1 PROC(4) File Formats and Configurations PROC(4)
2
3
4
5 NAME
6 proc - /proc, the process file system
7
8 DESCRIPTION
9 /proc is a file system that provides access to the state of each
10 process and light-weight process (lwp) in the system. The name of each
11 entry in the /proc directory is a decimal number corresponding to a
12 process-ID. These entries are themselves subdirectories. Access to
13 process state is provided by additional files contained within each
14 subdirectory; the hierarchy is described more completely below. In this
15 document, ``/proc file'' refers to a non-directory file within the
16 hierarchy rooted at /proc. The owner of each /proc file and
17 subdirectory is determined by the user-ID of the process.
18
19
20 /proc can be mounted on any mount point, in addition to the standard
21 /proc mount point, and can be mounted several places at once. Such
22 additional mounts are allowed in order to facilitate the confinement of
23 processes to subtrees of the file system via chroot(1M) and yet allow
24 such processes access to commands like ps(1).
25
26
27 Standard system calls are used to access /proc files: open(2),
28 close(2), read(2), and write(2) (including readv(2), writev(2),
29 pread(2), and pwrite(2)). Most files describe process state and can
30 only be opened for reading. ctl and lwpctl (control) files permit
31 manipulation of process state and can only be opened for writing. as
32 (address space) files contain the image of the running process and can
33 be opened for both reading and writing. An open for writing allows
34 process control; a read-only open allows inspection but not control. In
35 this document, we refer to the process as open for reading or writing
36 if any of its associated /proc files is open for reading or writing.
37
38
39 In general, more than one process can open the same /proc file at the
40 same time. Exclusive open is an advisory mechanism provided to allow
41 controlling processes to avoid collisions with each other. A process
42 can obtain exclusive control of a target process, with respect to other
43 cooperating processes, if it successfully opens any /proc file in the
44 target process for writing (the as or ctl files, or the lwpctl file of
45 any lwp) while specifying O_EXCL in the open(2). Such an open will fail
46 if the target process is already open for writing (that is, if an as,
47 ctl, or lwpctl file is already open for writing). There can be any
48 number of concurrent read-only opens; O_EXCL is ignored on opens for
49 reading. It is recommended that the first open for writing by a
50 controlling process use the O_EXCL flag; multiple controlling processes
51 usually result in chaos.
52
53
54 If a process opens one of its own /proc files for writing, the open
55 succeeds regardless of O_EXCL and regardless of whether some other
56 process has the process open for writing. Self-opens do not count when
57 another process attempts an exclusive open. (A process cannot exclude a
58 debugger by opening itself for writing and the application of a
59 debugger cannot prevent a process from opening itself.) All self-opens
60 for writing are forced to be close-on-exec (see the F_SETFD operation
61 of fcntl(2)).
62
63
64 Data may be transferred from or to any locations in the address space
65 of the traced process by applying lseek(2) to position the as file at
66 the virtual address of interest followed by read(2) or write(2) (or by
67 using pread(2) or pwrite(2) for the combined operation). The address-
68 map files /proc/pid/map and /proc/pid/xmap can be read to determine the
69 accessible areas (mappings) of the address space. I/O transfers may
70 span contiguous mappings. An I/O request extending into an unmapped
71 area is truncated at the boundary. A write request beginning at an
72 unmapped virtual address fails with EIO; a read request beginning at an
73 unmapped virtual address returns zero (an end-of-file indication).
74
75
76 Information and control operations are provided through additional
77 files. <procfs.h> contains definitions of data structures and message
78 formats used with these files. Some of these definitions involve the
79 use of sets of flags. The set types sigset_t, fltset_t, and sysset_t
80 correspond, respectively, to signal, fault, and system call
81 enumerations defined in <sys/signal.h>, <sys/fault.h>, and
82 <sys/syscall.h>. Each set type is large enough to hold flags for its
83 own enumeration. Although they are of different sizes, they have a
84 common structure and can be manipulated by these macros:
85
86 prfillset(&set); /* turn on all flags in set */
87 premptyset(&set); /* turn off all flags in set */
88 praddset(&set, flag); /* turn on the specified flag */
89 prdelset(&set, flag); /* turn off the specified flag */
90 r = prismember(&set, flag); /* != 0 iff flag is turned on */
91
92
93
94 One of prfillset() or premptyset() must be used to initialize set
95 before it is used in any other operation. flag must be a member of the
96 enumeration corresponding to set.
97
98
99 Every process contains at least one light-weight process, or lwp. Each
100 lwp represents a flow of execution that is independently scheduled by
101 the operating system. All lwps in a process share its address space as
102 well as many other attributes. Through the use of lwpctl and ctl files
103 as described below, it is possible to affect individual lwps in a
104 process or to affect all of them at once, depending on the operation.
105
106
107 When the process has more than one lwp, a representative lwp is chosen
108 by the system for certain process status files and control operations.
109 The representative lwp is a stopped lwp only if all of the process's
110 lwps are stopped; is stopped on an event of interest only if all of the
111 lwps are so stopped (excluding PR_SUSPENDED lwps); is in a PR_REQUESTED
112 stop only if there are no other events of interest to be found; or,
113 failing everything else, is in a PR_SUSPENDED stop (implying that the
114 process is deadlocked). See the description of the status file for
115 definitions of stopped states. See the PCSTOP control operation for the
116 definition of ``event of interest''.
117
118
119 The representative lwp remains fixed (it will be chosen again on the
120 next operation) as long as all of the lwps are stopped on events of
121 interest or are in a PR_SUSPENDED stop and the PCRUN control operation
122 is not applied to any of them.
123
124
125 When applied to the process control file, every /proc control operation
126 that must act on an lwp uses the same algorithm to choose which lwp to
127 act upon. Together with synchronous stopping (see PCSET), this enables
128 a debugger to control a multiple-lwp process using only the process-
129 level status and control files if it so chooses. More fine-grained
130 control can be achieved using the lwp-specific files.
131
132
133 The system supports two process data models, the traditional 32-bit
134 data model in which ints, longs and pointers are all 32 bits wide (the
135 ILP32 data model), and on some platforms the 64-bit data model in which
136 longs and pointers, but not ints, are 64 bits in width (the LP64 data
137 model). In the LP64 data model some system data types, notably size_t,
138 off_t, time_t and dev_t, grow from 32 bits to 64 bits as well.
139
140
141 The /proc interfaces described here are available to both 32-bit and
142 64-bit controlling processes. However, many operations attempted by a
143 32-bit controlling process on a 64-bit target process will fail with
144 EOVERFLOW because the address space range of a 32-bit process cannot
145 encompass a 64-bit process or because the data in some 64-bit system
146 data type cannot be compressed to fit into the corresponding 32-bit
147 type without loss of information. Operations that fail in this
148 circumstance include reading and writing the address space, reading the
149 address-map files, and setting the target process's registers. There is
150 no restriction on operations applied by a 64-bit process to either a
151 32-bit or a 64-bit target processes.
152
153
154 The format of the contents of any /proc file depends on the data model
155 of the observer (the controlling process), not on the data model of the
156 target process. A 64-bit debugger does not have to translate the
157 information it reads from a /proc file for a 32-bit process from 32-bit
158 format to 64-bit format. However, it usually has to be aware of the
159 data model of the target process. The pr_dmodel field of the status
160 files indicates the target process's data model.
161
162
163 To help deal with system data structures that are read from 32-bit
164 processes, a 64-bit controlling program can be compiled with the C
165 preprocessor symbol _SYSCALL32 defined before system header files are
166 included. This makes explicit 32-bit fixed-width data structures (like
167 cstruct stat32) visible to the 64-bit program. See types32.h(3HEAD).
168
169 DIRECTORY STRUCTURE
170 At the top level, the directory /proc contains entries each of which
171 names an existing process in the system. These entries are themselves
172 directories. Except where otherwise noted, the files described below
173 can be opened for reading only. In addition, if a process becomes a
174 zombie (one that has exited but whose parent has not yet performed a
175 wait(3C) upon it), most of its associated /proc files disappear from
176 the hierarchy; subsequent attempts to open them, or to read or write
177 files opened before the process exited, will elicit the error ENOENT.
178
179
180 Although process state and consequently the contents of /proc files can
181 change from instant to instant, a single read(2) of a /proc file is
182 guaranteed to return a sane representation of state; that is, the read
183 will be atomic with respect to the state of the process. No such
184 guarantee applies to successive reads applied to a /proc file for a
185 running process. In addition, atomicity is not guaranteed for I/O
186 applied to the as (address-space) file for a running process or for a
187 process whose address space contains memory shared by another running
188 process.
189
190
191 A number of structure definitions are used to describe the files. These
192 structures may grow by the addition of elements at the end in future
193 releases of the system and it is not legitimate for a program to assume
194 that they will not.
195
196 STRUCTURE OF /proc/pid
197 A given directory /proc/pid contains the following entries. A process
198 can use the invisible alias /proc/self if it wishes to open one of its
199 own /proc files (invisible in the sense that the name ``self'' does not
200 appear in a directory listing of /proc obtained from ls(1),
201 getdents(2), or readdir(3C)).
202
203 contracts
204 A directory containing references to the contracts held by the process.
205 Each entry is a symlink to the contract's directory under
206 /system/contract. See contract(4).
207
208 as
209 Contains the address-space image of the process; it can be opened for
210 both reading and writing. lseek(2) is used to position the file at the
211 virtual address of interest and then the address space can be examined
212 or changed through read(2) or write(2) (or by using pread(2) or
213 pwrite(2) for the combined operation).
214
215 ctl
216 A write-only file to which structured messages are written directing
217 the system to change some aspect of the process's state or control its
218 behavior in some way. The seek offset is not relevant when writing to
219 this file. Individual lwps also have associated lwpctl files in the lwp
220 subdirectories. A control message may be written either to the
221 process's ctl file or to a specific lwpctl file with operation-specific
222 effects. The effect of a control message is immediately reflected in
223 the state of the process visible through appropriate status and
224 information files. The types of control messages are described in
225 detail later. See CONTROL MESSAGES.
226
227 status
228 Contains state information about the process and the representative
229 lwp. The file contains a pstatus structure which contains an embedded
230 lwpstatus structure for the representative lwp, as follows:
231
232 typedef struct pstatus {
233 int pr_flags; /* flags (see below) */
234 int pr_nlwp; /* number of active lwps in the process */
235 int pr_nzomb; /* number of zombie lwps in the process */
236 pid_tpr_pid; /* process id */
237 pid_tpr_ppid; /* parent process id */
238 pid_tpr_pgid; /* process group id */
239 pid_tpr_sid; /* session id */
240 id_t pr_aslwpid; /* obsolete */
241 id_t pr_agentid; /* lwp-id of the agent lwp, if any */
242 sigset_t pr_sigpend; /* set of process pending signals */
243 uintptr_t pr_brkbase; /* virtual address of the process heap */
244 size_t pr_brksize; /* size of the process heap, in bytes */
245 uintptr_t pr_stkbase; /* virtual address of the process stack */
246 size_tpr_stksize; /* size of the process stack, in bytes */
247 timestruc_t pr_utime; /* process user cpu time */
248 timestruc_t pr_stime; /* process system cpu time */
249 timestruc_t pr_cutime; /* sum of children's user times */
250 timestruc_t pr_cstime; /* sum of children's system times */
251 sigset_t pr_sigtrace; /* set of traced signals */
252 fltset_t pr_flttrace; /* set of traced faults */
253 sysset_t pr_sysentry; /* set of system calls traced on entry */
254 sysset_t pr_sysexit; /* set of system calls traced on exit */
255 char pr_dmodel; /* data model of the process */
256 taskid_t pr_taskid; /* task id */
257 projid_t pr_projid; /* project id */
258 zoneid_t pr_zoneid; /* zone id */
259 lwpstatus_t pr_lwp; /* status of the representative lwp */
260 } pstatus_t;
261
262
263
264 pr_flags is a bit-mask holding the following process flags. For
265 convenience, it also contains the lwp flags for the representative lwp,
266 described later.
267
268 PR_ISSYS
269 process is a system process (see PCSTOP).
270
271
272 PR_VFORKP
273 process is the parent of a vforked child (see PCWATCH).
274
275
276 PR_FORK
277 process has its inherit-on-fork mode set (see PCSET).
278
279
280 PR_RLC
281 process has its run-on-last-close mode set (see PCSET).
282
283
284 PR_KLC
285 process has its kill-on-last-close mode set (see PCSET).
286
287
288 PR_ASYNC
289 process has its asynchronous-stop mode set (see PCSET).
290
291
292 PR_MSACCT
293 Set by default in all processes to indicate that
294 microstate accounting is enabled. However, this flag has
295 been deprecated and no longer has any effect. Microstate
296 accounting may not be disabled; however, it is still
297 possible to toggle the flag.
298
299
300 PR_MSFORK
301 Set by default in all processes to indicate that
302 microstate accounting will be enabled for processes that
303 this parent forks(). However, this flag has been
304 deprecated and no longer has any effect. It is possible to
305 toggle this flag; however, it is not possible to disable
306 microstate accounting.
307
308
309 PR_BPTADJ
310 process has its breakpoint adjustment mode set (see
311 PCSET).
312
313
314 PR_PTRACE
315 process has its ptrace-compatibility mode set (see PCSET).
316
317
318
319 pr_nlwp is the total number of active lwps in the process. pr_nzomb is
320 the total number of zombie lwps in the process. A zombie lwp is a non-
321 detached lwp that has terminated but has not been reaped with
322 thr_join(3C) or pthread_join(3C).
323
324
325 pr_pid, pr_ppid, pr_pgid, and pr_sid are, respectively, the process ID,
326 the ID of the process's parent, the process's process group ID, and the
327 process's session ID.
328
329
330 pr_aslwpid is obsolete and is always zero.
331
332
333 pr_agentid is the lwp-ID for the /proc agent lwp (see the PCAGENT
334 control operation). It is zero if there is no agent lwp in the process.
335
336
337 pr_sigpend identifies asynchronous signals pending for the process.
338
339
340 pr_brkbase is the virtual address of the process heap and pr_brksize is
341 its size in bytes. The address formed by the sum of these values is the
342 process break (see brk(2)). pr_stkbase and pr_stksize are,
343 respectively, the virtual address of the process stack and its size in
344 bytes. (Each lwp runs on a separate stack; the distinguishing
345 characteristic of the process stack is that the operating system will
346 grow it when necessary.)
347
348
349 pr_utime, pr_stime, pr_cutime, and pr_cstime are, respectively, the
350 user CPU and system CPU time consumed by the process, and the
351 cumulative user CPU and system CPU time consumed by the process's
352 children, in seconds and nanoseconds.
353
354
355 pr_sigtrace and pr_flttrace contain, respectively, the set of signals
356 and the set of hardware faults that are being traced (see PCSTRACE and
357 PCSFAULT).
358
359
360 pr_sysentry and pr_sysexit contain, respectively, the sets of system
361 calls being traced on entry and exit (see PCSENTRY and PCSEXIT).
362
363
364 pr_dmodel indicates the data model of the process. Possible values are:
365
366 PR_MODEL_ILP32
367 process data model is ILP32.
368
369
370 PR_MODEL_LP64
371 process data model is LP64.
372
373
374 PR_MODEL_NATIVE
375 process data model is native.
376
377
378
379 The pr_taskid, pr_projid, and pr_zoneid fields contain respectively,
380 the numeric IDs of the task, project, and zone in which the process was
381 running.
382
383
384 The constant PR_MODEL_NATIVE reflects the data model of the controlling
385 process, that is, its value is PR_MODEL_ILP32 or PR_MODEL_LP64
386 according to whether the controlling process has been compiled as a
387 32-bit program or a 64-bit program, respectively.
388
389
390 pr_lwp contains the status information for the representative lwp:
391
392 typedef struct lwpstatus {
393 int pr_flags; /* flags (see below) */
394 id_t pr_lwpid; /* specific lwp identifier */
395 short pr_why; /* reason for lwp stop, if stopped */
396 short pr_what; /* more detailed reason */
397 short pr_cursig; /* current signal, if any */
398 siginfo_t pr_info; /* info associated with signal or fault */
399 sigset_t pr_lwppend; /* set of signals pending to the lwp */
400 sigset_t pr_lwphold; /* set of signals blocked by the lwp */
401 struct sigaction pr_action;/* signal action for current signal */
402 stack_t pr_altstack; /* alternate signal stack info */
403 uintptr_t pr_oldcontext; /* address of previous ucontext */
404 short pr_syscall; /* system call number (if in syscall) */
405 short pr_nsysarg; /* number of arguments to this syscall */
406 int pr_errno; /* errno for failed syscall */
407 long pr_sysarg[PRSYSARGS]; /* arguments to this syscall */
408 long pr_rval1; /* primary syscall return value */
409 long pr_rval2; /* second syscall return value, if any */
410 char pr_clname[PRCLSZ]; /* scheduling class name */
411 timestruc_t pr_tstamp; /* real-time time stamp of stop */
412 timestruc_t pr_utime; /* lwp user cpu time */
413 timestruc_t pr_stime; /* lwp system cpu time */
414 uintptr_t pr_ustack; /* stack boundary data (stack_t) address */
415 ulong_t pr_instr; /* current instruction */
416 prgregset_t pr_reg; /* general registers */
417 prfpregset_t pr_fpreg; /* floating-point registers */
418 } lwpstatus_t;
419
420
421
422 pr_flags is a bit-mask holding the following lwp flags. For
423 convenience, it also contains the process flags, described previously.
424
425 PR_STOPPED
426 The lwp is stopped.
427
428
429 PR_ISTOP
430 The lwp is stopped on an event of interest (see PCSTOP).
431
432
433 PR_DSTOP
434 The lwp has a stop directive in effect (see PCSTOP).
435
436
437 PR_STEP
438 The lwp has a single-step directive in effect (see
439 PCRUN).
440
441
442 PR_ASLEEP
443 The lwp is in an interruptible sleep within a system
444 call.
445
446
447 PR_PCINVAL
448 The lwp's current instruction (pr_instr) is undefined.
449
450
451 PR_DETACH
452 This is a detached lwp (see pthread_create(3C) and
453 pthread_join(3C)).
454
455
456 PR_DAEMON
457 This is a daemon lwp (see pthread_create(3C)).
458
459
460 PR_ASLWP
461 This flag is obsolete and is never set.
462
463
464 PR_AGENT
465 This is the /proc agent lwp for the process.
466
467
468
469 pr_lwpid names the specific lwp.
470
471
472 pr_why and pr_what together describe, for a stopped lwp, the reason for
473 the stop. Possible values of pr_why and the associated pr_what are:
474
475 PR_REQUESTED
476 indicates that the stop occurred in response to a stop
477 directive, normally because PCSTOP was applied or
478 because another lwp stopped on an event of interest
479 and the asynchronous-stop flag (see PCSET) was not set
480 for the process. pr_what is unused in this case.
481
482
483 PR_SIGNALLED
484 indicates that the lwp stopped on receipt of a signal
485 (see PCSTRACE); pr_what holds the signal number that
486 caused the stop (for a newly-stopped lwp, the same
487 value is in pr_cursig).
488
489
490 PR_FAULTED
491 indicates that the lwp stopped on incurring a hardware
492 fault (see PCSFAULT); pr_what holds the fault number
493 that caused the stop.
494
495
496 PR_SYSENTRY
497 PR_SYSEXIT
498 indicate a stop on entry to or exit from a system call
499 (see PCSENTRY and PCSEXIT); pr_what holds the system
500 call number.
501
502
503 PR_JOBCONTROL
504 indicates that the lwp stopped due to the default
505 action of a job control stop signal (see
506 sigaction(2)); pr_what holds the stopping signal
507 number.
508
509
510 PR_SUSPENDED
511 indicates that the lwp stopped due to internal
512 synchronization of lwps within the process. pr_what is
513 unused in this case.
514
515
516
517 pr_cursig names the current signal, that is, the next signal to be
518 delivered to the lwp, if any. pr_info, when the lwp is in a
519 PR_SIGNALLED or PR_FAULTED stop, contains additional information
520 pertinent to the particular signal or fault (see <sys/siginfo.h>).
521
522
523 pr_lwppend identifies any synchronous or directed signals pending for
524 the lwp. pr_lwphold identifies those signals whose delivery is being
525 blocked by the lwp (the signal mask).
526
527
528 pr_action contains the signal action information pertaining to the
529 current signal (see sigaction(2)); it is undefined if pr_cursig is
530 zero. pr_altstack contains the alternate signal stack information for
531 the lwp (see sigaltstack(2)).
532
533
534 pr_oldcontext, if not zero, contains the address on the lwp stack of a
535 ucontext structure describing the previous user-level context (see
536 ucontext.h(3HEAD)). It is non-zero only if the lwp is executing in the
537 context of a signal handler.
538
539
540 pr_syscall is the number of the system call, if any, being executed by
541 the lwp; it is non-zero if and only if the lwp is stopped on
542 PR_SYSENTRY or PR_SYSEXIT, or is asleep within a system call (
543 PR_ASLEEP is set). If pr_syscall is non-zero, pr_nsysarg is the number
544 of arguments to the system call and pr_sysarg contains the actual
545 arguments.
546
547
548 pr_rval1, pr_rval2, and pr_errno are defined only if the lwp is stopped
549 on PR_SYSEXIT or if the PR_VFORKP flag is set. If pr_errno is zero,
550 pr_rval1 and pr_rval2 contain the return values from the system call.
551 Otherwise, pr_errno contains the error number for the failing system
552 call (see <sys/errno.h>).
553
554
555 pr_clname contains the name of the lwp's scheduling class.
556
557
558 pr_tstamp, if the lwp is stopped, contains a time stamp marking when
559 the lwp stopped, in real time seconds and nanoseconds since an
560 arbitrary time in the past.
561
562
563 pr_utime is the amount of user level CPU time used by this LWP.
564
565
566 pr_stime is the amount of system level CPU time used by this LWP.
567
568
569 pr_ustack is the virtual address of the stack_t that contains the stack
570 boundaries for this LWP. See getustack(2) and _stack_grow(3C).
571
572
573 pr_instr contains the machine instruction to which the lwp's program
574 counter refers. The amount of data retrieved from the process is
575 machine-dependent. On SPARC based machines, it is a 32-bit word. On
576 x86-based machines, it is a single byte. In general, the size is that
577 of the machine's smallest instruction. If PR_PCINVAL is set, pr_instr
578 is undefined; this occurs whenever the lwp is not stopped or when the
579 program counter refers to an invalid virtual address.
580
581
582 pr_reg is an array holding the contents of a stopped lwp's general
583 registers.
584
585 SPARC
586 On SPARC-based machines, the predefined constants
587 R_G0 ... R_G7, R_O0 ... R_O7, R_L0 ... R_L7, R_I0
588 ... R_I7, R_PC, R_nPC, and R_Y can be used as
589 indices to refer to the corresponding registers;
590 previous register windows can be read from their
591 overflow locations on the stack (however, see the
592 gwindows file in the /proc/pid/lwp/lwpid
593 subdirectory).
594
595
596 SPARC V8 (32-bit)
597 For SPARC V8 (32-bit) controlling processes, the
598 predefined constants R_PSR, R_WIM, and R_TBR can
599 be used as indices to refer to the corresponding
600 special registers. For SPARC V9 (64-bit)
601 controlling processes, the predefined constants
602 R_CCR, R_ASI, and R_FPRS can be used as indices to
603 refer to the corresponding special registers.
604
605
606 x86 (32-bit)
607 For 32-bit x86 processes, the predefined constants
608 listed belowcan be used as indices to refer to the
609 corresponding registers.
610
611 SS
612 UESP
613 EFL
614 CS
615 EIP
616 ERR
617 TRAPNO
618 EAX
619 ECX
620 EDX
621 EBX
622 ESP
623 EBP
624 ESI
625 EDI
626 DS
627 ES
628 GS
629
630 The preceding constants are listed in
631 <sys/regset.h>.
632
633 Note that a 32-bit process can run on an x86
634 64-bit system, using the constants listed above.
635
636
637 x86 (64-bit)
638 To read the registers of a 32- or a 64-bit
639 process, a 64-bit x86 process should use the
640 predefined constants listed below.
641
642 REG_GSBASE
643 REG_FSBASE
644 REG_DS
645 REG_ES
646 REG_GS
647 REG_FS
648 REG_SS
649 REG_RSP
650 REG_RFL
651 REG_CS
652 REG_RIP
653 REG_ERR
654 REG_TRAPNO
655 REG_RAX
656 REG_RCX
657 REG_RDX
658 REG_RBX
659 REG_RBP
660 REG_RSI
661 REG_RDI
662 REG_R8
663 REG_R9
664 REG_R10
665 REG_R11
666 REG_R12
667 REG_R13
668 REG_R14
669 REG_R15
670
671 The preceding constants are listed in
672 <sys/regset.h>.
673
674
675
676 pr_fpreg is a structure holding the contents of the floating-point
677 registers.
678
679
680 SPARC registers, both general and floating-point, as seen by a 64-bit
681 controlling process are the V9 versions of the registers, even if the
682 target process is a 32-bit (V8) process. V8 registers are a subset of
683 the V9 registers.
684
685
686 If the lwp is not stopped, all register values are undefined.
687
688 psinfo
689 Contains miscellaneous information about the process and the
690 representative lwp needed by the ps(1) command. psinfo remains
691 accessible after a process becomes a zombie. The file contains a psinfo
692 structure which contains an embedded lwpsinfo structure for the
693 representative lwp, as follows:
694
695 typedef struct psinfo {
696 int pr_flag; /* process flags (DEPRECATED: see below) */
697 int pr_nlwp; /* number of active lwps in the process */
698 int pr_nzomb; /* number of zombie lwps in the process */
699 pid_t pr_pid; /* process id */
700 pid_t pr_ppid; /* process id of parent */
701 pid_t pr_pgid; /* process id of process group leader */
702 pid_t pr_sid; /* session id */
703 uid_t pr_uid; /* real user id */
704 uid_t pr_euid; /* effective user id */
705 gid_t pr_gid; /* real group id */
706 gid_t pr_egid; /* effective group id */
707 uintptr_t pr_addr; /* address of process */
708 size_t pr_size; /* size of process image in Kbytes */
709 size_t pr_rssize; /* resident set size in Kbytes */
710 dev_t pr_ttydev; /* controlling tty device (or PRNODEV) */
711 ushort_t pr_pctcpu; /* % of recent cpu time used by all lwps */
712 ushort_t pr_pctmem; /* % of system memory used by process */
713 timestruc_t pr_start; /* process start time, from the epoch */
714 timestruc_t pr_time; /* cpu time for this process */
715 timestruc_t pr_ctime; /* cpu time for reaped children */
716 char pr_fname[PRFNSZ]; /* name of exec'ed file */
717 char pr_psargs[PRARGSZ]; /* initial characters of arg list */
718 int pr_wstat; /* if zombie, the wait() status */
719 int pr_argc; /* initial argument count */
720 uintptr_t pr_argv; /* address of initial argument vector */
721 uintptr_t pr_envp; /* address of initial environment vector */
722 char pr_dmodel; /* data model of the process */
723 lwpsinfo_t pr_lwp; /* information for representative lwp */
724 taskid_t pr_taskid; /* task id */
725 projid_t pr_projid; /* project id */
726 poolid_t pr_poolid; /* pool id */
727 zoneid_t pr_zoneid; /* zone id */
728 ctid_t pr_contract; /* process contract id */
729 } psinfo_t;
730
731
732
733 Some of the entries in psinfo, such as pr_addr, refer to internal
734 kernel data structures and should not be expected to retain their
735 meanings across different versions of the operating system.
736
737
738 psinfo_t.pr_flag is a deprecated interface that should no longer be
739 used. Applications currently relying on the SSYS bit in pr_flag should
740 migrate to checking PR_ISSYS in the pstatus structure's pr_flags field.
741
742
743 pr_pctcpu and pr_pctmem are 16-bit binary fractions in the range 0.0 to
744 1.0 with the binary point to the right of the high-order bit (1.0 ==
745 0x8000). pr_pctcpu is the summation over all lwps in the process.
746
747
748 pr_lwp contains the ps(1) information for the representative lwp. If
749 the process is a zombie, pr_nlwp, pr_nzomb, and pr_lwp.pr_lwpid are
750 zero and the other fields of pr_lwp are undefined:
751
752 typedef struct lwpsinfo {
753 int pr_flag; /* lwp flags (DEPRECATED: see below) */
754 id_t pr_lwpid; /* lwp id */
755 uintptr_t pr_addr; /* internal address of lwp */
756 uintptr_t pr_wchan; /* wait addr for sleeping lwp */
757 char pr_stype; /* synchronization event type */
758 char pr_state; /* numeric lwp state */
759 char pr_sname; /* printable character for pr_state */
760 char pr_nice; /* nice for cpu usage */
761 short pr_syscall; /* system call number (if in syscall) */
762 char pr_oldpri; /* pre-SVR4, low value is high priority */
763 char pr_cpu; /* pre-SVR4, cpu usage for scheduling */
764 int pr_pri; /* priority, high value = high priority */
765 ushort_t pr_pctcpu; /* % of recent cpu time used by this lwp */
766 timestruc_t pr_start; /* lwp start time, from the epoch */
767 timestruc_t pr_time; /* cpu time for this lwp */
768 char pr_clname[PRCLSZ]; /* scheduling class name */
769 char pr_name[PRFNSZ]; /* name of system lwp */
770 processorid_t pr_onpro; /* processor which last ran this lwp */
771 processorid_t pr_bindpro;/* processor to which lwp is bound */
772 psetid_t pr_bindpset; /* processor set to which lwp is bound */
773 lgrp_id_t pr_lgrp /* home lgroup */
774 } lwpsinfo_t;
775
776
777
778 Some of the entries in lwpsinfo, such as pr_addr, pr_wchan, pr_stype,
779 pr_state, and pr_name, refer to internal kernel data structures and
780 should not be expected to retain their meanings across different
781 versions of the operating system.
782
783
784 lwpsinfo_t.pr_flag is a deprecated interface that should no longer be
785 used.
786
787
788 pr_pctcpu is a 16-bit binary fraction, as described above. It
789 represents the CPU time used by the specific lwp. On a multi-processor
790 machine, the maximum value is 1/N, where N is the number of CPUs.
791
792
793 pr_contract is the id of the process contract of which the process is a
794 member. See contract(4) and process(4).
795
796 cred
797 Contains a description of the credentials associated with the process:
798
799 typedef struct prcred {
800 uid_t pr_euid; /* effective user id */
801 uid_t pr_ruid; /* real user id */
802 uid_t pr_suid; /* saved user id (from exec) */
803 gid_t pr_egid; /* effective group id */
804 gid_t pr_rgid; /* real group id */
805 gid_t pr_sgid; /* saved group id (from exec) */
806 int pr_ngroups; /* number of supplementary groups */
807 gid_t pr_groups[1]; /* array of supplementary groups */
808 } prcred_t;
809
810
811
812
813 The array of associated supplementary groups in pr_groups is of
814 variable length; the cred file contains all of the supplementary
815 groups. pr_ngroups indicates the number of supplementary groups. (See
816 also the PCSCRED and PCSCREDX control operations.)
817
818 priv
819 Contains a description of the privileges associated with the process:
820
821 typedef struct prpriv {
822 uint32_t pr_nsets; /* number of privilege set */
823 uint32_t pr_setsize; /* size of privilege set */
824 uint32_t pr_infosize; /* size of supplementary data */
825 priv_chunk_t pr_sets[1]; /* array of sets */
826 } prpriv_t;
827
828
829
830 The actual dimension of the pr_sets[] field is
831
832 pr_sets[pr_nsets][pr_setsize]
833
834
835
836 which is followed by additional information about the process state
837 pr_infosize bytes in size.
838
839
840 The full size of the structure can be computed using
841 PRIV_PRPRIV_SIZE(prpriv_t *).
842
843 secflags
844 This file contains the security-flags of the process. It contains a
845 description of the security flags associated with the process.
846
847 typedef struct prsecflags {
848 uint32_t pr_version; /* ABI Versioning of this structure */
849 secflagset_t pr_effective; /* Effective flags */
850 secflagset_t pr_inherit; /* Inheritable flags */
851 secflagset_t pr_lower; /* Lower flags */
852 secflagset_t pr_upper; /* Upper flags */
853 } prsecflags_t;
854
855
856
857 The pr_version field is a version number for the structure, currently
858 PRSECFLAGS_VERSION_1.
859
860 sigact
861 Contains an array of sigaction structures describing the current
862 dispositions of all signals associated with the traced process (see
863 sigaction(2)). Signal numbers are displaced by 1 from array indices, so
864 that the action for signal number n appears in position n-1 of the
865 array.
866
867 auxv
868 Contains the initial values of the process's aux vector in an array of
869 auxv_t structures (see <sys/auxv.h>). The values are those that were
870 passed by the operating system as startup information to the dynamic
871 linker.
872
873 ldt
874 This file exists only on x86-based machines. It is non-empty only if
875 the process has established a local descriptor table (LDT). If non-
876 empty, the file contains the array of currently active LDT entries in
877 an array of elements of type struct ssd, defined in <sys/sysi86.h>, one
878 element for each active LDT entry.
879
880 map, xmap
881 Contain information about the virtual address map of the process. The
882 map file contains an array of prmap structures while the xmap file
883 contains an array of prxmap structures. Each structure describes a
884 contiguous virtual address region in the address space of the traced
885 process:
886
887 typedef struct prmap {
888 uintptr_tpr_vaddr; /* virtual address of mapping */
889 size_t pr_size; /* size of mapping in bytes */
890 char pr_mapname[PRMAPSZ]; /* name in /proc/pid/object */
891 offset_t pr_offset; /* offset into mapped object, if any */
892 int pr_mflags; /* protection and attribute flags */
893 int pr_pagesize; /* pagesize for this mapping in bytes */
894 int pr_shmid; /* SysV shared memory identifier */
895 } prmap_t;
896
897
898
899 typedef struct prxmap {
900 uintptr_t pr_vaddr; /* virtual address of mapping */
901 size_t pr_size; /* size of mapping in bytes */
902 char pr_mapname[PRMAPSZ]; /* name in /proc/pid/object */
903 offset_t pr_offset; /* offset into mapped object, if any */
904 int pr_mflags; /* protection and attribute flags */
905 int pr_pagesize; /* pagesize for this mapping in bytes */
906 int pr_shmid; /* SysV shared memory identifier */
907 dev_t pr_dev; /* device of mapped object, if any */
908 uint64_t pr_ino; /* inode of mapped object, if any */
909 size_t pr_rss; /* pages of resident memory */
910 size_t pr_anon; /* pages of resident anonymous memory */
911 size_t pr_locked; /* pages of locked memory */
912 uint64_t pr_hatpagesize; /* pagesize of mapping */
913 } prxmap_t;
914
915
916
917
918 pr_vaddr is the virtual address of the mapping within the traced
919 process and pr_size is its size in bytes. pr_mapname, if it does not
920 contain a null string, contains the name of a file in the object
921 directory (see below) that can be opened read-only to obtain a file
922 descriptor for the mapped file associated with the mapping. This
923 enables a debugger to find object file symbol tables without having to
924 know the real path names of the executable file and shared libraries of
925 the process. pr_offset is the 64-bit offset within the mapped file (if
926 any) to which the virtual address is mapped.
927
928
929 pr_mflags is a bit-mask of protection and attribute flags:
930
931 MA_READ
932 mapping is readable by the traced process.
933
934
935 MA_WRITE
936 mapping is writable by the traced process.
937
938
939 MA_EXEC
940 mapping is executable by the traced process.
941
942
943 MA_SHARED
944 mapping changes are shared by the mapped object.
945
946
947 MA_ISM
948 mapping is intimate shared memory (shared MMU
949 resources)
950
951
952 MAP_NORESERVE
953 mapping does not have swap space reserved (mapped with
954 MAP_NORESERVE)
955
956
957 MA_SHM
958 mapping System V shared memory
959
960
961
962 A contiguous area of the address space having the same underlying
963 mapped object may appear as multiple mappings due to varying read,
964 write, and execute attributes. The underlying mapped object does not
965 change over the range of a single mapping. An I/O operation to a
966 mapping marked MA_SHARED fails if applied at a virtual address not
967 corresponding to a valid page in the underlying mapped object. A write
968 to a MA_SHARED mapping that is not marked MA_WRITE fails. Reads and
969 writes to private mappings always succeed. Reads and writes to unmapped
970 addresses fail.
971
972
973 pr_pagesize is the page size for the mapping, currently always the
974 system pagesize.
975
976
977 pr_shmid is the shared memory identifier, if any, for the mapping. Its
978 value is -1 if the mapping is not System V shared memory. See
979 shmget(2).
980
981
982 pr_dev is the device of the mapped object, if any, for the mapping. Its
983 value is PRNODEV (-1) if the mapping does not have a device.
984
985
986 pr_ino is the inode of the mapped object, if any, for the mapping. Its
987 contents are only valid if pr_dev is not PRNODEV.
988
989
990 pr_rss is the number of resident pages of memory for the mapping. The
991 number of resident bytes for the mapping may be determined by
992 multiplying pr_rss by the page size given by pr_pagesize.
993
994
995 pr_anon is the number of resident anonymous memory pages (pages which
996 are private to this process) for the mapping.
997
998
999 pr_locked is the number of locked pages for the mapping. Pages which
1000 are locked are always resident in memory.
1001
1002
1003 pr_hatpagesize is the size, in bytes, of the HAT (MMU) translation for
1004 the mapping. pr_hatpagesize may be different than pr_pagesize. The
1005 possible values are hardware architecture specific, and may change over
1006 a mapping's lifetime.
1007
1008 rmap
1009 Contains information about the reserved address ranges of the process.
1010 The file contains an array of prmap structures, as defined above for
1011 the map file. Each structure describes a contiguous virtual address
1012 region in the address space of the traced process that is reserved by
1013 the system in the sense that an mmap(2) system call that does not
1014 specify MAP_FIXED will not use any part of it for the new mapping.
1015 Examples of such reservations include the address ranges reserved for
1016 the process stack and the individual thread stacks of a multi-threaded
1017 process.
1018
1019 cwd
1020 A symbolic link to the process's current working directory. See
1021 chdir(2). A readlink(2) of /proc/pid/cwd yields a null string.
1022 However, it can be opened, listed, and searched as a directory, and can
1023 be the target of chdir(2).
1024
1025 root
1026 A symbolic link to the process's root directory. /proc/pid/root can
1027 differ from the system root directory if the process or one of its
1028 ancestors executed chroot(2) as super user. It has the same semantics
1029 as /proc/pid/cwd.
1030
1031 fd
1032 A directory containing references to the open files of the process.
1033 Each entry is a decimal number corresponding to an open file descriptor
1034 in the process.
1035
1036
1037 If an entry refers to a regular file, it can be opened with normal file
1038 system semantics but, to ensure that the controlling process cannot
1039 gain greater access than the controlled process, with no file access
1040 modes other than its read/write open modes in the controlled process.
1041 If an entry refers to a directory, it can be accessed with the same
1042 semantics as /proc/pid/cwd. An attempt to open any other type of entry
1043 fails with EACCES.
1044
1045 object
1046 A directory containing read-only files with names corresponding to the
1047 pr_mapname entries in the map and pagedata files. Opening such a file
1048 yields a file descriptor for the underlying mapped file associated with
1049 an address-space mapping in the process. The file name a.out appears in
1050 the directory as an alias for the process's executable file.
1051
1052
1053 The object directory makes it possible for a controlling process to
1054 gain access to the object file and any shared libraries (and
1055 consequently the symbol tables) without having to know the actual path
1056 names of the executable files.
1057
1058 path
1059 A directory containing symbolic links to files opened by the process.
1060 The directory includes one entry for cwd and root. The directory also
1061 contains a numerical entry for each file descriptor in the fd
1062 directory, and entries matching those in the object directory. If this
1063 information is not available, any attempt to read the contents of the
1064 symbolic link will fail. This is most common for files that do not
1065 exist in the filesystem namespace (such as FIFOs and sockets), but can
1066 also happen for regular files. For the file descriptor entries, the
1067 path may be different from the one used by the process to open the
1068 file.
1069
1070 pagedata
1071 Opening the page data file enables tracking of address space references
1072 and modifications on a per-page basis.
1073
1074
1075 A read(2) of the page data file descriptor returns structured page data
1076 and atomically clears the page data maintained for the file by the
1077 system. That is to say, each read returns data collected since the last
1078 read; the first read returns data collected since the file was opened.
1079 When the call completes, the read buffer contains the following
1080 structure as its header and thereafter contains a number of section
1081 header structures and associated byte arrays that must be accessed by
1082 walking linearly through the buffer.
1083
1084 typedef struct prpageheader {
1085 timestruc_t pr_tstamp; /* real time stamp, time of read() */
1086 ulong_t pr_nmap; /* number of address space mappings */
1087 ulong_t pr_npage; /* total number of pages */
1088 } prpageheader_t;
1089
1090
1091
1092 The header is followed by pr_nmap prasmap structures and associated
1093 data arrays. The prasmap structure contains the following elements:
1094
1095 typedef struct prasmap {
1096 uintptr_t pr_vaddr; /* virtual address of mapping */
1097 ulong_t pr_npage; /* number of pages in mapping */
1098 char pr_mapname[PRMAPSZ]; /* name in /proc/pid/object */
1099 offset_t pr_offset; /* offset into mapped object, if any */
1100 int pr_mflags; /* protection and attribute flags */
1101 int pr_pagesize; /* pagesize for this mapping in bytes */
1102 int pr_shmid; /* SysV shared memory identifier */
1103 } prasmap_t;
1104
1105
1106
1107 Each section header is followed by pr_npage bytes, one byte for each
1108 page in the mapping, plus 0-7 null bytes at the end so that the next
1109 prasmap structure begins on an eight-byte aligned boundary. Each data
1110 byte may contain these flags:
1111
1112 PG_REFERENCED
1113 page has been referenced.
1114
1115
1116 PG_MODIFIED
1117 page has been modified.
1118
1119
1120
1121 If the read buffer is not large enough to contain all of the page data,
1122 the read fails with E2BIG and the page data is not cleared. The
1123 required size of the read buffer can be determined through fstat(2).
1124 Application of lseek(2) to the page data file descriptor is
1125 ineffective; every read starts from the beginning of the file. Closing
1126 the page data file descriptor terminates the system overhead associated
1127 with collecting the data.
1128
1129
1130 More than one page data file descriptor for the same process can be
1131 opened, up to a system-imposed limit per traced process. A read of one
1132 does not affect the data being collected by the system for the others.
1133 An open of the page data file will fail with ENOMEM if the system-
1134 imposed limit would be exceeded.
1135
1136 watch
1137 Contains an array of prwatch structures, one for each watched area
1138 established by the PCWATCH control operation. See PCWATCH for details.
1139
1140 usage
1141 Contains process usage information described by a prusage structure
1142 which contains at least the following fields:
1143
1144 typedef struct prusage {
1145 id_t pr_lwpid; /* lwp id. 0: process or defunct */
1146 int pr_count; /* number of contributing lwps */
1147 timestruc_t pr_tstamp; /* real time stamp, time of read() */
1148 timestruc_t pr_create; /* process/lwp creation time stamp */
1149 timestruc_t pr_term; /* process/lwp termination time stamp */
1150 timestruc_t pr_rtime; /* total lwp real (elapsed) time */
1151 timestruc_t pr_utime; /* user level CPU time */
1152 timestruc_t pr_stime; /* system call CPU time */
1153 timestruc_t pr_ttime; /* other system trap CPU time */
1154 timestruc_t pr_tftime; /* text page fault sleep time */
1155 timestruc_t pr_dftime; /* data page fault sleep time */
1156 timestruc_t pr_kftime; /* kernel page fault sleep time */
1157 timestruc_t pr_ltime; /* user lock wait sleep time */
1158 timestruc_t pr_slptime; /* all other sleep time */
1159 timestruc_t pr_wtime; /* wait-cpu (latency) time */
1160 timestruc_t pr_stoptime; /* stopped time */
1161 ulong_t pr_minf; /* minor page faults */
1162 ulong_t pr_majf; /* major page faults */
1163 ulong_t pr_nswap; /* swaps */
1164 ulong_t pr_inblk; /* input blocks */
1165 ulong_t pr_oublk; /* output blocks */
1166 ulong_t pr_msnd; /* messages sent */
1167 ulong_t pr_mrcv; /* messages received */
1168 ulong_t pr_sigs; /* signals received */
1169 ulong_t pr_vctx; /* voluntary context switches */
1170 ulong_t pr_ictx; /* involuntary context switches */
1171 ulong_t pr_sysc; /* system calls */
1172 ulong_t pr_ioch; /* chars read and written */
1173 } prusage_t;
1174
1175
1176
1177 Microstate accounting is now continuously enabled. While this
1178 information was previously an estimate, if microstate accounting were
1179 not enabled, the current information is now never an estimate
1180 represents time the process has spent in various states.
1181
1182 lstatus
1183 Contains a prheader structure followed by an array of lwpstatus
1184 structures, one for each active lwp in the process (see also
1185 /proc/pid/lwp/lwpid/lwpstatus, below). The prheader structure describes
1186 the number and size of the array entries that follow.
1187
1188 typedef struct prheader {
1189 long pr_nent; /* number of entries */
1190 size_t pr_entsize; /* size of each entry, in bytes */
1191 } prheader_t;
1192
1193
1194
1195 The lwpstatus structure may grow by the addition of elements at the end
1196 in future releases of the system. Programs must use pr_entsize in the
1197 file header to index through the array. These comments apply to all
1198 /proc files that include a prheader structure (lpsinfo and lusage,
1199 below).
1200
1201 lpsinfo
1202 Contains a prheader structure followed by an array of lwpsinfo
1203 structures, one for eachactive and zombie lwp in the process. See also
1204 /proc/pid/lwp/lwpid/lwpsinfo, below.
1205
1206 lusage
1207 Contains a prheader structure followed by an array of prusage
1208 structures, one for each active lwp in the process, plus an additional
1209 element at the beginning that contains the summation over all defunct
1210 lwps (lwps that once existed but no longer exist in the process).
1211 Excluding the pr_lwpid, pr_tstamp, pr_create, and pr_term entries, the
1212 entry-by-entry summation over all these structures is the definition of
1213 the process usage information obtained from the usage file. (See also
1214 /proc/pid/lwp/lwpid/lwpusage, below.)
1215
1216 lwp
1217 A directory containing entries each of which names an active or zombie
1218 lwp within the process. These entries are themselves directories
1219 containing additional files as described below. Only the lwpsinfo file
1220 exists in the directory of a zombie lwp.
1221
1222 STRUCTURE OF /proc/pid/lwp/lwpid
1223 A given directory /proc/pid/lwp/lwpid contains the following entries:
1224
1225 lwpctl
1226 Write-only control file. The messages written to this file affect the
1227 specific lwp rather than the representative lwp, as is the case for the
1228 process's ctl file.
1229
1230 lwpstatus
1231 lwp-specific state information. This file contains the lwpstatus
1232 structure for the specific lwp as described above for the
1233 representative lwp in the process's status file.
1234
1235 lwpsinfo
1236 lwp-specific ps(1) information. This file contains the lwpsinfo
1237 structure for the specific lwp as described above for the
1238 representative lwp in the process's psinfo file. The lwpsinfo file
1239 remains accessible after an lwp becomes a zombie.
1240
1241 lwpusage
1242 This file contains the prusage structure for the specific lwp as
1243 described above for the process's usage file.
1244
1245 gwindows
1246 This file exists only on SPARC based machines. If it is non-empty, it
1247 contains a gwindows_t structure, defined in <sys/regset.h>, with the
1248 values of those SPARC register windows that could not be stored on the
1249 stack when the lwp stopped. Conditions under which register windows are
1250 not stored on the stack are: the stack pointer refers to nonexistent
1251 process memory or the stack pointer is improperly aligned. If the lwp
1252 is not stopped or if there are no register windows that could not be
1253 stored on the stack, the file is empty (the usual case).
1254
1255 xregs
1256 Extra state registers. The extra state register set is architecture
1257 dependent; this file is empty if the system does not support extra
1258 state registers. If the file is non-empty, it contains an architecture
1259 dependent structure of type prxregset_t, defined in <procfs.h>, with
1260 the values of the lwp's extra state registers. If the lwp is not
1261 stopped, all register values are undefined. See also the PCSXREG
1262 control operation, below.
1263
1264 asrs
1265 This file exists only for 64-bit SPARC V9 processes. It contains an
1266 asrset_t structure, defined in <sys/regset.h>, containing the values of
1267 the lwp's platform-dependent ancillary state registers. If the lwp is
1268 not stopped, all register values are undefined. See also the PCSASRS
1269 control operation, below.
1270
1271 spymaster
1272 For an agent lwp (see PCAGENT), this file contains a psinfo_t structure
1273 that corresponds to the process that created the agent lwp at the time
1274 the agent was created. This structure is identical to that retrieved
1275 via the psinfo file, with one modification: the pr_time field does not
1276 correspond to the CPU time for the process, but rather to the creation
1277 time of the agent lwp.
1278
1279 templates
1280 A directory which contains references to the active templates for the
1281 lwp, named by the contract type. Changes made to an active template
1282 descriptor do not affect the original template which was activated,
1283 though they do affect the active template. It is not possible to
1284 activate an active template descriptor. See contract(4).
1285
1286 CONTROL MESSAGES
1287 Process state changes are effected through messages written to a
1288 process's ctl file or to an individual lwp's lwpctl file. All control
1289 messages consist of a long that names the specific operation followed
1290 by additional data containing the operand, if any.
1291
1292
1293 Multiple control messages may be combined in a single write(2) (or
1294 writev(2)) to a control file, but no partial writes are permitted. That
1295 is, each control message, operation code plus operand, if any, must be
1296 presented in its entirety to the write(2) and not in pieces over
1297 several system calls. If a control operation fails, no subsequent
1298 operations contained in the same write(2) are attempted.
1299
1300
1301 Descriptions of the allowable control messages follow. In all cases,
1302 writing a message to a control file for a process or lwp that has
1303 terminated elicits the error ENOENT.
1304
1305 PCSTOP PCDSTOP PCWSTOP PCTWSTOP
1306 When applied to the process control file, PCSTOP directs all lwps to
1307 stop and waits for them to stop, PCDSTOP directs all lwps to stop
1308 without waiting for them to stop, and PCWSTOP simply waits for all lwps
1309 to stop. When applied to an lwp control file, PCSTOP directs the
1310 specific lwp to stop and waits until it has stopped, PCDSTOP directs
1311 the specific lwp to stop without waiting for it to stop, and PCWSTOP
1312 simply waits for the specific lwp to stop. When applied to an lwp
1313 control file, PCSTOP and PCWSTOP complete when the lwp stops on an
1314 event of interest, immediately if already so stopped; when applied to
1315 the process control file, they complete when every lwp has stopped
1316 either on an event of interest or on a PR_SUSPENDED stop.
1317
1318
1319 PCTWSTOP is identical to PCWSTOP except that it enables the operation
1320 to time out, to avoid waiting forever for a process or lwp that may
1321 never stop on an event of interest. PCTWSTOP takes a long operand
1322 specifying a number of milliseconds; the wait will terminate
1323 successfully after the specified number of milliseconds even if the
1324 process or lwp has not stopped; a timeout value of zero makes the
1325 operation identical to PCWSTOP.
1326
1327
1328 An ``event of interest'' is either a PR_REQUESTED stop or a stop that
1329 has been specified in the process's tracing flags (set by PCSTRACE,
1330 PCSFAULT, PCSENTRY, and PCSEXIT). PR_JOBCONTROL and PR_SUSPENDED stops
1331 are specifically not events of interest. (An lwp may stop twice due to
1332 a stop signal, first showing PR_SIGNALLED if the signal is traced and
1333 again showing PR_JOBCONTROL if the lwp is set running without clearing
1334 the signal.) If PCSTOP or PCDSTOP is applied to an lwp that is stopped,
1335 but not on an event of interest, the stop directive takes effect when
1336 the lwp is restarted by the competing mechanism. At that time, the lwp
1337 enters a PR_REQUESTED stop before executing any user-level code.
1338
1339
1340 A write of a control message that blocks is interruptible by a signal
1341 so that, for example, an alarm(2) can be set to avoid waiting forever
1342 for a process or lwp that may never stop on an event of interest. If
1343 PCSTOP is interrupted, the lwp stop directives remain in effect even
1344 though the write(2) returns an error. (Use of PCTWSTOP with a non-zero
1345 timeout is recommended over PCWSTOP with an alarm(2).)
1346
1347
1348 A system process (indicated by the PR_ISSYS flag) never executes at
1349 user level, has no user-level address space visible through /proc, and
1350 cannot be stopped. Applying one of these operations to a system process
1351 or any of its lwps elicits the error EBUSY.
1352
1353 PCRUN
1354 Make an lwp runnable again after a stop. This operation takes a long
1355 operand containing zero or more of the following flags:
1356
1357 PRCSIG
1358 clears the current signal, if any (see PCCSIG).
1359
1360
1361 PRCFAULT
1362 clears the current fault, if any (see PCCFAULT).
1363
1364
1365 PRSTEP
1366 directs the lwp to execute a single machine instruction. On
1367 completion of the instruction, a trace trap occurs. If
1368 FLTTRACE is being traced, the lwp stops; otherwise, it is
1369 sent SIGTRAP. If SIGTRAP is being traced and is not
1370 blocked, the lwp stops. When the lwp stops on an event of
1371 interest, the single-step directive is cancelled, even if
1372 the stop occurs before the instruction is executed. This
1373 operation requires hardware and operating system support
1374 and may not be implemented on all processors. It is
1375 implemented on SPARC and x86-based machines.
1376
1377
1378 PRSABORT
1379 is meaningful only if the lwp is in a PR_SYSENTRY stop or
1380 is marked PR_ASLEEP; it instructs the lwp to abort
1381 execution of the system call (see PCSENTRY and PCSEXIT).
1382
1383
1384 PRSTOP
1385 directs the lwp to stop again as soon as possible after
1386 resuming execution (see PCDSTOP). In particular, if the lwp
1387 is stopped on PR_SIGNALLED or PR_FAULTED, the next stop
1388 will show PR_REQUESTED, no other stop will have intervened,
1389 and the lwp will not have executed any user-level code.
1390
1391
1392
1393 When applied to an lwp control file, PCRUN clears any outstanding
1394 directed-stop request and makes the specific lwp runnable. The
1395 operation fails with EBUSY if the specific lwp is not stopped on an
1396 event of interest or has not been directed to stop or if the agent lwp
1397 exists and this is not the agent lwp (see PCAGENT).
1398
1399
1400 When applied to the process control file, a representative lwp is
1401 chosen for the operation as described for /proc/pid/status. The
1402 operation fails with EBUSY if the representative lwp is not stopped on
1403 an event of interest or has not been directed to stop or if the agent
1404 lwp exists. If PRSTEP or PRSTOP was requested, the representative lwp
1405 is made runnable and its outstanding directed-stop request is cleared;
1406 otherwise all outstanding directed-stop requests are cleared and, if it
1407 was stopped on an event of interest, the representative lwp is marked
1408 PR_REQUESTED. If, as a consequence, all lwps are in the PR_REQUESTED or
1409 PR_SUSPENDED stop state, all lwps showing PR_REQUESTED are made
1410 runnable.
1411
1412 PCSTRACE
1413 Define a set of signals to be traced in the process. The receipt of one
1414 of these signals by an lwp causes the lwp to stop. The set of signals
1415 is defined using an operand sigset_t contained in the control message.
1416 Receipt of SIGKILL cannot be traced; if specified, it is silently
1417 ignored.
1418
1419
1420 If a signal that is included in an lwp's held signal set (the signal
1421 mask) is sent to the lwp, the signal is not received and does not cause
1422 a stop until it is removed from the held signal set, either by the lwp
1423 itself or by setting the held signal set with PCSHOLD.
1424
1425 PCCSIG
1426 The current signal, if any, is cleared from the specific or
1427 representative lwp.
1428
1429 PCSSIG
1430 The current signal and its associated signal information for the
1431 specific or representative lwp are set according to the contents of the
1432 operand siginfo structure (see <sys/siginfo.h>). If the specified
1433 signal number is zero, the current signal is cleared. The semantics of
1434 this operation are different from those of kill(2) in that the signal
1435 is delivered to the lwp immediately after execution is resumed (even if
1436 it is being blocked) and an additional PR_SIGNALLED stop does not
1437 intervene even if the signal is traced. Setting the current signal to
1438 SIGKILL terminates the process immediately.
1439
1440 PCKILL
1441 If applied to the process control file, a signal is sent to the process
1442 with semantics identical to those of kill(2). If applied to an lwp
1443 control file, a directed signal is sent to the specific lwp. The signal
1444 is named in a long operand contained in the message. Sending SIGKILL
1445 terminates the process immediately.
1446
1447 PCUNKILL
1448 A signal is deleted, that is, it is removed from the set of pending
1449 signals. If applied to the process control file, the signal is deleted
1450 from the process's pending signals. If applied to an lwp control file,
1451 the signal is deleted from the lwp's pending signals. The current
1452 signal (if any) is unaffected. The signal is named in a long operand in
1453 the control message. It is an error (EINVAL) to attempt to delete
1454 SIGKILL.
1455
1456 PCSHOLD
1457 Set the set of held signals for the specific or representative lwp
1458 (signals whose delivery will be blocked if sent to the lwp). The set of
1459 signals is specified with a sigset_t operand. SIGKILL and SIGSTOP
1460 cannot be held; if specified, they are silently ignored.
1461
1462 PCSFAULT
1463 Define a set of hardware faults to be traced in the process. On
1464 incurring one of these faults, an lwp stops. The set is defined via the
1465 operand fltset_t structure. Fault names are defined in <sys/fault.h>
1466 and include the following. Some of these may not occur on all
1467 processors; there may be processor-specific faults in addition to
1468 these.
1469
1470 FLTILL
1471 illegal instruction
1472
1473
1474 FLTPRIV
1475 privileged instruction
1476
1477
1478 FLTBPT
1479 breakpoint trap
1480
1481
1482 FLTTRACE
1483 trace trap (single-step)
1484
1485
1486 FLTWATCH
1487 watchpoint trap
1488
1489
1490 FLTACCESS
1491 memory access fault (bus error)
1492
1493
1494 FLTBOUNDS
1495 memory bounds violation
1496
1497
1498 FLTIOVF
1499 integer overflow
1500
1501
1502 FLTIZDIV
1503 integer zero divide
1504
1505
1506 FLTFPE
1507 floating-point exception
1508
1509
1510 FLTSTACK
1511 unrecoverable stack fault
1512
1513
1514 FLTPAGE
1515 recoverable page fault
1516
1517
1518
1519 When not traced, a fault normally results in the posting of a signal to
1520 the lwp that incurred the fault. If an lwp stops on a fault, the signal
1521 is posted to the lwp when execution is resumed unless the fault is
1522 cleared by PCCFAULT or by the PRCFAULT option of PCRUN. FLTPAGE is an
1523 exception; no signal is posted. The pr_info field in the lwpstatus
1524 structure identifies the signal to be sent and contains machine-
1525 specific information about the fault.
1526
1527 PCCFAULT
1528 The current fault, if any, is cleared; the associated signal will not
1529 be sent to the specific or representative lwp.
1530
1531 PCSENTRY PCSEXIT
1532 These control operations instruct the process's lwps to stop on entry
1533 to or exit from specified system calls. The set of system calls to be
1534 traced is defined via an operand sysset_t structure.
1535
1536
1537 When entry to a system call is being traced, an lwp stops after having
1538 begun the call to the system but before the system call arguments have
1539 been fetched from the lwp. When exit from a system call is being
1540 traced, an lwp stops on completion of the system call just prior to
1541 checking for signals and returning to user level. At this point, all
1542 return values have been stored into the lwp's registers.
1543
1544
1545 If an lwp is stopped on entry to a system call (PR_SYSENTRY) or when
1546 sleeping in an interruptible system call (PR_ASLEEP is set), it may be
1547 instructed to go directly to system call exit by specifying the
1548 PRSABORT flag in a PCRUN control message. Unless exit from the system
1549 call is being traced, the lwp returns to user level showing EINTR.
1550
1551 PCWATCH
1552 Set or clear a watched area in the controlled process from a prwatch
1553 structure operand:
1554
1555 typedef struct prwatch {
1556 uintptr_t pr_vaddr; /* virtual address of watched area */
1557 size_t pr_size; /* size of watched area in bytes */
1558 int pr_wflags; /* watch type flags */
1559 } prwatch_t;
1560
1561
1562
1563 pr_vaddr specifies the virtual address of an area of memory to be
1564 watched in the controlled process. pr_size specifies the size of the
1565 area, in bytes. pr_wflags specifies the type of memory access to be
1566 monitored as a bit-mask of the following flags:
1567
1568 WA_READ
1569 read access
1570
1571
1572 WA_WRITE
1573 write access
1574
1575
1576 WA_EXEC
1577 execution access
1578
1579
1580 WA_TRAPAFTER
1581 trap after the instruction completes
1582
1583
1584
1585 If pr_wflags is non-empty, a watched area is established for the
1586 virtual address range specified by pr_vaddr and pr_size. If pr_wflags
1587 is empty, any previously-established watched area starting at the
1588 specified virtual address is cleared; pr_size is ignored.
1589
1590
1591 A watchpoint is triggered when an lwp in the traced process makes a
1592 memory reference that covers at least one byte of a watched area and
1593 the memory reference is as specified in pr_wflags. When an lwp triggers
1594 a watchpoint, it incurs a watchpoint trap. If FLTWATCH is being traced,
1595 the lwp stops; otherwise, it is sent a SIGTRAP signal; if SIGTRAP is
1596 being traced and is not blocked, the lwp stops.
1597
1598
1599 The watchpoint trap occurs before the instruction completes unless
1600 WA_TRAPAFTER was specified, in which case it occurs after the
1601 instruction completes. If it occurs before completion, the memory is
1602 not modified. If it occurs after completion, the memory is modified (if
1603 the access is a write access).
1604
1605
1606 Physical i/o is an exception for watchpoint traps. In this instance,
1607 there is no guarantee that memory before the watched area has already
1608 been modified (or in the case of WA_TRAPAFTER, that the memory
1609 following the watched area has not been modified) when the watchpoint
1610 trap occurs and the lwp stops.
1611
1612
1613 pr_info in the lwpstatus structure contains information pertinent to
1614 the watchpoint trap. In particular, the si_addr field contains the
1615 virtual address of the memory reference that triggered the watchpoint,
1616 and the si_code field contains one of TRAP_RWATCH, TRAP_WWATCH, or
1617 TRAP_XWATCH, indicating read, write, or execute access, respectively.
1618 The si_trapafter field is zero unless WA_TRAPAFTER is in effect for
1619 this watched area; non-zero indicates that the current instruction is
1620 not the instruction that incurred the watchpoint trap. The si_pc field
1621 contains the virtual address of the instruction that incurred the trap.
1622
1623
1624 A watchpoint trap may be triggered while executing a system call that
1625 makes reference to the traced process's memory. The lwp that is
1626 executing the system call incurs the watchpoint trap while still in the
1627 system call. If it stops as a result, the lwpstatus structure contains
1628 the system call number and its arguments. If the lwp does not stop, or
1629 if it is set running again without clearing the signal or fault, the
1630 system call fails with EFAULT. If WA_TRAPAFTER was specified, the
1631 memory reference will have completed and the memory will have been
1632 modified (if the access was a write access) when the watchpoint trap
1633 occurs.
1634
1635
1636 If more than one of WA_READ, WA_WRITE, and WA_EXEC is specified for a
1637 watched area, and a single instruction incurs more than one of the
1638 specified types, only one is reported when the watchpoint trap occurs.
1639 The precedence is WA_EXEC, WA_READ, WA_WRITE (WA_EXEC and WA_READ take
1640 precedence over WA_WRITE), unless WA_TRAPAFTER was specified, in which
1641 case it is WA_WRITE, WA_READ, WA_EXEC (WA_WRITE takes precedence).
1642
1643
1644 PCWATCH fails with EINVAL if an attempt is made to specify overlapping
1645 watched areas or if pr_wflags contains flags other than those specified
1646 above. It fails with ENOMEM if an attempt is made to establish more
1647 watched areas than the system can support (the system can support
1648 thousands).
1649
1650
1651 The child of a vfork(2) borrows the parent's address space. When a
1652 vfork(2) is executed by a traced process, all watched areas established
1653 for the parent are suspended until the child terminates or performs an
1654 exec(2). Any watched areas established independently in the child are
1655 cancelled when the parent resumes after the child's termination or
1656 exec(2). PCWATCH fails with EBUSY if applied to the parent of a
1657 vfork(2) before the child has terminated or performed an exec(2). The
1658 PR_VFORKP flag is set in the pstatus structure for such a parent
1659 process.
1660
1661
1662 Certain accesses of the traced process's address space by the operating
1663 system are immune to watchpoints. The initial construction of a signal
1664 stack frame when a signal is delivered to an lwp will not trigger a
1665 watchpoint trap even if the new frame covers watched areas of the
1666 stack. Once the signal handler is entered, watchpoint traps occur
1667 normally. On SPARC based machines, register window overflow and
1668 underflow will not trigger watchpoint traps, even if the register
1669 window save areas cover watched areas of the stack.
1670
1671
1672 Watched areas are not inherited by child processes, even if the traced
1673 process's inherit-on-fork mode, PR_FORK, is set (see PCSET, below).
1674 All watched areas are cancelled when the traced process performs a
1675 successful exec(2).
1676
1677 PCSET PCUNSET
1678 PCSET sets one or more modes of operation for the traced process.
1679 PCUNSET unsets these modes. The modes to be set or unset are specified
1680 by flags in an operand long in the control message:
1681
1682 PR_FORK
1683 (inherit-on-fork): When set, the process's tracing flags
1684 and its inherit-on-fork mode are inherited by the child of
1685 a fork(2), fork1(2), or vfork(2). When unset, child
1686 processes start with all tracing flags cleared.
1687
1688
1689 PR_RLC
1690 (run-on-last-close): When set and the last writable /proc
1691 file descriptor referring to the traced process or any of
1692 its lwps is closed, all of the process's tracing flags and
1693 watched areas are cleared, any outstanding stop directives
1694 are canceled, and if any lwps are stopped on events of
1695 interest, they are set running as though PCRUN had been
1696 applied to them. When unset, the process's tracing flags
1697 and watched areas are retained and lwps are not set
1698 running on last close.
1699
1700
1701 PR_KLC
1702 (kill-on-last-close): When set and the last writable /proc
1703 file descriptor referring to the traced process or any of
1704 its lwps is closed, the process is terminated with
1705 SIGKILL.
1706
1707
1708 PR_ASYNC
1709 (asynchronous-stop): When set, a stop on an event of
1710 interest by one lwp does not directly affect any other lwp
1711 in the process. When unset and an lwp stops on an event of
1712 interest other than PR_REQUESTED, all other lwps in the
1713 process are directed to stop.
1714
1715
1716 PR_MSACCT
1717 (microstate accounting): Microstate accounting is now
1718 continuously enabled. This flag is deprecated and no
1719 longer has any effect upon microstate accounting.
1720 Applications may toggle this flag; however, microstate
1721 accounting will remain enabled regardless.
1722
1723
1724 PR_MSFORK
1725 (inherit microstate accounting): All processes now inherit
1726 microstate accounting, as it is continuously enabled. This
1727 flag has been deprecated and its use no longer has any
1728 effect upon the behavior of microstate accounting.
1729
1730
1731 PR_BPTADJ
1732 (breakpoint trap pc adjustment): On x86-based machines, a
1733 breakpoint trap leaves the program counter (the EIP)
1734 referring to the breakpointed instruction plus one byte.
1735 When PR_BPTADJ is set, the system will adjust the program
1736 counter back to the location of the breakpointed
1737 instruction when the lwp stops on a breakpoint. This flag
1738 has no effect on SPARC based machines, where breakpoint
1739 traps leave the program counter referring to the
1740 breakpointed instruction.
1741
1742
1743 PR_PTRACE
1744 (ptrace-compatibility): When set, a stop on an event of
1745 interest by the traced process is reported to the parent
1746 of the traced process by wait(3C), SIGTRAP is sent to the
1747 traced process when it executes a successful exec(2),
1748 setuid/setgid flags are not honored for execs performed by
1749 the traced process, any exec of an object file that the
1750 traced process cannot read fails, and the process dies
1751 when its parent dies. This mode is deprecated; it is
1752 provided only to allow ptrace(3C) to be implemented as a
1753 library function using /proc.
1754
1755
1756
1757 It is an error (EINVAL) to specify flags other than those described
1758 above or to apply these operations to a system process. The current
1759 modes are reported in the pr_flags field of /proc/pid/status and
1760 /proc/pid/lwp/lwp/lwpstatus.
1761
1762 PCSREG
1763 Set the general registers for the specific or representative lwp
1764 according to the operand prgregset_t structure.
1765
1766
1767 On SPARC based systems, only the condition-code bits of the processor-
1768 status register (R_PSR) of SPARC V8 (32-bit) processes can be modified
1769 by PCSREG. Other privileged registers cannot be modified at all.
1770
1771
1772 On x86-based systems, only certain bits of the flags register (EFL) can
1773 be modified by PCSREG: these include the condition codes, direction-
1774 bit, and overflow-bit.
1775
1776
1777 PCSREG fails with EBUSY if the lwp is not stopped on an event of
1778 interest.
1779
1780 PCSVADDR
1781 Set the address at which execution will resume for the specific or
1782 representative lwp from the operand long. On SPARC based systems, both
1783 %pc and %npc are set, with %npc set to the instruction following the
1784 virtual address. On x86-based systems, only %eip is set. PCSVADDR fails
1785 with EBUSY if the lwp is not stopped on an event of interest.
1786
1787 PCSFPREG
1788 Set the floating-point registers for the specific or representative lwp
1789 according to the operand prfpregset_t structure. An error (EINVAL) is
1790 returned if the system does not support floating-point operations (no
1791 floating-point hardware and the system does not emulate floating-point
1792 machine instructions). PCSFPREG fails with EBUSY if the lwp is not
1793 stopped on an event of interest.
1794
1795 PCSXREG
1796 Set the extra state registers for the specific or representative lwp
1797 according to the architecture-dependent operand prxregset_t structure.
1798 An error (EINVAL) is returned if the system does not support extra
1799 state registers. PCSXREG fails with EBUSY if the lwp is not stopped on
1800 an event of interest.
1801
1802 PCSASRS
1803 Set the ancillary state registers for the specific or representative
1804 lwp according to the SPARC V9 platform-dependent operand asrset_t
1805 structure. An error (EINVAL) is returned if either the target process
1806 or the controlling process is not a 64-bit SPARC V9 process. Most of
1807 the ancillary state registers are privileged registers that cannot be
1808 modified. Only those that can be modified are set; all others are
1809 silently ignored. PCSASRS fails with EBUSY if the lwp is not stopped on
1810 an event of interest.
1811
1812 PCAGENT
1813 Create an agent lwp in the controlled process with register values from
1814 the operand prgregset_t structure (see PCSREG, above). The agent lwp is
1815 created in the stopped state showing PR_REQUESTED and with its held
1816 signal set (the signal mask) having all signals except SIGKILL and
1817 SIGSTOP blocked.
1818
1819
1820 The PCAGENT operation fails with EBUSY unless the process is fully
1821 stopped via /proc, that is, unless all of the lwps in the process are
1822 stopped either on events of interest or on PR_SUSPENDED, or are stopped
1823 on PR_JOBCONTROL and have been directed to stop via PCDSTOP. It fails
1824 with EBUSY if an agent lwp already exists. It fails with ENOMEM if
1825 system resources for creating new lwps have been exhausted.
1826
1827
1828 Any PCRUN operation applied to the process control file or to the
1829 control file of an lwp other than the agent lwp fails with EBUSY as
1830 long as the agent lwp exists. The agent lwp must be caused to terminate
1831 by executing the SYS_lwp_exit system call trap before the process can
1832 be restarted.
1833
1834
1835 Once the agent lwp is created, its lwp-ID can be found by reading the
1836 process status file. To facilitate opening the agent lwp's control and
1837 status files, the directory name /propc/pid/lwp/agent is accepted for
1838 lookup operations as an invisible alias for /proc/pid/lwp/lwpid, lwpid
1839 being the lwp-ID of the agent lwp (invisible in the sense that the name
1840 ``agent'' does not appear in a directory listing of /proc/pid/lwp
1841 obtained from ls(1), getdents(2), or readdir(3C)).
1842
1843
1844 The purpose of the agent lwp is to perform operations in the controlled
1845 process on behalf of the controlling process: to gather information not
1846 directly available via /proc files, or in general to make the process
1847 change state in ways not directly available via /proc control
1848 operations. To make use of an agent lwp, the controlling process must
1849 be capable of making it execute system calls (specifically, the
1850 SYS_lwp_exit system call trap). The register values given to the agent
1851 lwp on creation are typically the registers of the representative lwp,
1852 so that the agent lwp can use its stack.
1853
1854
1855 If the controlling process neglects to force the agent lwp to execute
1856 the SYS_lwp_exit system call (due to either logic error or fatal
1857 failure on the part of the controlling process), the agent lwp will
1858 remain in the target process. For purposes of being able to debug
1859 these otherwise rogue agents, information as to the creator of the
1860 agent lwp is reflected in that lwp's spymaster file in /proc. Should
1861 the target process generate a core dump with the agent lwp in place,
1862 this information will be available via the NT_SPYMASTER note in the
1863 core file (see core(4)).
1864
1865
1866 The agent lwp is not allowed to execute any variation of the SYS_fork
1867 or SYS_exec system call traps. Attempts to do so yield ENOTSUP to the
1868 agent lwp.
1869
1870
1871 Symbolic constants for system call trap numbers like SYS_lwp_exit and
1872 SYS_lwp_create can be found in the header file <sys/syscall.h>.
1873
1874 PCREAD PCWRITE
1875 Read or write the target process's address space via a priovec
1876 structure operand:
1877
1878 typedef struct priovec {
1879 void *pio_base; /* buffer in controlling process */
1880 size_t pio_len; /* size of read/write request in bytes */
1881 off_t pio_offset; /* virtual address in target process */
1882 } priovec_t;
1883
1884
1885
1886 These operations have the same effect as pread(2) and pwrite(2),
1887 respectively, of the target process's address space file. The
1888 difference is that more than one PCREAD or PCWRITE control operation
1889 can be written to the control file at once, and they can be
1890 interspersed with other control operations in a single write to the
1891 control file. This is useful, for example, when planting many
1892 breakpoint instructions in the process's address space, or when
1893 stepping over a breakpointed instruction. Unlike pread(2) and
1894 pwrite(2), no provision is made for partial reads or writes; if the
1895 operation cannot be performed completely, it fails with EIO.
1896
1897 PCNICE
1898 The traced process's nice(2) value is incremented by the amount in the
1899 operand long. Only a process with the {PRIV_PROC_PRIOCNTL} privilege
1900 asserted in its effective set can better a process's priority in this
1901 way, but any user may lower the priority. This operation is not
1902 meaningful for all scheduling classes.
1903
1904 PCSCRED
1905 Set the target process credentials to the values contained in the
1906 prcred_t structure operand (see /proc/pid/cred). The effective, real,
1907 and saved user-IDs and group-IDs of the target process are set. The
1908 target process's supplementary groups are not changed; the pr_ngroups
1909 and pr_groups members of the structure operand are ignored. Only the
1910 privileged processes can perform this operation; for all others it
1911 fails with EPERM.
1912
1913 PCSCREDX
1914 Operates like PCSCRED but also sets the supplementary groups; the
1915 length of the data written with this control operation should be
1916 "sizeof (prcred_t) + sizeof (gid_t) * (#groups - 1)".
1917
1918 PCSPRIV
1919 Set the target process privilege to the values contained in the
1920 prpriv_t operand (see /proc/pid/priv). The effective, permitted,
1921 inheritable, and limit sets are all changed. Privilege flags can also
1922 be set. The process is made privilege aware unless it can relinquish
1923 privilege awareness. See privileges(5).
1924
1925
1926 The limit set of the target process cannot be grown. The other
1927 privilege sets must be subsets of the intersection of the effective set
1928 of the calling process with the new limit set of the target process or
1929 subsets of the original values of the sets in the target process.
1930
1931
1932 If any of the above restrictions are not met, EPERM is returned. If the
1933 structure written is improperly formatted, EINVAL is returned.
1934
1935 PROGRAMMING NOTES
1936 For security reasons, except for the psinfo, usage, lpsinfo, lusage,
1937 lwpsinfo, and lwpusage files, which are world-readable, and except for
1938 privileged processes, an open of a /proc file fails unless both the
1939 user-ID and group-ID of the caller match those of the traced process
1940 and the process's object file is readable by the caller. The effective
1941 set of the caller is a superset of both the inheritable and the
1942 permitted set of the target process. The limit set of the caller is a
1943 superset of the limit set of the target process. Except for the world-
1944 readable files just mentioned, files corresponding to setuid and setgid
1945 processes can be opened only by the appropriately privileged process.
1946
1947
1948 A process that is missing the basic privilege {PRIV_PROC_INFO} cannot
1949 see any processes under /proc that it cannot send a signal to.
1950
1951
1952 A process that has {PRIV_PROC_OWNER} asserted in its effective set can
1953 open any file for reading. To manipulate or control a process, the
1954 controlling process must have at least as many privileges in its
1955 effective set as the target process has in its effective, inheritable,
1956 and permitted sets. The limit set of the controlling process must be a
1957 superset of the limit set of the target process. Additional
1958 restrictions apply if any of the uids of the target process are 0. See
1959 privileges(5).
1960
1961
1962 Even if held by a privileged process, an open process or lwp file
1963 descriptor (other than file descriptors for the world-readable files)
1964 becomes invalid if the traced process performs an exec(2) of a
1965 setuid/setgid object file or an object file that the traced process
1966 cannot read. Any operation performed on an invalid file descriptor,
1967 except close(2), fails with EAGAIN. In this situation, if any tracing
1968 flags are set and the process or any lwp file descriptor is open for
1969 writing, the process will have been directed to stop and its run-on-
1970 last-close flag will have been set (see PCSET). This enables a
1971 controlling process (if it has permission) to reopen the /proc files to
1972 get new valid file descriptors, close the invalid file descriptors,
1973 unset the run-on-last-close flag (if desired), and proceed. Just
1974 closing the invalid file descriptors causes the traced process to
1975 resume execution with all tracing flags cleared. Any process not
1976 currently open for writing via /proc, but that has left-over tracing
1977 flags from a previous open, and that executes a setuid/setgid or
1978 unreadable object file, will not be stopped but will have all its
1979 tracing flags cleared.
1980
1981
1982 To wait for one or more of a set of processes or lwps to stop or
1983 terminate, /proc file descriptors (other than those obtained by opening
1984 the cwd or root directories or by opening files in the fd or object
1985 directories) can be used in a poll(2) system call. When requested and
1986 returned, either of the polling events POLLPRI or POLLWRNORM indicates
1987 that the process or lwp stopped on an event of interest. Although they
1988 cannot be requested, the polling events POLLHUP, POLLERR, and POLLNVAL
1989 may be returned. POLLHUP indicates that the process or lwp has
1990 terminated. POLLERR indicates that the file descriptor has become
1991 invalid. POLLNVAL is returned immediately if POLLPRI or POLLWRNORM is
1992 requested on a file descriptor referring to a system process (see
1993 PCSTOP). The requested events may be empty to wait simply for
1994 termination.
1995
1996 FILES
1997 /proc
1998
1999 directory (list of processes)
2000
2001
2002 /proc/pid
2003
2004 specific process directory
2005
2006
2007 /proc/self
2008
2009 alias for a process's own directory
2010
2011
2012 /proc/pid/as
2013
2014 address space file
2015
2016
2017 /proc/pid/ctl
2018
2019 process control file
2020
2021
2022 /proc/pid/status
2023
2024 process status
2025
2026
2027 /proc/pid/lstatus
2028
2029 array of lwp status structs
2030
2031
2032 /proc/pid/psinfo
2033
2034 process ps(1) info
2035
2036
2037 /proc/pid/lpsinfo
2038
2039 array of lwp ps(1) info structs
2040
2041
2042 /proc/pid/map
2043
2044 address space map
2045
2046
2047 /proc/pid/xmap
2048
2049 extended address space map
2050
2051
2052 /proc/pid/rmap
2053
2054 reserved address map
2055
2056
2057 /proc/pid/cred
2058
2059 process credentials
2060
2061
2062 /proc/pid/priv
2063
2064 process privileges
2065
2066
2067 /proc/pid/sigact
2068
2069 process signal actions
2070
2071
2072 /proc/pid/auxv
2073
2074 process aux vector
2075
2076
2077 /proc/pid/ldt
2078
2079 process LDT (x86 only)
2080
2081
2082 /proc/pid/usage
2083
2084 process usage
2085
2086
2087 /proc/pid/lusage
2088
2089 array of lwp usage structs
2090
2091
2092 /proc/pid/path
2093
2094 symbolic links to process open files
2095
2096
2097 /proc/pid/pagedata
2098
2099 process page data
2100
2101
2102 /proc/pid/watch
2103
2104 active watchpoints
2105
2106
2107 /proc/pid/cwd
2108
2109 alias for the current working directory
2110
2111
2112 /proc/pid/root
2113
2114 alias for the root directory
2115
2116
2117 /proc/pid/fd
2118
2119 directory (list of open files)
2120
2121
2122 /proc/pid/fd/*
2123
2124 aliases for process's open files
2125
2126
2127 /proc/pid/object
2128
2129 directory (list of mapped files)
2130
2131
2132 /proc/pid/object/a.out
2133
2134 alias for process's executable file
2135
2136
2137 /proc/pid/object/*
2138
2139 aliases for other mapped files
2140
2141
2142 /proc/pid/lwp
2143
2144 directory (list of lwps)
2145
2146
2147 /proc/pid/lwp/lwpid
2148
2149 specific lwp directory
2150
2151
2152 /proc/pid/lwp/agent
2153
2154 alias for the agent lwp directory
2155
2156
2157 /proc/pid/lwp/lwpid/lwpctl
2158
2159 lwp control file
2160
2161
2162 /proc/pid/lwp/lwpid/lwpstatus
2163
2164 lwp status
2165
2166
2167 /proc/pid/lwp/lwpid/lwpsinfo
2168
2169 lwp ps(1) info
2170
2171
2172 /proc/pid/lwp/lwpid/lwpusage
2173
2174 lwp usage
2175
2176
2177 /proc/pid/lwp/lwpid/gwindows
2178
2179 register windows (SPARC only)
2180
2181
2182 /proc/pid/lwp/lwpid/xregs
2183
2184 extra state registers
2185
2186
2187 /proc/pid/lwp/lwpid/asrs
2188
2189 ancillary state registers (SPARC V9 only)
2190
2191
2192 /proc/pid/lwp/lwpid/spymaster
2193
2194 For an agent LWP, the controlling process
2195
2196
2197 SEE ALSO
2198 ls(1), ps(1), chroot(1M), alarm(2), brk(2), chdir(2), chroot(2),
2199 close(2), creat(2), dup(2), exec(2), fcntl(2), fork(2), fork1(2),
2200 fstat(2), getdents(2), getustack(2), kill(2), lseek(2), mmap(2),
2201 nice(2), open(2), poll(2), pread(2), ptrace(3C), pwrite(2), read(2),
2202 readlink(2), readv(2), shmget(2), sigaction(2), sigaltstack(2),
2203 vfork(2), write(2), writev(2), _stack_grow(3C), readdir(3C),
2204 pthread_create(3C), pthread_join(3C), siginfo.h(3HEAD),
2205 signal.h(3HEAD), thr_create(3C), thr_join(3C), types32.h(3HEAD),
2206 ucontext.h(3HEAD), wait(3C), contract(4), core(4), process(4),
2207 lfcompile(5), privileges(5), security-flags(5)
2208
2209 DIAGNOSTICS
2210 Errors that can occur in addition to the errors normally associated
2211 with file system access:
2212
2213 E2BIG
2214 Data to be returned in a read(2) of the page data file
2215 exceeds the size of the read buffer provided by the
2216 caller.
2217
2218
2219 EACCES
2220 An attempt was made to examine a process that ran under a
2221 different uid than the controlling process and
2222 {PRIV_PROC_OWNER} was not asserted in the effective set.
2223
2224
2225 EAGAIN
2226 The traced process has performed an exec(2) of a
2227 setuid/setgid object file or of an object file that it
2228 cannot read; all further operations on the process or lwp
2229 file descriptor (except close(2)) elicit this error.
2230
2231
2232 EBUSY
2233 PCSTOP, PCDSTOP, PCWSTOP, or PCTWSTOP was applied to a
2234 system process; an exclusive open(2) was attempted on a
2235 /proc file for a process already open for writing; PCRUN,
2236 PCSREG, PCSVADDR, PCSFPREG, or PCSXREG was applied to a
2237 process or lwp not stopped on an event of interest; an
2238 attempt was made to mount /proc when it was already
2239 mounted; PCAGENT was applied to a process that was not
2240 fully stopped or that already had an agent lwp.
2241
2242
2243 EINVAL
2244 In general, this means that some invalid argument was
2245 supplied to a system call. A non-exhaustive list of
2246 conditions eliciting this error includes: a control
2247 message operation code is undefined; an out-of-range
2248 signal number was specified with PCSSIG, PCKILL, or
2249 PCUNKILL; SIGKILL was specified with PCUNKILL; PCSFPREG
2250 was applied on a system that does not support floating-
2251 point operations; PCSXREG was applied on a system that
2252 does not support extra state registers.
2253
2254
2255 EINTR
2256 A signal was received by the controlling process while
2257 waiting for the traced process or lwp to stop via PCSTOP,
2258 PCWSTOP, or PCTWSTOP.
2259
2260
2261 EIO
2262 A write(2) was attempted at an illegal address in the
2263 traced process.
2264
2265
2266 ENOENT
2267 The traced process or lwp has terminated after being
2268 opened. The basic privilege {PRIV_PROC_INFO} is not
2269 asserted in the effective set of the calling process and
2270 the calling process cannot send a signal to the target
2271 process.
2272
2273
2274 ENOMEM
2275 The system-imposed limit on the number of page data file
2276 descriptors was reached on an open of /proc/pid/pagedata;
2277 an attempt was made with PCWATCH to establish more watched
2278 areas than the system can support; the PCAGENT operation
2279 was issued when the system was out of resources for
2280 creating lwps.
2281
2282
2283 ENOSYS
2284 An attempt was made to perform an unsupported operation
2285 (such as creat(2), link(2), or unlink(2)) on an entry in
2286 /proc.
2287
2288
2289 EOVERFLOW
2290 A 32-bit controlling process attempted to read or write
2291 the as file or attempted to read the map, rmap, or
2292 pagedata file of a 64-bit target process. A 32-bit
2293 controlling process attempted to apply one of the control
2294 operations PCSREG, PCSXREG, PCSVADDR, PCWATCH, PCAGENT,
2295 PCREAD, PCWRITE to a 64-bit target process.
2296
2297
2298 EPERM
2299 The process that issued the PCSCRED or PCSCREDX operation
2300 did not have the {PRIV_PROC_SETID} privilege asserted in
2301 its effective set, or the process that issued the PCNICE
2302 operation did not have the {PRIV_PROC_PRIOCNTL} in its
2303 effective set.
2304
2305 An attempt was made to control a process of which the E,
2306 P, and I privilege sets were not a subset of the effective
2307 set of the controlling process or the limit set of the
2308 controlling process is not a superset of limit set of the
2309 controlled process.
2310
2311 Any of the uids of the target process are 0 or an attempt
2312 was made to change any of the uids to 0 using PCSCRED and
2313 the security policy imposed additional restrictions. See
2314 privileges(5).
2315
2316
2317 NOTES
2318 Descriptions of structures in this document include only interesting
2319 structure elements, not filler and padding fields, and may show
2320 elements out of order for descriptive clarity. The actual structure
2321 definitions are contained in <procfs.h>.
2322
2323 BUGS
2324 Because the old ioctl(2)-based version of /proc is currently supported
2325 for binary compatibility with old applications, the top-level directory
2326 for a process, /proc/pid, is not world-readable, but it is world-
2327 searchable. Thus, anyone can open /proc/pid/psinfo even though ls(1)
2328 applied to /proc/pid will fail for anyone but the owner or an
2329 appropriately privileged process. Support for the old ioctl(2)-based
2330 version of /proc will be dropped in a future release, at which time the
2331 top-level directory for a process will be made world-readable.
2332
2333
2334 On SPARC based machines, the types gregset_t and fpregset_t defined in
2335 <sys/regset.h> are similar to but not the same as the types prgregset_t
2336 and prfpregset_t defined in <procfs.h>.
2337
2338
2339
2340 July 23, 2015 PROC(4)