pidin
Display system statistics (QNX Neutrino)
Syntax:
pidin [options] shorthand
Options:
The options are:
- -d delay
- Delay, in tenths of a second. The default is 10.
- -F formats
- A combination of format letters, like the
format string for
printf()
.
Like printf(), you can specify the width for a field
by including a number with the format, such as
"%I %60N".
Format Description A Arguments a Process ID B What you're blocked on; see Values in the Blocked column, below. b Thread ID c Code size of the process d Data size of the process E Environment e Parent PID f Process flags (see the flags shorthand below) H Scheduling-specific information for each thread. h Thread name; if a thread doesn't have a name, pidin displays the thread's ID (tid) instead. I PID and TID i Runmask and inherit mask. J State of the thread; see Thread life cycle in the QNX Neutrino Microkernel chapter of the System Architecture guide K What kernel call was executed last l (el) The last CPU the thread ran on M Memory owned by the PID m Stack size of the process N Short name of the process n Long name of the process o Connection IDs and file descriptors associated with the process (see the fds shorthand below) P (Uppercase P) parent group p (Lowercase p) thread priority Q Interrupt handlers q Backtrace of calling routines. For best results, use this format with the I format. R Timers S Signal ignore mask s Signal queued mask T Number of threads [ Lengths of the send, receive, reply and pulse queues. If you don't specify a format, the default is "%a %b %N %p %J %B".
The letter following the scheduling priority number stands for the scheduling policy used, as follows:
- f — FIFO scheduling
- r — round-robin scheduling
- o — other (currently, same as round-robin scheduling)
- s — sporadic scheduling
For more information on these scheduling policies, see Thread scheduling in the System Architecture guide.
- -f formats
- The same as the -F option, but the formats parameter is a contiguous string of format codes that gets expanded. For example, -f mbe is expanded to -F "%m %b %e".
- -h
- Display a brief usage message.
- -k
- Keep printing out data for PIDs and TIDs until an error occurs, for example, encountering a PID/TID in an unknown state (because the PID/TID is partially alive).
- -l
- Loop mode; display statistics every delay tenths of a second (specified with the -d option).
- -M formats
- A combination of format letters, like the
format string for
printf()
,
that controls the formatting of information about memory
regions.
The formats include:
Format Description < Memory object code size = Memory object data size > Memory object address ? Memory object offset M Memory owned by the PID : (colon) Memory object name ; (semicolon) Offset If you don't specify a format, the default is the empty string, "".
- -n node
- The name of the remote node from which to get the information.
- -o prio
- Run at prio priority.
- -P pid
- Show only the process family you're interested in (pid may be a name or number).
- -p pid
- Show only the process you're interested in (pid may be a name or number).
If the pid is a number, it's interpreted as a process ID; otherwise, it's interpreted as a name.
Don't assign a numerical name to a process.
The shorthand name is one of the following. You need to type only as many characters of the name as are required to uniquely identify it.
- arguments
- Show the arguments of the displayed processes.
- backtrace
- Display backtrace information for each thread in the displayed processes.
For example:
$ pidin -p devc-con-hid back pid-tid backtrace 4103-01 b033ab5b:b03323cb:b03324f3:804f6ed:804c120:804a285 4103-02 b033af63:805ca60:b031f0ad
The output includes the process ID hyphenated to the thread ID, followed by a backtrace of the addresses of the calling routines.
- channels
- Display the lengths of the send, receive, reply and pulse queues. This shorthand is useful if you're trying to track pulse leaks — that is, a process not receiving pulses. This can cause a growth in kernel memory usage, since pulse structures are allocated in the kernel.
- environment
- Show the environment of the displayed processes.
- extsched
- Display details of the active extended scheduler configuration.
- family
- Show the sessions, process groups, parents, siblings, and children of the displayed processes.
- fds
- Show information about the process's connections and file descriptors.
If you don't have permission to access them,
pidin provides only limited information.
The information for each connection and file descriptor includes the following:
- the file descriptor, followed by s if it's a side channel
- the ID of the process the connection is to
- open flags (r or -, followed by w or -), or MP for a mountpoint
- the offset
- the name of the file or device, if available.
- flags
- Show the processes' flags in hexadecimal, as follows:
Flag Value _NTO_PF_NOCLDSTOP 0x00000001 _NTO_PF_LOADING 0x00000002 _NTO_PF_TERMING 0x00000004 _NTO_PF_ZOMBIE 0x00000008 _NTO_PF_NOZOMBIE 0x00000010 _NTO_PF_FORKED 0x00000020 _NTO_PF_ORPHAN_PGRP 0x00000040 _NTO_PF_STOPPED 0x00000080 _NTO_PF_DEBUG_STOPPED 0x00000100 _NTO_PF_BKGND_PGRP 0x00000200 _NTO_PF_NO_LIMITS 0x00000400 _NTO_PF_CONTINUED 0x00000800 _NTO_PF_CHECK_INTR 0x00001000 _NTO_PF_COREDUMP 0x00002000 _NTO_PF_PTRACED 0x00004000 _NTO_PF_RING0 0x00008000 _NTO_PF_SLEADER 0x00010000 _NTO_PF_WAITINFO 0x00020000 _NTO_PF_VFORKED 0x00040000 _NTO_PF_DESTROYALL 0x00080000 _NTO_PF_NOCOREDUMP 0x00100000 _NTO_PF_NOCTTY 0x00200000 _NTO_PF_THREADWATCH 0x80000000 - info
- Display information about the system, such as the type of processor(s) and the amount of free memory.
- irqs
- Show the IRQ handlers owned by the process. For each handler,
pidin shows:
- id — the interrupt ID returned by InterruptAttach() or InterruptAttachEvent()
- vector — the interrupt vector passed to InterruptAttach() or InterruptAttachEvent()
- mask_count — the mask level count
- flags — the interrupt flags. Valid flags include:
- T—_NTO_INTR_FLAGS_TRK_MSK
- P—_NTO_INTR_FLAGS_PROCESS
- E—_NTO_INTR_FLAGS_END
- handler:area — the address of the interrupt handler and the interrupt handler area
- event — a description of the sigevent to be delivered. Valid descriptions include:
- SIGNAL— signo
- SIGNAL_CODE—signo code:value
- SIGNAL_THREAD—signo code:value
- PULSE—coid:priority code:value
- UNBLOCK
- INTR
- THREAD—code:value
- mapinfo
- Show information about memory mappings.
The output looks like this:
4101 8 proc/boot/io-usb 10o RECEIVE 80K 424K 4096(20K) libc.so.3 @b0300000 452K 16K devu-uhci.so @b8200000 24K 4096 devu-ohci.so @b8207000 24K 4096 devu-ehci.so @b820e000 28K 4096 Mapped Phys Memory @40100000 (ee000000) 12K S
It includes:
- the memory object's name, or Mapped Phys Memory for mapped physical memory
- the memory object's address, followed by the offset if applicable
- the object's code and data sizes
- the memory object's flags, which can include:
- ANON — MAP_PRIVATE, MAP_ANON
- E — MAP_ELF
- F — MAP_FIXED
- P — MAP_PRIVATE
- S — MAP_SHARED
For more information about these flags, see mmap() in the QNX Neutrino C Library Reference.
- memory
- Show the memory used by the displayed processes; pidin
displays the shared memory regions, including shared objects,
and stack usage for each thread. Shared code and data regions
are removed from the size of the process.
The stack numbers represent the amount of stack currently mapped and, in brackets, the maximum allowed for that process.
A * next to a stack size indicates that memory used in the stack isn't automatically returned to the system heap when the thread exits. The memory is returned when the process exits.
Entries for /dev/mem indicate shared memory that's mapped into the process address space. For example:
/dev/mem @38100000 ( 0) 172K
If the entries for different processes show the same object (@38100000 in this example), they all reference the same shared memory object. The processes can map that shared memory differently; the number in parentheses is the offset that was used in the mmap() call, and the last number is the size of the mapping.
- net
- Display system information about all the nodes on the Qnet network.
- pmem
- Display process memory only.
- rc
- Show the process name and arguments of all remote nodes connected to your machine.
- regs
- Show the values of the registers.
- rmasks
- Display runmasks and inherit masks.
- sched
- Display useful scheduling parameters for each thread. This shorthand is a synonym for -f abNplHJ.
- session
- Sort by session ID, then process ID. By default, pidin sorts the output by process ID.
- signals
- Show the signal state of the displayed processes.
- syspage
- Show the syspage entry. You can specify which section to print by indicating a name (e.g. the command pidin syspage=asinfo displays the asinfo section). The default is all.
- threads
- Show the thread name; if a thread doesn't have a name, pidin displays the thread's ID (tid) instead.
- timers
- Show the timers owned by the process. For each timer it shows:
- id — the timer ID returned by TimerCreate() .
- tid — the thread ID associated with this timer (0 for the entire process).
- overruns — the number of overruns.
- clock_type — the type of clock used. Valid descriptions include:
- REAL — CLOCK_REALTIME
- SOFT — CLOCK_SOFTTIME
- MONO — CLOCK_MONOTONIC
- flags — the timer flag. Valid flags include:
- X — _NTO_TI_EXPIRED
- A — _NTO_TI_ABSOLUTE
- a — _NTO_TI_ACTIVE
- msec — the time left before expiry, in milliseconds.
- interval_msec — the timer interval, in milliseconds.
- event — the description of the sigevent to be delivered on timer expiry. Valid event descriptions are listed for the irq shorthand.
- times
- For each process displayed, show:
- start time — the time and date that the process was started
- utime — the number of CPU seconds consumed by the process
- stime — the number of CPU seconds consumed by the kernel on behalf of the process
- cutime — the number of CPU seconds consumed by the children of the process
- cstime — the number of CPU seconds consumed by the kernel on behalf of the children of the process
The times for the child processes are added to cutime and cstime only after the children terminate.
CPU usage is calculated by sampling. When the timer interrupt occurs, the kernel determines which process is running, and adds the time to the total running times of the active thread and its process. If the kernel itself is active, it also adds the time to the system times (stime) of the active thread and its process. The utime is the total running time minus the system time.As a result, these times are approximate, and can be inaccurate (e.g. if a process is driven by the timer interrupt). To determine more accurate times, use the system profiler.
- ttimes
- Show thread times.
- users
- Display the real, effective, and saved user IDs and group IDs for the user who launched the processes. This option doesn't display the user name or group name, just the numerical IDs.
Description:
The pidin utility displays statistics about the processes running on a QNX Neutrino system.
By default, pidin displays the statistics once and then exits. If you specify the -l, pidin loops forever, displaying statistics after the delay specified by the -d option.
If you specify the -l and -k options, pidin loops until a error occurs, displaying statistics after the given delay. The most common error encountered is a race condition: procnto indicates that a process exists, but the process is gone when pidin queries it.
If you want to find out how much space the image file system (IFS) occupies in the memory, run the following command:
pidin syspage=asinfo
and look for the lines with imagefs. See the output in the display as shown in the example below.
Values in the Blocked column
If you specify the %B format, the output includes a Blocked column whose value depends on the thread's state:
State | Value |
---|---|
CONDVAR | Address of the condvar |
JOIN | Thread ID of the blocking thread |
MUTEX | The address of the mutex, or the IDs of the process and thread blocked on, followed by the number of times locked, in the form pid - tid # times |
RECEIVE | ID of the channel within the process that the thread is blocked on |
REPLY | Process IDa |
SEM | Address of the semaphore |
SEND | Process IDa |
STACK | Stack size |
WAITPAGE | Virtual address of the page |
WAITTHREAD | Thread ID of the blocking thread |
a If the process is running on a remote node, the process ID is followed by @ and the node name.
Examples:
The pidin command prints a listing similar to this:
pid tid name prio STATE Blocked 1 1 /sys/procnto-instr 0f READY 1 3 /sys/procnto-instr 10r RUNNING 1 4 /sys/procnto-instr 12r RECEIVE 1 1 5 /sys/procnto-instr 12r RECEIVE 1 1 6 /sys/procnto-instr 12r RECEIVE 1 1 11 /sys/procnto-instr 12r RECEIVE 1 1 12 /sys/procnto-instr 10r RECEIVE 1 1 13 /sys/procnto-instr 10r RECEIVE 1 1 15 /sys/procnto-instr 255r RECEIVE 1 1 16 /sys/procnto-instr 10r RECEIVE 1 1 17 /sys/procnto-instr 10r RECEIVE 1 2 1 sbin/tinit 10o REPLY 1 3 1 proc/boot/slogger 10o RECEIVE 1 5 1 proc/boot/pci-bios 10o RECEIVE 1 6 1 roc/boot/devb-eide 10o SIGWAITINFO 6 2 roc/boot/devb-eide 21r RECEIVE 1 …
Using pidin -F "%I %60N" displays the PID and TID, along with up to 60 characters of the processes' short name:
pid-tid name 1-01 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-03 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-04 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-05 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-06 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-11 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-12 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-13 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-15 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-16 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 1-17 rldbuild/cdr/qnx6/tmp/target/qnx6/x86/boot/sys/procnto-instr 2-01 sbin/tinit 3-01 proc/boot/slogger 5-01 proc/boot/pci-bios 6-01 proc/boot/devb-eide …
The pidin mem command displays:
pid tid name prio STATE code data stack 1 1 /sys/procnto-instr 0f READY 1812K 12K 0(320)* 1 3 /sys/procnto-instr 10r RUNNING 1812K 12K 0(8192) 1 4 /sys/procnto-instr 12r RECEIVE 1812K 12K 0(8192) 1 5 /sys/procnto-instr 12r RECEIVE 1812K 12K 0(8192) 1 6 /sys/procnto-instr 12r RECEIVE 1812K 12K 0(8192) 1 11 /sys/procnto-instr 12r RECEIVE 1812K 12K 0(8192) 1 12 /sys/procnto-instr 10r RECEIVE 1812K 12K 0(8192) 1 13 /sys/procnto-instr 10r RECEIVE 1812K 12K 0(8192) 1 15 /sys/procnto-instr 255r RECEIVE 1812K 12K 0(8192) 1 16 /sys/procnto-instr 10r RECEIVE 1812K 12K 0(8192) 1 17 /sys/procnto-instr 10r RECEIVE 1812K 12K 0(8192) procnto-instr @cfbe5000 12K 12K 2 1 sbin/tinit 10o REPLY 8192 36K 4096(516K)* ldqnx.so.2 @b0300000 344K 16K 3 1 proc/boot/slogger 10o RECEIVE 8192 104K 4096(516K)* ldqnx.so.2 @b0300000 344K 16K 5 1 proc/boot/pci-bios 10o RECEIVE 36K 40K 8192(516K)* ldqnx.so.2 @b0300000 344K 16K 6 1 roc/boot/devb-eide 10o SIGWAITINFO 52K 91M 8192(516K)* 6 2 roc/boot/devb-eide 21r RECEIVE 52K 91M 4096(12K) …
The pidin syspage=asinfo command displays:
Section:asinfo offset:0x00000568 size:0x00000240 0) 0-ffff o:ffff a:0000 p:100 n:io 20) 0-ffffffff o:ffff a:0010 p:100 n:memory 40) 0-ffffff o:0020 a:0010 p:100 n:memory/isa a0) 0-9fbff o:0040 a:0017 p:100 n:memory/isa/ram 180) 1000-cfff o:00a0 a:0007 p:100 n:memory/isa/ram/sysram 1a0) 20f98-9fbff o:00a0 a:0007 p:100 n:memory/isa/ram/sysram c0) 100000-ffffff o:0040 a:0037 p:100 n:memory/isa/ram 1c0) 100000-40e507 o:00c0 a:0007 p:100 n:memory/isa/ram/sysram 1e0) 5e533c-ffffff o:00c0 a:0027 p:100 n:memory/isa/ram/sysram 60) 6000000-ffefffff o:0020 a:0013 p:100 n:memory/device 100) 6000000-ffeafff o:0060 a:0017 p:100 n:memory/device/ram 220) 6000000-ffeafff o:0100 a:0007 p:100 n:memory/device/ram/sysram 80) fff00000-ffffffff o:0020 a:0005 p:100 n:memory/rom e0) 1000000-5ffffff o:0020 a:0037 p:100 n:memory/ram 200) 1000000-5ffffff o:00e0 a:0027 p:100 n:memory/ram/sysram 120) 40e508-5e533b o:0020 a:0005 p:100 n:memory/imagefs 140) 400400-40e507 o:0020 a:0007 p:100 n:memory/startup 160) 40e508-5e533b o:0020 a:0007 p:100 n:memory/bootram