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 * 3. All advertising materials mentioning features or use of this software
19 * must display the following acknowledgement:
20 * This product includes software developed by the University of
21 * California, Berkeley and its contributors.
22 * 4. Neither the name of the University nor the names of its contributors
23 * may be used to endorse or promote products derived from this software
24 * without specific prior written permission.
26 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
27 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
30 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
31 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
32 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
33 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
34 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
35 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
38 * @(#)proc.h 8.15 (Berkeley) 5/19/95
45 #include <sys/callout.h> /* For struct callout. */
46 #include <sys/event.h> /* For struct klist. */
47 #include <sys/filedesc.h>
48 #include <sys/queue.h>
49 #include <sys/priority.h>
50 #include <sys/rtprio.h> /* XXX */
52 #include <sys/signal.h>
54 #include <sys/time.h> /* For structs itimerval, timeval. */
58 #include <sys/ucontext.h>
59 #include <sys/ucred.h>
60 #include <machine/proc.h> /* Machine-dependent proc substruct. */
63 * One structure allocated per session.
66 * (m) locked by s_mtx mtx
67 * (e) locked by proctree_lock sx
68 * (c) const until freeing
71 int s_count; /* (m) Ref cnt; pgrps in session. */
72 struct proc *s_leader; /* (m + e) Session leader. */
73 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */
74 struct tty *s_ttyp; /* (m) Controlling tty. */
75 pid_t s_sid; /* (c) Session ID. */
76 /* (m) Setlogin() name: */
77 char s_login[roundup(MAXLOGNAME, sizeof(long))];
78 struct mtx s_mtx; /* Mutex to protect members */
82 * One structure allocated per process group.
85 * (m) locked by pg_mtx mtx
86 * (e) locked by proctree_lock sx
87 * (c) const until freeing
90 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */
91 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */
92 struct session *pg_session; /* (c) Pointer to session. */
93 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */
94 pid_t pg_id; /* (c) Pgrp id. */
95 int pg_jobc; /* (m) job cntl proc count */
96 struct mtx pg_mtx; /* Mutex to protect members */
100 sigset_t ps_sigignore; /* Signals being ignored. */
101 sigset_t ps_sigcatch; /* Signals being caught by user. */
103 struct sigacts *ps_sigacts; /* Signal actions, state. */
107 #define PS_NOCLDWAIT 0x0001 /* No zombies if child dies */
108 #define PS_NOCLDSTOP 0x0002 /* No SIGCHLD when children stop. */
109 #define PS_CLDSIGIGN 0x0004 /* The SIGCHLD handler is SIG_IGN. */
112 * pargs, used to hold a copy of the command line, if it had a sane length.
115 u_int ar_ref; /* Reference count. */
116 u_int ar_length; /* Length. */
117 u_char ar_args[1]; /* Arguments. */
121 * Description of a process.
123 * This structure contains the information needed to manage a thread of
124 * control, known in UN*X as a process; it has references to substructures
125 * containing descriptions of things that the process uses, but may share
126 * with related processes. The process structure and the substructures
127 * are always addressable except for those marked "(CPU)" below,
128 * which might be addressable only on a processor on which the process
131 * Below is a key of locks used to protect each member of struct proc. The
132 * lock is indicated by a reference to a specific character in parens in the
133 * associated comment.
134 * * - not yet protected
135 * a - only touched by curproc or parent during fork/wait
136 * b - created at fork, never changes
137 * (exception aiods switch vmspaces, but they are also
138 * marked 'P_SYSTEM' so hopefully it will be left alone)
139 * c - locked by proc mtx
140 * d - locked by allproc_lock lock
141 * e - locked by proctree_lock lock
143 * g - process group mtx
144 * h - callout_lock mtx
145 * i - by curproc or the master session mtx
146 * j - locked by sched_lock mtx
147 * k - only accessed by curthread
148 * l - the attaching proc or attaching proc parent
150 * n - not locked, lazy
152 * p - select lock (sellock)
155 * If the locking key specifies two identifiers (for example, p_pptr) then
156 * either lock is sufficient for read access, but both locks must be held
164 * Here we define the four structures used for process information.
166 * The first is the thread. It might be though of as a "Kernel
167 * Schedulable Entity Context".
168 * This structure contains all the information as to where a thread of
169 * execution is now, or was when it was suspended, why it was suspended,
170 * and anything else that will be needed to restart it when it is
171 * rescheduled. Always associated with a KSE when running, but can be
172 * reassigned to an equivalent KSE when being restarted for
173 * load balancing. Each of these is associated with a kernel stack
176 * It is important to remember that a particular thread structure only
177 * exists as long as the system call or kernel entrance (e.g. by pagefault)
178 * which it is currently executing. It should threfore NEVER be referenced
179 * by pointers in long lived structures that live longer than a single
180 * request. If several threads complete their work at the same time,
181 * they will all rewind their stacks to the user boundary, report their
182 * completion state, and all but one will be freed. That last one will
183 * be kept to provide a kernel stack and pcb for the NEXT syscall or kernel
184 * entrance. (basically to save freeing and then re-allocating it) The KSE
185 * keeps a cached thread available to allow it to quickly
186 * get one when it needs a new one. There is also a system
187 * cache of free threads. Threads have priority and partake in priority
188 * inherritance schemes.
193 * The second structure is the Kernel Schedulable Entity. (KSE)
194 * It represents the ability to take a slot in the scheduler queue.
195 * As long as this is scheduled, it could continue to run any threads that
196 * are assigned to the KSEGRP (see later) until either it runs out
197 * of runnable threads of high enough priority, or CPU.
198 * It runs on one CPU and is assigned a quantum of time. When a thread is
199 * blocked, The KSE continues to run and will search for another thread
200 * in a runnable state amongst those it has. It May decide to return to user
201 * mode with a new 'empty' thread if there are no runnable threads.
202 * Threads are temporarily associated with a KSE for scheduling reasons.
207 * The KSEGRP is allocated resources across a number of CPUs.
208 * (Including a number of CPUxQUANTA. It parcels these QUANTA up among
209 * Its KSEs, each of which should be running in a different CPU.
210 * BASE priority and total available quanta are properties of a KSEGRP.
211 * Multiple KSEGRPs in a single process compete against each other
212 * for total quanta in the same way that a forked child competes against
213 * it's parent process.
218 * A process is the owner of all system resources allocated to a task
220 * All KSEGs under one process see, and have the same access to, these
221 * resources (e.g. files, memory, sockets, permissions kqueues).
222 * A process may compete for CPU cycles on the same basis as a
223 * forked process cluster by spawning several KSEGRPs.
229 With a single run queue used by all processors:
231 RUNQ: --->KSE---KSE--... SLEEPQ:[]---THREAD---THREAD---THREAD
233 KSEG---THREAD--THREAD--THREAD []
236 (processors run THREADs from the KSEG until they are exhausted or
237 the KSEG exhausts its quantum)
239 With PER-CPU run queues:
240 KSEs on the separate run queues directly
241 They would be given priorities calculated from the KSEG.
247 * Kernel runnable context (thread).
248 * This is what is put to sleep and reactivated.
249 * The first KSE available in the correct group will run this thread.
250 * If several are available, use the one on the same CPU as last time.
251 * When waing to be run, threads are hung off the KSEGRP in priority order.
252 * with N runnable and queued KSEs in the KSEGRP, the first N threads
253 * are linked to them. Other threads are not yet assigned.
256 struct proc *td_proc; /* Associated process. */
257 struct ksegrp *td_ksegrp; /* Associated KSEG. */
258 TAILQ_ENTRY(thread) td_plist; /* All threads in this proc */
259 TAILQ_ENTRY(thread) td_kglist; /* All threads in this ksegrp */
261 /* The two queues below should someday be merged */
262 TAILQ_ENTRY(thread) td_slpq; /* (j) Sleep queue. XXXKSE */
263 TAILQ_ENTRY(thread) td_lockq; /* (j) Lock queue. XXXKSE */
264 TAILQ_ENTRY(thread) td_runq; /* (j) Run queue(s). XXXKSE */
266 TAILQ_HEAD(, selinfo) td_selq; /* (p) List of selinfos. */
268 /* Cleared during fork1() or thread_sched_upcall() */
269 #define td_startzero td_flags
270 int td_flags; /* (j) TDF_* flags. */
271 int td_inhibitors; /* (j) Why can not run */
272 struct kse *td_last_kse; /* (j) Previous value of td_kse */
273 struct kse *td_kse; /* (j) Current KSE if running. */
274 int td_dupfd; /* (k) Ret value from fdopen. XXX */
275 void *td_wchan; /* (j) Sleep address. */
276 const char *td_wmesg; /* (j) Reason for sleep. */
277 u_char td_lastcpu; /* (j) Last cpu we were on. */
278 u_char td_inktr; /* (k) Currently handling a KTR. */
279 u_char td_inktrace; /* (k) Currently handling a KTRACE. */
280 short td_locks; /* (k) DEBUG: lockmgr count of locks */
281 struct mtx *td_blocked; /* (j) Mutex process is blocked on. */
282 struct ithd *td_ithd; /* (b) For interrupt threads only. */
283 const char *td_lockname; /* (j) Name of lock blocked on. */
284 LIST_HEAD(, mtx) td_contested; /* (j) Contested locks. */
285 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
286 int td_intr_nesting_level; /* (k) Interrupt recursion. */
287 struct kse_thr_mailbox *td_mailbox; /* the userland mailbox address */
288 struct ucred *td_ucred; /* (k) Reference to credentials. */
289 void (*td_switchin)(void); /* (k) Switchin special func. */
290 struct thread *td_standin; /* (?) use this for an upcall */
291 u_int td_critnest; /* (k) Critical section nest level. */
292 #define td_endzero td_md
294 /* Copied during fork1() or thread_sched_upcall() */
295 #define td_startcopy td_endzero
296 /* XXXKSE just copying td_md needs checking! */
297 struct mdthread td_md; /* (k) Any machine-dependent fields. */
298 u_char td_base_pri; /* (j) Thread base kernel priority. */
299 u_char td_priority; /* (j) Thread active priority. */
300 #define td_endcopy td_pcb
303 * fields that must be manually set in fork1() or thread_sched_upcall()
304 * or already have been set in the allocator, contstructor, etc..
306 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */
314 register_t td_retval[2]; /* (k) Syscall aux returns. */
315 struct callout td_slpcallout; /* (h) Callout for sleep. */
316 struct trapframe *td_frame; /* (k) */
317 struct vm_object *td_kstack_obj;/* (a) Kstack object. */
318 vm_offset_t td_kstack; /* Kernel VA of kstack. */
319 int td_kstack_pages; /* Size of the kstack */
320 struct vm_object *td_altkstack_obj;/* (a) Alternate kstack object. */
321 vm_offset_t td_altkstack; /* Kernel VA of alternate kstack. */
322 int td_altkstack_pages; /* Size of the alternate kstack */
324 /* flags kept in td_flags */
325 #define TDF_UNBOUND 0x000001 /* May give away the kse, uses the kg runq. */
326 #define TDF_INPANIC 0x000002 /* Caused a panic, let it drive crashdump. */
327 #define TDF_SINTR 0x000008 /* Sleep is interruptible. */
328 #define TDF_TIMEOUT 0x000010 /* Timing out during sleep. */
329 #define TDF_SELECT 0x000040 /* Selecting; wakeup/waiting danger. */
330 #define TDF_CVWAITQ 0x000080 /* Thread is on a cv_waitq (not slpq). */
331 #define TDF_UPCALLING 0x000100 /* This thread is doing an upcall. */
332 #define TDF_ONSLEEPQ 0x000200 /* On the sleep queue. */
333 #define TDF_INMSLEEP 0x000400 /* Don't recurse in msleep(). */
334 #define TDF_TIMOFAIL 0x001000 /* Timeout from sleep after we were awake. */
335 #define TDF_DEADLKTREAT 0x800000 /* Lock aquisition - deadlock treatment. */
337 #define TDI_SUSPENDED 0x01 /* On suspension queue. */
338 #define TDI_SLEEPING 0x02 /* Actually asleep! (tricky). */
339 #define TDI_SWAPPED 0x04 /* Stack not in mem.. bad juju if run. */
340 #define TDI_LOCK 0x08 /* Stopped on a lock. */
341 #define TDI_IWAIT 0x10 /* Awaiting interrupt. */
342 #define TDI_LOAN 0x20 /* bound thread's KSE is lent */
344 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING)
345 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL)
346 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED)
347 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED)
348 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK)
349 #define TD_LENT(td) ((td)->td_inhibitors & TDI_LOAN)
350 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT)
351 #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING)
352 #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ)
353 #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN)
354 #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED)
356 #define TD_SET_INHIB(td, inhib) do { \
357 (td)->td_state = TDS_INHIBITED; \
358 (td)->td_inhibitors |= inhib; \
361 #define TD_CLR_INHIB(td, inhib) do { \
362 if (((td)->td_inhibitors & inhib) && \
363 (((td)->td_inhibitors &= ~inhib) == 0)) \
364 (td)->td_state = TDS_CAN_RUN; \
367 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING)
368 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED)
369 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK)
370 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED)
371 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT)
372 #define TD_SET_LOAN(td) TD_SET_INHIB((td), TDI_LOAN)
374 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING)
375 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED)
376 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK)
377 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED)
378 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT)
379 #define TD_CLR_LOAN(td) TD_CLR_INHIB((td), TDI_LOAN)
381 #define TD_SET_RUNNING(td) do {(td)->td_state = TDS_RUNNING; } while (0)
382 #define TD_SET_RUNQ(td) do {(td)->td_state = TDS_RUNQ; } while (0)
383 #define TD_SET_CAN_RUN(td) do {(td)->td_state = TDS_CAN_RUN; } while (0)
384 #define TD_SET_ON_SLEEPQ(td) do {(td)->td_flags |= TDF_ONSLEEPQ; } while (0)
385 #define TD_CLR_ON_SLEEPQ(td) do { \
386 (td)->td_flags &= ~TDF_ONSLEEPQ; \
387 (td)->td_wchan = NULL; \
392 * Traps for young players:
393 * The main thread variable that controls whether a thread acts as a threaded
394 * or unthreaded thread is the td_bound counter (0 == unbound).
395 * UPCALLS run with the UNBOUND flags clear, after they are first scheduled.
396 * i.e. they bind themselves to whatever thread thay are first scheduled with.
397 * You may see BOUND threads in KSE processes but you should never see
398 * UNBOUND threads in non KSE processes.
402 * The schedulable entity that can be given a context to run.
403 * A process may have several of these. Probably one per processor
404 * but posibly a few more. In this universe they are grouped
405 * with a KSEG that contains the priority and niceness
409 struct proc *ke_proc; /* Associated process. */
410 struct ksegrp *ke_ksegrp; /* Associated KSEG. */
411 TAILQ_ENTRY(kse) ke_kglist; /* Queue of all KSEs in ke_ksegrp. */
412 TAILQ_ENTRY(kse) ke_kgrlist; /* Queue of all KSEs in this state. */
413 TAILQ_ENTRY(kse) ke_procq; /* (j) Run queue. */
415 #define ke_startzero ke_flags
416 int ke_flags; /* (j) KEF_* flags. */
417 struct thread *ke_thread; /* Active associated thread. */
418 struct thread *ke_bound; /* Thread bound to this KSE (*) */
419 int ke_cpticks; /* (j) Ticks of cpu time. */
420 fixpt_t ke_pctcpu; /* (j) %cpu during p_swtime. */
421 u_int64_t ke_uu; /* (j) Previous user time in usec. */
422 u_int64_t ke_su; /* (j) Previous system time in usec. */
423 u_int64_t ke_iu; /* (j) Previous intr time in usec. */
424 u_int64_t ke_uticks; /* (j) Statclock hits in user mode. */
425 u_int64_t ke_sticks; /* (j) Statclock hits in system mode. */
426 u_int64_t ke_iticks; /* (j) Statclock hits in intr. */
427 u_char ke_oncpu; /* (j) Which cpu we are on. */
428 char ke_rqindex; /* (j) Run queue index. */
432 KES_UNQUEUED, /* in transit */
433 KES_THREAD /* slaved to thread state */
434 } ke_state; /* (j) S* process status. */
435 struct kse_mailbox *ke_mailbox; /* the userland mailbox address */
438 struct thread *ke_tdspare; /* spare thread for upcalls */
439 #define ke_endzero ke_dummy
443 /* flags kept in ke_flags */
444 #define KEF_OWEUPC 0x00002 /* Owe process an addupc() call at next ast. */
445 #define KEF_IDLEKSE 0x00004 /* A 'Per CPU idle process'.. has one thread */
446 #define KEF_LOANED 0x00008 /* On loan from the bound thread to another */
447 #define KEF_USER 0x00200 /* Process is not officially in the kernel */
448 #define KEF_ASTPENDING 0x00400 /* KSE has a pending ast. */
449 #define KEF_NEEDRESCHED 0x00800 /* Process needs to yield. */
450 #define KEF_ONLOANQ 0x01000 /* KSE is on loan queue. */
451 #define KEF_DIDRUN 0x02000 /* KSE actually ran. */
452 #define KEF_EXIT 0x04000 /* KSE is being killed. */
455 * (*) A bound KSE with a bound thread in a KSE process may be lent to
456 * Other threads, as long as those threads do not leave the kernel.
457 * The other threads must be either exiting, or be unbound with a valid
458 * mailbox so that they can save their state there rather than going
459 * to user space. While this happens the real bound thread is still linked
460 * to the kse via the ke_bound field, and the KSE has its "KEF_LOANED
465 * Kernel-scheduled entity group (KSEG). The scheduler considers each KSEG to
466 * be an indivisible unit from a time-sharing perspective, though each KSEG may
467 * contain multiple KSEs.
470 struct proc *kg_proc; /* Process that contains this KSEG. */
471 TAILQ_ENTRY(ksegrp) kg_ksegrp; /* Queue of KSEGs in kg_proc. */
472 TAILQ_HEAD(, kse) kg_kseq; /* (ke_kglist) All KSEs. */
473 TAILQ_HEAD(, kse) kg_iq; /* (ke_kgrlist) Idle KSEs. */
474 TAILQ_HEAD(, kse) kg_lq; /* (ke_kgrlist) Loan KSEs. */
475 TAILQ_HEAD(, thread) kg_threads;/* (td_kglist) All threads. */
476 TAILQ_HEAD(, thread) kg_runq; /* (td_runq) waiting RUNNABLE threads */
477 TAILQ_HEAD(, thread) kg_slpq; /* (td_runq) NONRUNNABLE threads. */
479 #define kg_startzero kg_estcpu
480 u_int kg_estcpu; /* Sum of the same field in KSEs. */
481 u_int kg_slptime; /* (j) How long completely blocked. */
482 struct thread *kg_last_assigned; /* Last thread assigned to a KSE */
483 int kg_runnable; /* Num runnable threads on queue. */
484 int kg_runq_kses; /* Num KSEs on runq. */
485 int kg_loan_kses; /* Num KSEs on loan queue. */
486 struct kse_thr_mailbox *kg_completed; /* (c) completed thread mboxes */
487 #define kg_endzero kg_pri_class
489 #define kg_startcopy kg_endzero
490 u_char kg_pri_class; /* (j) Scheduling class. */
491 u_char kg_user_pri; /* (j) User pri from estcpu and nice. */
492 char kg_nice; /* (j?/k?) Process "nice" value. */
493 #define kg_endcopy kg_numthreads
494 int kg_numthreads; /* Num threads in total */
495 int kg_idle_kses; /* num KSEs idle */
496 int kg_kses; /* Num KSEs in group. */
500 * The old fashionned process. May have multiple threads, KSEGRPs
501 * and KSEs. Starts off with a single embedded KSEGRP, KSE and THREAD.
504 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */
505 TAILQ_HEAD(, ksegrp) p_ksegrps; /* (kg_ksegrp) All KSEGs. */
506 TAILQ_HEAD(, thread) p_threads; /* (td_plist) Threads. (shortcut) */
507 TAILQ_HEAD(, thread) p_suspended; /* (td_runq) suspended threads */
508 struct ucred *p_ucred; /* (c) Process owner's identity. */
509 struct filedesc *p_fd; /* (b) Ptr to open files structure. */
510 /* Accumulated stats for all KSEs? */
511 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */
512 struct plimit *p_limit; /* (m) Process limits. */
513 struct vm_object *p_upages_obj; /* (a) Upages object. */
514 struct procsig *p_procsig; /* (c) Signal actions, state (CPU). */
516 /*struct ksegrp p_ksegrp;
520 * The following don't make too much sense..
521 * See the td_ or ke_ versions of the same flags
523 int p_flag; /* (c) P_* flags. */
524 int p_sflag; /* (j) PS_* flags. */
526 PRS_NEW = 0, /* In creation */
527 PRS_NORMAL, /* KSEs can be run */
528 PRS_WAIT, /* Waiting on interrupt ? */
530 } p_state; /* (j) S* process status. */
531 pid_t p_pid; /* (b) Process identifier. */
532 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */
533 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */
534 struct proc *p_pptr; /* (c + e) Pointer to parent process. */
535 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */
536 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */
537 struct mtx p_mtx; /* (k) Lock for this struct. */
539 /* The following fields are all zeroed upon creation in fork. */
540 #define p_startzero p_oppid
541 pid_t p_oppid; /* (c + e) Save ppid in ptrace. XXX */
542 struct vmspace *p_vmspace; /* (b) Address space. */
543 u_int p_swtime; /* (j) Time swapped in or out. */
544 struct itimerval p_realtimer; /* (h?/k?) Alarm timer. */
545 struct bintime p_runtime; /* (j) Real time. */
546 int p_traceflag; /* (o) Kernel trace points. */
547 struct vnode *p_tracep; /* (c + o) Trace to vnode. */
548 sigset_t p_siglist; /* (c) Sigs arrived, not delivered. */
549 struct vnode *p_textvp; /* (b) Vnode of executable. */
550 char p_lock; /* (c) Proclock (prevent swap) count. */
551 struct klist p_klist; /* (c) Knotes attached to this proc. */
552 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */
553 int p_sigparent; /* (c) Signal to parent on exit. */
554 sigset_t p_oldsigmask; /* (c) Saved mask from pre sigpause. */
555 int p_sig; /* (n) For core dump/debugger XXX. */
556 u_long p_code; /* (n) For core dump/debugger XXX. */
557 u_int p_stops; /* (c) Stop event bitmask. */
558 u_int p_stype; /* (c) Stop event type. */
559 char p_step; /* (c) Process is stopped. */
560 u_char p_pfsflags; /* (c) Procfs flags. */
561 struct nlminfo *p_nlminfo; /* (?) Only used by/for lockd. */
562 void *p_aioinfo; /* (c) ASYNC I/O info. */
563 struct thread *p_singlethread;/* (j) If single threading this is it */
564 int p_suspcount; /* (j) # threads in suspended mode */
565 int p_userthreads; /* (j) # threads in userland */
566 /* End area that is zeroed on creation. */
567 #define p_endzero p_sigmask
569 /* The following fields are all copied upon creation in fork. */
570 #define p_startcopy p_endzero
571 sigset_t p_sigmask; /* (c) Current signal mask. */
572 stack_t p_sigstk; /* (c) Stack ptr and on-stack flag. */
573 u_int p_magic; /* (b) Magic number. */
574 char p_comm[MAXCOMLEN + 1]; /* (b) Process name. */
575 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */
576 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */
577 struct pargs *p_args; /* (c) Process arguments. */
578 rlim_t p_cpulimit; /* (j) Current CPU limit in seconds. */
579 /* End area that is copied on creation. */
580 #define p_endcopy p_xstat
582 u_short p_xstat; /* (c) Exit status; also stop sig. */
583 int p_numthreads; /* (?) number of threads */
584 int p_numksegrps; /* (?) number of ksegrps */
585 struct mdproc p_md; /* (c) Any machine-dependent fields. */
586 struct callout p_itcallout; /* (h) Interval timer callout. */
587 struct user *p_uarea; /* (k) Kernel VA of u-area (CPU) */
588 u_short p_acflag; /* (c) Accounting flags. */
589 struct rusage *p_ru; /* (a) Exit information. XXX */
590 struct proc *p_peers; /* (r) */
591 struct proc *p_leader; /* (b) */
592 void *p_emuldata; /* (c) Emulator state data. */
595 #define p_rlimit p_limit->pl_rlimit
596 #define p_sigacts p_procsig->ps_sigacts
597 #define p_sigignore p_procsig->ps_sigignore
598 #define p_sigcatch p_procsig->ps_sigcatch
599 #define p_session p_pgrp->pg_session
600 #define p_pgid p_pgrp->pg_id
602 #define NOCPU 0xff /* For when we aren't on a CPU. (SMP) */
604 /* Status values (p_stat). */
606 /* These flags are kept in p_flag. */
607 #define P_ADVLOCK 0x00001 /* Process may hold a POSIX advisory lock. */
608 #define P_CONTROLT 0x00002 /* Has a controlling terminal. */
609 #define P_KTHREAD 0x00004 /* Kernel thread. (*)*/
610 #define P_NOLOAD 0x00008 /* Ignore during load avg calculations. */
611 #define P_PPWAIT 0x00010 /* Parent is waiting for child to exec/exit. */
612 #define P_SUGID 0x00100 /* Had set id privileges since last exec. */
613 #define P_SYSTEM 0x00200 /* System proc: no sigs, stats or swapping. */
614 #define P_WAITED 0x01000 /* Someone is waiting for us */
615 #define P_WEXIT 0x02000 /* Working on exiting. */
616 #define P_EXEC 0x04000 /* Process called exec. */
617 #define P_KSES 0x08000 /* Process is using KSEs. */
618 #define P_CONTINUED 0x10000 /* Proc has continued from a stopped state. */
620 /* flags that control how threads may be suspended for some reason */
621 #define P_STOPPED_SIG 0x20000 /* Stopped due to SIGSTOP/SIGTSTP */
622 #define P_STOPPED_TRACE 0x40000 /* Stopped because of tracing */
623 #define P_STOPPED_SINGLE 0x80000 /* Only one thread can continue */
625 #define P_SINGLE_EXIT 0x00400 /* Threads suspending should exit, */
627 #define P_TRACED 0x00800 /* Debugged process being traced. */
628 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
629 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED)
631 /* Should be moved to machine-dependent areas. */
632 #define P_UNUSED100000 0x100000
633 #define P_COWINPROGRESS 0x400000 /* Snapshot copy-on-write in progress. */
635 #define P_JAILED 0x1000000 /* Process is in jail. */
636 #define P_OLDMASK 0x2000000 /* Need to restore mask after suspend. */
637 #define P_ALTSTACK 0x4000000 /* Have alternate signal stack. */
638 #define P_INEXEC 0x8000000 /* Process is in execve(). */
640 /* These flags are kept in p_sflag and are protected with sched_lock. */
641 #define PS_INMEM 0x00001 /* Loaded into memory. */
642 #define PS_XCPU 0x00002 /* Exceeded CPU limit. */
643 #define PS_PROFIL 0x00004 /* Has started profiling. */
644 #define PS_ALRMPEND 0x00020 /* Pending SIGVTALRM needs to be posted. */
645 #define PS_PROFPEND 0x00040 /* Pending SIGPROF needs to be posted. */
646 #define PS_SWAPINREQ 0x00100 /* Swapin request due to wakeup. */
647 #define PS_SWAPPING 0x00200 /* Process is being swapped. */
648 #define PS_NEEDSIGCHK 0x02000 /* Process may need signal delivery. */
649 #define PS_SWAPPINGIN 0x04000 /* Swapin in progress. */
651 /* used only in legacy conversion code */
652 #define SIDL 1 /* Process being created by fork. */
653 #define SRUN 2 /* Currently runnable. */
654 #define SSLEEP 3 /* Sleeping on an address. */
655 #define SSTOP 4 /* Process debugging or suspension. */
656 #define SZOMB 5 /* Awaiting collection by parent. */
657 #define SWAIT 6 /* Waiting for interrupt. */
658 #define SLOCK 7 /* Blocked on a lock. */
660 #define P_MAGIC 0xbeefface
664 #ifdef MALLOC_DECLARE
665 MALLOC_DECLARE(M_PARGS);
666 MALLOC_DECLARE(M_PGRP);
667 MALLOC_DECLARE(M_SESSION);
668 MALLOC_DECLARE(M_SUBPROC);
669 MALLOC_DECLARE(M_ZOMBIE);
672 #define FOREACH_PROC_IN_SYSTEM(p) \
673 LIST_FOREACH((p), &allproc, p_list)
674 #define FOREACH_KSEGRP_IN_PROC(p, kg) \
675 TAILQ_FOREACH((kg), &(p)->p_ksegrps, kg_ksegrp)
676 #define FOREACH_THREAD_IN_GROUP(kg, td) \
677 TAILQ_FOREACH((td), &(kg)->kg_threads, td_kglist)
678 #define FOREACH_KSE_IN_GROUP(kg, ke) \
679 TAILQ_FOREACH((ke), &(kg)->kg_kseq, ke_kglist)
680 #define FOREACH_THREAD_IN_PROC(p, td) \
681 TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
683 /* XXXKSE the lines below should probably only be used in 1:1 code */
684 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&p->p_threads)
685 #define FIRST_KSEGRP_IN_PROC(p) TAILQ_FIRST(&p->p_ksegrps)
686 #define FIRST_KSE_IN_KSEGRP(kg) TAILQ_FIRST(&kg->kg_kseq)
687 #define FIRST_KSE_IN_PROC(p) FIRST_KSE_IN_KSEGRP(FIRST_KSEGRP_IN_PROC(p))
690 sigonstack(size_t sp)
692 register struct thread *td = curthread;
693 struct proc *p = td->td_proc;
695 return ((p->p_flag & P_ALTSTACK) ?
696 #if defined(COMPAT_43) || defined(COMPAT_SUNOS)
697 ((p->p_sigstk.ss_size == 0) ? (p->p_sigstk.ss_flags & SS_ONSTACK) :
698 ((sp - (size_t)p->p_sigstk.ss_sp) < p->p_sigstk.ss_size))
700 ((sp - (size_t)p->p_sigstk.ss_sp) < p->p_sigstk.ss_size)
705 /* Handy macro to determine if p1 can mangle p2. */
706 #define PRISON_CHECK(p1, p2) \
707 ((p1)->p_prison == NULL || (p1)->p_prison == (p2)->p_prison)
710 * We use process IDs <= PID_MAX; PID_MAX + 1 must also fit in a pid_t,
711 * as it is used to represent "no process group".
713 #define PID_MAX 99999
714 #define NO_PID 100000
716 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p))
717 #define SESSHOLD(s) ((s)->s_count++)
718 #define SESSRELE(s) { \
719 if (--(s)->s_count == 0) \
720 FREE(s, M_SESSION); \
723 #define STOPEVENT(p, e, v) do { \
725 _STOPEVENT((p), (e), (v)); \
728 #define _STOPEVENT(p, e, v) do { \
729 PROC_LOCK_ASSERT(p, MA_OWNED); \
730 if ((p)->p_stops & (e)) { \
731 stopevent((p), (e), (v)); \
735 /* Lock and unlock a process. */
736 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx)
737 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx)
738 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx)
739 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx)
740 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type))
742 /* Lock and unlock a process group. */
743 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx)
744 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx)
745 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx)
746 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type))
748 #define PGRP_LOCK_PGSIGNAL(pg) \
754 #define PGRP_UNLOCK_PGSIGNAL(pg) \
760 /* Lock and unlock a session. */
761 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx)
762 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx)
763 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx)
764 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type))
766 /* Hold process U-area in memory, normally for ptrace/procfs work. */
767 #define PHOLD(p) do { \
772 #define _PHOLD(p) do { \
773 PROC_LOCK_ASSERT((p), MA_OWNED); \
774 if ((p)->p_lock++ == 0) { \
775 mtx_lock_spin(&sched_lock); \
777 mtx_unlock_spin(&sched_lock); \
781 #define PRELE(p) do { \
786 #define _PRELE(p) do { \
787 PROC_LOCK_ASSERT((p), MA_OWNED); \
791 /* Check whether a thread is safe to be swapped out. */
792 #define thread_safetoswapout(td) (TD_IS_SLEEPING(td) || TD_IS_SUSPENDED(td))
794 /* Lock and unlock process arguments. */
795 #define PARGS_LOCK(p) mtx_lock(&pargs_ref_lock)
796 #define PARGS_UNLOCK(p) mtx_unlock(&pargs_ref_lock)
798 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash])
799 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
800 extern u_long pidhash;
802 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash])
803 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
804 extern u_long pgrphash;
806 extern struct sx allproc_lock;
807 extern struct sx proctree_lock;
808 extern struct mtx pargs_ref_lock;
809 extern struct mtx ppeers_lock;
810 extern struct proc proc0; /* Process slot for swapper. */
811 extern struct thread thread0; /* Primary thread in proc0 */
812 extern struct ksegrp ksegrp0; /* Primary ksegrp in proc0 */
813 extern struct kse kse0; /* Primary kse in proc0 */
814 extern int hogticks; /* Limit on kernel cpu hogs. */
815 extern int nprocs, maxproc; /* Current and max number of procs. */
816 extern int maxprocperuid; /* Max procs per uid. */
817 extern u_long ps_arg_cache_limit;
818 extern int ps_argsopen;
819 extern int ps_showallprocs;
820 extern int sched_quantum; /* Scheduling quantum in ticks. */
822 LIST_HEAD(proclist, proc);
823 TAILQ_HEAD(procqueue, proc);
824 TAILQ_HEAD(threadqueue, thread);
825 extern struct proclist allproc; /* List of all processes. */
826 extern struct proclist zombproc; /* List of zombie processes. */
827 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
828 extern struct proc *updateproc; /* Process slot for syncer (sic). */
830 extern struct uma_zone *proc_zone;
835 * XXX macros for scheduler. Shouldn't be here, but currently needed for
836 * bounding the dubious p_estcpu inheritance in wait1().
837 * INVERSE_ESTCPU_WEIGHT is only suitable for statclock() frequencies in
838 * the range 100-256 Hz (approximately).
840 #define ESTCPULIM(e) \
841 min((e), INVERSE_ESTCPU_WEIGHT * (NICE_WEIGHT * (PRIO_MAX - PRIO_MIN) - \
842 RQ_PPQ) + INVERSE_ESTCPU_WEIGHT - 1)
843 #define INVERSE_ESTCPU_WEIGHT 8 /* 1 / (priorities per estcpu level). */
844 #define NICE_WEIGHT 1 /* Priorities per nice level. */
846 struct proc *pfind(pid_t); /* Find process by id. */
847 struct pgrp *pgfind(pid_t); /* Find process group by id. */
848 struct proc *zpfind(pid_t); /* Find zombie process by id. */
850 void adjustrunqueue(struct thread *, int newpri);
851 void ast(struct trapframe *framep);
852 struct thread *choosethread(void);
853 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
854 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp, struct session *sess);
855 int enterthispgrp(struct proc *p, struct pgrp *pgrp);
856 void faultin(struct proc *p);
857 void fixjobc(struct proc *p, struct pgrp *pgrp, int entering);
858 int fork1(struct thread *, int, int, struct proc **);
859 void fork_exit(void (*)(void *, struct trapframe *), void *,
861 void fork_return(struct thread *, struct trapframe *);
862 int inferior(struct proc *p);
863 int leavepgrp(struct proc *p);
864 void mi_switch(void);
865 int p_candebug(struct thread *td, struct proc *p);
866 int p_cansee(struct thread *td, struct proc *p);
867 int p_cansched(struct thread *td, struct proc *p);
868 int p_cansignal(struct thread *td, struct proc *p, int signum);
869 struct pargs *pargs_alloc(int len);
870 void pargs_drop(struct pargs *pa);
871 void pargs_free(struct pargs *pa);
872 void pargs_hold(struct pargs *pa);
874 void threadinit(void);
875 void proc_linkup(struct proc *p, struct ksegrp *kg,
876 struct kse *ke, struct thread *td);
877 void proc_reparent(struct proc *child, struct proc *newparent);
878 int securelevel_ge(struct ucred *cr, int level);
879 int securelevel_gt(struct ucred *cr, int level);
880 void setrunnable(struct thread *);
881 void setrunqueue(struct thread *);
882 void setsugid(struct proc *p);
883 void sleepinit(void);
884 void stopevent(struct proc *, u_int, u_int);
886 void cpu_switch(void);
887 void cpu_throw(void) __dead2;
888 void unsleep(struct thread *);
889 void userret(struct thread *, struct trapframe *, u_int);
891 void cpu_exit(struct thread *);
892 void cpu_sched_exit(struct thread *);
893 void exit1(struct thread *, int) __dead2;
894 void cpu_fork(struct thread *, struct proc *, struct thread *, int);
895 void cpu_set_fork_handler(struct thread *, void (*)(void *), void *);
896 void cpu_wait(struct proc *);
899 struct ksegrp *ksegrp_alloc(void);
900 void ksegrp_free(struct ksegrp *td);
901 struct kse *kse_alloc(void);
902 void kse_free(struct kse *td);
903 struct thread *thread_alloc(void);
904 void thread_free(struct thread *td);
905 void cpu_set_upcall(struct thread *td, void *pcb);
906 void cpu_set_upcall_kse(struct thread *td, struct kse *ke);
907 void cpu_thread_exit(struct thread *);
908 void cpu_thread_setup(struct thread *td);
909 void kse_reassign(struct kse *ke);
910 void kse_link(struct kse *ke, struct ksegrp *kg);
911 void ksegrp_link(struct ksegrp *kg, struct proc *p);
912 void make_kse_runnable(struct kse *ke);
913 struct thread *signal_upcall(struct proc *p, int sig);
914 void thread_exit(void) __dead2;
915 int thread_export_context(struct thread *td);
916 void thread_link(struct thread *td, struct ksegrp *kg);
917 void thread_reap(void);
918 struct thread *thread_schedule_upcall(struct thread *td, struct kse *ke);
919 int thread_setcontext(struct thread *td, ucontext_t *uc);
920 void thread_getcontext(struct thread *td, ucontext_t *uc);
921 int thread_single(int how);
922 #define SINGLE_NO_EXIT 0 /* values for 'how' */
923 #define SINGLE_EXIT 1
924 void thread_single_end(void);
925 void thread_stash(struct thread *td);
926 int thread_suspend_check(int how);
927 void thread_unsuspend(struct proc *p);
928 void thread_suspend_one(struct thread *td);
929 void thread_unsuspend_one(struct thread *td);
930 int thread_userret(struct thread *td, struct trapframe *frame);
932 void thread_sanity_check(struct thread *td, char *);
935 #endif /* !_SYS_PROC_H_ */