STRACE(1) STRACE(1)
NAME
strace - trace(3x,n,3x _nc_tracebits) system calls and signals
SYNOPSIS
strace [ -dffhiqrtttTvxxX ] [ -acolumn ] [ -eexpr(1,3,n) ] ... [ -ofile(1,n) ] [
-ppid ] ... [ -sstrsize ] [ -uusername ] [ -Evar=val ] ... [ -Evar ]
... [ command [ arg ... ] ]
strace -c [ -eexpr(1,3,n) ] ... [ -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(8,n) or to the file(1,n) specified with the -o option.
strace is a useful diagnostic, instructional, and debugging tool. Sys-
tem 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(1,8) order
to trace(3x,n,3x _nc_tracebits) 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(2,7,n) examination of this boundary is very useful for bug
isolation, sanity checking and attempting to capture race conditions.
Each line in(1,8) the trace(3x,n,3x _nc_tracebits) contains the system call name, followed by its
arguments in(1,8) parentheses and its return value. An example from strac-
ing the command ``cat /dev/null'' is:
open(2,3,n)("/dev/null", O_RDONLY) = 3
Errors (typically a return value of -1) have the errno symbol and error(8,n)
string(3,n) appended.
open(2,3,n)("/foo/bar", O_RDONLY) = -1 ENOENT (No such file(1,n) or directory)
Signals are printed as a signal(2,7) symbol and a signal(2,7) string. An excerpt
from stracing and interrupting the command ``sleep(1,3) 666'' is:
sigsuspend([] <unfinished ...>
--- SIGINT (Interrupt) ---
+++ killed by SIGINT +++
Arguments are printed in(1,8) symbolic form with a passion. This example
shows the shell performing ``>>xyzzy'' output redirection:
open(2,3,n)("xyzzy", O_WRONLY|O_APPEND|O_CREAT, 0666) = 3
Here the three argument form of open(2,3,n) is decoded by breaking down the
flag argument into its three bitwise-OR constituents and printing the
mode value in(1,8) 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(1,8) 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(1,2)' argument is dereferenced and how each mem-
ber is displayed symbolically. In particular, observe how the st_mode
member is carefully decoded into a bitwise-OR of symbolic and numeric
values. Also notice in(1,8) 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(3,n) the system call fails, argu-
ments may not always be dereferenced. For example, retrying the ``ls
-l'' example with a non-existent file(1,n) produces the following line:
lstat("/foo/bar", 0xb004) = -1 ENOENT (No such file(1,n) 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(1,8) 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(2,n,1 builtins)(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 ele-
ments. 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(7,n,1 builtins) elements are separated only by a space. Here is the shell prepar-
ing to execute an external command:
sigprocmask(SIG_BLOCK, [CHLD TTOU], []) = 0
Here the second argument is a bit-set of two signals, SIGCHLD and SIGT-
TOU. 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(7,n,1 builtins) of all signals.
OPTIONS
-c Count time(1,2,n), calls, and errors for each system call
and report a summary on program exit. On Linux,
this attempts to show system time(1,2,n) (CPU time(1,2,n) spent
running in(1,8) the kernel) independent of wall clock(3,n)
time.
-d Show some debugging output of strace itself on the
standard error.
-f Trace child processes as they are created by cur-
rently 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(1,8) the parent process). This means that such
children may run uncontrolled for a while (espe-
cially in(1,8) the case of a vfork(2)), until the parent
is scheduled 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(2,7) that would cause it to terminate
(as determined from the child's current signal(2,7) dis-
position).
-ff If the -o filename option is in(1,8) effect, each pro-
cesses trace(3x,n,3x _nc_tracebits) 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(3,n) -f
has been given.
-h Print the help summary.
-i Print the instruction pointer at the time(1,2,n) of the
system call.
-q Suppress messages about attaching, detaching etc.
This happens automatically when output is redirected
to a file(1,n) and the command is run directly instead of
attaching.
-r Print a relative timestamp upon entry to each system
call. This records the time(1,2,n) difference between the
beginning of successive system calls.
-t Prefix each line of the trace(3x,n,3x _nc_tracebits) with the time(1,2,n) of day.
-tt If given twice, the time(1,2,n) printed will include the
microseconds.
-ttt If given thrice, the time(1,2,n) printed will include the
microseconds and the leading portion will be printed
as the number of seconds since the epoch.
-T Show the time(1,2,n) spent in(1,8) system calls. This records
the time(1,2,n) difference between the beginning and the
end of each system call.
-v Print unabbreviated versions of environment, stat(1,2),
termios, etc. calls. These structures are very
common in(1,8) calls and so the default behavior displays
a reasonable subset of structure members. Use this
option to get all of the gory details.
-V Print the version(1,3,5) number of strace.
-x Print all non-ASCII strings in(1,8) hexadecimal string(3,n)
format.
-xx Print all strings in(1,8) hexadecimal string(3,n) format.
-X Print SELinux specific security context information
when possible.
-a column Align return values in(1,8) a specific column (default
column 40).
-e expr(1,3,n) A qualifying expression which modifies which events
to trace(3x,n,3x _nc_tracebits) or how to trace(3x,n,3x _nc_tracebits) them. The format of the
expression is:
[qualifier=][!]value1[,value2]...
where qualifier is one of trace(3x,n,3x _nc_tracebits), abbrev, verbose,
raw(3x,7,8,3x cbreak), signal(2,7), read(2,n,1 builtins), or write(1,2) and value is a quali-
fier-dependent symbol or number. The default quali-
fier is trace(3x,n,3x _nc_tracebits). Using an exclamation mark negates
the set(7,n,1 builtins) of values. For example, -eopen means liter-
ally -e trace(3x,n,3x _nc_tracebits)=open(2,3,n) which in(1,8) turn means trace(3x,n,3x _nc_tracebits) only
the open(2,3,n) system call. By contrast, -etrace=!open(2,3,n)
means to trace(3x,n,3x _nc_tracebits) every system call except open(2,3,n). In
addition, the special values all and none have the
obvious meanings.
Note that some shells use the exclamation point for
history(1,3,n,1 builtins) expansion even inside quoted arguments. If
so, you must escape the exclamation point with a
backslash.
-e trace(3x,n,3x _nc_tracebits)=set(7,n,1 builtins)
Trace only the specified set(7,n,1 builtins) of system calls. The
-c option is useful for determining which system
calls might be useful to trace. For example,
trace(3x,n,3x _nc_tracebits)=open(2,3,n),close(2,7,n),read(2,n,1 builtins),write(1,2) means to only trace(3x,n,3x _nc_tracebits)
those four system calls. Be careful when making
inferences about the user/kernel boundary if(3,n) only a
subset of system calls are being monitored. The
default is trace(3x,n,3x _nc_tracebits)=all.
-e trace(3x,n,3x _nc_tracebits)=file(1,n)
Trace all system calls which take a file(1,n) name as an
argument. You can think of this as an abbreviation
for -e trace(3x,n,3x _nc_tracebits)=open(2,3,n),stat(1,2),chmod(1,2),unlink(1,2),... which is
useful to seeing what files the process is referenc-
ing. Furthermore, using the abbreviation will
ensure that you don't accidentally forget to include
a call like lstat in(1,8) the list. Betchya woulda for-
got that one.
-e trace(3x,n,3x _nc_tracebits)=process
Trace all system calls which involve process manage-
ment. This is useful for watching the fork, wait,
and exec(3,n,1 builtins) steps of a process.
-e trace(3x,n,3x _nc_tracebits)=network
Trace all the network related system calls.
-e trace(3x,n,3x _nc_tracebits)=signal(2,7)
Trace all signal(2,7) related system calls.
-e trace(3x,n,3x _nc_tracebits)=ipc(2,5)
Trace all IPC related system calls.
-e abbrev=set(7,n,1 builtins)
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(7,n,1 builtins)
Dereference structures for the specified set(7,n,1 builtins) of sys-
tem calls. The default is verbose=all.
-e raw(3x,7,8,3x cbreak)=set(7,n,1 builtins) Print raw(3x,7,8,3x cbreak), undecoded arguments for the specified set(7,n,1 builtins)
of system calls. This option has the effect of
causing all arguments to be printed in(1,8) hexadecimal.
This is mostly useful if(3,n) you don't trust the decod-
ing or you need to know the actual numeric value of
an argument.
-e signal(2,7)=set(7,n,1 builtins)
Trace only the specified subset of signals. The
default is signal(2,7)=all. For example, signal(2,7)=!SIGIO
(or signal(2,7)=!io) causes SIGIO signals not to be
traced.
-e read(2,n,1 builtins)=set(7,n,1 builtins) Perform a full hexadecimal and ASCII dump of all the
data read(2,n,1 builtins) from file(1,n) descriptors listed in(1,8) the speci-
fied set. For example, to see all input activity on
file(1,n) descriptors 3 and 5 use -e read(2,n,1 builtins)=3,5. Note that
this is independent from the normal tracing of the
read(2,n,1 builtins)(2) system call which is controlled by the
option -e trace(3x,n,3x _nc_tracebits)=read(2,n,1 builtins).
-e write(1,2)=set(7,n,1 builtins)
Perform a full hexadecimal and ASCII dump of all the
data written to file(1,n) descriptors listed in(1,8) the spec-
ified set. For example, to see all output activity
on file(1,n) descriptors 3 and 5 use -e write(1,2)=3,5. Note
that this is independent from the normal tracing of
the write(1,2)(2) system call which is controlled by the
option -e trace(3x,n,3x _nc_tracebits)=write(1,2).
-o filename Write the trace(3x,n,3x _nc_tracebits) output to the file(1,n) filename rather
than to stderr. Use filename.pid if(3,n) -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 pip-
ing the debugging output to a program without
affecting the redirections of executed programs.
-O overhead Set the overhead for tracing system calls to over-
head microseconds. This is useful for overriding
the default heuristic for guessing how much time(1,2,n) is
spent in(1,8) 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(1,2,n)(1)) and comparing the accumu-
lated system call time(1,2,n) to the total produced using
-c.
-p pid Attach to the process with the process ID pid and
begin tracing. The trace(3x,n,3x _nc_tracebits) may be terminated at any
time(1,2,n) by a keyboard interrupt signal(2,7) (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(1,8) addition to command (which
is optional if(3,n) at least one -p option is given).
-s strsize Specify the maximum string(3,n) size to print (the
default is 32). Note that filenames are not consid-
ered strings and are always printed in(1,8) full.
-S sortby Sort the output of the histogram printed by the -c
option by the specified criterion. Legal values are
time(1,2,n), calls, name, and nothing (default time(1,2,n)).
-u username Run command with the user ID, group ID, and supple-
mentary groups of username. This option is only
useful when running as root and enables the correct
execution of setuid and/or setgid binaries. Unless
this option is used setuid and setgid programs are
executed without effective privileges.
-E var=val Run command with var=val in(1,8) its list of environment
variables.
-E var Remove var from the inherited list of environment
variables before passing it on to the command.
SETUID INSTALLATION
If strace is installed setuid to root then the invoking user
will be able to attach to and trace(3x,n,3x _nc_tracebits) processes owned by any user.
In addition setuid and setgid programs will be executed and
traced with the correct effective privileges. Since only users(1,5)
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(1,5) who can execute it are restricted to those users(1,5)
who have this trust. For example, it makes sense to install a
special version(1,3,5) of strace with mode `rwsr-xr--', user root and
group trace(3x,n,3x _nc_tracebits), where members of the trace(3x,n,3x _nc_tracebits) group are trusted users.
If you do use this feature, please remember to install a non-
setuid version(1,3,5) of strace for ordinary lusers to use.
SEE ALSO
ptrace(2), proc(5,n)(4), time(1,2,n)(1), trace(3x,n,3x _nc_tracebits)(1), truss(1)
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 possible 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(1,2,n)(2) system call does not take an
argument and the stat(1,2) 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(3x,n,3x _nc_tracebits) applied
to it with the -p option will receive a SIGSTOP. This signal(2,7)
may interrupt a system call that is not restartable. This may
have an unpredictable effect on the process if(3,n) 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(1,8) 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(2,7) (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(3x,n,3x _nc_tracebits) utility. The SunOS version(1,3,5)
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(1,8) 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(1,8) 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)