proc.h revision 1.218
1/*	$NetBSD: proc.h,v 1.218 2006/03/29 23:02:31 cube Exp $	*/
2
3/*-
4 * Copyright (c) 1986, 1989, 1991, 1993
5 *	The Regents of the University of California.  All rights reserved.
6 * (c) UNIX System Laboratories, Inc.
7 * All or some portions of this file are derived from material licensed
8 * to the University of California by American Telephone and Telegraph
9 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
10 * the permission of UNIX System Laboratories, Inc.
11 *
12 * Redistribution and use in source and binary forms, with or without
13 * modification, are permitted provided that the following conditions
14 * are met:
15 * 1. Redistributions of source code must retain the above copyright
16 *    notice, this list of conditions and the following disclaimer.
17 * 2. Redistributions in binary form must reproduce the above copyright
18 *    notice, this list of conditions and the following disclaimer in the
19 *    documentation and/or other materials provided with the distribution.
20 * 3. Neither the name of the University nor the names of its contributors
21 *    may be used to endorse or promote products derived from this software
22 *    without specific prior written permission.
23 *
24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * SUCH DAMAGE.
35 *
36 *	@(#)proc.h	8.15 (Berkeley) 5/19/95
37 */
38
39#ifndef _SYS_PROC_H_
40#define	_SYS_PROC_H_
41
42#if defined(_KERNEL_OPT)
43#include "opt_multiprocessor.h"
44#include "opt_kstack.h"
45#endif
46
47#include <machine/proc.h>		/* Machine-dependent proc substruct */
48#include <sys/lock.h>
49#include <sys/lwp.h>
50#include <sys/queue.h>
51#include <sys/callout.h>
52#include <sys/signalvar.h>
53#include <sys/siginfo.h>
54#include <sys/event.h>
55
56/*
57 * One structure allocated per session.
58 */
59struct session {
60	int		s_count;	/* Ref cnt; pgrps in session */
61	u_int		s_flags;
62#define	S_LOGIN_SET	1		/* s_login set in this session */
63	struct proc	*s_leader;	/* Session leader */
64	struct vnode	*s_ttyvp;	/* Vnode of controlling terminal */
65	struct tty	*s_ttyp;	/* Controlling terminal */
66	char		s_login[MAXLOGNAME]; /* Setlogin() name */
67	pid_t		s_sid;		/* Session ID (pid of leader) */
68};
69
70/*
71 * One structure allocated per process group.
72 */
73struct pgrp {
74	LIST_HEAD(, proc) pg_members;	/* Pointer to pgrp members */
75	struct session	*pg_session;	/* Pointer to session */
76	pid_t		pg_id;		/* Pgrp id */
77	int		pg_jobc;	/*
78					 * Number of processes qualifying
79					 * pgrp for job control
80					 */
81};
82
83/*
84 * One structure allocated per emulation.
85 */
86struct exec_package;
87struct ps_strings;
88struct ras;
89struct sa_emul;
90
91struct emul {
92	const char	*e_name;	/* Symbolic name */
93	const char	*e_path;	/* Extra emulation path (NULL if none)*/
94#ifndef __HAVE_MINIMAL_EMUL
95	int		e_flags;	/* Miscellaneous flags, see above */
96					/* Syscall handling function */
97	const int	*e_errno;	/* Errno array */
98	int		e_nosys;	/* Offset of the nosys() syscall */
99	int		e_nsysent;	/* Number of system call entries */
100#endif
101	const struct sysent *e_sysent;	/* System call array */
102	const char * const *e_syscallnames; /* System call name array */
103					/* Signal sending function */
104	void		(*e_sendsig)(const struct ksiginfo *,
105					  const sigset_t *);
106	void		(*e_trapsignal)(struct lwp *, const struct ksiginfo *);
107	int		(*e_tracesig)(struct proc *, int);
108	char		*e_sigcode;	/* Start of sigcode */
109	char		*e_esigcode;	/* End of sigcode */
110					/* Set registers before execution */
111	struct uvm_object **e_sigobject;/* shared sigcode object */
112	void		(*e_setregs)(struct lwp *, struct exec_package *,
113					  u_long);
114
115					/* Per-process hooks */
116	void		(*e_proc_exec)(struct proc *, struct exec_package *);
117	void		(*e_proc_fork)(struct proc *, struct proc *, int);
118	void		(*e_proc_exit)(struct proc *);
119	void		(*e_lwp_fork)(struct lwp *, struct lwp *);
120	void		(*e_lwp_exit)(struct lwp *);
121
122#ifdef __HAVE_SYSCALL_INTERN
123	void		(*e_syscall_intern)(struct proc *);
124#else
125	void		(*e_syscall)(void);
126#endif
127					/* Emulation specific sysctl data */
128	struct sysctlnode *e_sysctlovly;
129	int		(*e_fault)(struct proc *, vaddr_t, int);
130
131	vaddr_t		(*e_vm_default_addr)(struct proc *, vaddr_t, vsize_t);
132
133	/* Emulation-specific hook for userspace page faults */
134	int		(*e_usertrap)(struct lwp *, vaddr_t, void *);
135
136	/* SA-related information */
137	const struct sa_emul *e_sa;
138};
139
140/*
141 * Emulation miscelaneous flags
142 */
143#define	EMUL_HAS_SYS___syscall	0x001	/* Has SYS___syscall */
144
145/*
146 * Description of a process.
147 *
148 * This structure contains the information needed to manage a thread of
149 * control, known in UN*X as a process; it has references to substructures
150 * containing descriptions of things that the process uses, but may share
151 * with related processes.  The process structure and the substructures
152 * are always addressible except for those marked "(PROC ONLY)" below,
153 * which might be addressible only on a processor on which the process
154 * is running.
155 *
156 * Fields marked 'p:' are protected by the process's own p_lock.
157 * Fields marked 'l:' are protected by the proclist_lock
158 * Fields marked 's:' are protected by the SCHED_LOCK.
159 */
160struct proc {
161	LIST_ENTRY(proc) p_list;	/* List of all processes */
162
163	/* Substructures: */
164	struct pcred	*p_cred;	/* Process owner's identity */
165	struct filedesc	*p_fd;		/* Ptr to open files structure */
166	struct cwdinfo	*p_cwdi;	/* cdir/rdir/cmask info */
167	struct pstats	*p_stats;	/* Accounting/statistics (PROC ONLY) */
168	struct plimit	*p_limit;	/* Process limits */
169	struct vmspace	*p_vmspace;	/* Address space */
170	struct sigacts	*p_sigacts;	/* Process sigactions (state is below)*/
171
172	void		*p_ksems;	/* p1003.1b semaphores */
173
174#define	p_ucred		p_cred->pc_ucred
175#define	p_rlimit	p_limit->pl_rlimit
176
177	int		p_exitsig;	/* signal to send to parent on exit */
178	int		p_flag;		/* P_* flags. */
179	char		p_stat;		/* S* process status. */
180	char		p_pad1[3];
181
182	pid_t		p_pid;		/* Process identifier. */
183	LIST_ENTRY(proc) p_pglist;	/* l: List of processes in pgrp. */
184	struct proc 	*p_pptr;	/* l: Pointer to parent process. */
185	LIST_ENTRY(proc) p_sibling;	/* l: List of sibling processes. */
186	LIST_HEAD(, proc) p_children;	/* l: Pointer to list of children. */
187
188	struct simplelock p_lock;	/* Lock on proc state (p:) */
189
190	/* XXX dsl: locking of LWP info is suspect in schedcpu and kpsignal2 */
191	LIST_HEAD(, lwp) p_lwps;	/* p: Pointer to list of LWPs. */
192
193	LIST_HEAD(, ras) p_raslist;	/* p: Pointer to RAS queue */
194
195/* The following fields are all zeroed upon creation in fork. */
196#define	p_startzero	p_nlwps
197
198	int 		p_nlwps;	/* p: Number of LWPs */
199	int 		p_nrlwps;	/* s: Number of running LWPs */
200	int 		p_nzlwps;	/* p: Number of zombie LWPs */
201	int 		p_nlwpid;	/* p: Next LWP ID */
202
203	u_int		p_nstopchild;	/* l: Count of stopped/dead children */
204
205	struct sadata 	*p_sa;		/* Scheduler activation information */
206
207	/* scheduling */
208	fixpt_t		p_estcpu;	/* Time averaged value of p_cpticks XXX belongs in p_startcopy section */
209	fixpt_t		p_estcpu_inherited;
210	unsigned int	p_forktime;
211	int		p_cpticks;	/* Ticks of CPU time */
212	fixpt_t		p_pctcpu;	/* %cpu for this process during p_swtime */
213
214	struct proc	*p_opptr;	/* Save parent during ptrace. */
215	struct ptimers	*p_timers;	/* Timers: real, virtual, profiling */
216	struct timeval 	p_rtime;	/* Real time */
217	u_quad_t 	p_uticks;	/* Statclock hits in user mode */
218	u_quad_t 	p_sticks;	/* Statclock hits in system mode */
219	u_quad_t 	p_iticks;	/* Statclock hits processing intr */
220
221	int		p_traceflag;	/* Kernel trace points */
222	void		*p_tracep;	/* Trace private data */
223	void		*p_systrace;	/* Back pointer to systrace */
224
225	struct vnode 	*p_textvp;	/* Vnode of executable */
226
227	const struct emul *p_emul;	/* Emulation information */
228	void		*p_emuldata;	/* Per-process emulation data, or NULL.
229					 * Malloc type M_EMULDATA
230					 */
231
232	void 		(*p_userret)(struct lwp *, void *);
233					/* Function to call at userret(). */
234	void		*p_userret_arg;
235
236	const struct execsw *p_execsw;	/* Exec package information */
237	struct klist	p_klist;	/* Knotes attached to this process */
238
239/*
240 * End area that is zeroed on creation
241 */
242#define	p_endzero	p_startcopy
243
244/*
245 * The following fields are all copied upon creation in fork.
246 */
247#define	p_startcopy	p_sigctx.ps_startcopy
248
249	struct sigctx 	p_sigctx;	/* Signal state */
250
251	u_char		p_nice;		/* Process "nice" value */
252	char		p_comm[MAXCOMLEN+1];	/* basename of last exec file */
253
254	struct pgrp 	*p_pgrp;	/* Pointer to process group */
255
256	struct ps_strings *p_psstr;	/* address of process's ps_strings */
257	size_t 		p_psargv;	/* offset of ps_argvstr in above */
258	size_t 		p_psnargv;	/* offset of ps_nargvstr in above */
259	size_t 		p_psenv;	/* offset of ps_envstr in above */
260	size_t 		p_psnenv;	/* offset of ps_nenvstr in above */
261
262/*
263 * End area that is copied on creation
264 */
265#define	p_endcopy	p_xstat
266
267	u_short		p_xstat;	/* Exit status for wait; also stop signal */
268	u_short		p_acflag;	/* Accounting flags */
269	struct rusage 	*p_ru;		/* Exit information. XXX */
270
271	struct mdproc	p_md;		/* Any machine-dependent fields */
272};
273
274#define	p_session	p_pgrp->pg_session
275#define	p_pgid		p_pgrp->pg_id
276
277/*
278 * Status values.
279 *
280 */
281#define	SIDL		1		/* Process being created by fork */
282#define	SACTIVE		2		/* Process is not stopped */
283#define	SSTOP		4		/* Process debugging or suspension */
284#define	SZOMB		5		/* Awaiting collection by parent */
285
286#define	P_ZOMBIE(p)	((p)->p_stat == SZOMB)
287
288/* These flags are kept in p_flag. */
289#define	P_ADVLOCK	0x00000001 /* Process may hold a POSIX advisory lock */
290#define	P_CONTROLT	0x00000002 /* Has a controlling terminal */
291#define	P_SYSCALL	0x00000004 /* process has PT_SYSCALL enabled */
292#define	P_NOCLDSTOP	0x00000008 /* No SIGCHLD when children stop */
293#define	P_PPWAIT	0x00000010 /* Parent is waiting for child exec/exit */
294#define	P_PROFIL	0x00000020 /* Has started profiling */
295#define	P_SUGID		0x00000100 /* Had set id privileges since last exec */
296#define	P_SYSTEM	0x00000200 /* System proc: no sigs, stats or swapping */
297#define	P_SA		0x00000400 /* Using scheduler activations */
298#define	P_TRACED	0x00000800 /* Debugged process being traced */
299#define	P_WAITED	0x00001000 /* Debugging process has waited for child */
300#define	P_WEXIT		0x00002000 /* Working on exiting */
301#define	P_EXEC		0x00004000 /* Process called exec */
302#define	P_OWEUPC	0x00008000 /* Owe process an addupc() at next ast */
303#define	P_FSTRACE	0x00010000 /* Debugger process being traced by procfs */
304#define	P_NOCLDWAIT	0x00020000 /* No zombies if child dies */
305#define	P_32		0x00040000 /* 32-bit process (used on 64-bit kernels) */
306#define	P_CLDSIGIGN	0x00080000 /* Process is ignoring SIGCHLD */
307#define	P_INEXEC	0x00100000 /* Process is exec'ing and can't be traced */
308#define	P_SYSTRACE	0x00200000 /* Process system call tracing active */
309#define	P_CHTRACED	0x00400000 /* Child has been traced & reparented */
310#define	P_STOPFORK	0x00800000 /* Child will be stopped on fork(2) */
311#define	P_STOPEXEC	0x01000000 /* Will be stopped on exec(2) */
312#define	P_STOPEXIT	0x02000000 /* Will be stopped at process exit */
313#define	P_MARKER	0x80000000 /* Is a dummy marker process */
314
315/*
316 * Macro to compute the exit signal to be delivered.
317 */
318#define	P_EXITSIG(p)	(((p)->p_flag & (P_TRACED|P_FSTRACE)) ? SIGCHLD : \
319			 p->p_exitsig)
320
321/*
322 * MOVE TO ucred.h?
323 *
324 * Shareable process credentials (always resident).  This includes a reference
325 * to the current user credentials as well as real and saved ids that may be
326 * used to change ids.
327 */
328struct pcred {
329	struct ucred	*pc_ucred;	/* Current credentials */
330	uid_t		p_ruid;		/* Real user id */
331	uid_t		p_svuid;	/* Saved effective user id */
332	gid_t		p_rgid;		/* Real group id */
333	gid_t		p_svgid;	/* Saved effective group id */
334	int		p_refcnt;	/* Number of references */
335};
336
337LIST_HEAD(proclist, proc);		/* A list of processes */
338
339/*
340 * This structure associates a proclist with its lock.
341 */
342struct proclist_desc {
343	struct proclist	*pd_list;	/* The list */
344	/*
345	 * XXX Add a pointer to the proclist's lock eventually.
346	 */
347};
348
349#ifdef _KERNEL
350#include <sys/mallocvar.h>
351MALLOC_DECLARE(M_EMULDATA);
352MALLOC_DECLARE(M_PROC);
353MALLOC_DECLARE(M_SESSION);
354MALLOC_DECLARE(M_SUBPROC);	/* XXX - only used by sparc/sparc64 */
355
356/*
357 * We use process IDs <= PID_MAX until there are > 16k processes.
358 * NO_PGID is used to represent "no process group" for a tty.
359 */
360#define	PID_MAX		30000
361#define	NO_PGID		((pid_t)-1)
362
363#define	SESS_LEADER(p)	((p)->p_session->s_leader == (p))
364#define	SESSHOLD(s)	((s)->s_count++)
365#define	SESSRELE(s)							\
366do {									\
367	if (--(s)->s_count == 0)					\
368		sessdelete(s);						\
369} while (/* CONSTCOND */ 0)
370
371
372/*
373 * Flags passed to fork1().
374 */
375#define	FORK_PPWAIT	0x01		/* Block parent until child exit */
376#define	FORK_SHAREVM	0x02		/* Share vmspace with parent */
377#define	FORK_SHARECWD	0x04		/* Share cdir/rdir/cmask */
378#define	FORK_SHAREFILES	0x08		/* Share file descriptors */
379#define	FORK_SHARESIGS	0x10		/* Share signal actions */
380#define	FORK_NOWAIT	0x20		/* Make init the parent of the child */
381#define	FORK_CLEANFILES	0x40		/* Start with a clean descriptor set */
382
383/*
384 * Allow machine-dependent code to override curproc in <machine/cpu.h> for
385 * its own convenience.  Otherwise, we declare it as appropriate.
386 */
387#if !defined(curlwp)
388#if defined(MULTIPROCESSOR)
389#define	curlwp		curcpu()->ci_curlwp	/* Current running LWP */
390#else
391extern struct lwp	*curlwp;		/* Current running LWP */
392#endif /* MULTIPROCESSOR */
393#endif /* ! curproc */
394
395static struct proc *__curproc(void);
396
397static __inline struct proc *
398__curproc()
399{
400	struct lwp *l = curlwp;
401
402	if (l == NULL)
403		return NULL;
404	return l->l_proc;
405}
406#define	curproc	__curproc()
407
408extern struct proc	proc0;		/* Process slot for swapper */
409extern int		nprocs, maxproc; /* Current and max number of procs */
410#define	vmspace_kernel()	(proc0.p_vmspace)
411
412/* Process list lock; see kern_proc.c for locking protocol details */
413extern struct lock	proclist_lock;
414
415extern struct proclist	allproc;	/* List of all processes */
416extern struct proclist	zombproc;	/* List of zombie processes */
417
418extern SLIST_HEAD(deadprocs, proc) deadprocs;	/* List of dead processes */
419extern struct simplelock deadproc_slock;
420
421extern struct proc	*initproc;	/* Process slots for init, pager */
422
423extern const struct proclist_desc proclists[];
424
425extern struct pool	pcred_pool;	/* Memory pool for pcreds */
426extern struct pool	plimit_pool;	/* Memory pool for plimits */
427extern struct pool 	pstats_pool;	/* memory pool for pstats */
428extern struct pool	rusage_pool;	/* Memory pool for rusages */
429extern struct pool	ptimer_pool;	/* Memory pool for ptimers */
430
431struct proc *p_find(pid_t, uint);	/* Find process by id */
432struct pgrp *pg_find(pid_t, uint);	/* Find process group by id */
433/* Flags values for p_find() and pg_find(). */
434#define PFIND_ZOMBIE		1	/* look for zombies as well */
435#define PFIND_LOCKED		2	/* proclist locked on entry */
436#define PFIND_UNLOCK_FAIL	4	/* unlock proclist on failure */
437#define PFIND_UNLOCK_OK		8	/* unlock proclist on success */
438#define PFIND_UNLOCK		(PFIND_UNLOCK_OK | PFIND_UNLOCK_FAIL)
439/* For source compatibility. but UNLOCK_OK gives a stale answer... */
440#define pfind(pid) p_find((pid), PFIND_UNLOCK)
441#define pgfind(pgid) pg_find((pgid), PFIND_UNLOCK)
442
443struct simplelock;
444int	enterpgrp(struct proc *, pid_t, int);
445void	fixjobc(struct proc *, struct pgrp *, int);
446int	inferior(struct proc *, struct proc *);
447int	leavepgrp(struct proc *);
448void	sessdelete(struct session *);
449void	yield(void);
450struct lwp *chooselwp(void);
451void	pgdelete(struct pgrp *);
452void	procinit(void);
453void	resetprocpriority(struct proc *);
454void	suspendsched(void);
455int	ltsleep(volatile const void *, int, const char *, int,
456	    volatile struct simplelock *);
457void	wakeup(volatile const void *);
458void	wakeup_one(volatile const void *);
459void	exit1(struct lwp *, int);
460int	find_stopped_child(struct proc *, pid_t, int, struct proc **);
461struct proc *proc_alloc(void);
462void	proc0_init(void);
463void	proc_free(struct proc *);
464void	proc_free_mem(struct proc *);
465void	exit_lwps(struct lwp *l);
466int	fork1(struct lwp *, int, int, void *, size_t,
467	    void (*)(void *), void *, register_t *, struct proc **);
468void	rqinit(void);
469int	groupmember(gid_t, const struct ucred *);
470int	pgid_in_session(struct proc *, pid_t);
471#ifndef cpu_idle
472void	cpu_idle(void);
473#endif
474void	cpu_exit(struct lwp *);
475void	cpu_lwp_fork(struct lwp *, struct lwp *, void *, size_t,
476	    void (*)(void *), void *);
477#ifndef cpu_lwp_free
478void	cpu_lwp_free(struct lwp *, int);
479#endif
480
481#ifdef __HAVE_SYSCALL_INTERN
482void	syscall_intern(struct proc *);
483#endif
484
485void	child_return(void *);
486
487int	proc_isunder(struct proc *, struct lwp *);
488void	proc_stop(struct proc *, int);
489
490void	proclist_lock_read(void);
491void	proclist_unlock_read(void);
492int	proclist_lock_write(void);
493void	proclist_unlock_write(int);
494void	p_sugid(struct proc *);
495
496int	proc_vmspace_getref(struct proc *, struct vmspace **);
497
498int	proclist_foreach_call(struct proclist *,
499    int (*)(struct proc *, void *arg), void *);
500static __inline struct proc *_proclist_skipmarker(struct proc *);
501
502static __inline struct proc *
503_proclist_skipmarker(struct proc *p0)
504{
505	struct proc *p = p0;
506
507	while (p != NULL && p->p_flag & P_MARKER)
508		p = LIST_NEXT(p, p_list);
509
510	return p;
511}
512#define	PROCLIST_FOREACH(var, head)					\
513	for ((var) = LIST_FIRST(head);					\
514		((var) = _proclist_skipmarker(var)) != NULL;		\
515		(var) = LIST_NEXT(var, p_list))
516
517/* Compatibility with old, non-interlocked tsleep call */
518#define	tsleep(chan, pri, wmesg, timo)					\
519	ltsleep(chan, pri, wmesg, timo, NULL)
520
521#if defined(MULTIPROCESSOR)
522void	proc_trampoline_mp(void);	/* XXX */
523#endif
524
525#ifdef KSTACK_CHECK_MAGIC
526void kstack_setup_magic(const struct lwp *);
527void kstack_check_magic(const struct lwp *);
528#endif
529
530/*
531 * kernel stack paramaters
532 * XXX require sizeof(struct user)
533 */
534/* the lowest address of kernel stack */
535#ifndef KSTACK_LOWEST_ADDR
536#define	KSTACK_LOWEST_ADDR(l)	((caddr_t)ALIGN((l)->l_addr + 1))
537#endif
538/* size of kernel stack */
539#ifndef KSTACK_SIZE
540#define	KSTACK_SIZE	(USPACE - ALIGN(sizeof(struct user)))
541#endif
542
543#endif	/* _KERNEL */
544#endif	/* !_SYS_PROC_H_ */
545