Chapter 25 proc Provider
The proc provider makes available probes pertaining to the following activities: process creation and termination, LWP creation and termination, executing new program images, and sending and handling signals.
Probes
The proc probes are described in Table 25–1.
Table 25–1 proc Probes|
Probe |
Description |
|---|---|
|
create |
Probe that fires when a process is created using fork(2), forkall(2), fork1(2), or vfork(2). The psinfo_t corresponding to the new child process is pointed to by args[0]. You can distinguish vfork from the other fork variants by checking for PR_VFORKP in the pr_flag member of the forking thread's lwpsinfo_t. You can distinguish fork1 from forkall by examining the pr_nlwp members of both the parent process's psinfo_t (curpsinfo) and the child process's psinfo_t (args[0]). Because the create probe only fires after the process has been successfully created, and because LWP creation is part of creating a process, lwp-create will fire for any LWPs created at process creation time before the create probe fires for the new process. |
|
exec |
Probe that fires whenever a process loads a new process image with a variant of the exec(2) system call: exec(2), execle(2), execlp(2), execv(2), execve(2), execvp(2). The exec probe fires before the process image is loaded. Process variables like execname and curpsinfo therefore contain the process state before the image is loaded. Some time after the exec probe fires, either the exec-failure probe or the exec-success probe will subsequently fire in the same thread. The path of the new process image is pointed to by args[0]. |
|
exec-failure |
Probe that fires when an exec(2) variant has failed. The exec-failure probe fires only after the exec probe has fired in the same thread. The errno(3C) value is provided in args[0]. |
|
exec-success |
Probe that fires when an exec(2) variant has succeeded. Like the exec-failure probe, the exec-success probe fires only after the exec probe has fired in the same thread. By the time the exec-success probe fires, process variables like execname and curpsinfo contain the process state after the new process image has been loaded. |
|
exit |
Probe that fires when the current process is exiting. The reason for exit, which is expressed as one of the SIGCHLD siginfo.h(3HEAD) codes, is contained in args[0]. |
|
fault |
Probe that fires when a thread experiences a machine fault. The fault code (as defined in proc(4)) is in args[0]. The siginfo structure corresponding to the fault is pointed to by args[1]. Only those faults that induce a signal can trigger the fault probe. |
|
lwp-create |
Probe that fires when an LWP is created, typically as a result of thr_create(3C). The lwpsinfo_t corresponding to the new thread is pointed to by args[0]. The psinfo_t of the process containing the thread is pointed to by args[1]. |
|
lwp-start |
Probe that fires within the context of a newly created LWP. The lwp-start probe will fire before any user-level instructions are executed. If the LWP is the first LWP in the process, the start probe will fire, followed by lwp-start. |
|
lwp-exit |
Probe that fires when an LWP is exiting, due either to a signal or to an explicit call to thr_exit(3C). |
|
signal-discard |
Probe that fires when a signal is sent to a single-threaded process, and the signal is both unblocked and ignored by the process. Under these conditions, the signal is discarded on generation. The lwpsinfo_t and psinfo_t of the target process and thread are in args[0] and args[1], respectively. The signal number is in args[2]. |
|
signal-send |
Probe that fires when a signal is sent to a thread or process. The signal-send probe fires in the context of the sending process and thread. The lwpsinfo_t and psinfo_t of the receiving process and thread are in args[0] and args[1], respectively. The signal number is in args[2]. signal-send is always followed by signal-handle or signal-clear in the receiving process and thread. |
|
signal-handle |
Probe that fires immediately before a thread handles a signal. The signal-handle probe fires in the context of the thread that will handle the signal. The signal number is in args[0]. A pointer to the siginfo_t structure that corresponds to the signal is in args[1]. The value of args[1] is NULL if there is no siginfo_t structure or if the signal handler does not have the SA_SIGINFO flag set. The address of the signal handler in the process is in args[2]. |
|
signal-clear |
Probes that fires when a pending signal is cleared because the target thread was waiting for the signal in sigwait(2), sigwaitinfo(3RT), or sigtimedwait(3RT). Under these conditions, the pending signal is cleared and the signal number is returned to the caller. The signal number is in args[0]. signal-clear fires in the context of the formerly waiting thread. |
|
start |
Probe that fires in the context of a newly created process. The start probe will fire before any user-level instructions are executed in the process. |
Arguments
The argument types for the proc probes are listed in Table 25–2. The arguments are described in Table 25–1.
Table 25–2 proc Probe Arguments|
Probe |
args[0] |
args[1] |
args[2] |
|---|---|---|---|
|
create |
psinfo_t * |
— |
— |
|
exec |
char * |
— |
— |
|
exec-failure |
int |
— |
— |
|
exit |
int |
— |
— |
|
fault |
int |
siginfo_t * |
— |
|
lwp-create |
lwpsinfo_t * |
psinfo_t * |
— |
|
lwp-start |
— |
— |
— |
|
lwp-exit |
— |
— |
— |
|
signal-discard |
lwpsinfo_t * |
psinfo_t * |
int |
|
signal-discard |
lwpsinfo_t * |
psinfo_t * |
int |
|
signal-send |
lwpsinfo_t * |
psinfo_t * |
int |
|
signal-handle |
int |
siginfo_t * |
void (*)(void) |
|
signal-clear |
int |
— |
— |
|
start |
— |
— |
— |
lwpsinfo_t
Several proc probes have arguments of type lwpsinfo_t, a structure that is documented in proc(4). The definition of the lwpsinfo_t structure as available to DTrace consumers is as follows:
typedef struct lwpsinfo {
int pr_flag; /* flags; see below */
id_t pr_lwpid; /* LWP id */
uintptr_t pr_addr; /* internal address of thread */
uintptr_t pr_wchan; /* wait addr for sleeping thread */
char pr_stype; /* synchronization event type */
char pr_state; /* numeric thread state */
char pr_sname; /* printable character for pr_state */
char pr_nice; /* nice for cpu usage */
short pr_syscall; /* system call number (if in syscall) */
int pr_pri; /* priority, high value = high priority */
char pr_clname[PRCLSZ]; /* scheduling class name */
processorid_t pr_onpro; /* processor which last ran this thread */
processorid_t pr_bindpro; /* processor to which thread is bound */
psetid_t pr_bindpset; /* processor set to which thread is bound */
} lwpsinfo_t;
The pr_flag field is a bit-mask holding flags describing the process. These flags and their meanings are described in Table 25–3.
Table 25–3 pr_flag Values|
PR_ISSYS |
The process is a system process. |
|
PR_VFORKP |
The process is the parent of a vfork(2)'d child. |
|
PR_FORK |
The process has its inherit-on-fork mode set. |
|
PR_RLC |
The process has its run-on-last-close mode set. |
|
PR_KLC |
The process has its kill-on-last-close mode set. |
|
PR_ASYNC |
The process has its asynchronous-stop mode set. |
|
PR_MSACCT |
The process has microstate accounting enabled. |
|
PR_MSFORK |
The process microstate accounting is inherited on fork. |
|
PR_BPTADJ |
The process has its breakpoint adjustment mode set. |
|
PR_PTRACE |
The process has its ptrace(3C)-compatibility mode set. |
|
PR_STOPPED |
The thread is an LWP that is stopped. |
|
PR_ISTOP |
The thread is an LWP stopped on an event of interest. |
|
PR_DSTOP |
The thread is an LWP that has a stop directive in effect. |
|
PR_STEP |
The thread is an LWP that has a single-step directive in effect. |
|
PR_ASLEEP |
The thread is an LWP in an interruptible sleep within a system call. |
|
PR_DETACH |
The thread is a detached LWP. See pthread_create(3C) and pthread_join(3C). |
|
PR_DAEMON |
The thread is a daemon LWP. See pthread_create(3C). |
|
PR_AGENT |
The thread is the agent LWP for the process. |
|
PR_IDLE |
The thread is the idle thread for a CPU. Idle threads only run on a CPU when the run queues for the CPU are empty. |
The pr_addr field is the address of a private, in-kernel data structure representing the thread. While the data structure is private, the pr_addr field may be used as a token unique to a thread for the thread's lifetime.
The pr_wchan field is set when the thread is sleeping on a synchronization object. The meaning of the pr_wchan field is private to the kernel implementation, but the field may be used as a token unique to the synchronization object.
The pr_stype field is set when the thread is sleeping on a synchronization object. The possible values for the pr_stype field are in Table 25–4.
Table 25–4 pr_stype Values|
SOBJ_MUTEX |
Kernel mutex synchronization object. Used to serialize access to shared data regions in the kernel. See Chapter 18, lockstat Provider and mutex_init(9F) for details on kernel mutex synchronization objects. |
|
SOBJ_RWLOCK |
Kernel readers/writer synchronization object. Used to synchronize access to shared objects in the kernel that can allow multiple concurrent readers or a single writer. See Chapter 18, lockstat Provider and rwlock(9F) for details on kernel readers/writer synchronization objects. |
|
SOBJ_CV |
Condition variable synchronization object. A condition variable is designed to wait indefinitely until some condition becomes true. Condition variables are typically used to synchronize for reasons other than access to a shared data region, and are the mechanism generally used when a process performs a program-directed indefinite wait. For example, blocking in poll(2), pause(2), wait(3C), and the like. |
|
SOBJ_SEMA |
Semaphore synchronization object. A general-purpose synchronization object that – like condition variable objects – does not track a notion of ownership. Because ownership is required to implement priority inheritance in the Solaris kernel, the lack of ownership inherent in semaphore objects inhibits their widespread use. See semaphore(9F) for details. |
|
SOBJ_USER |
A user-level synchronization object. All blocking on user-level synchronization objects is handled with SOBJ_USER synchronization objects. User-level synchronization objects include those created with mutex_init(3C), sema_init(3C), rwlock_init(3C), cond_init(3C) and their POSIX equivalents. |
|
SOBJ_USER_PI |
A user-level synchronization object that implements priority inheritance. Some user-level synchronization objects that track ownership additionally allow for priority inheritance. For example, mutex objects created with pthread_mutex_init(3C) may be made to inherit priority using pthread_mutexattr_setprotocol(3C). |
|
SOBJ_SHUTTLE |
A shuttle synchronization object. Shuttle objects are used to implement doors. See door_create(3DOOR) for more information. |
The pr_state field is set to one of the values in Table 25–5. The pr_sname field is set to a corresponding character shown in parentheses in the same table.
Table 25–5 pr_state Values|
SSLEEP (S) |
The thread is sleeping. The sched:::sleep probe will fire immediately before a thread's state is transitioned to SSLEEP. |
|
SRUN (R) |
The thread is runnable, but is not currently running. The sched:::enqueue probe will fire immediately before a thread's state is transitioned to SRUN. |
|
SZOMB (Z) |
The thread is a zombie LWP. |
|
SSTOP (T) |
The thread is stopped, either due to an explicit proc(4) directive or some other stopping mechanism. |
|
SIDL (I) |
The thread is an intermediate state during process creation. |
|
SONPROC (O) |
The thread is running on a CPU. The sched:::on-cpu probe will fire in the context of the SONPROC thread a short time after the thread's state is transitioned to SONPROC. |
psinfo_t
Several proc probes have an argument of type psinfo_t, a structure that is documented in proc(4). The definition of the psinfo_t structure as available to DTrace consumers is as follows:
typedef struct psinfo {
int pr_nlwp; /* number of active lwps in the process */
pid_t pr_pid; /* unique process id */
pid_t pr_ppid; /* process id of parent */
pid_t pr_pgid; /* pid of process group leader */
pid_t pr_sid; /* session id */
uid_t pr_uid; /* real user id */
uid_t pr_euid; /* effective user id */
gid_t pr_gid; /* real group id */
gid_t pr_egid; /* effective group id */
uintptr_t pr_addr; /* address of process */
dev_t pr_ttydev; /* controlling tty device (or PRNODEV) */
timestruc_t pr_start; /* process start time, from the epoch */
char pr_fname[PRFNSZ]; /* name of execed file */
char pr_psargs[PRARGSZ]; /* initial characters of arg list */
int pr_argc; /* initial argument count */
uintptr_t pr_argv; /* address of initial argument vector */
uintptr_t pr_envp; /* address of initial environment vector */
char pr_dmodel; /* data model of the process */
taskid_t pr_taskid; /* task id */
projid_t pr_projid; /* project id */
poolid_t pr_poolid; /* pool id */
zoneid_t pr_zoneid; /* zone id */
} psinfo_t;
The pr_dmodel field is set to either PR_MODEL_ILP32, denoting a 32–bit process, or PR_MODEL_LP64, denoting a 64–bit process.
Examples
exec
You can use the exec probe to easily determine which programs are being executed, and by whom, as shown in the following example:
#pragma D option quiet
proc:::exec
{
self->parent = execname;
}
proc:::exec-success
/self->parent != NULL/
{
@[self->parent, execname] = count();
self->parent = NULL;
}
proc:::exec-failure
/self->parent != NULL/
{
self->parent = NULL;
}
END
{
printf("%-20s %-20s %s\n", "WHO", "WHAT", "COUNT");
printa("%-20s %-20s %@d\n", @);
}
Running the example script for a short period of time on a build machine results in output similar to the following example:
# dtrace -s ./whoexec.d ^C WHO WHAT COUNT make.bin yacc 1 tcsh make 1 make.bin spec2map 1 sh grep 1 lint lint2 1 sh lint 1 sh ln 1 cc ld 1 make.bin cc 1 lint lint1 1 sh lex 1 make.bin mv 2 sh sh 3 sh make 3 sh sed 4 sh tr 4 make make.bin 4 sh install.bin 5 sh rm 6 cc ir2hf 33 cc ube 33 sh date 34 sh mcs 34 cc acomp 34 sh cc 34 sh basename 34 basename expr 34 make.bin sh 87 |
start and exit
If you want to know how long programs are running from creation to termination, you can enable the start and exit probes, as shown in the following example:
proc:::start
{
self->start = timestamp;
}
proc:::exit
/self->start/
{
@[execname] = quantize(timestamp - self->start);
self->start = 0;
}
Running the example script on the build server for several seconds results in output similar to the following example:
# dtrace -s ./progtime.d
dtrace: script './progtime.d' matched 2 probes
^C
ir2hf
value ------------- Distribution ------------- count
4194304 | 0
8388608 |@ 1
16777216 |@@@@@@@@@@@@@@@@ 14
33554432 |@@@@@@@@@@ 9
67108864 |@@@ 3
134217728 |@ 1
268435456 |@@@@ 4
536870912 |@ 1
1073741824 | 0
ube
value ------------- Distribution ------------- count
16777216 | 0
33554432 |@@@@@@@ 6
67108864 |@@@ 3
134217728 |@@ 2
268435456 |@@@@ 4
536870912 |@@@@@@@@@@@@ 10
1073741824 |@@@@@@@ 6
2147483648 |@@ 2
4294967296 | 0
acomp
value ------------- Distribution ------------- count
8388608 | 0
16777216 |@@ 2
33554432 | 0
67108864 |@ 1
134217728 |@@@ 3
268435456 | 0
536870912 |@@@@@ 5
1073741824 |@@@@@@@@@@@@@@@@@@@@@@@@@ 22
2147483648 |@ 1
4294967296 | 0
cc
value ------------- Distribution ------------- count
33554432 | 0
67108864 |@@@ 3
134217728 |@ 1
268435456 | 0
536870912 |@@@@ 4
1073741824 |@@@@@@@@@@@@@@ 13
2147483648 |@@@@@@@@@@@@ 11
4294967296 |@@@ 3
8589934592 | 0
sh
value ------------- Distribution ------------- count
262144 | 0
524288 |@ 5
1048576 |@@@@@@@ 29
2097152 | 0
4194304 | 0
8388608 |@@@ 12
16777216 |@@ 9
33554432 |@@ 9
67108864 |@@ 8
134217728 |@ 7
268435456 |@@@@@ 20
536870912 |@@@@@@ 26
1073741824 |@@@ 14
2147483648 |@@ 11
4294967296 | 3
8589934592 | 1
17179869184 | 0
make.bin
value ------------- Distribution ------------- count
16777216 | 0
33554432 |@ 1
67108864 |@ 1
134217728 |@@ 2
268435456 | 0
536870912 |@@ 2
1073741824 |@@@@@@@@@ 9
2147483648 |@@@@@@@@@@@@@@@ 14
4294967296 |@@@@@@ 6
8589934592 |@@ 2
17179869184 | 0
|
lwp-start and lwp-exit
Instead of knowing the amount of time that a particular process takes to run, you might want to know how long individual threads take to run. The following example shows how to use the lwp-start and lwp-exit probes for this purpose:
proc:::lwp-start
/tid != 1/
{
self->start = timestamp;
}
proc:::lwp-exit
/self->start/
{
@[execname] = quantize(timestamp - self->start);
self->start = 0;
}
Running the example script on an NFS and calendar server results in output similar to the following example:
# dtrace -s ./lwptime.d
dtrace: script './lwptime.d' matched 3 probes
^C
nscd
value ------------- Distribution ------------- count
131072 | 0
262144 |@ 18
524288 |@@ 24
1048576 |@@@@@@@ 75
2097152 |@@@@@@@@@@@@@@@@@@@@@@@ 245
4194304 |@@ 22
8388608 |@@ 24
16777216 | 6
33554432 | 3
67108864 | 1
134217728 | 1
268435456 | 0
mountd
value ------------- Distribution ------------- count
524288 | 0
1048576 |@ 15
2097152 |@ 24
4194304 |@@@ 51
8388608 |@ 17
16777216 |@ 24
33554432 |@ 15
67108864 |@@@@ 57
134217728 |@ 28
268435456 |@ 26
536870912 |@@ 39
1073741824 |@@@ 45
2147483648 |@@@@@ 72
4294967296 |@@@@@ 77
8589934592 |@@@ 55
17179869184 | 14
34359738368 | 2
68719476736 | 0
automountd
value ------------- Distribution ------------- count
1048576 | 0
2097152 | 3
4194304 |@@@@ 146
8388608 | 6
16777216 | 6
33554432 | 9
67108864 |@@@@@ 203
134217728 |@@ 87
268435456 |@@@@@@@@@@@@@@@ 534
536870912 |@@@@@@ 223
1073741824 |@ 45
2147483648 | 20
4294967296 | 26
8589934592 | 20
17179869184 | 19
34359738368 | 7
68719476736 | 2
137438953472 | 0
iCald
value ------------- Distribution ------------- count
8388608 | 0
16777216 |@@@@@@@ 20
33554432 |@@@ 9
67108864 |@@ 8
134217728 |@@@@@ 16
268435456 |@@@@ 11
536870912 |@@@@ 11
1073741824 |@ 4
2147483648 | 2
4294967296 | 0
8589934592 |@@ 8
17179869184 |@ 5
34359738368 |@ 4
68719476736 |@@ 6
137438953472 |@ 4
274877906944 | 2
549755813888 | 0
|
signal-send
You can use the signal-send probe to determine the sending and receiving process associated with any signal, as shown in the following example:
#pragma D option quiet
proc:::signal-send
{
@[execname, stringof(args[1]->pr_fname), args[2]] = count();
}
END
{
printf("%20s %20s %12s %s\n",
"SENDER", "RECIPIENT", "SIG", "COUNT");
printa("%20s %20s %12d %@d\n", @);
}
Running this script results in output similar to the following example:
# dtrace -s ./sig.d
^C
SENDER RECIPIENT SIG COUNT
xterm dtrace 2 1
xterm soffice.bin 2 1
tr init 18 1
sched test 18 1
sched fvwm2 18 1
bash bash 20 1
sed init 18 2
sched ksh 18 15
sched Xsun 22 471
|
Stability
The proc provider uses DTrace's stability mechanism to describe its stabilities, as shown in the following table. For more information about the stability mechanism, see Chapter 39, Stability.
|
Element |
Name stability |
Data stability |
Dependency class |
|---|---|---|---|
|
Provider |
Evolving |
Evolving |
ISA |
|
Module |
Private |
Private |
Unknown |
|
Function |
Private |
Private |
Unknown |
|
Name |
Evolving |
Evolving |
ISA |
|
Arguments |
Evolving |
Evolving |
ISA |
- © 2010, Oracle Corporation and/or its affiliates