xref: /freebsd/sys/sys/proc.h (revision acc1a9ef)
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  * 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.
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  *      l - the attaching proc or attaching proc parent
146  *      m - Giant
147  *      n - not locked, lazy
148  *      o - ktrace lock
149  *      q - td_contested lock
150  *      r - p_peers lock
151  *      t - thread lock
152  *	u - process stat lock
153  *	w - process timer lock
154  *      x - created at fork, only changes during single threading in exec
155  *      y - created at first aio, doesn't change until exit or exec at which
156  *          point we are single-threaded and only curthread changes it
157  *      z - zombie threads lock
158  *
159  * If the locking key specifies two identifiers (for example, p_pptr) then
160  * either lock is sufficient for read access, but both locks must be held
161  * for write access.
162  */
163 struct cpuset;
164 struct filecaps;
165 struct kaioinfo;
166 struct kaudit_record;
167 struct kdtrace_proc;
168 struct kdtrace_thread;
169 struct mqueue_notifier;
170 struct nlminfo;
171 struct p_sched;
172 struct proc;
173 struct procdesc;
174 struct racct;
175 struct sbuf;
176 struct sleepqueue;
177 struct syscall_args;
178 struct td_sched;
179 struct thread;
180 struct trapframe;
181 struct turnstile;
182 
183 /*
184  * XXX: Does this belong in resource.h or resourcevar.h instead?
185  * Resource usage extension.  The times in rusage structs in the kernel are
186  * never up to date.  The actual times are kept as runtimes and tick counts
187  * (with control info in the "previous" times), and are converted when
188  * userland asks for rusage info.  Backwards compatibility prevents putting
189  * this directly in the user-visible rusage struct.
190  *
191  * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux.
192  * Locking for td_rux: (t) for all fields.
193  */
194 struct rusage_ext {
195 	uint64_t	rux_runtime;    /* (cu) Real time. */
196 	uint64_t	rux_uticks;     /* (cu) Statclock hits in user mode. */
197 	uint64_t	rux_sticks;     /* (cu) Statclock hits in sys mode. */
198 	uint64_t	rux_iticks;     /* (cu) Statclock hits in intr mode. */
199 	uint64_t	rux_uu;         /* (c) Previous user time in usec. */
200 	uint64_t	rux_su;         /* (c) Previous sys time in usec. */
201 	uint64_t	rux_tu;         /* (c) Previous total time in usec. */
202 };
203 
204 /*
205  * Kernel runnable context (thread).
206  * This is what is put to sleep and reactivated.
207  * Thread context.  Processes may have multiple threads.
208  */
209 struct thread {
210 	struct mtx	*volatile td_lock; /* replaces sched lock */
211 	struct proc	*td_proc;	/* (*) Associated process. */
212 	TAILQ_ENTRY(thread) td_plist;	/* (*) All threads in this proc. */
213 	TAILQ_ENTRY(thread) td_runq;	/* (t) Run queue. */
214 	TAILQ_ENTRY(thread) td_slpq;	/* (t) Sleep queue. */
215 	TAILQ_ENTRY(thread) td_lockq;	/* (t) Lock queue. */
216 	LIST_ENTRY(thread) td_hash;	/* (d) Hash chain. */
217 	struct cpuset	*td_cpuset;	/* (t) CPU affinity mask. */
218 	struct seltd	*td_sel;	/* Select queue/channel. */
219 	struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */
220 	struct turnstile *td_turnstile;	/* (k) Associated turnstile. */
221 	struct rl_q_entry *td_rlqe;	/* (k) Associated range lock entry. */
222 	struct umtx_q   *td_umtxq;	/* (c?) Link for when we're blocked. */
223 	struct vm_domain_policy td_vm_dom_policy;	/* (c) current numa domain policy */
224 	lwpid_t		td_tid;		/* (b) Thread ID. */
225 	sigqueue_t	td_sigqueue;	/* (c) Sigs arrived, not delivered. */
226 #define	td_siglist	td_sigqueue.sq_signals
227 	u_char		td_lend_user_pri; /* (t) Lend user pri. */
228 
229 /* Cleared during fork1() */
230 #define	td_startzero td_flags
231 	int		td_flags;	/* (t) TDF_* flags. */
232 	int		td_inhibitors;	/* (t) Why can not run. */
233 	int		td_pflags;	/* (k) Private thread (TDP_*) flags. */
234 	int		td_dupfd;	/* (k) Ret value from fdopen. XXX */
235 	int		td_sqqueue;	/* (t) Sleepqueue queue blocked on. */
236 	void		*td_wchan;	/* (t) Sleep address. */
237 	const char	*td_wmesg;	/* (t) Reason for sleep. */
238 	volatile u_char td_owepreempt;  /* (k*) Preempt on last critical_exit */
239 	u_char		td_tsqueue;	/* (t) Turnstile queue blocked on. */
240 	short		td_locks;	/* (k) Debug: count of non-spin locks */
241 	short		td_rw_rlocks;	/* (k) Count of rwlock read locks. */
242 	short		td_lk_slocks;	/* (k) Count of lockmgr shared locks. */
243 	short		td_stopsched;	/* (k) Scheduler stopped. */
244 	struct turnstile *td_blocked;	/* (t) Lock thread is blocked on. */
245 	const char	*td_lockname;	/* (t) Name of lock blocked on. */
246 	LIST_HEAD(, turnstile) td_contested;	/* (q) Contested locks. */
247 	struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
248 	int		td_intr_nesting_level; /* (k) Interrupt recursion. */
249 	int		td_pinned;	/* (k) Temporary cpu pin count. */
250 	struct ucred	*td_ucred;	/* (k) Reference to credentials. */
251 	struct plimit	*td_limit;	/* (k) Resource limits. */
252 	u_int		td_estcpu;	/* (t) estimated cpu utilization */
253 	int		td_slptick;	/* (t) Time at sleep. */
254 	int		td_blktick;	/* (t) Time spent blocked. */
255 	int		td_swvoltick;	/* (t) Time at last SW_VOL switch. */
256 	u_int		td_cow;		/* (*) Number of copy-on-write faults */
257 	struct rusage	td_ru;		/* (t) rusage information. */
258 	struct rusage_ext td_rux;	/* (t) Internal rusage information. */
259 	uint64_t	td_incruntime;	/* (t) Cpu ticks to transfer to proc. */
260 	uint64_t	td_runtime;	/* (t) How many cpu ticks we've run. */
261 	u_int 		td_pticks;	/* (t) Statclock hits for profiling */
262 	u_int		td_sticks;	/* (t) Statclock hits in system mode. */
263 	u_int		td_iticks;	/* (t) Statclock hits in intr mode. */
264 	u_int		td_uticks;	/* (t) Statclock hits in user mode. */
265 	int		td_intrval;	/* (t) Return value for sleepq. */
266 	sigset_t	td_oldsigmask;	/* (k) Saved mask from pre sigpause. */
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_xsig;	/* (c) Signal for ptrace */
270 	u_long		td_profil_addr;	/* (k) Temporary addr until AST. */
271 	u_int		td_profil_ticks; /* (k) Temporary ticks until AST. */
272 	char		td_name[MAXCOMLEN + 1];	/* (*) Thread name. */
273 	struct file	*td_fpop;	/* (k) file referencing cdev under op */
274 	int		td_dbgflags;	/* (c) Userland debugger flags */
275 	struct ksiginfo td_dbgksi;	/* (c) ksi reflected to debugger. */
276 	int		td_ng_outbound;	/* (k) Thread entered ng from above. */
277 	struct osd	td_osd;		/* (k) Object specific data. */
278 	struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */
279 	pid_t		td_dbg_forked;	/* (c) Child pid for debugger. */
280 	u_int		td_vp_reserv;	/* (k) Count of reserved vnodes. */
281 	int		td_no_sleeping;	/* (k) Sleeping disabled count. */
282 	int		td_dom_rr_idx;	/* (k) RR Numa domain selection. */
283 	void		*td_su;		/* (k) FFS SU private */
284 #define	td_endzero td_sigmask
285 
286 /* Copied during fork1() or create_thread(). */
287 #define	td_startcopy td_endzero
288 	sigset_t	td_sigmask;	/* (c) Current signal mask. */
289 	u_char		td_rqindex;	/* (t) Run queue index. */
290 	u_char		td_base_pri;	/* (t) Thread base kernel priority. */
291 	u_char		td_priority;	/* (t) Thread active priority. */
292 	u_char		td_pri_class;	/* (t) Scheduling class. */
293 	u_char		td_user_pri;	/* (t) User pri from estcpu and nice. */
294 	u_char		td_base_user_pri; /* (t) Base user pri */
295 	u_int		td_dbg_sc_code;	/* (c) Syscall code to debugger. */
296 	u_int		td_dbg_sc_narg;	/* (c) Syscall arg count to debugger.*/
297 #define	td_endcopy td_pcb
298 
299 /*
300  * Fields that must be manually set in fork1() or create_thread()
301  * or already have been set in the allocator, constructor, etc.
302  */
303 	struct pcb	*td_pcb;	/* (k) Kernel VA of pcb and kstack. */
304 	enum {
305 		TDS_INACTIVE = 0x0,
306 		TDS_INHIBITED,
307 		TDS_CAN_RUN,
308 		TDS_RUNQ,
309 		TDS_RUNNING
310 	} td_state;			/* (t) thread state */
311 	union {
312 		register_t	tdu_retval[2];
313 		off_t		tdu_off;
314 	} td_uretoff;			/* (k) Syscall aux returns. */
315 #define td_retval	td_uretoff.tdu_retval
316 	u_int		td_cowgen;	/* (k) Generation of COW pointers. */
317 	struct callout	td_slpcallout;	/* (h) Callout for sleep. */
318 	struct trapframe *td_frame;	/* (k) */
319 	struct vm_object *td_kstack_obj;/* (a) Kstack object. */
320 	vm_offset_t	td_kstack;	/* (a) Kernel VA of kstack. */
321 	int		td_kstack_pages; /* (a) Size of the kstack. */
322 	volatile u_int	td_critnest;	/* (k*) Critical section nest level. */
323 	struct mdthread td_md;		/* (k) Any machine-dependent fields. */
324 	struct td_sched	*td_sched;	/* (*) Scheduler-specific data. */
325 	struct kaudit_record	*td_ar;	/* (k) Active audit record, if any. */
326 	struct lpohead	td_lprof[2];	/* (a) lock profiling objects. */
327 	struct kdtrace_thread	*td_dtrace; /* (*) DTrace-specific data. */
328 	int		td_errno;	/* Error returned by last syscall. */
329 	struct vnet	*td_vnet;	/* (k) Effective vnet. */
330 	const char	*td_vnet_lpush;	/* (k) Debugging vnet push / pop. */
331 	struct trapframe *td_intr_frame;/* (k) Frame of the current irq */
332 	struct proc	*td_rfppwait_p;	/* (k) The vforked child */
333 	struct vm_page	**td_ma;	/* (k) uio pages held */
334 	int		td_ma_cnt;	/* (k) size of *td_ma */
335 	void		*td_emuldata;	/* Emulator state data */
336 	int		td_lastcpu;	/* (t) Last cpu we were on. */
337 	int		td_oncpu;	/* (t) Which cpu we are on. */
338 };
339 
340 struct mtx *thread_lock_block(struct thread *);
341 void thread_lock_unblock(struct thread *, struct mtx *);
342 void thread_lock_set(struct thread *, struct mtx *);
343 #define	THREAD_LOCK_ASSERT(td, type)					\
344 do {									\
345 	struct mtx *__m = (td)->td_lock;				\
346 	if (__m != &blocked_lock)					\
347 		mtx_assert(__m, (type));				\
348 } while (0)
349 
350 #ifdef INVARIANTS
351 #define	THREAD_LOCKPTR_ASSERT(td, lock)					\
352 do {									\
353 	struct mtx *__m = (td)->td_lock;				\
354 	KASSERT((__m == &blocked_lock || __m == (lock)),		\
355 	    ("Thread %p lock %p does not match %p", td, __m, (lock)));	\
356 } while (0)
357 
358 #define	TD_LOCKS_INC(td)	((td)->td_locks++)
359 #define	TD_LOCKS_DEC(td)	((td)->td_locks--)
360 #else
361 #define	THREAD_LOCKPTR_ASSERT(td, lock)
362 
363 #define	TD_LOCKS_INC(td)
364 #define	TD_LOCKS_DEC(td)
365 #endif
366 
367 /*
368  * Flags kept in td_flags:
369  * To change these you MUST have the scheduler lock.
370  */
371 #define	TDF_BORROWING	0x00000001 /* Thread is borrowing pri from another. */
372 #define	TDF_INPANIC	0x00000002 /* Caused a panic, let it drive crashdump. */
373 #define	TDF_INMEM	0x00000004 /* Thread's stack is in memory. */
374 #define	TDF_SINTR	0x00000008 /* Sleep is interruptible. */
375 #define	TDF_TIMEOUT	0x00000010 /* Timing out during sleep. */
376 #define	TDF_IDLETD	0x00000020 /* This is a per-CPU idle thread. */
377 #define	TDF_CANSWAP	0x00000040 /* Thread can be swapped. */
378 #define	TDF_SLEEPABORT	0x00000080 /* sleepq_abort was called. */
379 #define	TDF_KTH_SUSP	0x00000100 /* kthread is suspended */
380 #define	TDF_ALLPROCSUSP	0x00000200 /* suspended by SINGLE_ALLPROC */
381 #define	TDF_BOUNDARY	0x00000400 /* Thread suspended at user boundary */
382 #define	TDF_ASTPENDING	0x00000800 /* Thread has some asynchronous events. */
383 #define	TDF_TIMOFAIL	0x00001000 /* Timeout from sleep after we were awake. */
384 #define	TDF_SBDRY	0x00002000 /* Stop only on usermode boundary. */
385 #define	TDF_UPIBLOCKED	0x00004000 /* Thread blocked on user PI mutex. */
386 #define	TDF_NEEDSUSPCHK	0x00008000 /* Thread may need to suspend. */
387 #define	TDF_NEEDRESCHED	0x00010000 /* Thread needs to yield. */
388 #define	TDF_NEEDSIGCHK	0x00020000 /* Thread may need signal delivery. */
389 #define	TDF_NOLOAD	0x00040000 /* Ignore during load avg calculations. */
390 #define	TDF_UNUSED19	0x00080000 /* --available-- */
391 #define	TDF_THRWAKEUP	0x00100000 /* Libthr thread must not suspend itself. */
392 #define	TDF_UNUSED21	0x00200000 /* --available-- */
393 #define	TDF_SWAPINREQ	0x00400000 /* Swapin request due to wakeup. */
394 #define	TDF_UNUSED23	0x00800000 /* --available-- */
395 #define	TDF_SCHED0	0x01000000 /* Reserved for scheduler private use */
396 #define	TDF_SCHED1	0x02000000 /* Reserved for scheduler private use */
397 #define	TDF_SCHED2	0x04000000 /* Reserved for scheduler private use */
398 #define	TDF_SCHED3	0x08000000 /* Reserved for scheduler private use */
399 #define	TDF_ALRMPEND	0x10000000 /* Pending SIGVTALRM needs to be posted. */
400 #define	TDF_PROFPEND	0x20000000 /* Pending SIGPROF needs to be posted. */
401 #define	TDF_MACPEND	0x40000000 /* AST-based MAC event pending. */
402 
403 /* Userland debug flags */
404 #define	TDB_SUSPEND	0x00000001 /* Thread is suspended by debugger */
405 #define	TDB_XSIG	0x00000002 /* Thread is exchanging signal under trace */
406 #define	TDB_USERWR	0x00000004 /* Debugger modified memory or registers */
407 #define	TDB_SCE		0x00000008 /* Thread performs syscall enter */
408 #define	TDB_SCX		0x00000010 /* Thread performs syscall exit */
409 #define	TDB_EXEC	0x00000020 /* TDB_SCX from exec(2) family */
410 #define	TDB_FORK	0x00000040 /* TDB_SCX from fork(2) that created new
411 				      process */
412 #define	TDB_STOPATFORK	0x00000080 /* Stop at the return from fork (child
413 				      only) */
414 #define	TDB_CHILD	0x00000100 /* New child indicator for ptrace() */
415 #define	TDB_BORN	0x00000200 /* New LWP indicator for ptrace() */
416 #define	TDB_EXIT	0x00000400 /* Exiting LWP indicator for ptrace() */
417 
418 /*
419  * "Private" flags kept in td_pflags:
420  * These are only written by curthread and thus need no locking.
421  */
422 #define	TDP_OLDMASK	0x00000001 /* Need to restore mask after suspend. */
423 #define	TDP_INKTR	0x00000002 /* Thread is currently in KTR code. */
424 #define	TDP_INKTRACE	0x00000004 /* Thread is currently in KTRACE code. */
425 #define	TDP_BUFNEED	0x00000008 /* Do not recurse into the buf flush */
426 #define	TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */
427 #define	TDP_ALTSTACK	0x00000020 /* Have alternate signal stack. */
428 #define	TDP_DEADLKTREAT	0x00000040 /* Lock aquisition - deadlock treatment. */
429 #define	TDP_NOFAULTING	0x00000080 /* Do not handle page faults. */
430 #define	TDP_UNUSED9	0x00000100 /* --available-- */
431 #define	TDP_OWEUPC	0x00000200 /* Call addupc() at next AST. */
432 #define	TDP_ITHREAD	0x00000400 /* Thread is an interrupt thread. */
433 #define	TDP_SYNCIO	0x00000800 /* Local override, disable async i/o. */
434 #define	TDP_SCHED1	0x00001000 /* Reserved for scheduler private use */
435 #define	TDP_SCHED2	0x00002000 /* Reserved for scheduler private use */
436 #define	TDP_SCHED3	0x00004000 /* Reserved for scheduler private use */
437 #define	TDP_SCHED4	0x00008000 /* Reserved for scheduler private use */
438 #define	TDP_GEOM	0x00010000 /* Settle GEOM before finishing syscall */
439 #define	TDP_SOFTDEP	0x00020000 /* Stuck processing softdep worklist */
440 #define	TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */
441 #define	TDP_WAKEUP	0x00080000 /* Don't sleep in umtx cond_wait */
442 #define	TDP_INBDFLUSH	0x00100000 /* Already in BO_BDFLUSH, do not recurse */
443 #define	TDP_KTHREAD	0x00200000 /* This is an official kernel thread */
444 #define	TDP_CALLCHAIN	0x00400000 /* Capture thread's callchain */
445 #define	TDP_IGNSUSP	0x00800000 /* Permission to ignore the MNTK_SUSPEND* */
446 #define	TDP_AUDITREC	0x01000000 /* Audit record pending on thread */
447 #define	TDP_RFPPWAIT	0x02000000 /* Handle RFPPWAIT on syscall exit */
448 #define	TDP_RESETSPUR	0x04000000 /* Reset spurious page fault history. */
449 #define	TDP_NERRNO	0x08000000 /* Last errno is already in td_errno */
450 #define	TDP_UIOHELD	0x10000000 /* Current uio has pages held in td_ma */
451 #define	TDP_FORKING	0x20000000 /* Thread is being created through fork() */
452 #define	TDP_EXECVMSPC	0x40000000 /* Execve destroyed old vmspace */
453 
454 /*
455  * Reasons that the current thread can not be run yet.
456  * More than one may apply.
457  */
458 #define	TDI_SUSPENDED	0x0001	/* On suspension queue. */
459 #define	TDI_SLEEPING	0x0002	/* Actually asleep! (tricky). */
460 #define	TDI_SWAPPED	0x0004	/* Stack not in mem.  Bad juju if run. */
461 #define	TDI_LOCK	0x0008	/* Stopped on a lock. */
462 #define	TDI_IWAIT	0x0010	/* Awaiting interrupt. */
463 
464 #define	TD_IS_SLEEPING(td)	((td)->td_inhibitors & TDI_SLEEPING)
465 #define	TD_ON_SLEEPQ(td)	((td)->td_wchan != NULL)
466 #define	TD_IS_SUSPENDED(td)	((td)->td_inhibitors & TDI_SUSPENDED)
467 #define	TD_IS_SWAPPED(td)	((td)->td_inhibitors & TDI_SWAPPED)
468 #define	TD_ON_LOCK(td)		((td)->td_inhibitors & TDI_LOCK)
469 #define	TD_AWAITING_INTR(td)	((td)->td_inhibitors & TDI_IWAIT)
470 #define	TD_IS_RUNNING(td)	((td)->td_state == TDS_RUNNING)
471 #define	TD_ON_RUNQ(td)		((td)->td_state == TDS_RUNQ)
472 #define	TD_CAN_RUN(td)		((td)->td_state == TDS_CAN_RUN)
473 #define	TD_IS_INHIBITED(td)	((td)->td_state == TDS_INHIBITED)
474 #define	TD_ON_UPILOCK(td)	((td)->td_flags & TDF_UPIBLOCKED)
475 #define TD_IS_IDLETHREAD(td)	((td)->td_flags & TDF_IDLETD)
476 
477 
478 #define	TD_SET_INHIB(td, inhib) do {			\
479 	(td)->td_state = TDS_INHIBITED;			\
480 	(td)->td_inhibitors |= (inhib);			\
481 } while (0)
482 
483 #define	TD_CLR_INHIB(td, inhib) do {			\
484 	if (((td)->td_inhibitors & (inhib)) &&		\
485 	    (((td)->td_inhibitors &= ~(inhib)) == 0))	\
486 		(td)->td_state = TDS_CAN_RUN;		\
487 } while (0)
488 
489 #define	TD_SET_SLEEPING(td)	TD_SET_INHIB((td), TDI_SLEEPING)
490 #define	TD_SET_SWAPPED(td)	TD_SET_INHIB((td), TDI_SWAPPED)
491 #define	TD_SET_LOCK(td)		TD_SET_INHIB((td), TDI_LOCK)
492 #define	TD_SET_SUSPENDED(td)	TD_SET_INHIB((td), TDI_SUSPENDED)
493 #define	TD_SET_IWAIT(td)	TD_SET_INHIB((td), TDI_IWAIT)
494 #define	TD_SET_EXITING(td)	TD_SET_INHIB((td), TDI_EXITING)
495 
496 #define	TD_CLR_SLEEPING(td)	TD_CLR_INHIB((td), TDI_SLEEPING)
497 #define	TD_CLR_SWAPPED(td)	TD_CLR_INHIB((td), TDI_SWAPPED)
498 #define	TD_CLR_LOCK(td)		TD_CLR_INHIB((td), TDI_LOCK)
499 #define	TD_CLR_SUSPENDED(td)	TD_CLR_INHIB((td), TDI_SUSPENDED)
500 #define	TD_CLR_IWAIT(td)	TD_CLR_INHIB((td), TDI_IWAIT)
501 
502 #define	TD_SET_RUNNING(td)	(td)->td_state = TDS_RUNNING
503 #define	TD_SET_RUNQ(td)		(td)->td_state = TDS_RUNQ
504 #define	TD_SET_CAN_RUN(td)	(td)->td_state = TDS_CAN_RUN
505 
506 /*
507  * Process structure.
508  */
509 struct proc {
510 	LIST_ENTRY(proc) p_list;	/* (d) List of all processes. */
511 	TAILQ_HEAD(, thread) p_threads;	/* (c) all threads. */
512 	struct mtx	p_slock;	/* process spin lock */
513 	struct ucred	*p_ucred;	/* (c) Process owner's identity. */
514 	struct filedesc	*p_fd;		/* (b) Open files. */
515 	struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */
516 	struct pstats	*p_stats;	/* (b) Accounting/statistics (CPU). */
517 	struct plimit	*p_limit;	/* (c) Resource limits. */
518 	struct callout	p_limco;	/* (c) Limit callout handle */
519 	struct sigacts	*p_sigacts;	/* (x) Signal actions, state (CPU). */
520 
521 	int		p_flag;		/* (c) P_* flags. */
522 	int		p_flag2;	/* (c) P2_* flags. */
523 	enum {
524 		PRS_NEW = 0,		/* In creation */
525 		PRS_NORMAL,		/* threads can be run. */
526 		PRS_ZOMBIE
527 	} p_state;			/* (j/c) Process status. */
528 	pid_t		p_pid;		/* (b) Process identifier. */
529 	LIST_ENTRY(proc) p_hash;	/* (d) Hash chain. */
530 	LIST_ENTRY(proc) p_pglist;	/* (g + e) List of processes in pgrp. */
531 	struct proc	*p_pptr;	/* (c + e) Pointer to parent process. */
532 	LIST_ENTRY(proc) p_sibling;	/* (e) List of sibling processes. */
533 	LIST_HEAD(, proc) p_children;	/* (e) Pointer to list of children. */
534 	struct proc	*p_reaper;	/* (e) My reaper. */
535 	LIST_HEAD(, proc) p_reaplist;	/* (e) List of my descendants
536 					       (if I am reaper). */
537 	LIST_ENTRY(proc) p_reapsibling;	/* (e) List of siblings - descendants of
538 					       the same reaper. */
539 	struct mtx	p_mtx;		/* (n) Lock for this struct. */
540 	struct mtx	p_statmtx;	/* Lock for the stats */
541 	struct mtx	p_itimmtx;	/* Lock for the virt/prof timers */
542 	struct mtx	p_profmtx;	/* Lock for the profiling */
543 	struct ksiginfo *p_ksi;	/* Locked by parent proc lock */
544 	sigqueue_t	p_sigqueue;	/* (c) Sigs not delivered to a td. */
545 #define p_siglist	p_sigqueue.sq_signals
546 
547 /* The following fields are all zeroed upon creation in fork. */
548 #define	p_startzero	p_oppid
549 	pid_t		p_oppid;	/* (c + e) Save ppid in ptrace. XXX */
550 	struct vmspace	*p_vmspace;	/* (b) Address space. */
551 	u_int		p_swtick;	/* (c) Tick when swapped in or out. */
552 	u_int		p_cowgen;	/* (c) Generation of COW pointers. */
553 	struct itimerval p_realtimer;	/* (c) Alarm timer. */
554 	struct rusage	p_ru;		/* (a) Exit information. */
555 	struct rusage_ext p_rux;	/* (cu) Internal resource usage. */
556 	struct rusage_ext p_crux;	/* (c) Internal child resource usage. */
557 	int		p_profthreads;	/* (c) Num threads in addupc_task. */
558 	volatile int	p_exitthreads;	/* (j) Number of threads exiting */
559 	int		p_traceflag;	/* (o) Kernel trace points. */
560 	struct vnode	*p_tracevp;	/* (c + o) Trace to vnode. */
561 	struct ucred	*p_tracecred;	/* (o) Credentials to trace with. */
562 	struct vnode	*p_textvp;	/* (b) Vnode of executable. */
563 	u_int		p_lock;		/* (c) Proclock (prevent swap) count. */
564 	struct sigiolst	p_sigiolst;	/* (c) List of sigio sources. */
565 	int		p_sigparent;	/* (c) Signal to parent on exit. */
566 	int		p_sig;		/* (n) For core dump/debugger XXX. */
567 	u_long		p_code;		/* (n) For core dump/debugger XXX. */
568 	u_int		p_stops;	/* (c) Stop event bitmask. */
569 	u_int		p_stype;	/* (c) Stop event type. */
570 	char		p_step;		/* (c) Process is stopped. */
571 	u_char		p_pfsflags;	/* (c) Procfs flags. */
572 	struct nlminfo	*p_nlminfo;	/* (?) Only used by/for lockd. */
573 	struct kaioinfo	*p_aioinfo;	/* (y) ASYNC I/O info. */
574 	struct thread	*p_singlethread;/* (c + j) If single threading this is it */
575 	int		p_suspcount;	/* (j) Num threads in suspended mode. */
576 	struct thread	*p_xthread;	/* (c) Trap thread */
577 	int		p_boundary_count;/* (j) Num threads at user boundary */
578 	int		p_pendingcnt;	/* how many signals are pending */
579 	struct itimers	*p_itimers;	/* (c) POSIX interval timers. */
580 	struct procdesc	*p_procdesc;	/* (e) Process descriptor, if any. */
581 	u_int		p_treeflag;	/* (e) P_TREE flags */
582 	int		p_pendingexits; /* (c) Count of pending thread exits. */
583 /* End area that is zeroed on creation. */
584 #define	p_endzero	p_magic
585 
586 /* The following fields are all copied upon creation in fork. */
587 #define	p_startcopy	p_endzero
588 	u_int		p_magic;	/* (b) Magic number. */
589 	int		p_osrel;	/* (x) osreldate for the
590 					       binary (from ELF note, if any) */
591 	char		p_comm[MAXCOMLEN + 1];	/* (b) Process name. */
592 	struct sysentvec *p_sysent;	/* (b) Syscall dispatch info. */
593 	struct pargs	*p_args;	/* (c) Process arguments. */
594 	rlim_t		p_cpulimit;	/* (c) Current CPU limit in seconds. */
595 	signed char	p_nice;		/* (c) Process "nice" value. */
596 	int		p_fibnum;	/* in this routing domain XXX MRT */
597 	pid_t		p_reapsubtree;	/* (e) Pid of the direct child of the
598 					       reaper which spawned
599 					       our subtree. */
600 	u_int		p_xexit;	/* (c) Exit code. */
601 	u_int		p_xsig;		/* (c) Stop/kill sig. */
602 /* End area that is copied on creation. */
603 #define	p_endcopy	p_xsig
604 	struct pgrp	*p_pgrp;	/* (c + e) Pointer to process group. */
605 	struct knlist	p_klist;	/* (c) Knotes attached to this proc. */
606 	int		p_numthreads;	/* (c) Number of threads. */
607 	struct mdproc	p_md;		/* Any machine-dependent fields. */
608 	struct callout	p_itcallout;	/* (h + c) Interval timer callout. */
609 	u_short		p_acflag;	/* (c) Accounting flags. */
610 	struct proc	*p_peers;	/* (r) */
611 	struct proc	*p_leader;	/* (b) */
612 	void		*p_emuldata;	/* (c) Emulator state data. */
613 	struct label	*p_label;	/* (*) Proc (not subject) MAC label. */
614 	struct p_sched	*p_sched;	/* (*) Scheduler-specific data. */
615 	STAILQ_HEAD(, ktr_request)	p_ktr;	/* (o) KTR event queue. */
616 	LIST_HEAD(, mqueue_notifier)	p_mqnotifier; /* (c) mqueue notifiers.*/
617 	struct kdtrace_proc	*p_dtrace; /* (*) DTrace-specific data. */
618 	struct cv	p_pwait;	/* (*) wait cv for exit/exec. */
619 	struct cv	p_dbgwait;	/* (*) wait cv for debugger attach
620 					   after fork. */
621 	uint64_t	p_prev_runtime;	/* (c) Resource usage accounting. */
622 	struct racct	*p_racct;	/* (b) Resource accounting. */
623 	u_char		p_throttled;	/* (c) Flag for racct pcpu throttling */
624 	struct vm_domain_policy p_vm_dom_policy;	/* (c) process default VM domain, or -1 */
625 	/*
626 	 * An orphan is the child that has beed re-parented to the
627 	 * debugger as a result of attaching to it.  Need to keep
628 	 * track of them for parent to be able to collect the exit
629 	 * status of what used to be children.
630 	 */
631 	LIST_ENTRY(proc) p_orphan;	/* (e) List of orphan processes. */
632 	LIST_HEAD(, proc) p_orphans;	/* (e) Pointer to list of orphans. */
633 };
634 
635 #define	p_session	p_pgrp->pg_session
636 #define	p_pgid		p_pgrp->pg_id
637 
638 #define	NOCPU		(-1)	/* For when we aren't on a CPU. */
639 #define	NOCPU_OLD	(255)
640 #define	MAXCPU_OLD	(254)
641 
642 #define	PROC_SLOCK(p)	mtx_lock_spin(&(p)->p_slock)
643 #define	PROC_SUNLOCK(p)	mtx_unlock_spin(&(p)->p_slock)
644 #define	PROC_SLOCK_ASSERT(p, type)	mtx_assert(&(p)->p_slock, (type))
645 
646 #define	PROC_STATLOCK(p)	mtx_lock_spin(&(p)->p_statmtx)
647 #define	PROC_STATUNLOCK(p)	mtx_unlock_spin(&(p)->p_statmtx)
648 #define	PROC_STATLOCK_ASSERT(p, type)	mtx_assert(&(p)->p_statmtx, (type))
649 
650 #define	PROC_ITIMLOCK(p)	mtx_lock_spin(&(p)->p_itimmtx)
651 #define	PROC_ITIMUNLOCK(p)	mtx_unlock_spin(&(p)->p_itimmtx)
652 #define	PROC_ITIMLOCK_ASSERT(p, type)	mtx_assert(&(p)->p_itimmtx, (type))
653 
654 #define	PROC_PROFLOCK(p)	mtx_lock_spin(&(p)->p_profmtx)
655 #define	PROC_PROFUNLOCK(p)	mtx_unlock_spin(&(p)->p_profmtx)
656 #define	PROC_PROFLOCK_ASSERT(p, type)	mtx_assert(&(p)->p_profmtx, (type))
657 
658 /* These flags are kept in p_flag. */
659 #define	P_ADVLOCK	0x00001	/* Process may hold a POSIX advisory lock. */
660 #define	P_CONTROLT	0x00002	/* Has a controlling terminal. */
661 #define	P_KPROC		0x00004	/* Kernel process. */
662 #define	P_FOLLOWFORK	0x00008	/* Attach parent debugger to children. */
663 #define	P_PPWAIT	0x00010	/* Parent is waiting for child to exec/exit. */
664 #define	P_PROFIL	0x00020	/* Has started profiling. */
665 #define	P_STOPPROF	0x00040	/* Has thread requesting to stop profiling. */
666 #define	P_HADTHREADS	0x00080	/* Has had threads (no cleanup shortcuts) */
667 #define	P_SUGID		0x00100	/* Had set id privileges since last exec. */
668 #define	P_SYSTEM	0x00200	/* System proc: no sigs, stats or swapping. */
669 #define	P_SINGLE_EXIT	0x00400	/* Threads suspending should exit, not wait. */
670 #define	P_TRACED	0x00800	/* Debugged process being traced. */
671 #define	P_WAITED	0x01000	/* Someone is waiting for us. */
672 #define	P_WEXIT		0x02000	/* Working on exiting. */
673 #define	P_EXEC		0x04000	/* Process called exec. */
674 #define	P_WKILLED	0x08000	/* Killed, go to kernel/user boundary ASAP. */
675 #define	P_CONTINUED	0x10000	/* Proc has continued from a stopped state. */
676 #define	P_STOPPED_SIG	0x20000	/* Stopped due to SIGSTOP/SIGTSTP. */
677 #define	P_STOPPED_TRACE	0x40000	/* Stopped because of tracing. */
678 #define	P_STOPPED_SINGLE 0x80000 /* Only 1 thread can continue (not to user). */
679 #define	P_PROTECTED	0x100000 /* Do not kill on memory overcommit. */
680 #define	P_SIGEVENT	0x200000 /* Process pending signals changed. */
681 #define	P_SINGLE_BOUNDARY 0x400000 /* Threads should suspend at user boundary. */
682 #define	P_HWPMC		0x800000 /* Process is using HWPMCs */
683 #define	P_JAILED	0x1000000 /* Process is in jail. */
684 #define	P_TOTAL_STOP	0x2000000 /* Stopped in stop_all_proc. */
685 #define	P_INEXEC	0x4000000 /* Process is in execve(). */
686 #define	P_STATCHILD	0x8000000 /* Child process stopped or exited. */
687 #define	P_INMEM		0x10000000 /* Loaded into memory. */
688 #define	P_SWAPPINGOUT	0x20000000 /* Process is being swapped out. */
689 #define	P_SWAPPINGIN	0x40000000 /* Process is being swapped in. */
690 #define	P_PPTRACE	0x80000000 /* PT_TRACEME by vforked child. */
691 
692 #define	P_STOPPED	(P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
693 #define	P_SHOULDSTOP(p)	((p)->p_flag & P_STOPPED)
694 #define	P_KILLED(p)	((p)->p_flag & P_WKILLED)
695 
696 /* These flags are kept in p_flag2. */
697 #define	P2_INHERIT_PROTECTED 0x00000001 /* New children get P_PROTECTED. */
698 #define	P2_NOTRACE	0x00000002	/* No ptrace(2) attach or coredumps. */
699 #define	P2_NOTRACE_EXEC 0x00000004	/* Keep P2_NOPTRACE on exec(2). */
700 #define	P2_AST_SU	0x00000008	/* Handles SU ast for kthreads. */
701 #define	P2_LWP_EVENTS	0x00000010	/* Report LWP events via ptrace(2). */
702 
703 /* Flags protected by proctree_lock, kept in p_treeflags. */
704 #define	P_TREE_ORPHANED		0x00000001	/* Reparented, on orphan list */
705 #define	P_TREE_FIRST_ORPHAN	0x00000002	/* First element of orphan
706 						   list */
707 #define	P_TREE_REAPER		0x00000004	/* Reaper of subtree */
708 
709 /*
710  * These were process status values (p_stat), now they are only used in
711  * legacy conversion code.
712  */
713 #define	SIDL	1		/* Process being created by fork. */
714 #define	SRUN	2		/* Currently runnable. */
715 #define	SSLEEP	3		/* Sleeping on an address. */
716 #define	SSTOP	4		/* Process debugging or suspension. */
717 #define	SZOMB	5		/* Awaiting collection by parent. */
718 #define	SWAIT	6		/* Waiting for interrupt. */
719 #define	SLOCK	7		/* Blocked on a lock. */
720 
721 #define	P_MAGIC		0xbeefface
722 
723 #ifdef _KERNEL
724 
725 /* Types and flags for mi_switch(). */
726 #define	SW_TYPE_MASK		0xff	/* First 8 bits are switch type */
727 #define	SWT_NONE		0	/* Unspecified switch. */
728 #define	SWT_PREEMPT		1	/* Switching due to preemption. */
729 #define	SWT_OWEPREEMPT		2	/* Switching due to opepreempt. */
730 #define	SWT_TURNSTILE		3	/* Turnstile contention. */
731 #define	SWT_SLEEPQ		4	/* Sleepq wait. */
732 #define	SWT_SLEEPQTIMO		5	/* Sleepq timeout wait. */
733 #define	SWT_RELINQUISH		6	/* yield call. */
734 #define	SWT_NEEDRESCHED		7	/* NEEDRESCHED was set. */
735 #define	SWT_IDLE		8	/* Switching from the idle thread. */
736 #define	SWT_IWAIT		9	/* Waiting for interrupts. */
737 #define	SWT_SUSPEND		10	/* Thread suspended. */
738 #define	SWT_REMOTEPREEMPT	11	/* Remote processor preempted. */
739 #define	SWT_REMOTEWAKEIDLE	12	/* Remote processor preempted idle. */
740 #define	SWT_COUNT		13	/* Number of switch types. */
741 /* Flags */
742 #define	SW_VOL		0x0100		/* Voluntary switch. */
743 #define	SW_INVOL	0x0200		/* Involuntary switch. */
744 #define SW_PREEMPT	0x0400		/* The invol switch is a preemption */
745 
746 /* How values for thread_single(). */
747 #define	SINGLE_NO_EXIT	0
748 #define	SINGLE_EXIT	1
749 #define	SINGLE_BOUNDARY	2
750 #define	SINGLE_ALLPROC	3
751 
752 #ifdef MALLOC_DECLARE
753 MALLOC_DECLARE(M_PARGS);
754 MALLOC_DECLARE(M_PGRP);
755 MALLOC_DECLARE(M_SESSION);
756 MALLOC_DECLARE(M_SUBPROC);
757 #endif
758 
759 #define	FOREACH_PROC_IN_SYSTEM(p)					\
760 	LIST_FOREACH((p), &allproc, p_list)
761 #define	FOREACH_THREAD_IN_PROC(p, td)					\
762 	TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
763 
764 #define	FIRST_THREAD_IN_PROC(p)	TAILQ_FIRST(&(p)->p_threads)
765 
766 /*
767  * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit
768  * in a pid_t, as it is used to represent "no process group".
769  */
770 #define	PID_MAX		99999
771 #define	NO_PID		100000
772 extern pid_t pid_max;
773 
774 #define	SESS_LEADER(p)	((p)->p_session->s_leader == (p))
775 
776 
777 #define	STOPEVENT(p, e, v) do {						\
778 	WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL,			\
779  	    "checking stopevent %d", (e));				\
780 	if ((p)->p_stops & (e))	{					\
781 		PROC_LOCK(p);						\
782 		stopevent((p), (e), (v));				\
783 		PROC_UNLOCK(p);						\
784 	}								\
785 } while (0)
786 #define	_STOPEVENT(p, e, v) do {					\
787 	PROC_LOCK_ASSERT(p, MA_OWNED);					\
788 	WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, &p->p_mtx.lock_object, \
789  	    "checking stopevent %d", (e));				\
790 	if ((p)->p_stops & (e))						\
791 		stopevent((p), (e), (v));				\
792 } while (0)
793 
794 /* Lock and unlock a process. */
795 #define	PROC_LOCK(p)	mtx_lock(&(p)->p_mtx)
796 #define	PROC_TRYLOCK(p)	mtx_trylock(&(p)->p_mtx)
797 #define	PROC_UNLOCK(p)	mtx_unlock(&(p)->p_mtx)
798 #define	PROC_LOCKED(p)	mtx_owned(&(p)->p_mtx)
799 #define	PROC_LOCK_ASSERT(p, type)	mtx_assert(&(p)->p_mtx, (type))
800 
801 /* Lock and unlock a process group. */
802 #define	PGRP_LOCK(pg)	mtx_lock(&(pg)->pg_mtx)
803 #define	PGRP_UNLOCK(pg)	mtx_unlock(&(pg)->pg_mtx)
804 #define	PGRP_LOCKED(pg)	mtx_owned(&(pg)->pg_mtx)
805 #define	PGRP_LOCK_ASSERT(pg, type)	mtx_assert(&(pg)->pg_mtx, (type))
806 
807 #define	PGRP_LOCK_PGSIGNAL(pg) do {					\
808 	if ((pg) != NULL)						\
809 		PGRP_LOCK(pg);						\
810 } while (0)
811 #define	PGRP_UNLOCK_PGSIGNAL(pg) do {					\
812 	if ((pg) != NULL)						\
813 		PGRP_UNLOCK(pg);					\
814 } while (0)
815 
816 /* Lock and unlock a session. */
817 #define	SESS_LOCK(s)	mtx_lock(&(s)->s_mtx)
818 #define	SESS_UNLOCK(s)	mtx_unlock(&(s)->s_mtx)
819 #define	SESS_LOCKED(s)	mtx_owned(&(s)->s_mtx)
820 #define	SESS_LOCK_ASSERT(s, type)	mtx_assert(&(s)->s_mtx, (type))
821 
822 /* Hold process U-area in memory, normally for ptrace/procfs work. */
823 #define	PHOLD(p) do {							\
824 	PROC_LOCK(p);							\
825 	_PHOLD(p);							\
826 	PROC_UNLOCK(p);							\
827 } while (0)
828 #define	_PHOLD(p) do {							\
829 	PROC_LOCK_ASSERT((p), MA_OWNED);				\
830 	KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc,		\
831 	    ("PHOLD of exiting process %p", p));			\
832 	(p)->p_lock++;							\
833 	if (((p)->p_flag & P_INMEM) == 0)				\
834 		faultin((p));						\
835 } while (0)
836 #define	PROC_ASSERT_HELD(p) do {					\
837 	KASSERT((p)->p_lock > 0, ("process %p not held", p));		\
838 } while (0)
839 
840 #define	PRELE(p) do {							\
841 	PROC_LOCK((p));							\
842 	_PRELE((p));							\
843 	PROC_UNLOCK((p));						\
844 } while (0)
845 #define	_PRELE(p) do {							\
846 	PROC_LOCK_ASSERT((p), MA_OWNED);				\
847 	PROC_ASSERT_HELD(p);						\
848 	(--(p)->p_lock);						\
849 	if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0)		\
850 		wakeup(&(p)->p_lock);					\
851 } while (0)
852 #define	PROC_ASSERT_NOT_HELD(p) do {					\
853 	KASSERT((p)->p_lock == 0, ("process %p held", p));		\
854 } while (0)
855 
856 #define	PROC_UPDATE_COW(p) do {						\
857 	PROC_LOCK_ASSERT((p), MA_OWNED);				\
858 	(p)->p_cowgen++;						\
859 } while (0)
860 
861 /* Check whether a thread is safe to be swapped out. */
862 #define	thread_safetoswapout(td)	((td)->td_flags & TDF_CANSWAP)
863 
864 /* Control whether or not it is safe for curthread to sleep. */
865 #define	THREAD_NO_SLEEPING()		((curthread)->td_no_sleeping++)
866 
867 #define	THREAD_SLEEPING_OK()		((curthread)->td_no_sleeping--)
868 
869 #define	THREAD_CAN_SLEEP()		((curthread)->td_no_sleeping == 0)
870 
871 #define	PIDHASH(pid)	(&pidhashtbl[(pid) & pidhash])
872 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
873 extern u_long pidhash;
874 #define	TIDHASH(tid)	(&tidhashtbl[(tid) & tidhash])
875 extern LIST_HEAD(tidhashhead, thread) *tidhashtbl;
876 extern u_long tidhash;
877 extern struct rwlock tidhash_lock;
878 
879 #define	PGRPHASH(pgid)	(&pgrphashtbl[(pgid) & pgrphash])
880 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
881 extern u_long pgrphash;
882 
883 extern struct sx allproc_lock;
884 extern int allproc_gen;
885 extern struct sx proctree_lock;
886 extern struct mtx ppeers_lock;
887 extern struct proc proc0;		/* Process slot for swapper. */
888 extern struct thread thread0;		/* Primary thread in proc0. */
889 extern struct vmspace vmspace0;		/* VM space for proc0. */
890 extern int hogticks;			/* Limit on kernel cpu hogs. */
891 extern int lastpid;
892 extern int nprocs, maxproc;		/* Current and max number of procs. */
893 extern int maxprocperuid;		/* Max procs per uid. */
894 extern u_long ps_arg_cache_limit;
895 
896 LIST_HEAD(proclist, proc);
897 TAILQ_HEAD(procqueue, proc);
898 TAILQ_HEAD(threadqueue, thread);
899 extern struct proclist allproc;		/* List of all processes. */
900 extern struct proclist zombproc;	/* List of zombie processes. */
901 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
902 
903 extern struct uma_zone *proc_zone;
904 
905 struct	proc *pfind(pid_t);		/* Find process by id. */
906 struct	proc *pfind_locked(pid_t pid);
907 struct	pgrp *pgfind(pid_t);		/* Find process group by id. */
908 struct	proc *zpfind(pid_t);		/* Find zombie process by id. */
909 
910 struct	fork_req {
911 	int		fr_flags;
912 	int		fr_pages;
913 	int 		*fr_pidp;
914 	struct proc 	**fr_procp;
915 	int 		*fr_pd_fd;
916 	int 		fr_pd_flags;
917 	struct filecaps	*fr_pd_fcaps;
918 };
919 
920 /*
921  * pget() flags.
922  */
923 #define	PGET_HOLD	0x00001	/* Hold the process. */
924 #define	PGET_CANSEE	0x00002	/* Check against p_cansee(). */
925 #define	PGET_CANDEBUG	0x00004	/* Check against p_candebug(). */
926 #define	PGET_ISCURRENT	0x00008	/* Check that the found process is current. */
927 #define	PGET_NOTWEXIT	0x00010	/* Check that the process is not in P_WEXIT. */
928 #define	PGET_NOTINEXEC	0x00020	/* Check that the process is not in P_INEXEC. */
929 #define	PGET_NOTID	0x00040	/* Do not assume tid if pid > PID_MAX. */
930 
931 #define	PGET_WANTREAD	(PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT)
932 
933 int	pget(pid_t pid, int flags, struct proc **pp);
934 
935 void	ast(struct trapframe *framep);
936 struct	thread *choosethread(void);
937 int	cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
938 int	enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp,
939 	    struct session *sess);
940 int	enterthispgrp(struct proc *p, struct pgrp *pgrp);
941 void	faultin(struct proc *p);
942 void	fixjobc(struct proc *p, struct pgrp *pgrp, int entering);
943 int	fork1(struct thread *, struct fork_req *);
944 void	fork_exit(void (*)(void *, struct trapframe *), void *,
945 	    struct trapframe *);
946 void	fork_return(struct thread *, struct trapframe *);
947 int	inferior(struct proc *p);
948 void	kern_yield(int);
949 void 	kick_proc0(void);
950 void	killjobc(void);
951 int	leavepgrp(struct proc *p);
952 int	maybe_preempt(struct thread *td);
953 void	maybe_yield(void);
954 void	mi_switch(int flags, struct thread *newtd);
955 int	p_candebug(struct thread *td, struct proc *p);
956 int	p_cansee(struct thread *td, struct proc *p);
957 int	p_cansched(struct thread *td, struct proc *p);
958 int	p_cansignal(struct thread *td, struct proc *p, int signum);
959 int	p_canwait(struct thread *td, struct proc *p);
960 struct	pargs *pargs_alloc(int len);
961 void	pargs_drop(struct pargs *pa);
962 void	pargs_hold(struct pargs *pa);
963 int	proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb);
964 int	proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb);
965 int	proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb);
966 void	procinit(void);
967 void	proc_linkup0(struct proc *p, struct thread *td);
968 void	proc_linkup(struct proc *p, struct thread *td);
969 struct proc *proc_realparent(struct proc *child);
970 void	proc_reap(struct thread *td, struct proc *p, int *status, int options);
971 void	proc_reparent(struct proc *child, struct proc *newparent);
972 struct	pstats *pstats_alloc(void);
973 void	pstats_fork(struct pstats *src, struct pstats *dst);
974 void	pstats_free(struct pstats *ps);
975 void	reaper_abandon_children(struct proc *p, bool exiting);
976 int	securelevel_ge(struct ucred *cr, int level);
977 int	securelevel_gt(struct ucred *cr, int level);
978 void	sess_hold(struct session *);
979 void	sess_release(struct session *);
980 int	setrunnable(struct thread *);
981 void	setsugid(struct proc *p);
982 int	should_yield(void);
983 int	sigonstack(size_t sp);
984 void	stopevent(struct proc *, u_int, u_int);
985 struct	thread *tdfind(lwpid_t, pid_t);
986 void	threadinit(void);
987 void	tidhash_add(struct thread *);
988 void	tidhash_remove(struct thread *);
989 void	cpu_idle(int);
990 int	cpu_idle_wakeup(int);
991 extern	void (*cpu_idle_hook)(sbintime_t);	/* Hook to machdep CPU idler. */
992 void	cpu_switch(struct thread *, struct thread *, struct mtx *);
993 void	cpu_throw(struct thread *, struct thread *) __dead2;
994 void	unsleep(struct thread *);
995 void	userret(struct thread *, struct trapframe *);
996 
997 void	cpu_exit(struct thread *);
998 void	exit1(struct thread *, int, int) __dead2;
999 int	cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa);
1000 void	cpu_fork(struct thread *, struct proc *, struct thread *, int);
1001 void	cpu_set_fork_handler(struct thread *, void (*)(void *), void *);
1002 void	cpu_set_syscall_retval(struct thread *, int);
1003 void	cpu_set_upcall(struct thread *td, struct thread *td0);
1004 void	cpu_set_upcall_kse(struct thread *, void (*)(void *), void *,
1005 	    stack_t *);
1006 int	cpu_set_user_tls(struct thread *, void *tls_base);
1007 void	cpu_thread_alloc(struct thread *);
1008 void	cpu_thread_clean(struct thread *);
1009 void	cpu_thread_exit(struct thread *);
1010 void	cpu_thread_free(struct thread *);
1011 void	cpu_thread_swapin(struct thread *);
1012 void	cpu_thread_swapout(struct thread *);
1013 struct	thread *thread_alloc(int pages);
1014 int	thread_alloc_stack(struct thread *, int pages);
1015 void	thread_cow_get_proc(struct thread *newtd, struct proc *p);
1016 void	thread_cow_get(struct thread *newtd, struct thread *td);
1017 void	thread_cow_free(struct thread *td);
1018 void	thread_cow_update(struct thread *td);
1019 int	thread_create(struct thread *td, struct rtprio *rtp,
1020 	    int (*initialize_thread)(struct thread *, void *), void *thunk);
1021 void	thread_exit(void) __dead2;
1022 void	thread_free(struct thread *td);
1023 void	thread_link(struct thread *td, struct proc *p);
1024 void	thread_reap(void);
1025 int	thread_single(struct proc *p, int how);
1026 void	thread_single_end(struct proc *p, int how);
1027 void	thread_stash(struct thread *td);
1028 void	thread_stopped(struct proc *p);
1029 void	childproc_stopped(struct proc *child, int reason);
1030 void	childproc_continued(struct proc *child);
1031 void	childproc_exited(struct proc *child);
1032 int	thread_suspend_check(int how);
1033 bool	thread_suspend_check_needed(void);
1034 void	thread_suspend_switch(struct thread *, struct proc *p);
1035 void	thread_suspend_one(struct thread *td);
1036 void	thread_unlink(struct thread *td);
1037 void	thread_unsuspend(struct proc *p);
1038 void	thread_wait(struct proc *p);
1039 struct thread	*thread_find(struct proc *p, lwpid_t tid);
1040 
1041 void	stop_all_proc(void);
1042 void	resume_all_proc(void);
1043 
1044 static __inline int
1045 curthread_pflags_set(int flags)
1046 {
1047 	struct thread *td;
1048 	int save;
1049 
1050 	td = curthread;
1051 	save = ~flags | (td->td_pflags & flags);
1052 	td->td_pflags |= flags;
1053 	return (save);
1054 }
1055 
1056 static __inline void
1057 curthread_pflags_restore(int save)
1058 {
1059 
1060 	curthread->td_pflags &= save;
1061 }
1062 
1063 #endif	/* _KERNEL */
1064 
1065 #endif	/* !_SYS_PROC_H_ */
1066