proc.h revision 333162
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: stable/11/sys/sys/proc.h 333162 2018-05-02 07:57:36Z kib $
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 */
77struct 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 */
97struct 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 */
110struct pargs {
111	u_int	ar_ref;		/* Reference count. */
112	u_int	ar_length;	/* Length. */
113	u_char	ar_args[1];	/* Arguments. */
114};
115
116/*-
117 * Description of a process.
118 *
119 * This structure contains the information needed to manage a thread of
120 * control, known in UN*X as a process; it has references to substructures
121 * containing descriptions of things that the process uses, but may share
122 * with related processes.  The process structure and the substructures
123 * are always addressable except for those marked "(CPU)" below,
124 * which might be addressable only on a processor on which the process
125 * is running.
126 *
127 * Below is a key of locks used to protect each member of struct proc.  The
128 * lock is indicated by a reference to a specific character in parens in the
129 * associated comment.
130 *      * - not yet protected
131 *      a - only touched by curproc or parent during fork/wait
132 *      b - created at fork, never changes
133 *		(exception aiods switch vmspaces, but they are also
134 *		marked 'P_SYSTEM' so hopefully it will be left alone)
135 *      c - locked by proc mtx
136 *      d - locked by allproc_lock lock
137 *      e - locked by proctree_lock lock
138 *      f - session mtx
139 *      g - process group mtx
140 *      h - callout_lock mtx
141 *      i - by curproc or the master session mtx
142 *      j - locked by proc slock
143 *      k - only accessed by curthread
144 *	k*- only accessed by curthread and from an interrupt
145 *	kx- only accessed by curthread and by debugger
146 *      l - the attaching proc or attaching proc parent
147 *      m - Giant
148 *      n - not locked, lazy
149 *      o - ktrace lock
150 *      q - td_contested lock
151 *      r - p_peers lock
152 *      s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9)
153 *      t - thread lock
154 *	u - process stat lock
155 *	w - process timer lock
156 *      x - created at fork, only changes during single threading in exec
157 *      y - created at first aio, doesn't change until exit or exec at which
158 *          point we are single-threaded and only curthread changes it
159 *      z - zombie threads lock
160 *
161 * If the locking key specifies two identifiers (for example, p_pptr) then
162 * either lock is sufficient for read access, but both locks must be held
163 * for write access.
164 */
165struct cpuset;
166struct filecaps;
167struct filemon;
168struct kaioinfo;
169struct kaudit_record;
170struct kdtrace_proc;
171struct kdtrace_thread;
172struct mqueue_notifier;
173struct nlminfo;
174struct p_sched;
175struct proc;
176struct procdesc;
177struct racct;
178struct sbuf;
179struct sleepqueue;
180struct syscall_args;
181struct td_sched;
182struct thread;
183struct trapframe;
184struct turnstile;
185struct vm_map;
186struct vm_map_entry;
187
188/*
189 * XXX: Does this belong in resource.h or resourcevar.h instead?
190 * Resource usage extension.  The times in rusage structs in the kernel are
191 * never up to date.  The actual times are kept as runtimes and tick counts
192 * (with control info in the "previous" times), and are converted when
193 * userland asks for rusage info.  Backwards compatibility prevents putting
194 * this directly in the user-visible rusage struct.
195 *
196 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux.
197 * Locking for td_rux: (t) for all fields.
198 */
199struct rusage_ext {
200	uint64_t	rux_runtime;    /* (cu) Real time. */
201	uint64_t	rux_uticks;     /* (cu) Statclock hits in user mode. */
202	uint64_t	rux_sticks;     /* (cu) Statclock hits in sys mode. */
203	uint64_t	rux_iticks;     /* (cu) Statclock hits in intr mode. */
204	uint64_t	rux_uu;         /* (c) Previous user time in usec. */
205	uint64_t	rux_su;         /* (c) Previous sys time in usec. */
206	uint64_t	rux_tu;         /* (c) Previous total time in usec. */
207};
208
209/*
210 * Kernel runnable context (thread).
211 * This is what is put to sleep and reactivated.
212 * Thread context.  Processes may have multiple threads.
213 */
214struct thread {
215	struct mtx	*volatile td_lock; /* replaces sched lock */
216	struct proc	*td_proc;	/* (*) Associated process. */
217	TAILQ_ENTRY(thread) td_plist;	/* (*) All threads in this proc. */
218	TAILQ_ENTRY(thread) td_runq;	/* (t) Run queue. */
219	TAILQ_ENTRY(thread) td_slpq;	/* (t) Sleep queue. */
220	TAILQ_ENTRY(thread) td_lockq;	/* (t) Lock queue. */
221	LIST_ENTRY(thread) td_hash;	/* (d) Hash chain. */
222	struct cpuset	*td_cpuset;	/* (t) CPU affinity mask. */
223	struct seltd	*td_sel;	/* Select queue/channel. */
224	struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */
225	struct turnstile *td_turnstile;	/* (k) Associated turnstile. */
226	struct rl_q_entry *td_rlqe;	/* (k) Associated range lock entry. */
227	struct umtx_q   *td_umtxq;	/* (c?) Link for when we're blocked. */
228	struct vm_domain_policy td_vm_dom_policy;	/* (c) current numa domain policy */
229	lwpid_t		td_tid;		/* (b) Thread ID. */
230	uint64_t	td_padding1[4];
231	void		*td_padding2[4];
232	u_char		td_lend_user_pri; /* (t) Lend user pri. */
233
234/* Cleared during fork1() */
235#define	td_startzero td_flags
236	int		td_flags;	/* (t) TDF_* flags. */
237	int		td_inhibitors;	/* (t) Why can not run. */
238	int		td_pflags;	/* (k) Private thread (TDP_*) flags. */
239	int		td_dupfd;	/* (k) Ret value from fdopen. XXX */
240	int		td_sqqueue;	/* (t) Sleepqueue queue blocked on. */
241	void		*td_wchan;	/* (t) Sleep address. */
242	const char	*td_wmesg;	/* (t) Reason for sleep. */
243	volatile u_char td_owepreempt;  /* (k*) Preempt on last critical_exit */
244	u_char		td_tsqueue;	/* (t) Turnstile queue blocked on. */
245	short		td_locks;	/* (k) Debug: count of non-spin locks */
246	short		td_rw_rlocks;	/* (k) Count of rwlock read locks. */
247	short		td_lk_slocks;	/* (k) Count of lockmgr shared locks. */
248	short		td_stopsched;	/* (k) Scheduler stopped. */
249	struct turnstile *td_blocked;	/* (t) Lock thread is blocked on. */
250	const char	*td_lockname;	/* (t) Name of lock blocked on. */
251	LIST_HEAD(, turnstile) td_contested;	/* (q) Contested locks. */
252	struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
253	int		td_intr_nesting_level; /* (k) Interrupt recursion. */
254	int		td_pinned;	/* (k) Temporary cpu pin count. */
255	struct ucred	*td_ucred;	/* (k) Reference to credentials. */
256	struct plimit	*td_limit;	/* (k) Resource limits. */
257	int		td_slptick;	/* (t) Time at sleep. */
258	int		td_blktick;	/* (t) Time spent blocked. */
259	int		td_swvoltick;	/* (t) Time at last SW_VOL switch. */
260	int		td_swinvoltick;	/* (t) Time at last SW_INVOL switch. */
261	u_int		td_cow;		/* (*) Number of copy-on-write faults */
262	struct rusage	td_ru;		/* (t) rusage information. */
263	struct rusage_ext td_rux;	/* (t) Internal rusage information. */
264	uint64_t	td_incruntime;	/* (t) Cpu ticks to transfer to proc. */
265	uint64_t	td_runtime;	/* (t) How many cpu ticks we've run. */
266	u_int 		td_pticks;	/* (t) Statclock hits for profiling */
267	u_int		td_sticks;	/* (t) Statclock hits in system mode. */
268	u_int		td_iticks;	/* (t) Statclock hits in intr mode. */
269	u_int		td_uticks;	/* (t) Statclock hits in user mode. */
270	int		td_intrval;	/* (t) Return value for sleepq. */
271	sigset_t	td_oldsigmask;	/* (k) Saved mask from pre sigpause. */
272	volatile u_int	td_generation;	/* (k) For detection of preemption */
273	stack_t		td_sigstk;	/* (k) Stack ptr and on-stack flag. */
274	int		td_xsig;	/* (c) Signal for ptrace */
275	u_long		td_profil_addr;	/* (k) Temporary addr until AST. */
276	u_int		td_profil_ticks; /* (k) Temporary ticks until AST. */
277	char		td_name[MAXCOMLEN + 1];	/* (*) Thread name. */
278	struct file	*td_fpop;	/* (k) file referencing cdev under op */
279	int		td_dbgflags;	/* (c) Userland debugger flags */
280	struct ksiginfo	td_padding3;
281	int		td_ng_outbound;	/* (k) Thread entered ng from above. */
282	struct osd	td_osd;		/* (k) Object specific data. */
283	struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */
284	pid_t		td_dbg_forked;	/* (c) Child pid for debugger. */
285	u_int		td_vp_reserv;	/* (k) Count of reserved vnodes. */
286	int		td_no_sleeping;	/* (k) Sleeping disabled count. */
287	int		td_dom_rr_idx;	/* (k) RR Numa domain selection. */
288	void		*td_su;		/* (k) FFS SU private */
289	int		td_rtcgen;	/* (s) rtc_generation of abs. sleep */
290#define	td_endzero td_sigmask
291
292/* Copied during fork1() or create_thread(). */
293#define	td_startcopy td_endzero
294	sigset_t	td_sigmask;	/* (c) Current signal mask. */
295	u_char		td_rqindex;	/* (t) Run queue index. */
296	u_char		td_base_pri;	/* (t) Thread base kernel priority. */
297	u_char		td_priority;	/* (t) Thread active priority. */
298	u_char		td_pri_class;	/* (t) Scheduling class. */
299	u_char		td_user_pri;	/* (t) User pri from estcpu and nice. */
300	u_char		td_base_user_pri; /* (t) Base user pri */
301	u_int		td_padding4;
302	u_int		td_padding5;
303	uintptr_t	td_rb_list;	/* (k) Robust list head. */
304	uintptr_t	td_rbp_list;	/* (k) Robust priv list head. */
305	uintptr_t	td_rb_inact;	/* (k) Current in-action mutex loc. */
306#define	td_endcopy td_pcb
307
308/*
309 * Fields that must be manually set in fork1() or create_thread()
310 * or already have been set in the allocator, constructor, etc.
311 */
312	struct pcb	*td_pcb;	/* (k) Kernel VA of pcb and kstack. */
313	enum {
314		TDS_INACTIVE = 0x0,
315		TDS_INHIBITED,
316		TDS_CAN_RUN,
317		TDS_RUNQ,
318		TDS_RUNNING
319	} td_state;			/* (t) thread state */
320	union {
321		register_t	tdu_retval[2];
322		off_t		tdu_off;
323	} td_uretoff;			/* (k) Syscall aux returns. */
324#define td_retval	td_uretoff.tdu_retval
325	u_int		td_cowgen;	/* (k) Generation of COW pointers. */
326	struct callout	td_slpcallout;	/* (h) Callout for sleep. */
327	struct trapframe *td_frame;	/* (k) */
328	struct vm_object *td_kstack_obj;/* (a) Kstack object. */
329	vm_offset_t	td_kstack;	/* (a) Kernel VA of kstack. */
330	int		td_kstack_pages; /* (a) Size of the kstack. */
331	volatile u_int	td_critnest;	/* (k*) Critical section nest level. */
332	struct mdthread td_md;		/* (k) Any machine-dependent fields. */
333	struct kaudit_record	*td_ar;	/* (k) Active audit record, if any. */
334	struct lpohead	td_lprof[2];	/* (a) lock profiling objects. */
335	struct kdtrace_thread	*td_dtrace; /* (*) DTrace-specific data. */
336	int		td_errno;	/* Error returned by last syscall. */
337	struct vnet	*td_vnet;	/* (k) Effective vnet. */
338	const char	*td_vnet_lpush;	/* (k) Debugging vnet push / pop. */
339	struct trapframe *td_intr_frame;/* (k) Frame of the current irq */
340	struct proc	*td_rfppwait_p;	/* (k) The vforked child */
341	struct vm_page	**td_ma;	/* (k) uio pages held */
342	int		td_ma_cnt;	/* (k) size of *td_ma */
343	void		*td_emuldata;	/* Emulator state data */
344	int		td_lastcpu;	/* (t) Last cpu we were on. */
345	int		td_oncpu;	/* (t) Which cpu we are on. */
346	sbintime_t	td_sleeptimo;	/* (t) Sleep timeout. */
347	sigqueue_t	td_sigqueue;	/* (c) Sigs arrived, not delivered. */
348#define	td_siglist	td_sigqueue.sq_signals
349	struct syscall_args td_sa;	/* (kx) Syscall parameters. Copied on
350					   fork for child tracing. */
351	siginfo_t	td_si;		/* (c) For debugger or core file */
352	void		*td_lkpi_task;	/* LinuxKPI task struct pointer */
353	size_t		td_vslock_sz;	/* (k) amount of vslock-ed space */
354};
355
356struct thread0_storage {
357	struct thread t0st_thread;
358	uint64_t t0st_sched[10];
359};
360
361struct mtx *thread_lock_block(struct thread *);
362void thread_lock_unblock(struct thread *, struct mtx *);
363void thread_lock_set(struct thread *, struct mtx *);
364#define	THREAD_LOCK_ASSERT(td, type)					\
365do {									\
366	struct mtx *__m = (td)->td_lock;				\
367	if (__m != &blocked_lock)					\
368		mtx_assert(__m, (type));				\
369} while (0)
370
371#ifdef INVARIANTS
372#define	THREAD_LOCKPTR_ASSERT(td, lock)					\
373do {									\
374	struct mtx *__m = (td)->td_lock;				\
375	KASSERT((__m == &blocked_lock || __m == (lock)),		\
376	    ("Thread %p lock %p does not match %p", td, __m, (lock)));	\
377} while (0)
378
379#define	TD_LOCKS_INC(td)	((td)->td_locks++)
380#define	TD_LOCKS_DEC(td) do {						\
381	KASSERT(SCHEDULER_STOPPED_TD(td) || (td)->td_locks > 0,		\
382	    ("thread %p owns no locks", (td)));				\
383	(td)->td_locks--;						\
384} while (0)
385#else
386#define	THREAD_LOCKPTR_ASSERT(td, lock)
387
388#define	TD_LOCKS_INC(td)
389#define	TD_LOCKS_DEC(td)
390#endif
391
392/*
393 * Flags kept in td_flags:
394 * To change these you MUST have the scheduler lock.
395 */
396#define	TDF_BORROWING	0x00000001 /* Thread is borrowing pri from another. */
397#define	TDF_INPANIC	0x00000002 /* Caused a panic, let it drive crashdump. */
398#define	TDF_INMEM	0x00000004 /* Thread's stack is in memory. */
399#define	TDF_SINTR	0x00000008 /* Sleep is interruptible. */
400#define	TDF_TIMEOUT	0x00000010 /* Timing out during sleep. */
401#define	TDF_IDLETD	0x00000020 /* This is a per-CPU idle thread. */
402#define	TDF_CANSWAP	0x00000040 /* Thread can be swapped. */
403#define	TDF_SLEEPABORT	0x00000080 /* sleepq_abort was called. */
404#define	TDF_KTH_SUSP	0x00000100 /* kthread is suspended */
405#define	TDF_ALLPROCSUSP	0x00000200 /* suspended by SINGLE_ALLPROC */
406#define	TDF_BOUNDARY	0x00000400 /* Thread suspended at user boundary */
407#define	TDF_ASTPENDING	0x00000800 /* Thread has some asynchronous events. */
408#define	TDF_UNUSED12	0x00001000 /* --available-- */
409#define	TDF_SBDRY	0x00002000 /* Stop only on usermode boundary. */
410#define	TDF_UPIBLOCKED	0x00004000 /* Thread blocked on user PI mutex. */
411#define	TDF_NEEDSUSPCHK	0x00008000 /* Thread may need to suspend. */
412#define	TDF_NEEDRESCHED	0x00010000 /* Thread needs to yield. */
413#define	TDF_NEEDSIGCHK	0x00020000 /* Thread may need signal delivery. */
414#define	TDF_NOLOAD	0x00040000 /* Ignore during load avg calculations. */
415#define	TDF_SERESTART	0x00080000 /* ERESTART on stop attempts. */
416#define	TDF_THRWAKEUP	0x00100000 /* Libthr thread must not suspend itself. */
417#define	TDF_SEINTR	0x00200000 /* EINTR on stop attempts. */
418#define	TDF_SWAPINREQ	0x00400000 /* Swapin request due to wakeup. */
419#define	TDF_UNUSED23	0x00800000 /* --available-- */
420#define	TDF_SCHED0	0x01000000 /* Reserved for scheduler private use */
421#define	TDF_SCHED1	0x02000000 /* Reserved for scheduler private use */
422#define	TDF_SCHED2	0x04000000 /* Reserved for scheduler private use */
423#define	TDF_SCHED3	0x08000000 /* Reserved for scheduler private use */
424#define	TDF_ALRMPEND	0x10000000 /* Pending SIGVTALRM needs to be posted. */
425#define	TDF_PROFPEND	0x20000000 /* Pending SIGPROF needs to be posted. */
426#define	TDF_MACPEND	0x40000000 /* AST-based MAC event pending. */
427
428/* Userland debug flags */
429#define	TDB_SUSPEND	0x00000001 /* Thread is suspended by debugger */
430#define	TDB_XSIG	0x00000002 /* Thread is exchanging signal under trace */
431#define	TDB_USERWR	0x00000004 /* Debugger modified memory or registers */
432#define	TDB_SCE		0x00000008 /* Thread performs syscall enter */
433#define	TDB_SCX		0x00000010 /* Thread performs syscall exit */
434#define	TDB_EXEC	0x00000020 /* TDB_SCX from exec(2) family */
435#define	TDB_FORK	0x00000040 /* TDB_SCX from fork(2) that created new
436				      process */
437#define	TDB_STOPATFORK	0x00000080 /* Stop at the return from fork (child
438				      only) */
439#define	TDB_CHILD	0x00000100 /* New child indicator for ptrace() */
440#define	TDB_BORN	0x00000200 /* New LWP indicator for ptrace() */
441#define	TDB_EXIT	0x00000400 /* Exiting LWP indicator for ptrace() */
442#define	TDB_VFORK	0x00000800 /* vfork indicator for ptrace() */
443#define	TDB_FSTP	0x00001000 /* The thread is PT_ATTACH leader */
444
445/*
446 * "Private" flags kept in td_pflags:
447 * These are only written by curthread and thus need no locking.
448 */
449#define	TDP_OLDMASK	0x00000001 /* Need to restore mask after suspend. */
450#define	TDP_INKTR	0x00000002 /* Thread is currently in KTR code. */
451#define	TDP_INKTRACE	0x00000004 /* Thread is currently in KTRACE code. */
452#define	TDP_BUFNEED	0x00000008 /* Do not recurse into the buf flush */
453#define	TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */
454#define	TDP_ALTSTACK	0x00000020 /* Have alternate signal stack. */
455#define	TDP_DEADLKTREAT	0x00000040 /* Lock acquisition - deadlock treatment. */
456#define	TDP_NOFAULTING	0x00000080 /* Do not handle page faults. */
457#define	TDP_UNUSED9	0x00000100 /* --available-- */
458#define	TDP_OWEUPC	0x00000200 /* Call addupc() at next AST. */
459#define	TDP_ITHREAD	0x00000400 /* Thread is an interrupt thread. */
460#define	TDP_SYNCIO	0x00000800 /* Local override, disable async i/o. */
461#define	TDP_SCHED1	0x00001000 /* Reserved for scheduler private use */
462#define	TDP_SCHED2	0x00002000 /* Reserved for scheduler private use */
463#define	TDP_SCHED3	0x00004000 /* Reserved for scheduler private use */
464#define	TDP_SCHED4	0x00008000 /* Reserved for scheduler private use */
465#define	TDP_GEOM	0x00010000 /* Settle GEOM before finishing syscall */
466#define	TDP_SOFTDEP	0x00020000 /* Stuck processing softdep worklist */
467#define	TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */
468#define	TDP_WAKEUP	0x00080000 /* Don't sleep in umtx cond_wait */
469#define	TDP_INBDFLUSH	0x00100000 /* Already in BO_BDFLUSH, do not recurse */
470#define	TDP_KTHREAD	0x00200000 /* This is an official kernel thread */
471#define	TDP_CALLCHAIN	0x00400000 /* Capture thread's callchain */
472#define	TDP_IGNSUSP	0x00800000 /* Permission to ignore the MNTK_SUSPEND* */
473#define	TDP_AUDITREC	0x01000000 /* Audit record pending on thread */
474#define	TDP_RFPPWAIT	0x02000000 /* Handle RFPPWAIT on syscall exit */
475#define	TDP_RESETSPUR	0x04000000 /* Reset spurious page fault history. */
476#define	TDP_NERRNO	0x08000000 /* Last errno is already in td_errno */
477#define	TDP_UIOHELD	0x10000000 /* Current uio has pages held in td_ma */
478#define	TDP_FORKING	0x20000000 /* Thread is being created through fork() */
479#define	TDP_EXECVMSPC	0x40000000 /* Execve destroyed old vmspace */
480
481/*
482 * Reasons that the current thread can not be run yet.
483 * More than one may apply.
484 */
485#define	TDI_SUSPENDED	0x0001	/* On suspension queue. */
486#define	TDI_SLEEPING	0x0002	/* Actually asleep! (tricky). */
487#define	TDI_SWAPPED	0x0004	/* Stack not in mem.  Bad juju if run. */
488#define	TDI_LOCK	0x0008	/* Stopped on a lock. */
489#define	TDI_IWAIT	0x0010	/* Awaiting interrupt. */
490
491#define	TD_IS_SLEEPING(td)	((td)->td_inhibitors & TDI_SLEEPING)
492#define	TD_ON_SLEEPQ(td)	((td)->td_wchan != NULL)
493#define	TD_IS_SUSPENDED(td)	((td)->td_inhibitors & TDI_SUSPENDED)
494#define	TD_IS_SWAPPED(td)	((td)->td_inhibitors & TDI_SWAPPED)
495#define	TD_ON_LOCK(td)		((td)->td_inhibitors & TDI_LOCK)
496#define	TD_AWAITING_INTR(td)	((td)->td_inhibitors & TDI_IWAIT)
497#define	TD_IS_RUNNING(td)	((td)->td_state == TDS_RUNNING)
498#define	TD_ON_RUNQ(td)		((td)->td_state == TDS_RUNQ)
499#define	TD_CAN_RUN(td)		((td)->td_state == TDS_CAN_RUN)
500#define	TD_IS_INHIBITED(td)	((td)->td_state == TDS_INHIBITED)
501#define	TD_ON_UPILOCK(td)	((td)->td_flags & TDF_UPIBLOCKED)
502#define TD_IS_IDLETHREAD(td)	((td)->td_flags & TDF_IDLETD)
503
504#define	KTDSTATE(td)							\
505	(((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep"  :		\
506	((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" :	\
507	((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" :		\
508	((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" :		\
509	((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding")
510
511#define	TD_SET_INHIB(td, inhib) do {			\
512	(td)->td_state = TDS_INHIBITED;			\
513	(td)->td_inhibitors |= (inhib);			\
514} while (0)
515
516#define	TD_CLR_INHIB(td, inhib) do {			\
517	if (((td)->td_inhibitors & (inhib)) &&		\
518	    (((td)->td_inhibitors &= ~(inhib)) == 0))	\
519		(td)->td_state = TDS_CAN_RUN;		\
520} while (0)
521
522#define	TD_SET_SLEEPING(td)	TD_SET_INHIB((td), TDI_SLEEPING)
523#define	TD_SET_SWAPPED(td)	TD_SET_INHIB((td), TDI_SWAPPED)
524#define	TD_SET_LOCK(td)		TD_SET_INHIB((td), TDI_LOCK)
525#define	TD_SET_SUSPENDED(td)	TD_SET_INHIB((td), TDI_SUSPENDED)
526#define	TD_SET_IWAIT(td)	TD_SET_INHIB((td), TDI_IWAIT)
527#define	TD_SET_EXITING(td)	TD_SET_INHIB((td), TDI_EXITING)
528
529#define	TD_CLR_SLEEPING(td)	TD_CLR_INHIB((td), TDI_SLEEPING)
530#define	TD_CLR_SWAPPED(td)	TD_CLR_INHIB((td), TDI_SWAPPED)
531#define	TD_CLR_LOCK(td)		TD_CLR_INHIB((td), TDI_LOCK)
532#define	TD_CLR_SUSPENDED(td)	TD_CLR_INHIB((td), TDI_SUSPENDED)
533#define	TD_CLR_IWAIT(td)	TD_CLR_INHIB((td), TDI_IWAIT)
534
535#define	TD_SET_RUNNING(td)	(td)->td_state = TDS_RUNNING
536#define	TD_SET_RUNQ(td)		(td)->td_state = TDS_RUNQ
537#define	TD_SET_CAN_RUN(td)	(td)->td_state = TDS_CAN_RUN
538
539#define	TD_SBDRY_INTR(td) \
540    (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0)
541#define	TD_SBDRY_ERRNO(td) \
542    (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART)
543
544/*
545 * Process structure.
546 */
547struct proc {
548	LIST_ENTRY(proc) p_list;	/* (d) List of all processes. */
549	TAILQ_HEAD(, thread) p_threads;	/* (c) all threads. */
550	struct mtx	p_slock;	/* process spin lock */
551	struct ucred	*p_ucred;	/* (c) Process owner's identity. */
552	struct filedesc	*p_fd;		/* (b) Open files. */
553	struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */
554	struct pstats	*p_stats;	/* (b) Accounting/statistics (CPU). */
555	struct plimit	*p_limit;	/* (c) Resource limits. */
556	struct callout	p_limco;	/* (c) Limit callout handle */
557	struct sigacts	*p_sigacts;	/* (x) Signal actions, state (CPU). */
558
559	int		p_flag;		/* (c) P_* flags. */
560	int		p_flag2;	/* (c) P2_* flags. */
561	enum {
562		PRS_NEW = 0,		/* In creation */
563		PRS_NORMAL,		/* threads can be run. */
564		PRS_ZOMBIE
565	} p_state;			/* (j/c) Process status. */
566	pid_t		p_pid;		/* (b) Process identifier. */
567	LIST_ENTRY(proc) p_hash;	/* (d) Hash chain. */
568	LIST_ENTRY(proc) p_pglist;	/* (g + e) List of processes in pgrp. */
569	struct proc	*p_pptr;	/* (c + e) Pointer to parent process. */
570	LIST_ENTRY(proc) p_sibling;	/* (e) List of sibling processes. */
571	LIST_HEAD(, proc) p_children;	/* (e) Pointer to list of children. */
572	struct proc	*p_reaper;	/* (e) My reaper. */
573	LIST_HEAD(, proc) p_reaplist;	/* (e) List of my descendants
574					       (if I am reaper). */
575	LIST_ENTRY(proc) p_reapsibling;	/* (e) List of siblings - descendants of
576					       the same reaper. */
577	struct mtx	p_mtx;		/* (n) Lock for this struct. */
578	struct mtx	p_statmtx;	/* Lock for the stats */
579	struct mtx	p_itimmtx;	/* Lock for the virt/prof timers */
580	struct mtx	p_profmtx;	/* Lock for the profiling */
581	struct ksiginfo *p_ksi;	/* Locked by parent proc lock */
582	uint64_t	padding1[4];
583	void		*padding2[4];
584
585/* The following fields are all zeroed upon creation in fork. */
586#define	p_startzero	p_oppid
587	pid_t		p_oppid;	/* (c + e) Save ppid in ptrace. XXX */
588	struct vmspace	*p_vmspace;	/* (b) Address space. */
589	u_int		p_swtick;	/* (c) Tick when swapped in or out. */
590	u_int		p_cowgen;	/* (c) Generation of COW pointers. */
591	struct itimerval p_realtimer;	/* (c) Alarm timer. */
592	struct rusage	p_ru;		/* (a) Exit information. */
593	struct rusage_ext p_rux;	/* (cu) Internal resource usage. */
594	struct rusage_ext p_crux;	/* (c) Internal child resource usage. */
595	int		p_profthreads;	/* (c) Num threads in addupc_task. */
596	volatile int	p_exitthreads;	/* (j) Number of threads exiting */
597	int		p_traceflag;	/* (o) Kernel trace points. */
598	struct vnode	*p_tracevp;	/* (c + o) Trace to vnode. */
599	struct ucred	*p_tracecred;	/* (o) Credentials to trace with. */
600	struct vnode	*p_textvp;	/* (b) Vnode of executable. */
601	u_int		p_lock;		/* (c) Proclock (prevent swap) count. */
602	struct sigiolst	p_sigiolst;	/* (c) List of sigio sources. */
603	int		p_sigparent;	/* (c) Signal to parent on exit. */
604	int		p_sig;		/* (n) For core dump/debugger XXX. */
605	u_long		p_code;		/* (n) For core dump/debugger XXX. */
606	u_int		p_stops;	/* (c) Stop event bitmask. */
607	u_int		p_stype;	/* (c) Stop event type. */
608	char		p_step;		/* (c) Process is stopped. */
609	u_char		p_pfsflags;	/* (c) Procfs flags. */
610	struct nlminfo	*p_nlminfo;	/* (?) Only used by/for lockd. */
611	struct kaioinfo	*p_aioinfo;	/* (y) ASYNC I/O info. */
612	struct thread	*p_singlethread;/* (c + j) If single threading this is it */
613	int		p_suspcount;	/* (j) Num threads in suspended mode. */
614	struct thread	*p_xthread;	/* (c) Trap thread */
615	int		p_boundary_count;/* (j) Num threads at user boundary */
616	int		p_pendingcnt;	/* how many signals are pending */
617	struct itimers	*p_itimers;	/* (c) POSIX interval timers. */
618	struct procdesc	*p_procdesc;	/* (e) Process descriptor, if any. */
619	u_int		p_treeflag;	/* (e) P_TREE flags */
620	int		p_pendingexits; /* (c) Count of pending thread exits. */
621	struct filemon	*p_filemon;	/* (c) filemon-specific data. */
622/* End area that is zeroed on creation. */
623#define	p_endzero	p_magic
624
625/* The following fields are all copied upon creation in fork. */
626#define	p_startcopy	p_endzero
627	u_int		p_magic;	/* (b) Magic number. */
628	int		p_osrel;	/* (x) osreldate for the
629					       binary (from ELF note, if any) */
630	char		p_comm[MAXCOMLEN + 1];	/* (x) Process name. */
631	struct sysentvec *p_sysent;	/* (b) Syscall dispatch info. */
632	struct pargs	*p_args;	/* (c) Process arguments. */
633	rlim_t		p_cpulimit;	/* (c) Current CPU limit in seconds. */
634	signed char	p_nice;		/* (c) Process "nice" value. */
635	int		p_fibnum;	/* in this routing domain XXX MRT */
636	pid_t		p_reapsubtree;	/* (e) Pid of the direct child of the
637					       reaper which spawned
638					       our subtree. */
639/* End area that is copied on creation. */
640#define	p_endcopy	p_xexit
641
642	u_int		p_xexit;	/* (c) Exit code. */
643	u_int		p_xsig;		/* (c) Stop/kill sig. */
644	struct pgrp	*p_pgrp;	/* (c + e) Pointer to process group. */
645	struct knlist	*p_klist;	/* (c) Knotes attached to this proc. */
646	int		p_numthreads;	/* (c) Number of threads. */
647	struct mdproc	p_md;		/* Any machine-dependent fields. */
648	struct callout	p_itcallout;	/* (h + c) Interval timer callout. */
649	u_short		p_acflag;	/* (c) Accounting flags. */
650	struct proc	*p_peers;	/* (r) */
651	struct proc	*p_leader;	/* (b) */
652	void		*p_emuldata;	/* (c) Emulator state data. */
653	struct label	*p_label;	/* (*) Proc (not subject) MAC label. */
654	STAILQ_HEAD(, ktr_request)	p_ktr;	/* (o) KTR event queue. */
655	LIST_HEAD(, mqueue_notifier)	p_mqnotifier; /* (c) mqueue notifiers.*/
656	struct kdtrace_proc	*p_dtrace; /* (*) DTrace-specific data. */
657	struct cv	p_pwait;	/* (*) wait cv for exit/exec. */
658	struct cv	p_dbgwait;	/* (*) wait cv for debugger attach
659					   after fork. */
660	uint64_t	p_prev_runtime;	/* (c) Resource usage accounting. */
661	struct racct	*p_racct;	/* (b) Resource accounting. */
662	int		p_throttled;	/* (c) Flag for racct pcpu throttling */
663	struct vm_domain_policy p_vm_dom_policy;	/* (c) process default VM domain, or -1 */
664	/*
665	 * An orphan is the child that has beed re-parented to the
666	 * debugger as a result of attaching to it.  Need to keep
667	 * track of them for parent to be able to collect the exit
668	 * status of what used to be children.
669	 */
670	LIST_ENTRY(proc) p_orphan;	/* (e) List of orphan processes. */
671	LIST_HEAD(, proc) p_orphans;	/* (e) Pointer to list of orphans. */
672	u_int		p_ptevents;	/* (c) ptrace() event mask. */
673	uint16_t	p_elf_machine;	/* (x) ELF machine type */
674	uint64_t	p_elf_flags;	/* (x) ELF flags */
675	sigqueue_t	p_sigqueue;	/* (c) Sigs not delivered to a td. */
676#define p_siglist	p_sigqueue.sq_signals
677	int		p_pdeathsig;	/* (c) Signal from parent on exit. */
678};
679
680#define	p_session	p_pgrp->pg_session
681#define	p_pgid		p_pgrp->pg_id
682
683#define	NOCPU		(-1)	/* For when we aren't on a CPU. */
684#define	NOCPU_OLD	(255)
685#define	MAXCPU_OLD	(254)
686
687#define	PROC_SLOCK(p)	mtx_lock_spin(&(p)->p_slock)
688#define	PROC_SUNLOCK(p)	mtx_unlock_spin(&(p)->p_slock)
689#define	PROC_SLOCK_ASSERT(p, type)	mtx_assert(&(p)->p_slock, (type))
690
691#define	PROC_STATLOCK(p)	mtx_lock_spin(&(p)->p_statmtx)
692#define	PROC_STATUNLOCK(p)	mtx_unlock_spin(&(p)->p_statmtx)
693#define	PROC_STATLOCK_ASSERT(p, type)	mtx_assert(&(p)->p_statmtx, (type))
694
695#define	PROC_ITIMLOCK(p)	mtx_lock_spin(&(p)->p_itimmtx)
696#define	PROC_ITIMUNLOCK(p)	mtx_unlock_spin(&(p)->p_itimmtx)
697#define	PROC_ITIMLOCK_ASSERT(p, type)	mtx_assert(&(p)->p_itimmtx, (type))
698
699#define	PROC_PROFLOCK(p)	mtx_lock_spin(&(p)->p_profmtx)
700#define	PROC_PROFUNLOCK(p)	mtx_unlock_spin(&(p)->p_profmtx)
701#define	PROC_PROFLOCK_ASSERT(p, type)	mtx_assert(&(p)->p_profmtx, (type))
702
703/* These flags are kept in p_flag. */
704#define	P_ADVLOCK	0x00001	/* Process may hold a POSIX advisory lock. */
705#define	P_CONTROLT	0x00002	/* Has a controlling terminal. */
706#define	P_KPROC		0x00004	/* Kernel process. */
707#define	P_UNUSED3	0x00008	/* --available-- */
708#define	P_PPWAIT	0x00010	/* Parent is waiting for child to exec/exit. */
709#define	P_PROFIL	0x00020	/* Has started profiling. */
710#define	P_STOPPROF	0x00040	/* Has thread requesting to stop profiling. */
711#define	P_HADTHREADS	0x00080	/* Has had threads (no cleanup shortcuts) */
712#define	P_SUGID		0x00100	/* Had set id privileges since last exec. */
713#define	P_SYSTEM	0x00200	/* System proc: no sigs, stats or swapping. */
714#define	P_SINGLE_EXIT	0x00400	/* Threads suspending should exit, not wait. */
715#define	P_TRACED	0x00800	/* Debugged process being traced. */
716#define	P_WAITED	0x01000	/* Someone is waiting for us. */
717#define	P_WEXIT		0x02000	/* Working on exiting. */
718#define	P_EXEC		0x04000	/* Process called exec. */
719#define	P_WKILLED	0x08000	/* Killed, go to kernel/user boundary ASAP. */
720#define	P_CONTINUED	0x10000	/* Proc has continued from a stopped state. */
721#define	P_STOPPED_SIG	0x20000	/* Stopped due to SIGSTOP/SIGTSTP. */
722#define	P_STOPPED_TRACE	0x40000	/* Stopped because of tracing. */
723#define	P_STOPPED_SINGLE 0x80000 /* Only 1 thread can continue (not to user). */
724#define	P_PROTECTED	0x100000 /* Do not kill on memory overcommit. */
725#define	P_SIGEVENT	0x200000 /* Process pending signals changed. */
726#define	P_SINGLE_BOUNDARY 0x400000 /* Threads should suspend at user boundary. */
727#define	P_HWPMC		0x800000 /* Process is using HWPMCs */
728#define	P_JAILED	0x1000000 /* Process is in jail. */
729#define	P_TOTAL_STOP	0x2000000 /* Stopped in stop_all_proc. */
730#define	P_INEXEC	0x4000000 /* Process is in execve(). */
731#define	P_STATCHILD	0x8000000 /* Child process stopped or exited. */
732#define	P_INMEM		0x10000000 /* Loaded into memory. */
733#define	P_SWAPPINGOUT	0x20000000 /* Process is being swapped out. */
734#define	P_SWAPPINGIN	0x40000000 /* Process is being swapped in. */
735#define	P_PPTRACE	0x80000000 /* PT_TRACEME by vforked child. */
736
737#define	P_STOPPED	(P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
738#define	P_SHOULDSTOP(p)	((p)->p_flag & P_STOPPED)
739#define	P_KILLED(p)	((p)->p_flag & P_WKILLED)
740
741/* These flags are kept in p_flag2. */
742#define	P2_INHERIT_PROTECTED 0x00000001 /* New children get P_PROTECTED. */
743#define	P2_NOTRACE	0x00000002	/* No ptrace(2) attach or coredumps. */
744#define	P2_NOTRACE_EXEC 0x00000004	/* Keep P2_NOPTRACE on exec(2). */
745#define	P2_AST_SU	0x00000008	/* Handles SU ast for kthreads. */
746#define	P2_PTRACE_FSTP	0x00000010 /* SIGSTOP from PT_ATTACH not yet handled. */
747#define	P2_TRAPCAP	0x00000020	/* SIGTRAP on ENOTCAPABLE */
748
749/* Flags protected by proctree_lock, kept in p_treeflags. */
750#define	P_TREE_ORPHANED		0x00000001	/* Reparented, on orphan list */
751#define	P_TREE_FIRST_ORPHAN	0x00000002	/* First element of orphan
752						   list */
753#define	P_TREE_REAPER		0x00000004	/* Reaper of subtree */
754
755/*
756 * These were process status values (p_stat), now they are only used in
757 * legacy conversion code.
758 */
759#define	SIDL	1		/* Process being created by fork. */
760#define	SRUN	2		/* Currently runnable. */
761#define	SSLEEP	3		/* Sleeping on an address. */
762#define	SSTOP	4		/* Process debugging or suspension. */
763#define	SZOMB	5		/* Awaiting collection by parent. */
764#define	SWAIT	6		/* Waiting for interrupt. */
765#define	SLOCK	7		/* Blocked on a lock. */
766
767#define	P_MAGIC		0xbeefface
768
769#ifdef _KERNEL
770
771/* Types and flags for mi_switch(). */
772#define	SW_TYPE_MASK		0xff	/* First 8 bits are switch type */
773#define	SWT_NONE		0	/* Unspecified switch. */
774#define	SWT_PREEMPT		1	/* Switching due to preemption. */
775#define	SWT_OWEPREEMPT		2	/* Switching due to owepreempt. */
776#define	SWT_TURNSTILE		3	/* Turnstile contention. */
777#define	SWT_SLEEPQ		4	/* Sleepq wait. */
778#define	SWT_SLEEPQTIMO		5	/* Sleepq timeout wait. */
779#define	SWT_RELINQUISH		6	/* yield call. */
780#define	SWT_NEEDRESCHED		7	/* NEEDRESCHED was set. */
781#define	SWT_IDLE		8	/* Switching from the idle thread. */
782#define	SWT_IWAIT		9	/* Waiting for interrupts. */
783#define	SWT_SUSPEND		10	/* Thread suspended. */
784#define	SWT_REMOTEPREEMPT	11	/* Remote processor preempted. */
785#define	SWT_REMOTEWAKEIDLE	12	/* Remote processor preempted idle. */
786#define	SWT_COUNT		13	/* Number of switch types. */
787/* Flags */
788#define	SW_VOL		0x0100		/* Voluntary switch. */
789#define	SW_INVOL	0x0200		/* Involuntary switch. */
790#define SW_PREEMPT	0x0400		/* The invol switch is a preemption */
791
792/* How values for thread_single(). */
793#define	SINGLE_NO_EXIT	0
794#define	SINGLE_EXIT	1
795#define	SINGLE_BOUNDARY	2
796#define	SINGLE_ALLPROC	3
797
798#ifdef MALLOC_DECLARE
799MALLOC_DECLARE(M_PARGS);
800MALLOC_DECLARE(M_PGRP);
801MALLOC_DECLARE(M_SESSION);
802MALLOC_DECLARE(M_SUBPROC);
803#endif
804
805#define	FOREACH_PROC_IN_SYSTEM(p)					\
806	LIST_FOREACH((p), &allproc, p_list)
807#define	FOREACH_THREAD_IN_PROC(p, td)					\
808	TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
809
810#define	FIRST_THREAD_IN_PROC(p)	TAILQ_FIRST(&(p)->p_threads)
811
812/*
813 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit
814 * in a pid_t, as it is used to represent "no process group".
815 */
816#define	PID_MAX		99999
817#define	NO_PID		100000
818extern pid_t pid_max;
819
820#define	SESS_LEADER(p)	((p)->p_session->s_leader == (p))
821
822
823#define	STOPEVENT(p, e, v) do {						\
824	WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL,			\
825 	    "checking stopevent %d", (e));				\
826	if ((p)->p_stops & (e))	{					\
827		PROC_LOCK(p);						\
828		stopevent((p), (e), (v));				\
829		PROC_UNLOCK(p);						\
830	}								\
831} while (0)
832#define	_STOPEVENT(p, e, v) do {					\
833	PROC_LOCK_ASSERT(p, MA_OWNED);					\
834	WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, &p->p_mtx.lock_object, \
835 	    "checking stopevent %d", (e));				\
836	if ((p)->p_stops & (e))						\
837		stopevent((p), (e), (v));				\
838} while (0)
839
840/* Lock and unlock a process. */
841#define	PROC_LOCK(p)	mtx_lock(&(p)->p_mtx)
842#define	PROC_TRYLOCK(p)	mtx_trylock(&(p)->p_mtx)
843#define	PROC_UNLOCK(p)	mtx_unlock(&(p)->p_mtx)
844#define	PROC_LOCKED(p)	mtx_owned(&(p)->p_mtx)
845#define	PROC_LOCK_ASSERT(p, type)	mtx_assert(&(p)->p_mtx, (type))
846
847/* Lock and unlock a process group. */
848#define	PGRP_LOCK(pg)	mtx_lock(&(pg)->pg_mtx)
849#define	PGRP_UNLOCK(pg)	mtx_unlock(&(pg)->pg_mtx)
850#define	PGRP_LOCKED(pg)	mtx_owned(&(pg)->pg_mtx)
851#define	PGRP_LOCK_ASSERT(pg, type)	mtx_assert(&(pg)->pg_mtx, (type))
852
853#define	PGRP_LOCK_PGSIGNAL(pg) do {					\
854	if ((pg) != NULL)						\
855		PGRP_LOCK(pg);						\
856} while (0)
857#define	PGRP_UNLOCK_PGSIGNAL(pg) do {					\
858	if ((pg) != NULL)						\
859		PGRP_UNLOCK(pg);					\
860} while (0)
861
862/* Lock and unlock a session. */
863#define	SESS_LOCK(s)	mtx_lock(&(s)->s_mtx)
864#define	SESS_UNLOCK(s)	mtx_unlock(&(s)->s_mtx)
865#define	SESS_LOCKED(s)	mtx_owned(&(s)->s_mtx)
866#define	SESS_LOCK_ASSERT(s, type)	mtx_assert(&(s)->s_mtx, (type))
867
868/*
869 * Non-zero p_lock ensures that:
870 * - exit1() is not performed until p_lock reaches zero;
871 * - the process' threads stack are not swapped out if they are currently
872 *   not (P_INMEM).
873 *
874 * PHOLD() asserts that the process (except the current process) is
875 * not exiting, increments p_lock and swaps threads stacks into memory,
876 * if needed.
877 * _PHOLD() is same as PHOLD(), it takes the process locked.
878 * _PHOLD_LITE() also takes the process locked, but comparing with
879 * _PHOLD(), it only guarantees that exit1() is not executed,
880 * faultin() is not called.
881 */
882#define	PHOLD(p) do {							\
883	PROC_LOCK(p);							\
884	_PHOLD(p);							\
885	PROC_UNLOCK(p);							\
886} while (0)
887#define	_PHOLD(p) do {							\
888	PROC_LOCK_ASSERT((p), MA_OWNED);				\
889	KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc,		\
890	    ("PHOLD of exiting process %p", p));			\
891	(p)->p_lock++;							\
892	if (((p)->p_flag & P_INMEM) == 0)				\
893		faultin((p));						\
894} while (0)
895#define	_PHOLD_LITE(p) do {						\
896	PROC_LOCK_ASSERT((p), MA_OWNED);				\
897	KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc,		\
898	    ("PHOLD of exiting process %p", p));			\
899	(p)->p_lock++;							\
900} while (0)
901#define	PROC_ASSERT_HELD(p) do {					\
902	KASSERT((p)->p_lock > 0, ("process %p not held", p));		\
903} while (0)
904
905#define	PRELE(p) do {							\
906	PROC_LOCK((p));							\
907	_PRELE((p));							\
908	PROC_UNLOCK((p));						\
909} while (0)
910#define	_PRELE(p) do {							\
911	PROC_LOCK_ASSERT((p), MA_OWNED);				\
912	PROC_ASSERT_HELD(p);						\
913	(--(p)->p_lock);						\
914	if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0)		\
915		wakeup(&(p)->p_lock);					\
916} while (0)
917#define	PROC_ASSERT_NOT_HELD(p) do {					\
918	KASSERT((p)->p_lock == 0, ("process %p held", p));		\
919} while (0)
920
921#define	PROC_UPDATE_COW(p) do {						\
922	PROC_LOCK_ASSERT((p), MA_OWNED);				\
923	(p)->p_cowgen++;						\
924} while (0)
925
926/* Check whether a thread is safe to be swapped out. */
927#define	thread_safetoswapout(td)	((td)->td_flags & TDF_CANSWAP)
928
929/* Control whether or not it is safe for curthread to sleep. */
930#define	THREAD_NO_SLEEPING()		((curthread)->td_no_sleeping++)
931
932#define	THREAD_SLEEPING_OK()		((curthread)->td_no_sleeping--)
933
934#define	THREAD_CAN_SLEEP()		((curthread)->td_no_sleeping == 0)
935
936#define	PIDHASH(pid)	(&pidhashtbl[(pid) & pidhash])
937extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
938extern u_long pidhash;
939#define	TIDHASH(tid)	(&tidhashtbl[(tid) & tidhash])
940extern LIST_HEAD(tidhashhead, thread) *tidhashtbl;
941extern u_long tidhash;
942extern struct rwlock tidhash_lock;
943
944#define	PGRPHASH(pgid)	(&pgrphashtbl[(pgid) & pgrphash])
945extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
946extern u_long pgrphash;
947
948extern struct sx allproc_lock;
949extern int allproc_gen;
950extern struct sx proctree_lock;
951extern struct mtx ppeers_lock;
952extern struct proc proc0;		/* Process slot for swapper. */
953extern struct thread0_storage thread0_st;	/* Primary thread in proc0. */
954#define	thread0 (thread0_st.t0st_thread)
955extern struct vmspace vmspace0;		/* VM space for proc0. */
956extern int hogticks;			/* Limit on kernel cpu hogs. */
957extern int lastpid;
958extern int nprocs, maxproc;		/* Current and max number of procs. */
959extern int maxprocperuid;		/* Max procs per uid. */
960extern u_long ps_arg_cache_limit;
961
962LIST_HEAD(proclist, proc);
963TAILQ_HEAD(procqueue, proc);
964TAILQ_HEAD(threadqueue, thread);
965extern struct proclist allproc;		/* List of all processes. */
966extern struct proclist zombproc;	/* List of zombie processes. */
967extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
968
969extern struct uma_zone *proc_zone;
970
971struct	proc *pfind(pid_t);		/* Find process by id. */
972struct	proc *pfind_locked(pid_t pid);
973struct	pgrp *pgfind(pid_t);		/* Find process group by id. */
974struct	proc *zpfind(pid_t);		/* Find zombie process by id. */
975
976struct	fork_req {
977	int		fr_flags;
978	int		fr_pages;
979	int 		*fr_pidp;
980	struct proc 	**fr_procp;
981	int 		*fr_pd_fd;
982	int 		fr_pd_flags;
983	struct filecaps	*fr_pd_fcaps;
984};
985
986/*
987 * pget() flags.
988 */
989#define	PGET_HOLD	0x00001	/* Hold the process. */
990#define	PGET_CANSEE	0x00002	/* Check against p_cansee(). */
991#define	PGET_CANDEBUG	0x00004	/* Check against p_candebug(). */
992#define	PGET_ISCURRENT	0x00008	/* Check that the found process is current. */
993#define	PGET_NOTWEXIT	0x00010	/* Check that the process is not in P_WEXIT. */
994#define	PGET_NOTINEXEC	0x00020	/* Check that the process is not in P_INEXEC. */
995#define	PGET_NOTID	0x00040	/* Do not assume tid if pid > PID_MAX. */
996
997#define	PGET_WANTREAD	(PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT)
998
999int	pget(pid_t pid, int flags, struct proc **pp);
1000
1001void	ast(struct trapframe *framep);
1002struct	thread *choosethread(void);
1003int	cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
1004int	enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp,
1005	    struct session *sess);
1006int	enterthispgrp(struct proc *p, struct pgrp *pgrp);
1007void	faultin(struct proc *p);
1008void	fixjobc(struct proc *p, struct pgrp *pgrp, int entering);
1009int	fork1(struct thread *, struct fork_req *);
1010void	fork_exit(void (*)(void *, struct trapframe *), void *,
1011	    struct trapframe *);
1012void	fork_return(struct thread *, struct trapframe *);
1013int	inferior(struct proc *p);
1014void	kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry,
1015	    int *resident_count, bool *super);
1016void	kern_yield(int);
1017void 	kick_proc0(void);
1018void	killjobc(void);
1019int	leavepgrp(struct proc *p);
1020int	maybe_preempt(struct thread *td);
1021void	maybe_yield(void);
1022void	mi_switch(int flags, struct thread *newtd);
1023int	p_candebug(struct thread *td, struct proc *p);
1024int	p_cansee(struct thread *td, struct proc *p);
1025int	p_cansched(struct thread *td, struct proc *p);
1026int	p_cansignal(struct thread *td, struct proc *p, int signum);
1027int	p_canwait(struct thread *td, struct proc *p);
1028struct	pargs *pargs_alloc(int len);
1029void	pargs_drop(struct pargs *pa);
1030void	pargs_hold(struct pargs *pa);
1031int	proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb);
1032int	proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb);
1033int	proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb);
1034void	procinit(void);
1035void	proc_linkup0(struct proc *p, struct thread *td);
1036void	proc_linkup(struct proc *p, struct thread *td);
1037struct proc *proc_realparent(struct proc *child);
1038void	proc_reap(struct thread *td, struct proc *p, int *status, int options);
1039void	proc_reparent(struct proc *child, struct proc *newparent);
1040void	proc_set_traced(struct proc *p, bool stop);
1041struct	pstats *pstats_alloc(void);
1042void	pstats_fork(struct pstats *src, struct pstats *dst);
1043void	pstats_free(struct pstats *ps);
1044void	reaper_abandon_children(struct proc *p, bool exiting);
1045int	securelevel_ge(struct ucred *cr, int level);
1046int	securelevel_gt(struct ucred *cr, int level);
1047void	sess_hold(struct session *);
1048void	sess_release(struct session *);
1049int	setrunnable(struct thread *);
1050void	setsugid(struct proc *p);
1051int	should_yield(void);
1052int	sigonstack(size_t sp);
1053void	stopevent(struct proc *, u_int, u_int);
1054struct	thread *tdfind(lwpid_t, pid_t);
1055void	threadinit(void);
1056void	tidhash_add(struct thread *);
1057void	tidhash_remove(struct thread *);
1058void	cpu_idle(int);
1059int	cpu_idle_wakeup(int);
1060extern	void (*cpu_idle_hook)(sbintime_t);	/* Hook to machdep CPU idler. */
1061void	cpu_switch(struct thread *, struct thread *, struct mtx *);
1062void	cpu_throw(struct thread *, struct thread *) __dead2;
1063void	unsleep(struct thread *);
1064void	userret(struct thread *, struct trapframe *);
1065
1066void	cpu_exit(struct thread *);
1067void	exit1(struct thread *, int, int) __dead2;
1068void	cpu_copy_thread(struct thread *td, struct thread *td0);
1069int	cpu_fetch_syscall_args(struct thread *td);
1070void	cpu_fork(struct thread *, struct proc *, struct thread *, int);
1071void	cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *);
1072void	cpu_set_syscall_retval(struct thread *, int);
1073void	cpu_set_upcall(struct thread *, void (*)(void *), void *,
1074	    stack_t *);
1075int	cpu_set_user_tls(struct thread *, void *tls_base);
1076void	cpu_thread_alloc(struct thread *);
1077void	cpu_thread_clean(struct thread *);
1078void	cpu_thread_exit(struct thread *);
1079void	cpu_thread_free(struct thread *);
1080void	cpu_thread_swapin(struct thread *);
1081void	cpu_thread_swapout(struct thread *);
1082struct	thread *thread_alloc(int pages);
1083int	thread_alloc_stack(struct thread *, int pages);
1084void	thread_cow_get_proc(struct thread *newtd, struct proc *p);
1085void	thread_cow_get(struct thread *newtd, struct thread *td);
1086void	thread_cow_free(struct thread *td);
1087void	thread_cow_update(struct thread *td);
1088int	thread_create(struct thread *td, struct rtprio *rtp,
1089	    int (*initialize_thread)(struct thread *, void *), void *thunk);
1090void	thread_exit(void) __dead2;
1091void	thread_free(struct thread *td);
1092void	thread_link(struct thread *td, struct proc *p);
1093void	thread_reap(void);
1094int	thread_single(struct proc *p, int how);
1095void	thread_single_end(struct proc *p, int how);
1096void	thread_stash(struct thread *td);
1097void	thread_stopped(struct proc *p);
1098void	childproc_stopped(struct proc *child, int reason);
1099void	childproc_continued(struct proc *child);
1100void	childproc_exited(struct proc *child);
1101int	thread_suspend_check(int how);
1102bool	thread_suspend_check_needed(void);
1103void	thread_suspend_switch(struct thread *, struct proc *p);
1104void	thread_suspend_one(struct thread *td);
1105void	thread_unlink(struct thread *td);
1106void	thread_unsuspend(struct proc *p);
1107void	thread_wait(struct proc *p);
1108struct thread	*thread_find(struct proc *p, lwpid_t tid);
1109
1110void	stop_all_proc(void);
1111void	resume_all_proc(void);
1112
1113static __inline int
1114curthread_pflags_set(int flags)
1115{
1116	struct thread *td;
1117	int save;
1118
1119	td = curthread;
1120	save = ~flags | (td->td_pflags & flags);
1121	td->td_pflags |= flags;
1122	return (save);
1123}
1124
1125static __inline void
1126curthread_pflags_restore(int save)
1127{
1128
1129	curthread->td_pflags &= save;
1130}
1131
1132static __inline __pure2 struct td_sched *
1133td_get_sched(struct thread *td)
1134{
1135
1136	return ((struct td_sched *)&td[1]);
1137}
1138
1139extern void (*softdep_ast_cleanup)(struct thread *);
1140static __inline void
1141td_softdep_cleanup(struct thread *td)
1142{
1143
1144	if (td->td_su != NULL && softdep_ast_cleanup != NULL)
1145		softdep_ast_cleanup(td);
1146}
1147
1148#endif	/* _KERNEL */
1149
1150#endif	/* !_SYS_PROC_H_ */
1151