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.
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
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 * 4. 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.
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
34 * @(#)proc.h 8.15 (Berkeley) 5/19/95
41 #include <sys/callout.h> /* For struct callout. */
42 #include <sys/event.h> /* For struct klist. */
44 #include <sys/filedesc.h>
46 #include <sys/queue.h>
47 #include <sys/_lock.h>
48 #include <sys/_mutex.h>
49 #include <sys/priority.h>
50 #include <sys/rtprio.h> /* XXX. */
52 #include <sys/sigio.h>
53 #include <sys/signal.h>
54 #include <sys/signalvar.h>
56 #include <sys/time.h> /* For structs itimerval, timeval. */
60 #include <sys/ucontext.h>
61 #include <sys/ucred.h>
62 #include <machine/proc.h> /* Machine-dependent proc substruct. */
65 * One structure allocated per session.
68 * (m) locked by s_mtx mtx
69 * (e) locked by proctree_lock sx
70 * (c) const until freeing
73 int s_count; /* (m) Ref cnt; pgrps in session. */
74 struct proc *s_leader; /* (m + e) Session leader. */
75 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */
76 struct tty *s_ttyp; /* (m) Controlling tty. */
77 pid_t s_sid; /* (c) Session ID. */
78 /* (m) Setlogin() name: */
79 char s_login[roundup(MAXLOGNAME, sizeof(long))];
80 struct mtx s_mtx; /* Mutex to protect members. */
84 * One structure allocated per process group.
87 * (m) locked by pg_mtx mtx
88 * (e) locked by proctree_lock sx
89 * (c) const until freeing
92 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */
93 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */
94 struct session *pg_session; /* (c) Pointer to session. */
95 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */
96 pid_t pg_id; /* (c) Process group id. */
97 int pg_jobc; /* (m) Job control process count. */
98 struct mtx pg_mtx; /* Mutex to protect members */
102 * pargs, used to hold a copy of the command line, if it had a sane length.
105 u_int ar_ref; /* Reference count. */
106 u_int ar_length; /* Length. */
107 u_char ar_args[1]; /* Arguments. */
111 * Description of a process.
113 * This structure contains the information needed to manage a thread of
114 * control, known in UN*X as a process; it has references to substructures
115 * containing descriptions of things that the process uses, but may share
116 * with related processes. The process structure and the substructures
117 * are always addressable except for those marked "(CPU)" below,
118 * which might be addressable only on a processor on which the process
121 * Below is a key of locks used to protect each member of struct proc. The
122 * lock is indicated by a reference to a specific character in parens in the
123 * associated comment.
124 * * - not yet protected
125 * a - only touched by curproc or parent during fork/wait
126 * b - created at fork, never changes
127 * (exception aiods switch vmspaces, but they are also
128 * marked 'P_SYSTEM' so hopefully it will be left alone)
129 * c - locked by proc mtx
130 * d - locked by allproc_lock lock
131 * e - locked by proctree_lock lock
133 * g - process group mtx
134 * h - callout_lock mtx
135 * i - by curproc or the master session mtx
136 * j - locked by sched_lock mtx
137 * k - only accessed by curthread
138 * k*- only accessed by curthread and from an interrupt
139 * l - the attaching proc or attaching proc parent
141 * n - not locked, lazy
143 * p - select lock (sellock)
144 * q - td_contested lock
146 * x - created at fork, only changes during single threading in exec
147 * z - zombie threads lock
149 * If the locking key specifies two identifiers (for example, p_pptr) then
150 * either lock is sufficient for read access, but both locks must be held
154 struct kaudit_record;
164 struct mqueue_notifier;
167 * Here we define the two structures used for process information.
169 * The first is the thread. It might be thought of as a "Kernel
170 * Schedulable Entity Context".
171 * This structure contains all the information as to where a thread of
172 * execution is now, or was when it was suspended, why it was suspended,
173 * and anything else that will be needed to restart it when it is
174 * rescheduled. It includes a scheduler specific substructure that is different
175 * for each scheduler.
178 * It is important to remember that when using M:N threading,
179 * a particular thread structure may only exist as long as
180 * the system call or kernel entrance (e.g. by pagefault)
181 * which it is currently executing. It should therefore NEVER be referenced
182 * by pointers in long lived structures that live longer than a single
183 * request. If several threads complete their work at the same time,
184 * they will all rewind their stacks to the user boundary, report their
185 * completion state, and all but one will be freed. That last one will
186 * be kept to provide a kernel stack and pcb for the NEXT syscall or kernel
187 * entrance (basically to save freeing and then re-allocating it). The existing
188 * thread keeps a cached spare thread available to allow it to quickly
189 * get one when it needs a new one. There is also a system
190 * cache of free threads. Threads have priority and partake in priority
191 * inheritance schemes.
193 * The second is the proc (process) which owns all the resources of a process
194 * other than CPU cycles, which are parceled out to the threads.
198 * Threads are the unit of execution
199 With a single run queue used by all processors:
201 RUNQ: --->THREAD---THREAD--... SLEEPQ:[]---THREAD---THREAD---THREAD
206 With PER-CPU run queues:
207 it gets more complicated.
212 * Kernel runnable context (thread).
213 * This is what is put to sleep and reactivated.
214 * Thread context. Processes may have multiple threads.
217 struct proc *td_proc; /* (*) Associated process. */
218 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */
220 /* The two queues below should someday be merged. */
221 TAILQ_ENTRY(thread) td_slpq; /* (j) Sleep queue. */
222 TAILQ_ENTRY(thread) td_lockq; /* (j) Lock queue. */
224 TAILQ_HEAD(, selinfo) td_selq; /* (p) List of selinfos. */
225 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */
226 struct turnstile *td_turnstile; /* (k) Associated turnstile. */
227 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */
228 lwpid_t td_tid; /* (b) Thread ID. */
229 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */
230 #define td_siglist td_sigqueue.sq_signals
232 /* Cleared during fork1() or thread_schedule_upcall(). */
233 #define td_startzero td_flags
234 int td_flags; /* (j) TDF_* flags. */
235 int td_inhibitors; /* (j) 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; /* (j) Sleepqueue queue blocked on. */
239 void *td_wchan; /* (j) Sleep address. */
240 const char *td_wmesg; /* (j) Reason for sleep. */
241 u_char td_lastcpu; /* (j) Last cpu we were on. */
242 u_char td_oncpu; /* (j) Which cpu we are on. */
243 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */
244 short td_locks; /* (k) Count of non-spin locks. */
245 u_char td_tsqueue; /* (j) Turnstile queue blocked on. */
246 struct turnstile *td_blocked; /* (j) Lock thread is blocked on. */
247 const char *td_lockname; /* (j) Name of lock blocked on. */
248 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */
249 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
250 int td_intr_nesting_level; /* (k) Interrupt recursion. */
251 int td_pinned; /* (k) Temporary cpu pin count. */
252 struct kse_thr_mailbox *td_mailbox; /* (*) Userland mailbox address. */
253 struct ucred *td_ucred; /* (k) Reference to credentials. */
254 struct thread *td_standin; /* (k + a) Use this for an upcall. */
255 struct kse_upcall *td_upcall; /* (k + j) Upcall structure. */
256 u_int td_estcpu; /* (j) Sum of the same field in KSEs. */
257 u_int td_slptime; /* (j) How long completely blocked. */
258 u_int td_pticks; /* (k) Statclock hits for profiling */
259 u_int td_sticks; /* (k) Statclock hits in system mode. */
260 u_int td_iticks; /* (k) Statclock hits in intr mode. */
261 u_int td_uticks; /* (k) Statclock hits in user mode. */
262 u_int td_uuticks; /* (k) Statclock hits (usr), for UTS. */
263 u_int td_usticks; /* (k) Statclock hits (sys), for UTS. */
264 int td_intrval; /* (j) Return value of TDF_INTERRUPT. */
265 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */
266 sigset_t td_sigmask; /* (c) Current signal mask. */
267 volatile u_int td_generation; /* (k) For detection of preemption */
268 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */
269 int td_kflags; /* (c) Flags for KSE threading. */
270 int td_xsig; /* (c) Signal for ptrace */
271 u_long td_profil_addr; /* (k) Temporary addr until AST. */
272 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */
273 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */
274 #define td_endzero td_base_pri
276 /* Copied during fork1() or thread_sched_upcall(). */
277 #define td_startcopy td_endzero
278 u_char td_base_pri; /* (j) Thread base kernel priority. */
279 u_char td_priority; /* (j) Thread active priority. */
280 u_char td_pri_class; /* (j) Scheduling class. */
281 u_char td_user_pri; /* (j) User pri from estcpu and nice. */
282 u_char td_base_user_pri; /* (j) Base user pri */
283 #define td_endcopy td_pcb
286 * Fields that must be manually set in fork1() or thread_sched_upcall()
287 * or already have been set in the allocator, constructor, etc.
289 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */
297 register_t td_retval[2]; /* (k) Syscall aux returns. */
298 struct callout td_slpcallout; /* (h) Callout for sleep. */
299 struct trapframe *td_frame; /* (k) */
300 struct vm_object *td_kstack_obj;/* (a) Kstack object. */
301 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */
302 int td_kstack_pages; /* (a) Size of the kstack. */
303 struct vm_object *td_altkstack_obj;/* (a) Alternate kstack object. */
304 vm_offset_t td_altkstack; /* (a) Kernel VA of alternate kstack. */
305 int td_altkstack_pages; /* (a) Size of alternate kstack. */
306 volatile u_int td_critnest; /* (k*) Critical section nest level. */
307 struct mdthread td_md; /* (k) Any machine-dependent fields. */
308 struct td_sched *td_sched; /* (*) Scheduler-specific data. */
309 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */
313 * Flags kept in td_flags:
314 * To change these you MUST have the scheduler lock.
316 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */
317 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */
318 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */
319 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */
320 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */
321 #define TDF_SELECT 0x00000040 /* Selecting; wakeup/waiting danger. */
322 #define TDF_SLEEPABORT 0x00000080 /* sleepq_abort was called. */
323 #define TDF_TSNOBLOCK 0x00000100 /* Don't block on a turnstile due to race. */
324 #define TDF_UBORROWING 0x00000200 /* Thread is borrowing user pri. */
325 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */
326 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */
327 #define TDF_TIMOFAIL 0x00001000 /* Timeout from sleep after we were awake. */
328 #define TDF_INTERRUPT 0x00002000 /* Thread is marked as interrupted. */
329 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */
330 #define TDF_UNUSED15 0x00008000 /* --available -- */
331 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */
332 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */
333 #define TDF_XSIG 0x00040000 /* Thread is exchanging signal under trace */
334 #define TDF_UNUSED19 0x00080000 /* Thread is sleeping on a umtx. */
335 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */
336 #define TDF_DBSUSPEND 0x00200000 /* Thread is suspended by debugger */
337 #define TDF_UNUSED22 0x00400000 /* --available -- */
338 #define TDF_UNUSED23 0x00800000 /* --available -- */
339 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */
340 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */
341 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */
342 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */
345 * "Private" flags kept in td_pflags:
346 * These are only accessed by curthread and thus need no locking.
348 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */
349 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */
350 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */
351 #define TDP_UPCALLING 0x00000008 /* This thread is doing an upcall. */
352 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */
353 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */
354 #define TDP_DEADLKTREAT 0x00000040 /* Lock aquisition - deadlock treatment. */
355 #define TDP_SA 0x00000080 /* A scheduler activation based thread. */
356 #define TDP_NOSLEEPING 0x00000100 /* Thread is not allowed to sleep on a sq. */
357 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */
358 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */
359 #define TDP_CAN_UNBIND 0x00000800 /* Only temporarily bound. */
360 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */
361 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */
362 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */
363 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */
364 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */
365 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */
366 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */
367 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */
370 * Reasons that the current thread can not be run yet.
371 * More than one may apply.
373 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */
374 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */
375 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */
376 #define TDI_LOCK 0x0008 /* Stopped on a lock. */
377 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */
380 * flags (in kflags) related to M:N threading.
382 #define TDK_KSEREL 0x0001 /* Blocked in msleep on p->p_completed. */
383 #define TDK_KSERELSIG 0x0002 /* Blocked in msleep on p->p_siglist. */
384 #define TDK_WAKEUP 0x0004 /* Thread has been woken by kse_wakeup. */
386 #define TD_CAN_UNBIND(td) \
387 (((td)->td_pflags & TDP_CAN_UNBIND) && \
388 ((td)->td_upcall != NULL))
390 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING)
391 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL)
392 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED)
393 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED)
394 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK)
395 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT)
396 #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING)
397 #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ)
398 #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN)
399 #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED)
400 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED)
402 #define TD_SET_INHIB(td, inhib) do { \
403 (td)->td_state = TDS_INHIBITED; \
404 (td)->td_inhibitors |= (inhib); \
407 #define TD_CLR_INHIB(td, inhib) do { \
408 if (((td)->td_inhibitors & (inhib)) && \
409 (((td)->td_inhibitors &= ~(inhib)) == 0)) \
410 (td)->td_state = TDS_CAN_RUN; \
413 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING)
414 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED)
415 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK)
416 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED)
417 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT)
418 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING)
420 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING)
421 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED)
422 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK)
423 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED)
424 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT)
426 #define TD_SET_RUNNING(td) (td)->td_state = TDS_RUNNING
427 #define TD_SET_RUNQ(td) (td)->td_state = TDS_RUNQ
428 #define TD_SET_CAN_RUN(td) (td)->td_state = TDS_CAN_RUN
431 * An upcall is used when returning to userland. If a thread does not have
432 * an upcall on return to userland the thread exports its context and exits.
435 TAILQ_ENTRY(kse_upcall) ku_link; /* List of upcalls in proc. */
436 struct proc *ku_proc; /* Associated proc. */
437 struct thread *ku_owner; /* Owning thread. */
438 int ku_flags; /* KUF_* flags. */
439 struct kse_mailbox *ku_mailbox; /* Userland mailbox address. */
440 stack_t ku_stack; /* Userland upcall stack. */
441 void *ku_func; /* Userland upcall function. */
442 unsigned int ku_mflags; /* Cached upcall mbox flags. */
445 #define KUF_DOUPCALL 0x00001 /* Do upcall now; don't wait. */
446 #define KUF_EXITING 0x00002 /* Upcall structure is exiting. */
449 * XXX: Does this belong in resource.h or resourcevar.h instead?
450 * Resource usage extension. The times in rusage structs in the kernel are
451 * never up to date. The actual times are kept as runtimes and tick counts
452 * (with control info in the "previous" times), and are converted when
453 * userland asks for rusage info. Backwards compatibility prevents putting
454 * this directly in the user-visible rusage struct.
456 * Locking: (cj) means (j) for p_rux and (c) for p_crux.
459 u_int64_t rux_runtime; /* (cj) Real time. */
460 u_int64_t rux_uticks; /* (cj) Statclock hits in user mode. */
461 u_int64_t rux_sticks; /* (cj) Statclock hits in sys mode. */
462 u_int64_t rux_iticks; /* (cj) Statclock hits in intr mode. */
463 u_int64_t rux_uu; /* (c) Previous user time in usec. */
464 u_int64_t rux_su; /* (c) Previous sys time in usec. */
465 u_int64_t rux_tu; /* (c) Previous total time in usec. */
469 * The old fashionned process. May have multiple threads.
470 * Starts off with a single embedded THREAD.
473 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */
474 TAILQ_HEAD(, thread) p_threads; /* (j)(td_plist) Threads. (shortcut) */
475 struct ucred *p_ucred; /* (c) Process owner's identity. */
476 struct filedesc *p_fd; /* (b) Open files. */
477 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */
478 /* Accumulated stats for all threads? */
479 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */
480 struct plimit *p_limit; /* (c) Process limits. */
481 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */
482 TAILQ_HEAD(, kse_upcall) p_upcalls; /* All upcalls in the proc. */
485 * The following don't make too much sense.
486 * See the td_ or ke_ versions of the same flags.
488 int p_flag; /* (c) P_* flags. */
489 int p_sflag; /* (j) PS_* flags. */
491 PRS_NEW = 0, /* In creation */
492 PRS_NORMAL, /* threads can be run. */
494 } p_state; /* (j/c) S* process status. */
496 pid_t p_pid; /* (b) Process identifier. */
497 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */
498 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */
499 struct proc *p_pptr; /* (c + e) Pointer to parent process. */
500 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */
501 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */
502 struct mtx p_mtx; /* (n) Lock for this struct. */
503 struct ksiginfo *p_ksi; /* Locked by parent proc lock */
504 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */
505 #define p_siglist p_sigqueue.sq_signals
507 /* The following fields are all zeroed upon creation in fork. */
508 #define p_startzero p_oppid
509 pid_t p_oppid; /* (c + e) Save ppid in ptrace. XXX */
510 struct vmspace *p_vmspace; /* (b) Address space. */
511 u_int p_swtime; /* (j) Time swapped in or out. */
512 struct itimerval p_realtimer; /* (c) Alarm timer. */
513 struct rusage_ext p_rux; /* (cj) Internal resource usage. */
514 struct rusage_ext p_crux; /* (c) Internal child resource usage. */
515 int p_profthreads; /* (c) Num threads in addupc_task. */
516 int p_maxthrwaits; /* (c) Max threads num waiters */
517 int p_traceflag; /* (o) Kernel trace points. */
518 struct vnode *p_tracevp; /* (c + o) Trace to vnode. */
519 struct ucred *p_tracecred; /* (o) Credentials to trace with. */
520 struct vnode *p_textvp; /* (b) Vnode of executable. */
521 char p_lock; /* (c) Proclock (prevent swap) count. */
522 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */
523 int p_sigparent; /* (c) Signal to parent on exit. */
524 int p_sig; /* (n) For core dump/debugger XXX. */
525 u_long p_code; /* (n) For core dump/debugger XXX. */
526 u_int p_stops; /* (c) Stop event bitmask. */
527 u_int p_stype; /* (c) Stop event type. */
528 char p_step; /* (c) Process is stopped. */
529 u_char p_pfsflags; /* (c) Procfs flags. */
530 struct nlminfo *p_nlminfo; /* (?) Only used by/for lockd. */
531 struct kaioinfo *p_aioinfo; /* (c) ASYNC I/O info. */
532 struct thread *p_singlethread;/* (c + j) If single threading this is it */
533 int p_suspcount; /* (c) Num threads in suspended mode. */
534 struct thread *p_xthread; /* (c) Trap thread */
535 int p_boundary_count;/* (c) Num threads at user boundary */
536 int p_pendingcnt; /* how many signals are pending */
537 struct itimers *p_itimers; /* (c) POSIX interval timers. */
539 u_int p_estcpu; /* (j) Sum of the field in threads. */
540 u_int p_slptime; /* (j) How long completely blocked. */
541 int p_numupcalls; /* (j) Num upcalls. */
542 int p_upsleeps; /* (c) Num threads in kse_release(). */
543 struct kse_thr_mailbox *p_completed; /* (c) Completed thread mboxes. */
544 int p_nextupcall; /* (n) Next upcall time. */
545 int p_upquantum; /* (n) Quantum to schedule an upcall. */
546 /* End area that is zeroed on creation. */
547 #define p_endzero p_magic
549 /* The following fields are all copied upon creation in fork. */
550 #define p_startcopy p_endzero
551 u_int p_magic; /* (b) Magic number. */
552 char p_comm[MAXCOMLEN + 1]; /* (b) Process name. */
553 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */
554 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */
555 struct pargs *p_args; /* (c) Process arguments. */
556 rlim_t p_cpulimit; /* (j) Current CPU limit in seconds. */
557 signed char p_nice; /* (c + j) Process "nice" value. */
558 /* End area that is copied on creation. */
559 #define p_endcopy p_xstat
561 u_short p_xstat; /* (c) Exit status; also stop sig. */
562 struct knlist p_klist; /* (c) Knotes attached to this proc. */
563 int p_numthreads; /* (j) Number of threads. */
564 struct mdproc p_md; /* Any machine-dependent fields. */
565 struct callout p_itcallout; /* (h + c) Interval timer callout. */
566 u_short p_acflag; /* (c) Accounting flags. */
567 struct rusage *p_ru; /* (a) Exit information. XXX */
568 struct proc *p_peers; /* (r) */
569 struct proc *p_leader; /* (b) */
570 void *p_emuldata; /* (c) Emulator state data. */
571 struct label *p_label; /* (*) Proc (not subject) MAC label. */
572 struct p_sched *p_sched; /* (*) Scheduler-specific data. */
573 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */
574 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/
575 struct auditinfo *p_au; /* (c) Process audit properties. */
578 #define p_session p_pgrp->pg_session
579 #define p_pgid p_pgrp->pg_id
581 #define NOCPU 0xff /* For when we aren't on a CPU. */
584 /* These flags are kept in p_flag. */
585 #define P_ADVLOCK 0x00001 /* Process may hold a POSIX advisory lock. */
586 #define P_CONTROLT 0x00002 /* Has a controlling terminal. */
587 #define P_KTHREAD 0x00004 /* Kernel thread (*). */
588 #define P_NOLOAD 0x00008 /* Ignore during load avg calculations. */
589 #define P_PPWAIT 0x00010 /* Parent is waiting for child to exec/exit. */
590 #define P_PROFIL 0x00020 /* Has started profiling. */
591 #define P_STOPPROF 0x00040 /* Has thread requesting to stop profiling. */
592 #define P_HADTHREADS 0x00080 /* Has had threads (no cleanup shortcuts) */
593 #define P_SUGID 0x00100 /* Had set id privileges since last exec. */
594 #define P_SYSTEM 0x00200 /* System proc: no sigs, stats or swapping. */
595 #define P_SINGLE_EXIT 0x00400 /* Threads suspending should exit, not wait. */
596 #define P_TRACED 0x00800 /* Debugged process being traced. */
597 #define P_WAITED 0x01000 /* Someone is waiting for us. */
598 #define P_WEXIT 0x02000 /* Working on exiting. */
599 #define P_EXEC 0x04000 /* Process called exec. */
600 #define P_SA 0x08000 /* Using scheduler activations. */
601 #define P_CONTINUED 0x10000 /* Proc has continued from a stopped state. */
602 #define P_STOPPED_SIG 0x20000 /* Stopped due to SIGSTOP/SIGTSTP. */
603 #define P_STOPPED_TRACE 0x40000 /* Stopped because of tracing. */
604 #define P_STOPPED_SINGLE 0x80000 /* Only 1 thread can continue (not to user). */
605 #define P_PROTECTED 0x100000 /* Do not kill on memory overcommit. */
606 #define P_SIGEVENT 0x200000 /* Process pending signals changed. */
607 #define P_SINGLE_BOUNDARY 0x400000 /* Threads should suspend at user boundary. */
608 #define P_HWPMC 0x800000 /* Process is using HWPMCs */
610 #define P_JAILED 0x1000000 /* Process is in jail. */
611 #define P_INEXEC 0x4000000 /* Process is in execve(). */
612 #define P_STATCHILD 0x8000000 /* Child process stopped or exited. */
614 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
615 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED)
617 /* These flags are kept in p_sflag and are protected with sched_lock. */
618 #define PS_INMEM 0x00001 /* Loaded into memory. */
619 #define PS_XCPU 0x00002 /* Exceeded CPU limit. */
620 #define PS_ALRMPEND 0x00020 /* Pending SIGVTALRM needs to be posted. */
621 #define PS_PROFPEND 0x00040 /* Pending SIGPROF needs to be posted. */
622 #define PS_SWAPINREQ 0x00100 /* Swapin request due to wakeup. */
623 #define PS_SWAPPINGOUT 0x00200 /* Process is being swapped out. */
624 #define PS_SWAPPINGIN 0x04000 /* Process is being swapped in. */
625 #define PS_MACPEND 0x08000 /* AST-based MAC event pending. */
628 * These were process status values (p_stat), now they are only used in
629 * legacy conversion code.
631 #define SIDL 1 /* Process being created by fork. */
632 #define SRUN 2 /* Currently runnable. */
633 #define SSLEEP 3 /* Sleeping on an address. */
634 #define SSTOP 4 /* Process debugging or suspension. */
635 #define SZOMB 5 /* Awaiting collection by parent. */
636 #define SWAIT 6 /* Waiting for interrupt. */
637 #define SLOCK 7 /* Blocked on a lock. */
639 #define P_MAGIC 0xbeefface
643 /* Flags for mi_switch(). */
644 #define SW_VOL 0x0001 /* Voluntary switch. */
645 #define SW_INVOL 0x0002 /* Involuntary switch. */
646 #define SW_PREEMPT 0x0004 /* The invol switch is a preemption */
648 /* How values for thread_single(). */
649 #define SINGLE_NO_EXIT 0
650 #define SINGLE_EXIT 1
651 #define SINGLE_BOUNDARY 2
653 /* XXXKSE: Missing values for thread_suspend_check(). */
655 #ifdef MALLOC_DECLARE
656 MALLOC_DECLARE(M_PARGS);
657 MALLOC_DECLARE(M_PGRP);
658 MALLOC_DECLARE(M_SESSION);
659 MALLOC_DECLARE(M_SUBPROC);
660 MALLOC_DECLARE(M_ZOMBIE);
663 #define FOREACH_PROC_IN_SYSTEM(p) \
664 LIST_FOREACH((p), &allproc, p_list)
665 #define FOREACH_THREAD_IN_PROC(p, td) \
666 TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
667 #define FOREACH_UPCALL_IN_PROC(p, ku) \
668 TAILQ_FOREACH((ku), &(p)->p_upcalls, ku_link)
670 /* XXXKSE the following lines should probably only be used in 1:1 code: */
671 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads)
674 * We use process IDs <= PID_MAX; PID_MAX + 1 must also fit in a pid_t,
675 * as it is used to represent "no process group".
677 #define PID_MAX 99999
678 #define NO_PID 100000
680 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p))
681 #define SESSHOLD(s) ((s)->s_count++)
682 #define SESSRELE(s) sessrele(s)
685 #define STOPEVENT(p, e, v) do { \
686 if ((p)->p_stops & (e)) { \
688 stopevent((p), (e), (v)); \
692 #define _STOPEVENT(p, e, v) do { \
693 PROC_LOCK_ASSERT(p, MA_OWNED); \
694 WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, &p->p_mtx.mtx_object, \
695 "checking stopevent %d", (e)); \
696 if ((p)->p_stops & (e)) \
697 stopevent((p), (e), (v)); \
700 /* Lock and unlock a process. */
701 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx)
702 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx)
703 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx)
704 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx)
705 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type))
707 /* Lock and unlock a process group. */
708 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx)
709 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx)
710 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx)
711 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type))
713 #define PGRP_LOCK_PGSIGNAL(pg) do { \
717 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \
722 /* Lock and unlock a session. */
723 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx)
724 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx)
725 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx)
726 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type))
728 /* Hold process U-area in memory, normally for ptrace/procfs work. */
729 #define PHOLD(p) do { \
734 #define _PHOLD(p) do { \
735 PROC_LOCK_ASSERT((p), MA_OWNED); \
736 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
737 ("PHOLD of exiting process")); \
739 if (((p)->p_sflag & PS_INMEM) == 0) \
743 #define PRELE(p) do { \
748 #define _PRELE(p) do { \
749 PROC_LOCK_ASSERT((p), MA_OWNED); \
751 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \
752 wakeup(&(p)->p_lock); \
755 /* Check whether a thread is safe to be swapped out. */
756 #define thread_safetoswapout(td) (TD_IS_SLEEPING(td) || TD_IS_SUSPENDED(td))
758 /* Control whether or not it is safe for curthread to sleep. */
759 #define THREAD_NO_SLEEPING() do { \
760 KASSERT(!(curthread->td_pflags & TDP_NOSLEEPING), \
761 ("nested no sleeping")); \
762 curthread->td_pflags |= TDP_NOSLEEPING; \
765 #define THREAD_SLEEPING_OK() do { \
766 KASSERT((curthread->td_pflags & TDP_NOSLEEPING), \
767 ("nested sleeping ok")); \
768 curthread->td_pflags &= ~TDP_NOSLEEPING; \
771 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash])
772 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
773 extern u_long pidhash;
775 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash])
776 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
777 extern u_long pgrphash;
779 extern struct sx allproc_lock;
780 extern struct sx proctree_lock;
781 extern struct mtx ppeers_lock;
782 extern struct proc proc0; /* Process slot for swapper. */
783 extern struct thread thread0; /* Primary thread in proc0. */
784 extern struct vmspace vmspace0; /* VM space for proc0. */
785 extern int hogticks; /* Limit on kernel cpu hogs. */
787 extern int nprocs, maxproc; /* Current and max number of procs. */
788 extern int maxprocperuid; /* Max procs per uid. */
789 extern u_long ps_arg_cache_limit;
791 LIST_HEAD(proclist, proc);
792 TAILQ_HEAD(procqueue, proc);
793 TAILQ_HEAD(threadqueue, thread);
794 extern struct proclist allproc; /* List of all processes. */
795 extern struct proclist zombproc; /* List of zombie processes. */
796 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
798 extern struct uma_zone *proc_zone;
800 struct proc *pfind(pid_t); /* Find process by id. */
801 struct pgrp *pgfind(pid_t); /* Find process group by id. */
802 struct proc *zpfind(pid_t); /* Find zombie process by id. */
804 void ast(struct trapframe *framep);
805 struct thread *choosethread(void);
806 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
807 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp,
808 struct session *sess);
809 int enterthispgrp(struct proc *p, struct pgrp *pgrp);
810 void faultin(struct proc *p);
811 void fixjobc(struct proc *p, struct pgrp *pgrp, int entering);
812 int fork1(struct thread *, int, int, struct proc **);
813 void fork_exit(void (*)(void *, struct trapframe *), void *,
815 void fork_return(struct thread *, struct trapframe *);
816 int inferior(struct proc *p);
817 void kick_proc0(void);
818 int leavepgrp(struct proc *p);
819 int maybe_preempt(struct thread *td);
820 void mi_switch(int flags, struct thread *newtd);
821 int p_candebug(struct thread *td, struct proc *p);
822 int p_cansee(struct thread *td, struct proc *p);
823 int p_cansched(struct thread *td, struct proc *p);
824 int p_cansignal(struct thread *td, struct proc *p, int signum);
825 int p_canwait(struct thread *td, struct proc *p);
826 struct pargs *pargs_alloc(int len);
827 void pargs_drop(struct pargs *pa);
828 void pargs_free(struct pargs *pa);
829 void pargs_hold(struct pargs *pa);
831 void proc_linkup(struct proc *p, struct thread *td);
832 void proc_reparent(struct proc *child, struct proc *newparent);
833 struct pstats *pstats_alloc(void);
834 void pstats_fork(struct pstats *src, struct pstats *dst);
835 void pstats_free(struct pstats *ps);
836 int securelevel_ge(struct ucred *cr, int level);
837 int securelevel_gt(struct ucred *cr, int level);
838 void sessrele(struct session *);
839 void setrunnable(struct thread *);
840 void setsugid(struct proc *p);
841 int sigonstack(size_t sp);
842 void sleepinit(void);
843 void stopevent(struct proc *, u_int, u_int);
844 void threadinit(void);
846 extern void (*cpu_idle_hook)(void); /* Hook to machdep CPU idler. */
847 void cpu_switch(struct thread *old, struct thread *new);
848 void cpu_throw(struct thread *old, struct thread *new) __dead2;
849 void unsleep(struct thread *);
850 void userret(struct thread *, struct trapframe *);
852 void cpu_exit(struct thread *);
853 void exit1(struct thread *, int) __dead2;
854 void cpu_fork(struct thread *, struct proc *, struct thread *, int);
855 void cpu_set_fork_handler(struct thread *, void (*)(void *), void *);
860 void cpu_set_upcall(struct thread *td, struct thread *td0);
861 void cpu_set_upcall_kse(struct thread *, void (*)(void *), void *, stack_t *);
862 int cpu_set_user_tls(struct thread *, void *tls_base);
863 void cpu_thread_clean(struct thread *);
864 void cpu_thread_exit(struct thread *);
865 void cpu_thread_setup(struct thread *td);
866 void cpu_thread_swapin(struct thread *);
867 void cpu_thread_swapout(struct thread *);
868 struct thread *thread_alloc(void);
869 void thread_continued(struct proc *p);
870 void thread_exit(void) __dead2;
871 int thread_export_context(struct thread *td, int willexit);
872 void thread_free(struct thread *td);
873 void thread_link(struct thread *td, struct proc *p);
874 void thread_reap(void);
875 struct thread *thread_schedule_upcall(struct thread *td, struct kse_upcall *ku);
876 void thread_signal_add(struct thread *td, ksiginfo_t *);
877 int thread_single(int how);
878 void thread_single_end(void);
879 void thread_stash(struct thread *td);
880 int thread_statclock(int user);
881 void thread_stopped(struct proc *p);
882 void childproc_stopped(struct proc *child, int reason);
883 void childproc_continued(struct proc *child);
884 void childproc_exited(struct proc *child);
885 int thread_suspend_check(int how);
886 void thread_suspend_one(struct thread *td);
887 struct thread *thread_switchout(struct thread *td, int flags,
888 struct thread *newtd);
889 void thread_unlink(struct thread *td);
890 void thread_unsuspend(struct proc *p);
891 void thread_unsuspend_one(struct thread *td);
892 void thread_unthread(struct thread *td);
893 int thread_userret(struct thread *td, struct trapframe *frame);
894 void thread_user_enter(struct thread *td);
895 void thread_wait(struct proc *p);
896 struct thread *thread_find(struct proc *p, lwpid_t tid);
897 void thr_exit1(void);
898 struct kse_upcall *upcall_alloc(void);
899 void upcall_free(struct kse_upcall *ku);
900 void upcall_link(struct kse_upcall *ku, struct proc *p);
901 void upcall_unlink(struct kse_upcall *ku);
902 void upcall_remove(struct thread *td);
903 void upcall_stash(struct kse_upcall *ke);
907 #endif /* !_SYS_PROC_H_ */