2 * Copyright (c) 1986, 1989, 1991, 1993
3 * The Regents of the University of California. All rights reserved.
4 * (c) UNIX System Laboratories, Inc.
5 * All or some portions of this file are derived from material licensed
6 * to the University of California by American Telephone and Telegraph
7 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
8 * the permission of UNIX System Laboratories, Inc.
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
13 * 1. Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 * notice, this list of conditions and the following disclaimer in the
17 * documentation and/or other materials provided with the distribution.
18 * 4. Neither the name of the University nor the names of its contributors
19 * may be used to endorse or promote products derived from this software
20 * without specific prior written permission.
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * @(#)proc.h 8.15 (Berkeley) 5/19/95
41 #include <sys/callout.h> /* For struct callout. */
42 #include <sys/event.h> /* For struct klist. */
43 #include <sys/condvar.h>
45 #include <sys/filedesc.h>
47 #include <sys/queue.h>
48 #include <sys/_lock.h>
49 #include <sys/lock_profile.h>
50 #include <sys/_mutex.h>
52 #include <sys/priority.h>
53 #include <sys/rtprio.h> /* XXX. */
55 #include <sys/resource.h>
56 #include <sys/sigio.h>
57 #include <sys/signal.h>
58 #include <sys/signalvar.h>
60 #include <sys/time.h> /* For structs itimerval, timeval. */
64 #include <sys/ucontext.h>
65 #include <sys/ucred.h>
66 #include <machine/proc.h> /* Machine-dependent proc substruct. */
69 * One structure allocated per session.
72 * (m) locked by s_mtx mtx
73 * (e) locked by proctree_lock sx
74 * (c) const until freeing
77 u_int s_count; /* Ref cnt; pgrps in session - atomic. */
78 struct proc *s_leader; /* (m + e) Session leader. */
79 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */
80 struct tty *s_ttyp; /* (e) Controlling tty. */
81 pid_t s_sid; /* (c) Session ID. */
82 /* (m) Setlogin() name: */
83 char s_login[roundup(MAXLOGNAME, sizeof(long))];
84 struct mtx s_mtx; /* Mutex to protect members. */
88 * One structure allocated per process group.
91 * (m) locked by pg_mtx mtx
92 * (e) locked by proctree_lock sx
93 * (c) const until freeing
96 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */
97 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */
98 struct session *pg_session; /* (c) Pointer to session. */
99 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */
100 pid_t pg_id; /* (c) Process group id. */
101 int pg_jobc; /* (m) Job control process count. */
102 struct mtx pg_mtx; /* Mutex to protect members */
106 * pargs, used to hold a copy of the command line, if it had a sane length.
109 u_int ar_ref; /* Reference count. */
110 u_int ar_length; /* Length. */
111 u_char ar_args[1]; /* Arguments. */
115 * Description of a process.
117 * This structure contains the information needed to manage a thread of
118 * control, known in UN*X as a process; it has references to substructures
119 * containing descriptions of things that the process uses, but may share
120 * with related processes. The process structure and the substructures
121 * are always addressable except for those marked "(CPU)" below,
122 * which might be addressable only on a processor on which the process
125 * Below is a key of locks used to protect each member of struct proc. The
126 * lock is indicated by a reference to a specific character in parens in the
127 * associated comment.
128 * * - not yet protected
129 * a - only touched by curproc or parent during fork/wait
130 * b - created at fork, never changes
131 * (exception aiods switch vmspaces, but they are also
132 * marked 'P_SYSTEM' so hopefully it will be left alone)
133 * c - locked by proc mtx
134 * d - locked by allproc_lock lock
135 * e - locked by proctree_lock lock
137 * g - process group mtx
138 * h - callout_lock mtx
139 * i - by curproc or the master session mtx
140 * j - locked by proc slock
141 * k - only accessed by curthread
142 * k*- only accessed by curthread and from an interrupt
143 * l - the attaching proc or attaching proc parent
145 * n - not locked, lazy
147 * q - td_contested lock
150 * x - created at fork, only changes during single threading in exec
151 * y - created at first aio, doesn't change until exit or exec at which
152 * point we are single-threaded and only curthread changes it
153 * z - zombie threads lock
155 * If the locking key specifies two identifiers (for example, p_pptr) then
156 * either lock is sufficient for read access, but both locks must be held
159 struct kaudit_record;
169 struct mqueue_notifier;
171 struct kdtrace_thread;
175 * XXX: Does this belong in resource.h or resourcevar.h instead?
176 * Resource usage extension. The times in rusage structs in the kernel are
177 * never up to date. The actual times are kept as runtimes and tick counts
178 * (with control info in the "previous" times), and are converted when
179 * userland asks for rusage info. Backwards compatibility prevents putting
180 * this directly in the user-visible rusage struct.
182 * Locking for p_rux: (cj) means (j) for p_rux and (c) for p_crux.
183 * Locking for td_rux: (t) for all fields.
186 u_int64_t rux_runtime; /* (cj) Real time. */
187 u_int64_t rux_uticks; /* (cj) Statclock hits in user mode. */
188 u_int64_t rux_sticks; /* (cj) Statclock hits in sys mode. */
189 u_int64_t rux_iticks; /* (cj) Statclock hits in intr mode. */
190 u_int64_t rux_uu; /* (c) Previous user time in usec. */
191 u_int64_t rux_su; /* (c) Previous sys time in usec. */
192 u_int64_t rux_tu; /* (c) Previous total time in usec. */
196 * Kernel runnable context (thread).
197 * This is what is put to sleep and reactivated.
198 * Thread context. Processes may have multiple threads.
201 struct mtx *volatile td_lock; /* replaces sched lock */
202 struct proc *td_proc; /* (*) Associated process. */
203 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */
204 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */
205 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */
206 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */
207 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */
208 struct seltd *td_sel; /* Select queue/channel. */
209 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */
210 struct turnstile *td_turnstile; /* (k) Associated turnstile. */
211 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */
212 lwpid_t td_tid; /* (b) Thread ID. */
213 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */
214 #define td_siglist td_sigqueue.sq_signals
216 /* Cleared during fork1() */
217 #define td_startzero td_flags
218 int td_flags; /* (t) TDF_* flags. */
219 int td_inhibitors; /* (t) Why can not run. */
220 int td_pflags; /* (k) Private thread (TDP_*) flags. */
221 int td_dupfd; /* (k) Ret value from fdopen. XXX */
222 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */
223 void *td_wchan; /* (t) Sleep address. */
224 const char *td_wmesg; /* (t) Reason for sleep. */
225 u_char td_lastcpu; /* (t) Last cpu we were on. */
226 u_char td_oncpu; /* (t) Which cpu we are on. */
227 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */
228 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */
229 short td_locks; /* (k) Count of non-spin locks. */
230 short td_rw_rlocks; /* (k) Count of rwlock read locks. */
231 short td_lk_slocks; /* (k) Count of lockmgr shared locks. */
232 short td_stopsched; /* (k) Scheduler stopped. */
233 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */
234 const char *td_lockname; /* (t) Name of lock blocked on. */
235 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */
236 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
237 int td_intr_nesting_level; /* (k) Interrupt recursion. */
238 int td_pinned; /* (k) Temporary cpu pin count. */
239 struct ucred *td_ucred; /* (k) Reference to credentials. */
240 u_int td_estcpu; /* (t) estimated cpu utilization */
241 int td_slptick; /* (t) Time at sleep. */
242 int td_blktick; /* (t) Time spent blocked. */
243 struct rusage td_ru; /* (t) rusage information. */
244 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */
245 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */
246 u_int td_pticks; /* (t) Statclock hits for profiling */
247 u_int td_sticks; /* (t) Statclock hits in system mode. */
248 u_int td_iticks; /* (t) Statclock hits in intr mode. */
249 u_int td_uticks; /* (t) Statclock hits in user mode. */
250 int td_intrval; /* (t) Return value for sleepq. */
251 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */
252 sigset_t td_sigmask; /* (c) Current signal mask. */
253 volatile u_int td_generation; /* (k) For detection of preemption */
254 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */
255 int td_xsig; /* (c) Signal for ptrace */
256 u_long td_profil_addr; /* (k) Temporary addr until AST. */
257 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */
258 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */
259 struct file *td_fpop; /* (k) file referencing cdev under op */
260 int td_dbgflags; /* (c) Userland debugger flags */
261 struct ksiginfo td_dbgksi; /* (c) ksi reflected to debugger. */
262 int td_ng_outbound; /* (k) Thread entered ng from above. */
263 struct osd td_osd; /* (k) Object specific data. */
264 #define td_endzero td_rqindex
266 /* Copied during fork1() or thread_sched_upcall(). */
267 #define td_startcopy td_endzero
268 u_char td_rqindex; /* (t) Run queue index. */
269 u_char td_base_pri; /* (t) Thread base kernel priority. */
270 u_char td_priority; /* (t) Thread active priority. */
271 u_char td_pri_class; /* (t) Scheduling class. */
272 u_char td_user_pri; /* (t) User pri from estcpu and nice. */
273 u_char td_base_user_pri; /* (t) Base user pri */
274 #define td_endcopy td_pcb
277 * Fields that must be manually set in fork1() or thread_sched_upcall()
278 * or already have been set in the allocator, constructor, etc.
280 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */
287 } td_state; /* (t) thread state */
288 register_t td_retval[2]; /* (k) Syscall aux returns. */
289 struct callout td_slpcallout; /* (h) Callout for sleep. */
290 struct trapframe *td_frame; /* (k) */
291 struct vm_object *td_kstack_obj;/* (a) Kstack object. */
292 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */
293 int td_kstack_pages; /* (a) Size of the kstack. */
295 vm_offset_t td_unused2;
297 volatile u_int td_critnest; /* (k*) Critical section nest level. */
298 struct mdthread td_md; /* (k) Any machine-dependent fields. */
299 struct td_sched *td_sched; /* (*) Scheduler-specific data. */
300 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */
301 int td_syscalls; /* per-thread syscall count (used by NFS :)) */
302 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */
303 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */
304 int td_errno; /* Error returned by last syscall. */
305 struct vnet *td_vnet; /* (k) Effective vnet. */
306 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */
307 struct rusage_ext td_rux; /* (t) Internal rusage information. */
308 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */
309 pid_t td_dbg_forked; /* (c) Child pid for debugger. */
312 struct mtx *thread_lock_block(struct thread *);
313 void thread_lock_unblock(struct thread *, struct mtx *);
314 void thread_lock_set(struct thread *, struct mtx *);
315 #define THREAD_LOCK_ASSERT(td, type) \
317 struct mtx *__m = (td)->td_lock; \
318 if (__m != &blocked_lock) \
319 mtx_assert(__m, (type)); \
323 #define THREAD_LOCKPTR_ASSERT(td, lock) \
325 struct mtx *__m = (td)->td_lock; \
326 KASSERT((__m == &blocked_lock || __m == (lock)), \
327 ("Thread %p lock %p does not match %p", td, __m, (lock))); \
330 #define THREAD_LOCKPTR_ASSERT(td, lock)
333 #define CRITICAL_ASSERT(td) \
334 KASSERT((td)->td_critnest >= 1, ("Not in critical section"));
337 * Flags kept in td_flags:
338 * To change these you MUST have the scheduler lock.
340 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */
341 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */
342 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */
343 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */
344 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */
345 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */
346 #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */
347 #define TDF_SLEEPABORT 0x00000080 /* sleepq_abort was called. */
348 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */
349 #define TDF_UBORROWING 0x00000200 /* Thread is borrowing user pri. */
350 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */
351 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */
352 #define TDF_TIMOFAIL 0x00001000 /* Timeout from sleep after we were awake. */
353 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */
354 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */
355 #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */
356 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */
357 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */
358 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */
359 #define TDF_UNUSED19 0x00080000 /* Thread is sleeping on a umtx. */
360 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */
361 #define TDF_UNUSED21 0x00200000 /* --available-- */
362 #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */
363 #define TDF_UNUSED23 0x00800000 /* --available-- */
364 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */
365 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */
366 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */
367 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */
368 #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */
369 #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */
370 #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */
372 /* Userland debug flags */
373 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */
374 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */
375 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */
376 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */
377 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */
378 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */
379 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new
381 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child
385 * "Private" flags kept in td_pflags:
386 * These are only written by curthread and thus need no locking.
388 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */
389 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */
390 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */
391 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */
392 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */
393 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */
394 #define TDP_DEADLKTREAT 0x00000040 /* Lock aquisition - deadlock treatment. */
395 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */
396 #define TDP_NOSLEEPING 0x00000100 /* Thread is not allowed to sleep on a sq. */
397 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */
398 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */
399 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */
400 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */
401 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */
402 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */
403 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */
404 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */
405 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */
406 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */
407 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */
408 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */
409 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */
410 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */
411 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */
412 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */
413 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */
414 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */
417 * Reasons that the current thread can not be run yet.
418 * More than one may apply.
420 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */
421 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */
422 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */
423 #define TDI_LOCK 0x0008 /* Stopped on a lock. */
424 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */
426 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING)
427 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL)
428 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED)
429 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED)
430 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK)
431 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT)
432 #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING)
433 #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ)
434 #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN)
435 #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED)
436 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED)
437 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD)
440 #define TD_SET_INHIB(td, inhib) do { \
441 (td)->td_state = TDS_INHIBITED; \
442 (td)->td_inhibitors |= (inhib); \
445 #define TD_CLR_INHIB(td, inhib) do { \
446 if (((td)->td_inhibitors & (inhib)) && \
447 (((td)->td_inhibitors &= ~(inhib)) == 0)) \
448 (td)->td_state = TDS_CAN_RUN; \
451 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING)
452 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED)
453 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK)
454 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED)
455 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT)
456 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING)
458 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING)
459 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED)
460 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK)
461 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED)
462 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT)
464 #define TD_SET_RUNNING(td) (td)->td_state = TDS_RUNNING
465 #define TD_SET_RUNQ(td) (td)->td_state = TDS_RUNQ
466 #define TD_SET_CAN_RUN(td) (td)->td_state = TDS_CAN_RUN
472 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */
473 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */
474 struct mtx p_slock; /* process spin lock */
475 struct ucred *p_ucred; /* (c) Process owner's identity. */
476 struct filedesc *p_fd; /* (b) Open files. */
477 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */
478 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */
479 struct plimit *p_limit; /* (c) Process limits. */
480 struct callout p_limco; /* (c) Limit callout handle */
481 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */
484 * The following don't make too much sense.
485 * See the td_ or ke_ versions of the same flags.
487 int p_flag; /* (c) P_* flags. */
489 PRS_NEW = 0, /* In creation */
490 PRS_NORMAL, /* threads can be run. */
492 } p_state; /* (j/c) S* process status. */
493 pid_t p_pid; /* (b) Process identifier. */
494 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */
495 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */
496 struct proc *p_pptr; /* (c + e) Pointer to parent process. */
497 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */
498 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */
499 struct mtx p_mtx; /* (n) Lock for this struct. */
500 struct ksiginfo *p_ksi; /* Locked by parent proc lock */
501 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */
502 #define p_siglist p_sigqueue.sq_signals
504 /* The following fields are all zeroed upon creation in fork. */
505 #define p_startzero p_oppid
506 pid_t p_oppid; /* (c + e) Save ppid in ptrace. XXX */
507 struct vmspace *p_vmspace; /* (b) Address space. */
508 u_int p_swtick; /* (c) Tick when swapped in or out. */
509 struct itimerval p_realtimer; /* (c) Alarm timer. */
510 struct rusage p_ru; /* (a) Exit information. */
511 struct rusage_ext p_rux; /* (cj) Internal resource usage. */
512 struct rusage_ext p_crux; /* (c) Internal child resource usage. */
513 int p_profthreads; /* (c) Num threads in addupc_task. */
514 volatile int p_exitthreads; /* (j) Number of threads exiting */
515 int p_traceflag; /* (o) Kernel trace points. */
516 struct vnode *p_tracevp; /* (c + o) Trace to vnode. */
517 struct ucred *p_tracecred; /* (o) Credentials to trace with. */
518 struct vnode *p_textvp; /* (b) Vnode of executable. */
519 u_int p_lock; /* (c) Proclock (prevent swap) count. */
520 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */
521 int p_sigparent; /* (c) Signal to parent on exit. */
522 int p_sig; /* (n) For core dump/debugger XXX. */
523 u_long p_code; /* (n) For core dump/debugger XXX. */
524 u_int p_stops; /* (c) Stop event bitmask. */
525 u_int p_stype; /* (c) Stop event type. */
526 char p_step; /* (c) Process is stopped. */
527 u_char p_pfsflags; /* (c) Procfs flags. */
528 struct nlminfo *p_nlminfo; /* (?) Only used by/for lockd. */
529 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */
530 struct thread *p_singlethread;/* (c + j) If single threading this is it */
531 int p_suspcount; /* (j) Num threads in suspended mode. */
532 struct thread *p_xthread; /* (c) Trap thread */
533 int p_boundary_count;/* (j) Num threads at user boundary */
534 int p_pendingcnt; /* how many signals are pending */
535 struct itimers *p_itimers; /* (c) POSIX interval timers. */
536 /* End area that is zeroed on creation. */
537 #define p_endzero p_magic
539 /* The following fields are all copied upon creation in fork. */
540 #define p_startcopy p_endzero
541 u_int p_magic; /* (b) Magic number. */
542 int p_osrel; /* (x) osreldate for the
543 binary (from ELF note, if any) */
544 char p_comm[MAXCOMLEN + 1]; /* (b) Process name. */
545 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */
546 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */
547 struct pargs *p_args; /* (c) Process arguments. */
548 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */
549 signed char p_nice; /* (c) Process "nice" value. */
550 int p_fibnum; /* in this routing domain XXX MRT */
551 /* End area that is copied on creation. */
552 #define p_endcopy p_xstat
554 u_short p_xstat; /* (c) Exit status; also stop sig. */
555 struct knlist p_klist; /* (c) Knotes attached to this proc. */
556 int p_numthreads; /* (c) Number of threads. */
557 struct mdproc p_md; /* Any machine-dependent fields. */
558 struct callout p_itcallout; /* (h + c) Interval timer callout. */
559 u_short p_acflag; /* (c) Accounting flags. */
560 struct proc *p_peers; /* (r) */
561 struct proc *p_leader; /* (b) */
562 void *p_emuldata; /* (c) Emulator state data. */
563 struct label *p_label; /* (*) Proc (not subject) MAC label. */
564 struct p_sched *p_sched; /* (*) Scheduler-specific data. */
565 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */
566 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/
567 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */
568 struct cv p_pwait; /* (*) wait cv for exit/exec. */
569 struct cv p_dbgwait; /* (*) wait cv for debugger attach
573 #define p_session p_pgrp->pg_session
574 #define p_pgid p_pgrp->pg_id
576 #define NOCPU 0xff /* For when we aren't on a CPU. */
578 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock)
579 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock)
580 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type))
582 /* These flags are kept in p_flag. */
583 #define P_ADVLOCK 0x00001 /* Process may hold a POSIX advisory lock. */
584 #define P_CONTROLT 0x00002 /* Has a controlling terminal. */
585 #define P_KTHREAD 0x00004 /* Kernel thread (*). */
586 #define P_FOLLOWFORK 0x00008 /* Attach parent debugger to children. */
587 #define P_PPWAIT 0x00010 /* Parent is waiting for child to exec/exit. */
588 #define P_PROFIL 0x00020 /* Has started profiling. */
589 #define P_STOPPROF 0x00040 /* Has thread requesting to stop profiling. */
590 #define P_HADTHREADS 0x00080 /* Has had threads (no cleanup shortcuts) */
591 #define P_SUGID 0x00100 /* Had set id privileges since last exec. */
592 #define P_SYSTEM 0x00200 /* System proc: no sigs, stats or swapping. */
593 #define P_SINGLE_EXIT 0x00400 /* Threads suspending should exit, not wait. */
594 #define P_TRACED 0x00800 /* Debugged process being traced. */
595 #define P_WAITED 0x01000 /* Someone is waiting for us. */
596 #define P_WEXIT 0x02000 /* Working on exiting. */
597 #define P_EXEC 0x04000 /* Process called exec. */
598 #define P_WKILLED 0x08000 /* Killed, go to kernel/user boundary ASAP. */
599 #define P_CONTINUED 0x10000 /* Proc has continued from a stopped state. */
600 #define P_STOPPED_SIG 0x20000 /* Stopped due to SIGSTOP/SIGTSTP. */
601 #define P_STOPPED_TRACE 0x40000 /* Stopped because of tracing. */
602 #define P_STOPPED_SINGLE 0x80000 /* Only 1 thread can continue (not to user). */
603 #define P_PROTECTED 0x100000 /* Do not kill on memory overcommit. */
604 #define P_SIGEVENT 0x200000 /* Process pending signals changed. */
605 #define P_SINGLE_BOUNDARY 0x400000 /* Threads should suspend at user boundary. */
606 #define P_HWPMC 0x800000 /* Process is using HWPMCs */
608 #define P_JAILED 0x1000000 /* Process is in jail. */
609 #define P_INEXEC 0x4000000 /* Process is in execve(). */
610 #define P_STATCHILD 0x8000000 /* Child process stopped or exited. */
611 #define P_INMEM 0x10000000 /* Loaded into memory. */
612 #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */
613 #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */
615 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
616 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED)
617 #define P_KILLED(p) ((p)->p_flag & P_WKILLED)
620 * These were process status values (p_stat), now they are only used in
621 * legacy conversion code.
623 #define SIDL 1 /* Process being created by fork. */
624 #define SRUN 2 /* Currently runnable. */
625 #define SSLEEP 3 /* Sleeping on an address. */
626 #define SSTOP 4 /* Process debugging or suspension. */
627 #define SZOMB 5 /* Awaiting collection by parent. */
628 #define SWAIT 6 /* Waiting for interrupt. */
629 #define SLOCK 7 /* Blocked on a lock. */
631 #define P_MAGIC 0xbeefface
635 /* Types and flags for mi_switch(). */
636 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */
637 #define SWT_NONE 0 /* Unspecified switch. */
638 #define SWT_PREEMPT 1 /* Switching due to preemption. */
639 #define SWT_OWEPREEMPT 2 /* Switching due to opepreempt. */
640 #define SWT_TURNSTILE 3 /* Turnstile contention. */
641 #define SWT_SLEEPQ 4 /* Sleepq wait. */
642 #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */
643 #define SWT_RELINQUISH 6 /* yield call. */
644 #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */
645 #define SWT_IDLE 8 /* Switching from the idle thread. */
646 #define SWT_IWAIT 9 /* Waiting for interrupts. */
647 #define SWT_SUSPEND 10 /* Thread suspended. */
648 #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */
649 #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */
650 #define SWT_COUNT 13 /* Number of switch types. */
652 #define SW_VOL 0x0100 /* Voluntary switch. */
653 #define SW_INVOL 0x0200 /* Involuntary switch. */
654 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */
656 /* How values for thread_single(). */
657 #define SINGLE_NO_EXIT 0
658 #define SINGLE_EXIT 1
659 #define SINGLE_BOUNDARY 2
661 #ifdef MALLOC_DECLARE
662 MALLOC_DECLARE(M_PARGS);
663 MALLOC_DECLARE(M_PGRP);
664 MALLOC_DECLARE(M_SESSION);
665 MALLOC_DECLARE(M_SUBPROC);
668 #define FOREACH_PROC_IN_SYSTEM(p) \
669 LIST_FOREACH((p), &allproc, p_list)
670 #define FOREACH_THREAD_IN_PROC(p, td) \
671 TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
673 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads)
676 * We use process IDs <= PID_MAX; PID_MAX + 1 must also fit in a pid_t,
677 * as it is used to represent "no process group".
679 #define PID_MAX 99999
680 #define NO_PID 100000
682 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p))
685 #define STOPEVENT(p, e, v) do { \
686 if ((p)->p_stops & (e)) { \
688 stopevent((p), (e), (v)); \
692 #define _STOPEVENT(p, e, v) do { \
693 PROC_LOCK_ASSERT(p, MA_OWNED); \
694 WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, &p->p_mtx.lock_object, \
695 "checking stopevent %d", (e)); \
696 if ((p)->p_stops & (e)) \
697 stopevent((p), (e), (v)); \
700 /* Lock and unlock a process. */
701 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx)
702 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx)
703 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx)
704 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx)
705 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type))
707 /* Lock and unlock a process group. */
708 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx)
709 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx)
710 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx)
711 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type))
713 #define PGRP_LOCK_PGSIGNAL(pg) do { \
717 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \
722 /* Lock and unlock a session. */
723 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx)
724 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx)
725 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx)
726 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type))
728 /* Hold process U-area in memory, normally for ptrace/procfs work. */
729 #define PHOLD(p) do { \
734 #define _PHOLD(p) do { \
735 PROC_LOCK_ASSERT((p), MA_OWNED); \
736 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
737 ("PHOLD of exiting process")); \
739 if (((p)->p_flag & P_INMEM) == 0) \
742 #define PROC_ASSERT_HELD(p) do { \
743 KASSERT((p)->p_lock > 0, ("process not held")); \
746 #define PRELE(p) do { \
751 #define _PRELE(p) do { \
752 PROC_LOCK_ASSERT((p), MA_OWNED); \
753 PROC_ASSERT_HELD(p); \
755 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \
756 wakeup(&(p)->p_lock); \
758 #define PROC_ASSERT_NOT_HELD(p) do { \
759 KASSERT((p)->p_lock == 0, ("process held")); \
762 /* Check whether a thread is safe to be swapped out. */
763 #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP)
765 /* Control whether or not it is safe for curthread to sleep. */
766 #define THREAD_NO_SLEEPING() do { \
767 KASSERT(!(curthread->td_pflags & TDP_NOSLEEPING), \
768 ("nested no sleeping")); \
769 curthread->td_pflags |= TDP_NOSLEEPING; \
772 #define THREAD_SLEEPING_OK() do { \
773 KASSERT((curthread->td_pflags & TDP_NOSLEEPING), \
774 ("nested sleeping ok")); \
775 curthread->td_pflags &= ~TDP_NOSLEEPING; \
778 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash])
779 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
780 extern u_long pidhash;
782 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash])
783 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
784 extern u_long pgrphash;
786 extern struct sx allproc_lock;
787 extern struct sx proctree_lock;
788 extern struct mtx ppeers_lock;
789 extern struct proc proc0; /* Process slot for swapper. */
790 extern struct thread thread0; /* Primary thread in proc0. */
791 extern struct vmspace vmspace0; /* VM space for proc0. */
792 extern int hogticks; /* Limit on kernel cpu hogs. */
794 extern int nprocs, maxproc; /* Current and max number of procs. */
795 extern int maxprocperuid; /* Max procs per uid. */
796 extern u_long ps_arg_cache_limit;
798 LIST_HEAD(proclist, proc);
799 TAILQ_HEAD(procqueue, proc);
800 TAILQ_HEAD(threadqueue, thread);
801 extern struct proclist allproc; /* List of all processes. */
802 extern struct proclist zombproc; /* List of zombie processes. */
803 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
805 extern struct uma_zone *proc_zone;
807 struct proc *pfind(pid_t); /* Find process by id. */
808 struct pgrp *pgfind(pid_t); /* Find process group by id. */
809 struct proc *zpfind(pid_t); /* Find zombie process by id. */
814 #define PGET_HOLD 0x00001 /* Hold the process. */
815 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */
816 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */
817 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */
818 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */
819 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */
820 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */
822 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT)
824 int pget(pid_t pid, int flags, struct proc **pp);
826 void ast(struct trapframe *framep);
827 struct thread *choosethread(void);
828 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
829 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp,
830 struct session *sess);
831 int enterthispgrp(struct proc *p, struct pgrp *pgrp);
832 void faultin(struct proc *p);
833 void fixjobc(struct proc *p, struct pgrp *pgrp, int entering);
834 int fork1(struct thread *, int, int, struct proc **);
835 void fork_exit(void (*)(void *, struct trapframe *), void *,
837 void fork_return(struct thread *, struct trapframe *);
838 int inferior(struct proc *p);
839 void kern_yield(int);
840 void kick_proc0(void);
841 int leavepgrp(struct proc *p);
842 int maybe_preempt(struct thread *td);
843 void maybe_yield(void);
844 void mi_switch(int flags, struct thread *newtd);
845 int p_candebug(struct thread *td, struct proc *p);
846 int p_cansee(struct thread *td, struct proc *p);
847 int p_cansched(struct thread *td, struct proc *p);
848 int p_cansignal(struct thread *td, struct proc *p, int signum);
849 int p_canwait(struct thread *td, struct proc *p);
850 struct pargs *pargs_alloc(int len);
851 void pargs_drop(struct pargs *pa);
852 void pargs_hold(struct pargs *pa);
854 void proc_linkup0(struct proc *p, struct thread *td);
855 void proc_linkup(struct proc *p, struct thread *td);
856 void proc_reparent(struct proc *child, struct proc *newparent);
857 struct pstats *pstats_alloc(void);
858 void pstats_fork(struct pstats *src, struct pstats *dst);
859 void pstats_free(struct pstats *ps);
860 int securelevel_ge(struct ucred *cr, int level);
861 int securelevel_gt(struct ucred *cr, int level);
862 void sess_hold(struct session *);
863 void sess_release(struct session *);
864 int setrunnable(struct thread *);
865 void setsugid(struct proc *p);
866 int should_yield(void);
867 int sigonstack(size_t sp);
868 void sleepinit(void);
869 void stopevent(struct proc *, u_int, u_int);
870 void threadinit(void);
872 int cpu_idle_wakeup(int);
873 extern void (*cpu_idle_hook)(void); /* Hook to machdep CPU idler. */
874 void cpu_switch(struct thread *, struct thread *, struct mtx *);
875 void cpu_throw(struct thread *, struct thread *) __dead2;
876 void unsleep(struct thread *);
877 void userret(struct thread *, struct trapframe *);
879 void cpu_exit(struct thread *);
880 void exit1(struct thread *, int) __dead2;
882 int cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa);
883 void cpu_fork(struct thread *, struct proc *, struct thread *, int);
884 void cpu_set_fork_handler(struct thread *, void (*)(void *), void *);
885 void cpu_set_syscall_retval(struct thread *, int);
886 void cpu_set_upcall(struct thread *td, struct thread *td0);
887 void cpu_set_upcall_kse(struct thread *, void (*)(void *), void *,
889 int cpu_set_user_tls(struct thread *, void *tls_base);
890 void cpu_thread_alloc(struct thread *);
891 void cpu_thread_clean(struct thread *);
892 void cpu_thread_exit(struct thread *);
893 void cpu_thread_free(struct thread *);
894 void cpu_thread_swapin(struct thread *);
895 void cpu_thread_swapout(struct thread *);
896 struct thread *thread_alloc(int pages);
897 int thread_alloc_stack(struct thread *, int pages);
898 void thread_exit(void) __dead2;
899 void thread_free(struct thread *td);
900 void thread_link(struct thread *td, struct proc *p);
901 void thread_reap(void);
902 int thread_single(int how);
903 void thread_single_end(void);
904 void thread_stash(struct thread *td);
905 void thread_stopped(struct proc *p);
906 void childproc_stopped(struct proc *child, int reason);
907 void childproc_continued(struct proc *child);
908 void childproc_exited(struct proc *child);
909 int thread_suspend_check(int how);
910 void thread_suspend_switch(struct thread *);
911 void thread_suspend_one(struct thread *td);
912 void thread_unlink(struct thread *td);
913 void thread_unsuspend(struct proc *p);
914 int thread_unsuspend_one(struct thread *td);
915 void thread_unthread(struct thread *td);
916 void thread_wait(struct proc *p);
917 struct thread *thread_find(struct proc *p, lwpid_t tid);
920 curthread_pflags_set(int flags)
926 save = ~flags | (td->td_pflags & flags);
927 td->td_pflags |= flags;
932 curthread_pflags_restore(int save)
935 curthread->td_pflags &= save;
940 #endif /* !_SYS_PROC_H_ */