struct trapframe *); 1053 1054 void cpu_exit(struct thread *); 1055 void exit1(struct thread *, int, int) __dead2; 1056 void cpu_copy_thread(struct thread *td, struct thread *td0); 1057 int cpu_fetch_syscall_args(struct thread *td); 1058 void cpu_fork(struct thread *, struct proc *, struct thread *, int); 1059 void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *); 1060 void cpu_set_syscall_retval(struct thread *, int); 1061 void cpu_set_upcall(struct thread *, void (*)(void *), void *, 1062 stack_t *); 1063 int cpu_set_user_tls(struct thread *, void *tls_base); 1064 void cpu_thread_alloc(struct thread *); 1065 void cpu_thread_clean(struct thread *); 1066 void cpu_thread_exit(struct thread *); 1067 void cpu_thread_free(struct thread *); 1068 void cpu_thread_swapin(struct thread *); 1069 void cpu_thread_swapout(struct thread *); 1070 struct thread *thread_alloc(int pages); 1071 int thread_alloc_stack(struct thread *, int pages); 1072 void thread_cow_get_proc(struct thread *newtd, struct proc *p); 1073 void thread_cow_get(struct thread *newtd, struct thread *td); 1074 void thread_cow_free(struct thread *td); 1075 void thread_cow_update(struct thread *td); 1076 int thread_create(struct thread *td, struct rtprio *rtp, 1077 int (*initialize_thread)(struct thread *, void *), void *thunk); 1078 void thread_exit(void) __dead2; 1079 void thread_free(struct thread *td); 1080 void thread_link(struct thread *td, struct proc *p); 1081 void thread_reap(void); 1082 int thread_single(struct proc *p, int how); 1083 void thread_single_end(struct proc *p, int how); 1084 void thread_stash(struct thread *td); 1085 void thread_stopped(struct proc *p); 1086 void childproc_stopped(struct proc *child, int reason); 1087 void childproc_continued(struct proc *child); 1088 void childproc_exited(struct proc *child); 1089 int thread_suspend_check(int how); 1090 bool thread_suspend_check_needed(void); 1091 void thread_suspend_switch(struct thread *, struct proc *p); 1092 void thread_suspend_one(struct thread *td); 1093 void thread_unlink(struct thread *td); 1094 void thread_unsuspend(struct proc *p); 1095 void thread_wait(struct proc *p); 1096 struct thread *thread_find(struct proc *p, lwpid_t tid); 1097 1098 void stop_all_proc(void); 1099 void resume_all_proc(void); 1100 1101 static __inline int 1102 curthread_pflags_set(int flags) 1103 { 1104 struct thread *td; 1105 int save; 1106 1107 td = curthread; 1108 save = ~flags | (td->td_pflags & flags); 1109 td->td_pflags |= flags; 1110 return (save); 1111 } 1112 1113 static __inline void 1114 curthread_pflags_restore(int save) 1115 { 1116 1117 curthread->td_pflags &= save; 1118 } 1119 1120 static __inline __pure2 struct td_sched * 1121 td_get_sched(struct thread *td) 1122 { 1123 1124 return ((struct td_sched *)&td[1]); 1125 } 1126 1127 extern void (*softdep_ast_cleanup)(struct thread *); 1128 static __inline void 1129 td_softdep_cleanup(struct thread *td) 1130 { 1131 1132 if (td->td_su != NULL && softdep_ast_cleanup != NULL) 1133 softdep_ast_cleanup(td); 1134 } 1135 1136 #endif /* _KERNEL */ 1137 1138 #endif /* !_SYS_PROC_H_ */ 1139