본문 바로가기

Programming/Linux_Kernel

strace man page

STRACE(1)                                                            STRACE(1)

NAME
       strace - trace system calls and signals

SYNOPSIS
       strace  [ -dffhiqrtttTvxx ] [ -acolumn ] [ -eexpr ] ...  [ -ofile ] [ -ppid ] ...  [ -sstrsize ] [ -uusername ]
       [ -Evar=val ] ...  [ -Evar ] ...  [ command [ arg ...  ] ]

       strace -c [ -eexpr ] ...  [ -Ooverhead ] [ -Ssortby ] [ command [ arg ...  ] ]

DESCRIPTION
       In the simplest case strace runs the specified command until it exits.  It intercepts and  records  the  system
       calls  which  are called by a process and the signals which are received by a process.  The name of each system
       call, its arguments and its return value are printed on standard error or to the file  specified  with  the  -o
       option.

       strace  is  a  useful diagnostic, instructional, and debugging tool.  System administrators, diagnosticians and
       trouble-shooters will find it invaluable for solving problems with programs for which the source is not readily
       available  since  they  do not need to be recompiled in order to trace them.  Students, hackers and the overly-
       curious will find that a great deal can be learned about a system and its system calls by tracing even ordinary
       programs.   And  programmers  will  find  that  since  system  calls  and signals are events that happen at the
       user/kernel interface, a close examination of this boundary is very useful for bug isolation,  sanity  checking
       and attempting to capture race conditions.

       Each  line  in the trace contains the system call name, followed by its arguments in parentheses and its return
       value.  An example from stracing the command ‘‘cat /dev/null’’ is:

       open("/dev/null", O_RDONLY) = 3

       Errors (typically a return value of -1) have the errno symbol and error string appended.

       open("/foo/bar", O_RDONLY) = -1 ENOENT (No such file or directory)

       Signals are printed as a signal symbol and a signal string.  An excerpt from stracing and interrupting the com-
       mand ‘‘sleep 666’’ is:

       sigsuspend([] <unfinished ...>
       --- SIGINT (Interrupt) ---
       +++ killed by SIGINT +++

       Arguments  are  printed  in  symbolic form with a passion.  This example shows the shell performing ‘‘>>xyzzy’’
       output redirection:

       open("xyzzy", O_WRONLY|O_APPEND|O_CREAT, 0666) = 3

       Here the three argument form of open is decoded by breaking down the flag argument into  its  three  bitwise-OR
       constituents and printing the mode value in octal by tradition.  Where traditional or native usage differs from
       ANSI or POSIX, the latter forms are preferred.  In some cases, strace output has proven  to  be  more  readable
       than the source.

       Structure  pointers  are dereferenced and the members are displayed as appropriate.  In all cases arguments are
       formatted in the most C-like fashion possible.  For example, the essence of the command ‘‘ls -l /dev/null’’  is
       captured as:

       lstat("/dev/null", {st_mode=S_IFCHR|0666, st_rdev=makedev(1, 3), ...}) = 0

       Notice  how  the ‘struct stat’ argument is dereferenced and how each member is displayed symbolically.  In par-
       ticular, observe how the st_mode member is carefully decoded into a bitwise-OR of symbolic and numeric  values.
       Also  notice  in  this  example  that the first argument to lstat is an input to the system call and the second
       argument is an output.  Since output arguments are not modified if the system call  fails,  arguments  may  not
       always be dereferenced.  For example, retrying the ‘‘ls -l’’ example with a non-existent file produces the fol-
       lowing line:

       lstat("/foo/bar", 0xb004) = -1 ENOENT (No such file or directory)

       In this case the porch light is on but nobody is home.

       Character pointers are dereferenced and printed as C strings.  Non-printing characters in strings are  normally
       represented  by  ordinary C escape codes.  Only the first strsize (32 by default) bytes of strings are printed;
       longer strings have an ellipsis appended following the closing quote.  Here is a line from ‘‘ls -l’’ where  the
       getpwuid library routine is reading the password file:

       read(3, "root::0:0:System Administrator:/"..., 1024) = 422

       While structures are annotated using curly braces, simple pointers and arrays are printed using square brackets
       with commas separating elements.  Here is an example from the command ‘‘id’’ on  a  system  with  supplementary
       group ids:

       getgroups(32, [100, 0]) = 2

       On  the  other  hand,  bit-sets  are  also shown using square brackets but set elements are separated only by a
       space.  Here is the shell preparing to execute an external command:

       sigprocmask(SIG_BLOCK, [CHLD TTOU], []) = 0

       Here the second argument is a bit-set of two signals, SIGCHLD and SIGTTOU.  In some cases  the  bit-set  is  so
       full  that  printing out the unset elements is more valuable.  In that case, the bit-set is prefixed by a tilde
       like this:

       sigprocmask(SIG_UNBLOCK, ~[], NULL) = 0

       Here the second argument represents the full set of all signals.

OPTIONS
       -c          Count time, calls, and errors for each system call and report a summary on program exit.  On
                   Linux,  this attempts to show system time (CPU time spent running in the kernel) independent
                   of wall clock time.


strace -c -p 555
Process 555 attached - interrupt to quit
Process 555 detached
% time     seconds  usecs/call     calls    errors syscall
------ ----------- ----------- --------- --------- ----------------
   nan    0.000000           0        48           read
   nan    0.000000           0       141           ioctl
   nan    0.000000           0         2           setitimer
   nan    0.000000           0         2         2 sigreturn
   nan    0.000000           0        51         2 poll
------ ----------- ----------- --------- --------- ----------------
100.00    0.000000                   244         4 total

       -d          Show some debugging output of strace itself on the standard error.

       -f          Trace child processes as they are created by currently traced processes as a result  of  the
                   fork(2)  system  call.   The new process is attached to as soon as its pid is known (through
                   the return value of fork(2) in the parent process). This means that such  children  may  run
                   uncontrolled  for a while (especially in the case of a vfork(2)), until the parent is sched-
                   uled again to complete its (v)fork(2) call.  If the parent process decides to wait(2) for  a
                   child  that  is  currently  being traced, it is suspended until an appropriate child process
                   either terminates or incurs a signal that would cause it to terminate  (as  determined  from
                   the child’s current signal disposition).

       -ff         If  the  -o  filename  option  is in effect, each processes trace is written to filename.pid
                   where pid is the numeric process id of each process.

       -F          Attempt to follow vforks.  (On SunOS 4.x, this is accomplished  with  some  dynamic  linking
                   trickery.)  Otherwise, vforks will not be followed even if -f has been given.

       -h          Print the help summary.

       -i          Print the instruction pointer at the time of the system call.

       -q          Suppress messages about attaching, detaching etc.  This happens automatically when output is
                   redirected to a file and the command is run directly instead of attaching.

       -r          Print a relative timestamp upon entry to each system call.  This records the time difference
                   between the beginning of successive system calls.


# strace -p 555 -r
Process 555 attached - interrupt to quit
     0.000000 --- SIGSTOP (Stopped (signal)) @ 0 (0) ---
     0.001174 --- SIGSTOP (Stopped (signal)) @ 0 (0) ---
     0.001161 poll([{fd=4, events=POLLIN}, {fd=13, events=POLLIN|POLLPRI}, {fd=16, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=6, events=POLLIN}, {fd=15, events=POLLIN}, {fd=14, events=POLLIN}, {fd=12, events=POLLIN|POLLPRI, revents=POLLIN}], 8, -1) = 1
     0.153307 read(12, "\7\0\0\0\3\0\0\0\0\0\0\0\356\365\n\1\0\0\0\0\f\372y\0\0"..., 136) = 136
     0.002980 ioctl(12, 0x80044b03, 0xbe970424) = 0
     0.001676 ioctl(12, 0x80044b03, 0xbe97040c) = 0
     0.001447 ioctl(12, 0x40044b14, 0x3ce28) = 0

       -t          Prefix each line of the trace with the time of day.


strace -p 555 -t
Process 555 attached - interrupt to quit
09:50:25 --- SIGSTOP (Stopped (signal)) @ 0 (0) ---
09:50:25 --- SIGSTOP (Stopped (signal)) @ 0 (0) ---
09:50:25 poll([{fd=4, events=POLLIN}, {fd=13, events=POLLIN|POLLPRI}, {fd=16, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=6, events=POLLIN}, {fd=15, events=POLLIN}, {fd=14, events=POLLIN}, {fd=12, events=POLLIN|POLLPRI, revents=POLLIN}], 8, -1) = 1
09:50:25 read(12, "\7\0\0\0\3\0\0\0\0\0\0\0\"\262\20\1\0\0\0\0\304:}\0\0\0"..., 136) = 136
09:50:25 ioctl(12, 0x80044b03, 0xbe970424) = 0
09:50:25 ioctl(12, 0x80044b03, 0xbe97040c) = 0
09:50:25 ioctl(12, 0x40044b14, 0x3ce28) = 0

       -tt         If given twice, the time printed will include the microseconds.


strace -p 555 -tt
Process 555 attached - interrupt to quit
09:50:46.462134 --- SIGSTOP (Stopped (signal)) @ 0 (0) ---
09:50:46.465769 --- SIGSTOP (Stopped (signal)) @ 0 (0) ---
09:50:46.467256 poll([{fd=4, events=POLLIN}, {fd=13, events=POLLIN|POLLPRI}, {fd=16, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=6, events=POLLIN}, {fd=15, events=POLLIN}, {fd=14, events=POLLIN}, {fd=12, events=POLLIN|POLLPRI, revents=POLLIN}], 8, -1) = 1
09:50:46.617210 read(12, "\7\0\0\0\3\0\0\0\0\0\0\0>I\21\1\0\0\0\0\3628~\0\0\0\0\0"..., 136) = 136
09:50:46.620281 ioctl(12, 0x80044b03, 0xbe970424) = 0
09:50:46.622395 ioctl(12, 0x80044b03, 0xbe97040c) = 0
09:50:46.624289 ioctl(12, 0x40044b14, 0x3ce28) = 0

       -ttt        If given thrice, the time printed will include the microseconds and the leading portion will
                   be printed as the number of seconds since the epoch.


strace -p 555 -ttt
Process 555 attached - interrupt to quit
1222768294.783245 --- SIGSTOP (Stopped (signal)) @ 0 (0) ---
1222768294.784455 --- SIGSTOP (Stopped (signal)) @ 0 (0) ---
1222768294.785453 poll([{fd=4, events=POLLIN}, {fd=13, events=POLLIN|POLLPRI}, {fd=16, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=6, events=POLLIN}, {fd=15, events=POLLIN}, {fd=14, events=POLLIN}, {fd=12, events=POLLIN|POLLPRI, revents=POLLIN}], 8, -1) = 1
1222768294.857007 read(12, "\7\0\0\0\3\0\0\0\0\0\0\0I\320\f\1\0\0\0\0k9{\0\0\0\0\0"..., 136) = 136
1222768294.859535 ioctl(12, 0x80044b03, 0xbe970424) = 0
1222768294.861090 ioctl(12, 0x80044b03, 0xbe97040c) = 0
1222768294.862416 ioctl(12, 0x40044b14, 0x3ce28) = 0

       -T          Show the time spent in system calls. This records the time difference between the  beginning
                   and the end of each system call.


strace -p 555 -T 
Process 555 attached - interrupt to quit
--- SIGSTOP (Stopped (signal)) @ 0 (0) ---
--- SIGSTOP (Stopped (signal)) @ 0 (0) ---
poll([{fd=4, events=POLLIN}, {fd=13, events=POLLIN|POLLPRI}, {fd=16, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=6, events=POLLIN}, {fd=15, events=POLLIN}, {fd=14, events=POLLIN}, {fd=12, events=POLLIN|POLLPRI, revents=POLLIN}], 8, -1) = 1 <0.029597>
read(12, "\7\0\0\0\3\0\0\0\0\0\0\0Bu\16\1\0\0\0\0\202\266{\0\0\0"..., 136) = 136 <0.000364>
ioctl(12, 0x80044b03, 0xbe970424)       = 0 <0.000353>
ioctl(12, 0x80044b03, 0xbe97040c)       = 0 <0.000353>
ioctl(12, 0x40044b14, 0x3ce28)          = 0 <0.000550>

       -v          Print  unabbreviated  versions of environment, stat, termios, etc.  calls.  These structures
                   are very common in calls and so the default behavior displays a reasonable subset of  struc-
                   ture members.  Use this option to get all of the gory details.

       -V          Print the version number of strace.

       -x          Print all non-ASCII strings in hexadecimal string format.

       -xx         Print all strings in hexadecimal string format.

       -a column   Align return values in a specific column (default column 40).

       -e expr     A qualifying expression which modifies which events to trace or how to trace them.  The format of the expression is:

                             [qualifier=][!]value1[,value2]...

                   where qualifier is one of trace, abbrev, verbose, raw, signal, read, or write and value is a
                   qualifier-dependent symbol or number.  The default qualifier is trace.  Using an exclamation
                   mark negates the set of values.  For example, -eopen means literally -e trace=open which  in
                   turn means trace only the open system call.  By contrast, -etrace=!open means to trace every
                   system call except open.  In addition, the special values all  and  none  have  the  obvious
                   meanings.

                   Note  that  some  shells  use the exclamation point for history expansion even inside quoted
                   arguments.  If so, you must escape the exclamation point with a backslash.

       -e trace=set
                   Trace only the specified set of system calls.  The -c option is useful for determining which
                   system  calls  might  be useful to trace.  For example, trace=open,close,read,write means to
                   only trace those four system calls.  Be careful when making inferences about the user/kernel
                   boundary if only a subset of system calls are being monitored.  The default is trace=all.

       -e trace=file
                   Trace  all  system calls which take a file name as an argument.  You can think of this as an
                   abbreviation for -e trace=open,stat,chmod,unlink,...  which is useful to seeing  what  files
                   the  process is referencing.  Furthermore, using the abbreviation will ensure that you don’t
                   accidentally forget to include a call like lstat in the list.  Betchya  woulda  forgot  that
                   one.

       -e trace=process
                   Trace  all  system  calls which involve process management.  This is useful for watching the
                   fork, wait, and exec steps of a process.

       -e trace=network
                   Trace all the network related system calls.

       -e trace=signal
                   Trace all signal related system calls.

       -e trace=ipc
                   Trace all IPC related system calls.

       -e trace=desc
                   Trace all file descriptor related system calls.

       -e abbrev=set
                   Abbreviate the output from printing  each  member  of  large  structures.   The  default  is
                   abbrev=all.  The -v option has the effect of abbrev=none.

       -e verbose=set
                   Dereference structures for the specified set of system calls.  The default is verbose=all.

       -e raw=set  Print  raw,  undecoded arguments for the specified set of system calls.  This option has the
                   effect of causing all arguments to be printed in hexadecimal.  This is mostly useful if  you
                   don’t trust the decoding or you need to know the actual numeric value of an argument.


# strace -p 555 -e raw=read
Process 555 attached - interrupt to quit
poll([{fd=4, events=POLLIN}, {fd=13, events=POLLIN|POLLPRI}, {fd=16, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=6, events=POLLIN}, {fd=15, events=POLLIN}, {fd=14, events=POLLIN}, {fd=12, events=POLLIN|POLLPRI, revents=POLLIN}], 8, -1) = 1
read(0xc, 0xbe970588, 0x88)             = 0x88
ioctl(12, 0x80044b03, 0xbe970424)       = 0

       -e signal=set
                   Trace  only  the specified subset of signals.  The default is signal=all.  For example, signal=!SIGIO (or signal=!io) causes SIGIO signals not to be traced.

       -e read=set Perform a full hexadecimal and ASCII dump of all the data read from file descriptors  listed
                   in  the  specified  set.  For example, to see all input activity on file descriptors 3 and 5
                   use -e read=3,5.  Note that this is independent from the normal tracing of the read(2)  sys-
                   tem call which is controlled by the option -e trace=read.


strace -p 555 -e read=12
Process 555 attached - interrupt to quit
--- SIGSTOP (Stopped (signal)) @ 0 (0) ---
--- SIGSTOP (Stopped (signal)) @ 0 (0) ---
poll([{fd=4, events=POLLIN}, {fd=13, events=POLLIN|POLLPRI}, {fd=16, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=6, events=POLLIN}, {fd=15, events=POLLIN}, {fd=14, events=POLLIN}, {fd=12, events=POLLIN|POLLPRI, revents=POLLIN}], 8, -1) = 1
read(12, "\7\0\0\0\3\0\0\0\0\0\0\0j\312\f\1\0\0\0\0\202M{\0\0\0\0"..., 136) = 136
 | 00000  07 00 00 00 03 00 00 00  00 00 00 00 6a ca 0c 01  ........ ....j... |
 | 00010  00 00 00 00 82 4d 7b 00  00 00 00 00 f4 81 38 00  .....M{. ......8. |
 | 00020  00 00 00 00 5a 4b 0c 00  00 00 00 00 f5 f9 01 00  ....ZK.. ........ |
 | 00030  13 00 00 00 c8 00 00 00  00 00 00 00 00 00 00 00  ........ ........ |
 | 00040  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  ........ ........ |
 | 00050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  ........ ........ |
 | 00060  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  ........ ........ |
 | 00070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  ........ ........ |
 | 00080  00 00 00 00 00 00 00 00                           ........          |
ioctl(12, 0x80044b03, 0xbe970424)       = 0
ioctl(12, 0x80044b03, 0xbe97040c)       = 0
ioctl(12, 0x40044b14, 0x3ce28)          = 0

       -e write=set
                   Perform a full hexadecimal and ASCII dump of all the data written to file descriptors listed
                   in the specified set.  For example, to see all output activity on file descriptors 3  and  5
                   use  -e  write=3,5.   Note  that this is independent from the normal tracing of the write(2)
                   system call which is controlled by the option -e trace=write.

       -o filename Write the trace output to the file filename rather than to stderr.  Use filename.pid if  -ff
                   is  used.
   If  the  argument  begins  with ‘|’ or with ‘!’ then the rest of the argument is
                   treated as a command and all output is piped to it.   This  is  convenient  for  piping  the
                   debugging output to a program without affecting the redirections of executed programs.

       -O overhead Set  the  overhead  for  tracing  system calls to overhead microseconds.  This is useful for
                   overriding the default heuristic for guessing how much time is spent in mere measuring  when
                   timing  system  calls  using  the -c option.  The accuracy of the heuristic can be gauged by
                   timing a given program run without tracing (using time(1))  and  comparing  the  accumulated
                   system call time to the total produced using -c.

       -p pid      Attach  to  the  process with the process ID pid and begin tracing.  The trace may be termi-
                   nated at any time by a keyboard interrupt signal (CTRL-C).  strace will respond by detaching
                   itself  from  the  traced  process(es)  leaving  it (them) to continue running.  Multiple -p
                   options can be used to attach to up to  32  processes  in  addition  to  command 
(which  is
                   optional if at least one -p option is given).

       -s strsize  Specify  the  maximum string size to print (the default is 32).  Note that filenames are not
                   considered strings and are always printed in full.

       -S sortby   Sort the output of the histogram printed by the -c option by the specified criterion.  Legal
                   values are time, calls, name, and nothing (default time).

       -u username Run  command  with the user ID, group ID, and supplementary groups of username.  This option
                   is only useful when running as root and enables the correct execution of setuid and/or  set-
                   gid  binaries.   Unless  this option is used setuid and setgid programs are executed without
                   effective privileges.

       -E var=val  Run command with var=val in its list of environment variables.

       -E var      Remove var from the inherited list of environment variables before passing it on to the com-
                   mand.

SETUID INSTALLATION
       If  strace  is  installed setuid to root then the invoking user will be able to attach to and trace pro-
       cesses owned by any user.  In addition setuid and setgid programs will be executed and traced  with  the
       correct  effective  privileges.  Since only users trusted with full root privileges should be allowed to
       do these things, it only makes sense to install strace as setuid to root when the users who can  execute
       it  are restricted to those users who have this trust.  For example, it makes sense to install a special
       version of strace with mode ‘rwsr-xr--’, user root and group trace, where members of the trace group are
       trusted  users.   If  you do use this feature, please remember to install a non-setuid version of strace
       for ordinary lusers to use.

SEE ALSO
       ltrace(1), time(1), ptrace(2), proc(5)

NOTES
       It is a pity that so much tracing clutter is produced by systems employing shared libraries.

       It is instructive to think about system call inputs and outputs  as  data-flow  across  the  user/kernel
       boundary.   Because user-space and kernel-space are separate and address-protected, it is sometimes pos-
       sible to make deductive inferences about process behavior using inputs and outputs as propositions.

       In some cases, a system call will differ from the documented behavior or have  a  different  name.   For
       example, on System V-derived systems the true time(2) system call does not take an argument and the stat
       function is called xstat and takes an extra  leading  argument.   These  discrepancies  are  normal  but
       idiosyncratic  characteristics  of  the system call interface and are accounted for by C library wrapper
       functions.

       On some platforms a process that has a system call trace applied to it with the -p option will receive a
       SIGSTOP.   This  signal  may  interrupt  a system call that is not restartable.  This may have an unpre-
       dictable effect on the process if the process takes no action to restart the system call.

BUGS
       Programs that use the setuid bit do not have effective user ID privileges while being traced.

       A traced process ignores SIGSTOP except on SVR4 platforms.

       A traced process which tries to block SIGTRAP will be sent a SIGSTOP in an attempt to force continuation
       of tracing.

       A traced process runs slowly.

       Traced  processes  which are descended from command may be left running after an interrupt signal (CTRL-
       C).

       On Linux, exciting as it would be, tracing the init process is forbidden.

       The -i option is weakly supported.

HISTORY
       strace The original strace was written by Paul Kranenburg for SunOS and was inspired by its trace  util-
       ity.   The  SunOS version of strace was ported to Linux and enhanced by Branko Lankester, who also wrote
       the Linux kernel support.  Even though Paul released strace 2.5 in 1992,  Branko’s  work  was  based  on
       Paul’s  strace  1.5 release from 1991.  In 1993, Rick Sladkey merged strace 2.5 for SunOS and the second
       release of strace for Linux, added many of the features of truss(1) from SVR4, and  produced  an  strace
       that  worked  on both platforms.  In 1994 Rick ported strace to SVR4 and Solaris and wrote the automatic
       configuration support.  In 1995 he ported strace to Irix and tired of writing about himself in the third
       person.

PROBLEMS
       Problems  with  strace  should  be reported via the Debian Bug Tracking System, or to the strace mailing
       list at <strace-devel@lists.sourceforge.net>.

                                  2003-01-21                         STRACE(1)