{ "content": [ { "type": "text", "text": "# strace (man)\n\n## NAME\n\nstrace - trace system calls and signals\n\n## SYNOPSIS\n\nstrace [-ACdffhikqqrtttTvVwxxyyzZ] [-I n] [-b execve] [-e expr]... [-O overhead] [-S sortby]\n[-U columns] [-a column] [-o file] [-s strsize] [-X format] [-P path]... [-p pid]...\n[--seccomp-bpf] { -p pid | [-DDD] [-E var[=val]]... [-u username] command [args] }\nstrace -c [-dfwzZ] [-I n] [-b execve] [-e expr]... [-O overhead] [-S sortby] [-U columns]\n[-P path]... [-p pid]... [--seccomp-bpf] { -p pid | [-DDD] [-E var[=val]]...\n[-u username] command [args] }\n\n## DESCRIPTION\n\nIn the simplest case strace runs the specified command until it exits. It intercepts and\nrecords the system calls which are called by a process and the signals which are received by\na process. The name of each system call, its arguments and its return value are printed on\nstandard error or to the file specified with the -o option.\n\n## TLDR\n\n> Troubleshooting tool for tracing system calls.\n\n- Start tracing a specific process by its PID:\n `sudo strace {{-p|--attach}} {{pid}}`\n- Trace a process and filter output by system call [e]xpression:\n `sudo strace {{-p|--attach}} {{pid}} -e {{system_call,system_call2,...}}`\n- Count time, calls, and errors for each system call and report a summary on program exit:\n `sudo strace {{-p|--attach}} {{pid}} {{-c|--summary-only}}`\n- Show the time spent in every system call and specify the maximum string size to print:\n `sudo strace {{-p|--attach}} {{pid}} {{-T|--syscall-times}} {{-s|--string-limit}} {{32}}`\n- Start tracing a program by executing it:\n `strace {{program}}`\n- Start tracing file operations of a program:\n `strace -e trace=file {{program}}`\n- Start tracing network operations of a program as well as all its forked and child processes, saving the output to a file:\n `strace {{-f|--follow-forks}} -e trace=network {{-o|--output}} {{trace.txt}} {{program}}`\n\n*Source: tldr-pages*\n\n## Sections\n\n- **NAME**\n- **SYNOPSIS**\n- **DESCRIPTION**\n- **OPTIONS** (90 subsections)\n- **DIAGNOSTICS**\n- **SETUID INSTALLATION**\n- **MULTIPLE PERSONALITIES SUPPORT**\n- **NOTES**\n- **BUGS**\n- **HISTORY**\n- **REPORTING BUGS**\n- **SEE ALSO**\n- **AUTHORS**\n\nUse structuredContent.sections for detailed options, examples, and full documentation.\n" } ], "structuredContent": { "command": "strace", "section": "", "mode": "man", "summary": "strace - trace system calls and signals", "synopsis": "strace [-ACdffhikqqrtttTvVwxxyyzZ] [-I n] [-b execve] [-e expr]... [-O overhead] [-S sortby]\n[-U columns] [-a column] [-o file] [-s strsize] [-X format] [-P path]... [-p pid]...\n[--seccomp-bpf] { -p pid | [-DDD] [-E var[=val]]... [-u username] command [args] }\nstrace -c [-dfwzZ] [-I n] [-b execve] [-e expr]... [-O overhead] [-S sortby] [-U columns]\n[-P path]... [-p pid]... [--seccomp-bpf] { -p pid | [-DDD] [-E var[=val]]...\n[-u username] command [args] }", "tldr_summary": "Troubleshooting tool for tracing system calls.", "tldr_examples": [ { "description": "Start tracing a specific process by its PID", "command": "sudo strace {{-p|--attach}} {{pid}}" }, { "description": "Trace a process and filter output by system call [e]xpression", "command": "sudo strace {{-p|--attach}} {{pid}} -e {{system_call,system_call2,...}}" }, { "description": "Count time, calls, and errors for each system call and report a summary on program exit", "command": "sudo strace {{-p|--attach}} {{pid}} {{-c|--summary-only}}" }, { "description": "Show the time spent in every system call and specify the maximum string size to print", "command": "sudo strace {{-p|--attach}} {{pid}} {{-T|--syscall-times}} {{-s|--string-limit}} {{32}}" }, { "description": "Start tracing a program by executing it", "command": "strace {{program}}" }, { "description": "Start tracing file operations of a program", "command": "strace -e trace=file {{program}}" }, { "description": "Start tracing network operations of a program as well as all its forked and child processes, saving the output to a file", "command": "strace {{-f|--follow-forks}} -e trace=network {{-o|--output}} {{trace.txt}} {{program}}" } ], "tldr_source": "official", "flags": [ { "flag": "-e", "long": null, "arg": null, "description": "them. The format of the expression is: [qualifier=][!]value[,value]... where qualifier is one of trace (or t), abbrev (or a), verbose (or v), raw (or x), signal (or signals or s), read (or reads or r), write (or writes or w), fault, inject, status, quiet (or silent or silence or q), decode-fds (or de‐‐ code-fd), decode-pids (or decode-pid), or kvm, 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, -e open means literally -e trace=open which in turn means trace only the open system call. By contrast, -e trace=!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." }, { "flag": "-E", "long": null, "arg": null, "description": "--env=var=val Run command with var=val in its list of environment variables." }, { "flag": "-E", "long": null, "arg": null, "description": "--env=var Remove var from the inherited list of environment variables before passing it on to the command." }, { "flag": "-p", "long": null, "arg": null, "description": "--attach=pid Attach to the process with the process ID pid and begin tracing. The trace may be terminated 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 con‐ tinue running. Multiple -p options can be used to attach to many processes in addition to command (which is optional if at least one -p option is given). Mul‐ tiple process IDs, separated by either comma (“,”), space (“ ”), tab, or newline character, can be provided as an argument to a single -p option, so, for example, -p \"$(pidof PROG)\" and -p \"$(pgrep PROG)\" syntaxes are supported." }, { "flag": "-u", "long": null, "arg": null, "description": "--user=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 setgid binaries. Unless this option is used setuid and setgid programs are executed without effective privileges." }, { "flag": "-b", "long": null, "arg": null, "description": "--detach-on=syscall If specified syscall is reached, detach from traced process. Currently, only ex‐‐ ecve(2) syscall is supported. This option is useful if you want to trace multi- threaded process and therefore require -f, but don't want to trace its (poten‐ tially very complex) children." }, { "flag": "-D", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--daemonize", "arg": null, "description": "--daemonize=grandchild Run tracer process as a grandchild, not as the parent of the tracee. This re‐ duces the visible effect of strace by keeping the tracee a direct child of the calling process." }, { "flag": "", "long": null, "arg": null, "description": "--daemonize=pgroup --daemonize=pgrp Run tracer process as tracee's grandchild in a separate process group. In addi‐ tion to reduction of the visible effect of strace, it also avoids killing of strace with kill(2) issued to the whole process group." }, { "flag": "", "long": null, "arg": null, "description": "--daemonize=session Run tracer process as tracee's grandchild in a separate session (\"true daemonisa‐ tion\"). In addition to reduction of the visible effect of strace, it also avoids killing of strace upon session termination." }, { "flag": "-f", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--follow-forks", "arg": null, "description": "Trace child processes as they are created by currently traced processes as a re‐ sult of the fork(2), vfork(2) and clone(2) system calls. Note that -p PID -f will attach all threads of process PID if it is multi-threaded, not only thread with threadid = PID." }, { "flag": "", "long": "--output-separately", "arg": null, "description": "If the --output=filename option is in effect, each processes trace is written to filename.pid where pid is the numeric process id of each process." }, { "flag": "", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--output-separately", "arg": null, "description": "Combine the effects of --follow-forks and --output-separately options. This is incompatible with -c, since no per-process counts are kept. One might want to consider using strace-log-merge(1) to obtain a combined strace log view." }, { "flag": "-I", "long": null, "arg": null, "description": "--interruptible=interruptible When strace can be interrupted by signals (such as pressing CTRL-C). 1, anywhere no signals are blocked; 2, waiting fatal signals are blocked while decoding syscall (default); 3, never fatal signals are always blocked (default if -o FILE PROG); 4, nevertstp fatal signals and SIGTSTP (CTRL-Z) are always blocked (useful to make strace -o FILE PROG not stop on CTRL-Z, default if -D)." }, { "flag": "-e", "long": null, "arg": null, "description": "--trace=syscallset Trace only the specified set of system calls. syscallset is defined as [!]value[,value], and value can be one of the following: syscall Trace specific syscall, specified by its name (but see NOTES). ?value Question mark before the syscall qualification allows suppression of error in case no syscalls matched the qualification provided. /regex Trace only those system calls that match the regex. You can use POSIX Extended Regular Expression syntax (see regex(7)). syscall@64 Trace syscall only for the 64-bit personality. syscall@32 Trace syscall only for the 32-bit personality. syscall@x32 Trace syscall only for the 32-on-64-bit personality. %file 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 abbrevia‐ tion will ensure that you don't accidentally forget to include a call like lstat(2) in the list. Betchya woulda forgot that one. The syntax without a preceding percent sign (\"-e trace=file\") is deprecated. %process process Trace system calls associated with process lifecycle (creation, exec, termination). The syntax without a preceding percent sign (\"-e trace=process\") is deprecated. %net %network network Trace all the network related system calls. The syntax without a preceding percent sign (\"-e trace=network\") is deprecated. %signal signal Trace all signal related system calls. The syntax without a preced‐ ing percent sign (\"-e trace=signal\") is deprecated. %ipc ipc Trace all IPC related system calls. The syntax without a preceding percent sign (\"-e trace=ipc\") is deprecated. %desc desc Trace all file descriptor related system calls. The syntax without a preceding percent sign (\"-e trace=desc\") is deprecated. %memory memory Trace all memory mapping related system calls. The syntax without a preceding percent sign (\"-e trace=memory\") is deprecated. %creds Trace system calls that read or modify user and group identifiers or capability sets. %stat Trace stat syscall variants. %lstat Trace lstat syscall variants. %fstat Trace fstat, fstatat, and statx syscall variants. %%stat Trace syscalls used for requesting file status (stat, lstat, fstat, fstatat, statx, and their variants). %statfs Trace statfs, statfs64, statvfs, osfstatfs, and osfstatfs64 system calls. The same effect can be achieved with -e trace=/^(.*)?statv?fs regular expression. %fstatfs Trace fstatfs, fstatfs64, fstatvfs, osffstatfs, and osffstatfs64 system calls. The same effect can be achieved with -e trace=/fs‐‐ tatv?fs regular expression. %%statfs Trace syscalls related to file system statistics (statfs-like, fs‐ tatfs-like, and ustat). The same effect can be achieved with -e trace=/statv?fs|fsstat|ustat regular expression. %clock Trace system calls that read or modify system clocks. %pure Trace syscalls that always succeed and have no arguments. Cur‐ rently, this list includes arcgettls(2), getdtablesize(2), gete‐‐ gid(2), getegid32(2), geteuid(2), geteuid32(2), getgid(2), get‐‐ gid32(2), getpagesize(2), getpgrp(2), getpid(2), getppid(2), getthreadarea(2) (on architectures other than x86), gettid(2), gettls(2), getuid(2), getuid32(2), getxgid(2), getxpid(2), getxuid(2), kernfeatures(2), and metaggettls(2) syscalls. 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." }, { "flag": "-e", "long": null, "arg": null, "description": "--signal=set Trace only the specified subset of signals. The default is signal=all. For ex‐ ample, signal=!SIGIO (or signal=!io) causes SIGIO signals not to be traced." }, { "flag": "-e", "long": null, "arg": null, "description": "--status=set Print only system calls with the specified return status. The default is sta‐‐ tus=all. When using the status qualifier, because strace waits for system calls to return before deciding whether they should be printed or not, the traditional order of events may not be preserved anymore. If two system calls are executed by concurrent threads, strace will first print both the entry and exit of the first system call to exit, regardless of their respective entry time. The entry and exit of the second system call to exit will be printed afterwards. Here is an example when select(2) is called, but a different thread calls clockget‐‐ time(2) before select(2) finishes: [pid 28779] 1130322148.939977 clockgettime(CLOCKREALTIME, {1130322148, 939977000}) = 0 [pid 28772] 1130322148.438139 select(4, [3], NULL, NULL, NULL) = 1 (in [3]) set can include the following elements: successful Trace system calls that returned without an error code. The -z op‐ tion has the effect of status=successful. failed Trace system calls that returned with an error code. The -Z option has the effect of status=failed. unfinished Trace system calls that did not return. This might happen, for ex‐ ample, due to an execve call in a neighbour thread. unavailable Trace system calls that returned but strace failed to fetch the er‐ ror status. detached Trace system calls for which strace detached before the return." }, { "flag": "-P", "long": null, "arg": null, "description": "--trace-path=path Trace only system calls accessing path. Multiple -P options can be used to spec‐ ify several paths." }, { "flag": "-z", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--successful-only", "arg": null, "description": "Print only syscalls that returned without an error code." }, { "flag": "-Z", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--failed-only", "arg": null, "description": "Print only syscalls that returned with an error code." }, { "flag": "-a", "long": null, "arg": null, "description": "--columns=column Align return values in a specific column (default column 40)." }, { "flag": "-e", "long": null, "arg": null, "description": "--abbrev=syscallset Abbreviate the output from printing each member of large structures. The syntax of the syscallset specification is the same as in the -e trace option. The de‐ fault is abbrev=all. The -v option has the effect of abbrev=none." }, { "flag": "-e", "long": null, "arg": null, "description": "--verbose=syscallset Dereference structures for the specified set of system calls. The syntax of the syscallset specification is the same as in the -e trace option. The default is verbose=all." }, { "flag": "-e", "long": null, "arg": null, "description": "--raw=syscallset Print raw, undecoded arguments for the specified set of system calls. The syntax of the syscallset specification is the same as in the -e trace option. This op‐ tion 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. See also -X raw option." }, { "flag": "-e", "long": null, "arg": null, "description": "--read=set Perform a full hexadecimal and ASCII dump of all the data read from file descrip‐ tors 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 nor‐ mal tracing of the read(2) system call which is controlled by the option -e trace=read." }, { "flag": "-e", "long": null, "arg": null, "description": "--write=set Perform a full hexadecimal and ASCII dump of all the data written to file de‐ scriptors 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." }, { "flag": "-e", "long": null, "arg": null, "description": "--quiet=set --silent=set --silence=set Suppress various information messages. The default is quiet=none. set can in‐ clude the following elements: attach Suppress messages about attaching and detaching (\"[ Process NNNN attached ]\", \"[ Process NNNN detached ]\"). exit Suppress messages about process exits (\"+++ exited with SSS +++\"). path-resolution Suppress messages about resolution of paths provided via the -P option (\"Requested path \"...\" resolved into \"...\"\"). personality Suppress messages about process personality changes (\"[ Process PID=NNNN runs in PPP mode. ]\"). thread-execve superseded Suppress messages about process being superseded by execve(2) in another thread (\"+++ superseded by execve in pid NNNN +++\")." }, { "flag": "-e", "long": null, "arg": null, "description": "--decode-fds=set Decode various information associated with file descriptors. The default is de‐‐ code-fds=none. set can include the following elements: path Print file paths. Also enables printing of tracee's current working di‐ rectory when ATFDCWD constant is used. socket Print socket protocol-specific information, dev Print character/block device numbers. pidfd Print PIDs associated with pidfd file descriptors." }, { "flag": "-e", "long": null, "arg": null, "description": "--decode-pids=set Decode various information associated with process IDs (and also thread IDs, process group IDs, and session IDs). The default is decode-pids=none. set can include the following elements: comm Print command names associated with thread or process IDs. pidns Print thread, process, process group, and session IDs in strace's PID namespace if the tracee is in a different PID namespace." }, { "flag": "-e", "long": null, "arg": null, "description": "--kvm=vcpu Print the exit reason of kvm vcpu. Requires Linux kernel version 4.16.0 or higher." }, { "flag": "-i", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--instruction-pointer", "arg": null, "description": "Print the instruction pointer at the time of the system call." }, { "flag": "-n", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--syscall-number", "arg": null, "description": "Print the syscall number." }, { "flag": "-k", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--stack-traces", "arg": null, "description": "Print the execution stack trace of the traced processes after each system call." }, { "flag": "-o", "long": null, "arg": null, "description": "--output=filename Write the trace output to the file filename rather than to stderr. filename.pid form is used if -ff option is supplied. If the argument begins with '|' or '!', 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. The latter is not compatible with -ff op‐ tion currently." }, { "flag": "-A", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--output-append-mode", "arg": null, "description": "Open the file provided in the -o option in append mode." }, { "flag": "-q", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--quiet", "arg": null, "description": "--quiet=attach,personality Suppress messages about attaching, detaching, and personality changes. This hap‐ pens automatically when output is redirected to a file and the command is run di‐ rectly instead of attaching." }, { "flag": "", "long": null, "arg": null, "description": "--quiet=attach,personality,exit Suppress messages attaching, detaching, personality changes, and about process exit status." }, { "flag": "", "long": null, "arg": null, "description": "--quiet=all Suppress all suppressible messages (please refer to the -e quiet option descrip‐ tion for the full list of suppressible messages)." }, { "flag": "-r", "long": null, "arg": null, "description": "--relative-timestamps[=precision] Print a relative timestamp upon entry to each system call. This records the time difference between the beginning of successive system calls. precision can be one of s (for seconds), ms (milliseconds), us (microseconds), or ns (nanosec‐ onds), and allows setting the precision of time value being printed. Default is" }, { "flag": "-s", "long": null, "arg": null, "description": "--string-limit=strsize Specify the maximum string size to print (the default is 32). Note that file‐ names are not considered strings and are always printed in full. --absolute-timestamps[=[[format:]format],[[precision:]precision]] --timestamps[=[[format:]format],[[precision:]precision]] Prefix each line of the trace with the wall clock time in the specified format with the specified precision. format can be one of the following: none No time stamp is printed. Can be used to override the previous setting. time Wall clock time (strftime(3) format string is %T). unix Number of seconds since the epoch (strftime(3) format string is %s). precision can be one of s (for seconds), ms (milliseconds), us (microseconds), or" }, { "flag": "-t", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--absolute-timestamps", "arg": null, "description": "Prefix each line of the trace with the wall clock time." }, { "flag": "", "long": null, "arg": null, "description": "--absolute-timestamps=precision:us If given twice, the time printed will include the microseconds." }, { "flag": "", "long": null, "arg": null, "description": "--absolute-timestamps=format:unix,precision:us 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." }, { "flag": "-T", "long": null, "arg": null, "description": "--syscall-times[=precision] Show the time spent in system calls. This records the time difference between the beginning and the end of each system call. precision can be one of s (for seconds), ms (milliseconds), us (microseconds), or ns (nanoseconds), and allows setting the precision of time value being printed. Default is us (microseconds)." }, { "flag": "-v", "long": null, "arg": null, "description": "--no-abbrev Print unabbreviated versions of environment, stat, termios, etc. calls. These structures are very common in calls and so the default behavior displays a rea‐ sonable subset of structure members. Use this option to get all of the gory de‐ tails. --strings-in-hex[=option] Control usage of escape sequences with hexadecimal numbers in the printed strings. Normally (when no --strings-in-hex or -x option is supplied), escape sequences are used to print non-printable and non-ASCII characters (that is, characters with a character code less than 32 or greater than 127), or to disam‐ biguate the output (so, for quotes and other characters that encase the printed string, for example, angle brackets, in case of file descriptor path output); for the former use case, unless it is a white space character that has a symbolic es‐ cape sequence defined in the C standard (that is, “\\t” for a horizontal tab, “\\n” for a newline, “\\v” for a vertical tab, “\\f” for a form feed page break, and “\\r” for a carriage return) are printed using escape sequences with numbers that cor‐ respond to their byte values, with octal number format being the default. option can be one of the following: none Hexadecimal numbers are not used in the output at all. When there is a need to emit an escape sequence, octal numbers are used. non-ascii-chars Hexadecimal numbers are used instead of octal in the escape se‐ quences. non-ascii Strings that contain non-ASCII characters are printed using es‐ cape sequences with hexadecimal numbers. all All strings are printed using escape sequences with hexadecimal numbers. When the option is supplied without an argument, all is assumed." }, { "flag": "-x", "long": null, "arg": null, "description": "--strings-in-hex=non-ascii Print all non-ASCII strings in hexadecimal string format." }, { "flag": "", "long": null, "arg": null, "description": "--strings-in-hex[=all] Print all strings in hexadecimal string format." }, { "flag": "-X", "long": null, "arg": null, "description": "--const-print-style=format Set the format for printing of named constants and flags. Supported format val‐ ues are: raw Raw number output, without decoding. abbrev Output a named constant or a set of flags instead of the raw number if they are found. This is the default strace behaviour. verbose Output both the raw value and the decoded string (as a comment)." }, { "flag": "-y", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--decode-fds", "arg": null, "description": "--decode-fds=path Print paths associated with file descriptor arguments and with the ATFDCWD con‐ stant." }, { "flag": "", "long": null, "arg": null, "description": "--decode-fds=all Print all available information associated with file descriptors: protocol-spe‐ cific information associated with socket file descriptors, block/character device number associated with device file descriptors, and PIDs associated with pidfd file descriptors." }, { "flag": "", "long": "--pidns-translation", "arg": null, "description": "--decode-pids=pidns If strace and tracee are in different PID namespaces, print PIDs in strace's namespace, too." }, { "flag": "-Y", "long": null, "arg": null, "description": "--decode-pids=comm Print command names for PIDs." }, { "flag": "-c", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--summary-only", "arg": null, "description": "Count time, calls, and errors for each system call and report a summary on pro‐ gram exit, suppressing the regular output. This attempts to show system time (CPU time spent running in the kernel) independent of wall clock time. If -c is used with -f, only aggregate totals for all traced processes are kept." }, { "flag": "-C", "long": null, "arg": null, "description": "--summary Like -c but also print regular output while processes are running." }, { "flag": "-O", "long": null, "arg": null, "description": "--summary-syscall-overhead=overhead Set the overhead for tracing system calls to overhead. This is useful for over‐ riding the default heuristic for guessing how much time is spent in mere measur‐ ing 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. The format of overhead specification is described in section Time specification format description." }, { "flag": "-S", "long": null, "arg": null, "description": "--summary-sort-by=sortby Sort the output of the histogram printed by the -c option by the specified crite‐ rion. Legal values are time (or time-percent or time-total or total-time), min-time (or shortest or time-min), max-time (or longest or time-max), avg-time (or time-avg), calls (or count), errors (or error), name (or syscall or syscall-name), and nothing (or none); default is time." }, { "flag": "-U", "long": null, "arg": null, "description": "--summary-columns=columns Configure a set (and order) of columns being shown in the call summary. The col‐ umns argument is a comma-separated list with items being one of the following: time-percent (or time) Percentage of cumulative time consumed by a specific system call. total-time (or time-total) Total system (or wall clock, if -w option is provided) time consumed by a specific system call. min-time (or shortest or time-min) Minimum observed call duration. max-time (or longest or time-max) Maximum observed call duration. avg-time (or time-avg) Average call duration. calls (or count) Call count. errors (or error) Error count. name (or syscall or syscall-name) Syscall name. The default value is time-percent,total-time,avg-time,calls,errors,name. If the name field is not supplied explicitly, it is added as the last column." }, { "flag": "-w", "long": null, "arg": null, "description": "" }, { "flag": "", "long": "--summary-wall-clock", "arg": null, "description": "Summarise the time difference between the beginning and end of each system call. The default is to summarise the system time." }, { "flag": "-e", "long": null, "arg": null, "description": "ter=delay][:delayexit=delay][:pokeen‐‐ ter=@argN=DATAN,@argM=DATAM...][:pokeexit=@argN=DATAN,@argM=DATAM...][:when=expr] --inject=syscallset[:error=errno|:retval=value][:signal=sig][:syscall=syscall][:delayen‐‐ ter=delay][:de‐‐ layexit=delay][:pokeen‐‐ ter=@argN=DATAN,@argM=DATAM...][:pokeexit=@argN=DATAN,@argM=DATAM...][:when=expr] Perform syscall tampering for the specified set of syscalls. The syntax of the syscallset specification is the same as in the -e trace option. At least one of error, retval, signal, delayenter, delayexit, pokeenter, or pokeexit options has to be specified. error and retval are mutually exclusive. If :error=errno option is specified, a fault is injected into a syscall invoca‐ tion: the syscall number is replaced by -1 which corresponds to an invalid syscall (unless a syscall is specified with :syscall= option), and the error code is specified using a symbolic errno value like ENOSYS or a numeric value within 1..4095 range. If :retval=value option is specified, success injection is performed: the syscall number is replaced by -1, but a bogus success value is returned to the callee. If :signal=sig option is specified with either a symbolic value like SIGSEGV or a numeric value within 1..SIGRTMAX range, that signal is delivered on entering ev‐ ery syscall specified by the set. If :delayenter=delay or :delayexit=delay options are specified, delay injection is performed: the tracee is delayed by time period specified by delay on entering or exiting the syscall, respectively. The format of delay specification is de‐ scribed in section Time specification format description. If :pokeenter=@argN=DATAN,@argM=DATAM... or :pokeexit=@argN=DATAN,@argM=DATAM... options are specified, tracee's memory at locations, pointed to by system call arguments argN and argM (going from arg1 to arg7) is overwritten by data DATAN and DATAM (specified in hexadecimal format; for example :pokeenter=@arg1=0000DEAD0000BEEF). :pokeenter modifies memory on syscall enter, and :pokeexit - on exit. If :signal=sig option is specified without :error=errno, :retval=value or :de‐‐ lay{enter,exit}=usecs options, then only a signal sig is delivered without a syscall fault or delay injection. Conversely, :error=errno or :retval=value op‐ tion without :delayenter=delay, :delayexit=delay or :signal=sig options injects a fault without delivering a signal or injecting a delay, etc. If :signal=sig option is specified together with :error=errno or :retval=value, then both injection of a fault or success and signal delivery are performed. if :syscall=syscall option is specified, the corresponding syscall with no side effects is injected instead of -1. Currently, only \"pure\" (see -e trace=%pure description) syscalls can be specified there. Unless a :when=expr subexpression is specified, an injection is being made into every invocation of each syscall from the set. The format of the subexpression is: first[..last][+[step]] Number first stands for the first invocation number in the range, number last stands for the last invocation number in the range, and step stands for the step between two consecutive invocations. The following combinations are useful: first For every syscall from the set, perform an injection for the syscall invocation number first only. first..last For every syscall from the set, perform an injection for the syscall invocation number first and all subsequent invocations until the invocation number last (inclusive). first+ For every syscall from the set, perform injections for the syscall invocation number first and all subsequent invocations. first..last+ For every syscall from the set, perform injections for the syscall invocation number first and all subsequent invocations until the invocation number last (inclusive). first+step For every syscall from the set, perform injections for syscall invocations number first, first+step, first+step+step, and so on. first..last+step Same as the previous, but consider only syscall invocations with numbers up to last (inclusive). For example, to fail each third and subsequent chdir syscalls with ENOENT, use -e inject=chdir:error=ENOENT:when=3+. The valid range for numbers first and step is 1..65535, and for number last is 1..65534. An injection expression can contain only one error= or retval= specification, and only one signal= specification. If an injection expression contains multiple when= specifications, the last one takes precedence. Accounting of syscalls that are subject to injection is done per syscall and per tracee. Specification of syscall injection can be combined with other syscall filtering options, for example, -P /dev/urandom -e inject=file:error=ENOENT." }, { "flag": "-e", "long": null, "arg": null, "description": "--fault=syscallset[:error=errno][:when=expr] Perform syscall fault injection for the specified set of syscalls. This is equivalent to more generic -e inject= expression with default value of errno option set to ENOSYS." }, { "flag": "-d", "long": null, "arg": null, "description": "--debug Show some debugging output of strace itself on the standard error." }, { "flag": "-F", "long": null, "arg": null, "description": "may be removed in future releases. Usage of multiple instances of -F option is still equivalent to a single -f, and it is ignored at all if used along with one or more instances of -f option." }, { "flag": "-h", "long": null, "arg": null, "description": "--help Print the help summary." }, { "flag": "", "long": "--seccomp-bpf", "arg": null, "description": "Try to enable use of seccomp-bpf (see seccomp(2)) to have ptrace(2)-stops only when system calls that are being traced occur in the traced processes. This op‐ tion has no effect unless -f/--follow-forks is also specified. --seccomp-bpf is also not applicable to processes attached using -p/--attach option. An attempt to enable system calls filtering using seccomp-bpf may fail for various reasons, e.g. there are too many system calls to filter, the seccomp API is not available, or strace itself is being traced. In cases when seccomp-bpf filter setup failed, strace proceeds as usual and stops traced processes on every system call." }, { "flag": "-V", "long": null, "arg": null, "description": "--version Print the version number of strace." } ], "examples": [], "see_also": [ { "name": "strace-log-merge", "section": "1", "url": "https://www.chedong.com/phpMan.php/man/strace-log-merge/1/json" }, { "name": "ltrace", "section": "1", "url": "https://www.chedong.com/phpMan.php/man/ltrace/1/json" }, { "name": "perf-trace", "section": "1", "url": "https://www.chedong.com/phpMan.php/man/perf-trace/1/json" }, { "name": "trace-cmd", "section": "1", "url": "https://www.chedong.com/phpMan.php/man/trace-cmd/1/json" }, { "name": "time", "section": "1", "url": "https://www.chedong.com/phpMan.php/man/time/1/json" }, { "name": "ptrace", "section": "2", "url": "https://www.chedong.com/phpMan.php/man/ptrace/2/json" }, { "name": "proc", "section": "5", "url": "https://www.chedong.com/phpMan.php/man/proc/5/json" } ], "section_outline": [ { "name": "NAME", "lines": 2, "subsections": [] }, { "name": "SYNOPSIS", "lines": 8, "subsections": [] }, { "name": "DESCRIPTION", "lines": 108, "subsections": [] }, { "name": "OPTIONS", "lines": 1, "subsections": [ { "name": "General", "lines": 1 }, { "name": "-e", "lines": 17, "flag": "-e" }, { "name": "Startup", "lines": 1 }, { "name": "-E", "lines": 3, "flag": "-E" }, { "name": "-E", "lines": 3, "flag": "-E" }, { "name": "-p", "lines": 10, "flag": "-p" }, { "name": "-u", "lines": 6, "flag": "-u" }, { "name": "Tracing", "lines": 1 }, { "name": "-b", "lines": 6, "flag": "-b" }, { "name": "-D", "lines": 1, "flag": "-D" }, { "name": "--daemonize", "lines": 5, "long": "--daemonize" }, { "name": "-DD", "lines": 6 }, { "name": "-DDD", "lines": 5 }, { "name": "-f", "lines": 1, "flag": "-f" }, { "name": "--follow-forks", "lines": 5, "long": "--follow-forks" }, { "name": "--output-separately", "lines": 3, "long": "--output-separately" }, { "name": "-ff", "lines": 1 }, { "name": "--follow-forks --output-separately", "lines": 6, "long": "--output-separately" }, { "name": "-I", "lines": 9, "flag": "-I" }, { "name": "Filtering", "lines": 1 }, { "name": "-e", "lines": 93, "flag": "-e" }, { "name": "-e", "lines": 4, "flag": "-e" }, { "name": "-e", "lines": 26, "flag": "-e" }, { "name": "-P", "lines": 4, "flag": "-P" }, { "name": "-z", "lines": 1, "flag": "-z" }, { "name": "--successful-only", "lines": 2, "long": "--successful-only" }, { "name": "-Z", "lines": 1, "flag": "-Z" }, { "name": "--failed-only", "lines": 2, "long": "--failed-only" }, { "name": "Output format", "lines": 1 }, { "name": "-a", "lines": 3, "flag": "-a" }, { "name": "-e", "lines": 5, "flag": "-e" }, { "name": "-e", "lines": 5, "flag": "-e" }, { "name": "-e", "lines": 7, "flag": "-e" }, { "name": "-e", "lines": 6, "flag": "-e" }, { "name": "-e", "lines": 6, "flag": "-e" }, { "name": "-e", "lines": 18, "flag": "-e" }, { "name": "-e", "lines": 10, "flag": "-e" }, { "name": "-e", "lines": 9, "flag": "-e" }, { "name": "-e", "lines": 3, "flag": "-e" }, { "name": "-i", "lines": 1, "flag": "-i" }, { "name": "--instruction-pointer", "lines": 2, "long": "--instruction-pointer" }, { "name": "-n", "lines": 1, "flag": "-n" }, { "name": "--syscall-number", "lines": 2, "long": "--syscall-number" }, { "name": "-k", "lines": 1, "flag": "-k" }, { "name": "--stack-traces", "lines": 2, "long": "--stack-traces" }, { "name": "-o", "lines": 8, "flag": "-o" }, { "name": "-A", "lines": 1, "flag": "-A" }, { "name": "--output-append-mode", "lines": 2, "long": "--output-append-mode" }, { "name": "-q", "lines": 1, "flag": "-q" }, { "name": "--quiet", "lines": 5, "long": "--quiet" }, { "name": "-qq", "lines": 4 }, { "name": "-qqq", "lines": 3 }, { "name": "-r", "lines": 5, "flag": "-r" }, { "name": "us (microseconds)", "lines": 3 }, { "name": "-s", "lines": 16, "flag": "-s" }, { "name": "ns (nanoseconds)", "lines": 1 }, { "name": "-t", "lines": 1, "flag": "-t" }, { "name": "--absolute-timestamps", "lines": 2, "long": "--absolute-timestamps" }, { "name": "-tt", "lines": 3 }, { "name": "-ttt", "lines": 4 }, { "name": "-T", "lines": 6, "flag": "-T" }, { "name": "-v", "lines": 31, "flag": "-v" }, { "name": "-x", "lines": 3, "flag": "-x" }, { "name": "-xx", "lines": 3 }, { "name": "-X", "lines": 9, "flag": "-X" }, { "name": "-y", "lines": 1, "flag": "-y" }, { "name": "--decode-fds", "lines": 4, "long": "--decode-fds" }, { "name": "-yy", "lines": 6 }, { "name": "--pidns-translation", "lines": 4, "long": "--pidns-translation" }, { "name": "-Y", "lines": 3, "flag": "-Y" }, { "name": "Statistics", "lines": 1 }, { "name": "-c", "lines": 1, "flag": "-c" }, { "name": "--summary-only", "lines": 5, "long": "--summary-only" }, { "name": "-C", "lines": 2, "flag": "-C" }, { "name": "-O", "lines": 10, "flag": "-O" }, { "name": "-S", "lines": 7, "flag": "-S" }, { "name": "-U", "lines": 19, "flag": "-U" }, { "name": "-w", "lines": 1, "flag": "-w" }, { "name": "--summary-wall-clock", "lines": 3, "long": "--summary-wall-clock" }, { "name": "Tampering", "lines": 1 }, { "name": "-e", "lines": 93, "flag": "-e" }, { "name": "-e", "lines": 6, "flag": "-e" }, { "name": "Miscellaneous", "lines": 1 }, { "name": "-d", "lines": 2, "flag": "-d" }, { "name": "-F", "lines": 4, "flag": "-F" }, { "name": "-h", "lines": 2, "flag": "-h" }, { "name": "--seccomp-bpf", "lines": 9, "long": "--seccomp-bpf" }, { "name": "-V", "lines": 2, "flag": "-V" }, { "name": "Time specification format description", "lines": 2 }, { "name": "s (seconds), ms (milliseconds), us (microseconds),", "lines": 4 } ] }, { "name": "DIAGNOSTICS", "lines": 9, "subsections": [] }, { "name": "SETUID INSTALLATION", "lines": 10, "subsections": [] }, { "name": "MULTIPLE PERSONALITIES SUPPORT", "lines": 39, "subsections": [] }, { "name": "NOTES", "lines": 36, "subsections": [] }, { "name": "BUGS", "lines": 7, "subsections": [] }, { "name": "HISTORY", "lines": 25, "subsections": [] }, { "name": "REPORTING BUGS", "lines": 3, "subsections": [] }, { "name": "SEE ALSO", "lines": 4, "subsections": [] }, { "name": "AUTHORS", "lines": 5, "subsections": [] } ], "sections": { "NAME": { "content": "strace - trace system calls and signals\n", "subsections": [] }, "SYNOPSIS": { "content": "strace [-ACdffhikqqrtttTvVwxxyyzZ] [-I n] [-b execve] [-e expr]... [-O overhead] [-S sortby]\n[-U columns] [-a column] [-o file] [-s strsize] [-X format] [-P path]... [-p pid]...\n[--seccomp-bpf] { -p pid | [-DDD] [-E var[=val]]... [-u username] command [args] }\n\nstrace -c [-dfwzZ] [-I n] [-b execve] [-e expr]... [-O overhead] [-S sortby] [-U columns]\n[-P path]... [-p pid]... [--seccomp-bpf] { -p pid | [-DDD] [-E var[=val]]...\n[-u username] command [args] }\n", "subsections": [] }, "DESCRIPTION": { "content": "In the simplest case strace runs the specified command until it exits. It intercepts and\nrecords the system calls which are called by a process and the signals which are received by\na process. The name of each system call, its arguments and its return value are printed on\nstandard error or to the file specified with the -o option.\n\nstrace is a useful diagnostic, instructional, and debugging tool. System administrators, di‐\nagnosticians and trouble-shooters will find it invaluable for solving problems with programs\nfor which the source is not readily available since they do not need to be recompiled in or‐\nder to trace them. Students, hackers and the overly-curious will find that a great deal can\nbe learned about a system and its system calls by tracing even ordinary programs. And pro‐\ngrammers will find that since system calls and signals are events that happen at the\nuser/kernel interface, a close examination of this boundary is very useful for bug isolation,\nsanity checking and attempting to capture race conditions.\n\nEach line in the trace contains the system call name, followed by its arguments in parenthe‐\nses and its return value. An example from stracing the command \"cat /dev/null\" is:\n\nopen(\"/dev/null\", ORDONLY) = 3\n\nErrors (typically a return value of -1) have the errno symbol and error string appended.\n\nopen(\"/foo/bar\", ORDONLY) = -1 ENOENT (No such file or directory)\n\nSignals are printed as signal symbol and decoded siginfo structure. An excerpt from stracing\nand interrupting the command \"sleep 666\" is:\n\nsigsuspend([] \n--- SIGINT {sisigno=SIGINT, sicode=SIUSER, sipid=...} ---\n+++ killed by SIGINT +++\n\nIf a system call is being executed and meanwhile another one is being called from a different\nthread/process then strace will try to preserve the order of those events and mark the ongo‐\ning call as being unfinished. When the call returns it will be marked as resumed.\n\n[pid 28772] select(4, [3], NULL, NULL, NULL \n[pid 28779] clockgettime(CLOCKREALTIME, {1130322148, 939977000}) = 0\n[pid 28772] <... select resumed> ) = 1 (in [3])\n\nInterruption of a (restartable) system call by a signal delivery is processed differently as\nkernel terminates the system call and also arranges its immediate reexecution after the sig‐\nnal handler completes.\n\nread(0, 0x7ffff72cf5cf, 1) = ? ERESTARTSYS (To be restarted)\n--- SIGALRM ... ---\nrtsigreturn(0xe) = 0\nread(0, \"\", 1) = 0\n\nArguments are printed in symbolic form with passion. This example shows the shell performing\n\">>xyzzy\" output redirection:\n\nopen(\"xyzzy\", OWRONLY|OAPPEND|OCREAT, 0666) = 3\n\nHere, the second and the third argument of open(2) are decoded by breaking down the flag ar‐\ngument into its three bitwise-OR constituents and printing the mode value in octal by tradi‐\ntion. Where the traditional or native usage differs from ANSI or POSIX, the latter forms are\npreferred. In some cases, strace output is proven to be more readable than the source.\n\nStructure pointers are dereferenced and the members are displayed as appropriate. In most\ncases, arguments are formatted in the most C-like fashion possible. For example, the essence\nof the command \"ls -l /dev/null\" is captured as:\n\nlstat(\"/dev/null\", {stmode=SIFCHR|0666, strdev=makedev(0x1, 0x3), ...}) = 0\n\nNotice how the 'struct stat' argument is dereferenced and how each member is displayed sym‐\nbolically. In particular, observe how the stmode member is carefully decoded into a bit‐\nwise-OR of symbolic and numeric values. Also notice in this example that the first argument\nto lstat(2) is an input to the system call and the second argument is an output. Since out‐\nput arguments are not modified if the system call fails, arguments may not always be derefer‐\nenced. For example, retrying the \"ls -l\" example with a non-existent file produces the fol‐\nlowing line:\n\nlstat(\"/foo/bar\", 0xb004) = -1 ENOENT (No such file or directory)\n\nIn this case the porch light is on but nobody is home.\n\nSyscalls unknown to strace are printed raw, with the unknown system call number printed in\nhexadecimal form and prefixed with \"syscall\":\n\nsyscall0xbad(0x1, 0x2, 0x3, 0x4, 0x5, 0x6) = -1 ENOSYS (Function not implemented)\n\n\nCharacter pointers are dereferenced and printed as C strings. Non-printing characters in\nstrings are normally represented by ordinary C escape codes. Only the first strsize (32 by\ndefault) bytes of strings are printed; longer strings have an ellipsis appended following the\nclosing quote. Here is a line from \"ls -l\" where the getpwuid(3) library routine is reading\nthe password file:\n\nread(3, \"root::0:0:System Administrator:/\"..., 1024) = 422\n\nWhile structures are annotated using curly braces, simple pointers and arrays are printed us‐\ning square brackets with commas separating elements. Here is an example from the command\nid(1) on a system with supplementary group ids:\n\ngetgroups(32, [100, 0]) = 2\n\nOn the other hand, bit-sets are also shown using square brackets, but set elements are sepa‐\nrated only by a space. Here is the shell, preparing to execute an external command:\n\nsigprocmask(SIGBLOCK, [CHLD TTOU], []) = 0\n\nHere, the second argument is a bit-set of two signals, SIGCHLD and SIGTTOU. In some cases,\nthe bit-set is so full that printing out the unset elements is more valuable. In that case,\nthe bit-set is prefixed by a tilde like this:\n\nsigprocmask(SIGUNBLOCK, ~[], NULL) = 0\n\nHere, the second argument represents the full set of all signals.\n", "subsections": [] }, "OPTIONS": { "content": "", "subsections": [ { "name": "General", "content": "" }, { "name": "-e", "content": "them. The format of the expression is:\n\n[qualifier=][!]value[,value]...\n\nwhere qualifier is one of trace (or t), abbrev (or a), verbose (or v), raw (or\nx), signal (or signals or s), read (or reads or r), write (or writes or w),\nfault, inject, status, quiet (or silent or silence or q), decode-fds (or de‐‐\ncode-fd), decode-pids (or decode-pid), or kvm, and value is a qualifier-dependent\nsymbol or number. The default qualifier is trace. Using an exclamation mark\nnegates the set of values. For example, -e open means literally -e trace=open\nwhich in turn means trace only the open system call. By contrast, -e trace=!open\nmeans to trace every system call except open. In addition, the special values\nall and none have the obvious meanings.\n\nNote that some shells use the exclamation point for history expansion even inside\nquoted arguments. If so, you must escape the exclamation point with a backslash.\n", "flag": "-e" }, { "name": "Startup", "content": "" }, { "name": "-E", "content": "--env=var=val\nRun command with var=val in its list of environment variables.\n", "flag": "-E" }, { "name": "-E", "content": "--env=var Remove var from the inherited list of environment variables before passing it on\nto the command.\n", "flag": "-E" }, { "name": "-p", "content": "--attach=pid\nAttach to the process with the process ID pid and begin tracing. The trace may\nbe terminated at any time by a keyboard interrupt signal (CTRL-C). strace will\nrespond by detaching itself from the traced process(es) leaving it (them) to con‐\ntinue running. Multiple -p options can be used to attach to many processes in\naddition to command (which is optional if at least one -p option is given). Mul‐\ntiple process IDs, separated by either comma (“,”), space (“ ”), tab, or newline\ncharacter, can be provided as an argument to a single -p option, so, for example,\n-p \"$(pidof PROG)\" and -p \"$(pgrep PROG)\" syntaxes are supported.\n", "flag": "-p" }, { "name": "-u", "content": "--user=username\nRun command with the user ID, group ID, and supplementary groups of username.\nThis option is only useful when running as root and enables the correct execution\nof setuid and/or setgid binaries. Unless this option is used setuid and setgid\nprograms are executed without effective privileges.\n", "flag": "-u" }, { "name": "Tracing", "content": "" }, { "name": "-b", "content": "--detach-on=syscall\nIf specified syscall is reached, detach from traced process. Currently, only ex‐‐\necve(2) syscall is supported. This option is useful if you want to trace multi-\nthreaded process and therefore require -f, but don't want to trace its (poten‐\ntially very complex) children.\n", "flag": "-b" }, { "name": "-D", "content": "", "flag": "-D" }, { "name": "--daemonize", "content": "--daemonize=grandchild\nRun tracer process as a grandchild, not as the parent of the tracee. This re‐\nduces the visible effect of strace by keeping the tracee a direct child of the\ncalling process.\n", "long": "--daemonize" }, { "name": "-DD", "content": "--daemonize=pgroup\n--daemonize=pgrp\nRun tracer process as tracee's grandchild in a separate process group. In addi‐\ntion to reduction of the visible effect of strace, it also avoids killing of\nstrace with kill(2) issued to the whole process group.\n" }, { "name": "-DDD", "content": "--daemonize=session\nRun tracer process as tracee's grandchild in a separate session (\"true daemonisa‐\ntion\"). In addition to reduction of the visible effect of strace, it also avoids\nkilling of strace upon session termination.\n" }, { "name": "-f", "content": "", "flag": "-f" }, { "name": "--follow-forks", "content": "Trace child processes as they are created by currently traced processes as a re‐\nsult of the fork(2), vfork(2) and clone(2) system calls. Note that -p PID -f\nwill attach all threads of process PID if it is multi-threaded, not only thread\nwith threadid = PID.\n", "long": "--follow-forks" }, { "name": "--output-separately", "content": "If the --output=filename option is in effect, each processes trace is written to\nfilename.pid where pid is the numeric process id of each process.\n", "long": "--output-separately" }, { "name": "-ff", "content": "" }, { "name": "--follow-forks --output-separately", "content": "Combine the effects of --follow-forks and --output-separately options. This is\nincompatible with -c, since no per-process counts are kept.\n\nOne might want to consider using strace-log-merge(1) to obtain a combined strace\nlog view.\n", "long": "--output-separately" }, { "name": "-I", "content": "--interruptible=interruptible\nWhen strace can be interrupted by signals (such as pressing CTRL-C).\n\n1, anywhere no signals are blocked;\n2, waiting fatal signals are blocked while decoding syscall (default);\n3, never fatal signals are always blocked (default if -o FILE PROG);\n4, nevertstp fatal signals and SIGTSTP (CTRL-Z) are always blocked (useful to\nmake strace -o FILE PROG not stop on CTRL-Z, default if -D).\n", "flag": "-I" }, { "name": "Filtering", "content": "" }, { "name": "-e", "content": "--trace=syscallset\nTrace only the specified set of system calls. syscallset is defined as\n[!]value[,value], and value can be one of the following:\n\nsyscall Trace specific syscall, specified by its name (but see NOTES).\n\n?value Question mark before the syscall qualification allows suppression of\nerror in case no syscalls matched the qualification provided.\n\n/regex Trace only those system calls that match the regex. You can use\nPOSIX Extended Regular Expression syntax (see regex(7)).\n\nsyscall@64 Trace syscall only for the 64-bit personality.\n\nsyscall@32 Trace syscall only for the 32-bit personality.\n\nsyscall@x32 Trace syscall only for the 32-on-64-bit personality.\n\n%file\nfile Trace all system calls which take a file name as an argument. You\ncan think of this as an abbreviation for\n-e trace=open,stat,chmod,unlink,... which is useful to seeing what\nfiles the process is referencing. Furthermore, using the abbrevia‐\ntion will ensure that you don't accidentally forget to include a\ncall like lstat(2) in the list. Betchya woulda forgot that one.\nThe syntax without a preceding percent sign (\"-e trace=file\") is\ndeprecated.\n\n%process\nprocess Trace system calls associated with process lifecycle (creation,\nexec, termination). The syntax without a preceding percent sign\n(\"-e trace=process\") is deprecated.\n\n%net\n%network\nnetwork Trace all the network related system calls. The syntax without a\npreceding percent sign (\"-e trace=network\") is deprecated.\n\n%signal\nsignal Trace all signal related system calls. The syntax without a preced‐\ning percent sign (\"-e trace=signal\") is deprecated.\n\n%ipc\nipc Trace all IPC related system calls. The syntax without a preceding\npercent sign (\"-e trace=ipc\") is deprecated.\n\n%desc\ndesc Trace all file descriptor related system calls. The syntax without\na preceding percent sign (\"-e trace=desc\") is deprecated.\n\n%memory\nmemory Trace all memory mapping related system calls. The syntax without a\npreceding percent sign (\"-e trace=memory\") is deprecated.\n\n%creds Trace system calls that read or modify user and group identifiers or\ncapability sets.\n\n%stat Trace stat syscall variants.\n\n%lstat Trace lstat syscall variants.\n\n%fstat Trace fstat, fstatat, and statx syscall variants.\n\n%%stat Trace syscalls used for requesting file status (stat, lstat, fstat,\nfstatat, statx, and their variants).\n\n%statfs Trace statfs, statfs64, statvfs, osfstatfs, and osfstatfs64 system\ncalls. The same effect can be achieved with\n-e trace=/^(.*)?statv?fs regular expression.\n\n%fstatfs Trace fstatfs, fstatfs64, fstatvfs, osffstatfs, and osffstatfs64\nsystem calls. The same effect can be achieved with -e trace=/fs‐‐\ntatv?fs regular expression.\n\n%%statfs Trace syscalls related to file system statistics (statfs-like, fs‐\ntatfs-like, and ustat). The same effect can be achieved with\n-e trace=/statv?fs|fsstat|ustat regular expression.\n\n%clock Trace system calls that read or modify system clocks.\n\n%pure Trace syscalls that always succeed and have no arguments. Cur‐\nrently, this list includes arcgettls(2), getdtablesize(2), gete‐‐\ngid(2), getegid32(2), geteuid(2), geteuid32(2), getgid(2), get‐‐\ngid32(2), getpagesize(2), getpgrp(2), getpid(2), getppid(2),\ngetthreadarea(2) (on architectures other than x86), gettid(2),\ngettls(2), getuid(2), getuid32(2), getxgid(2), getxpid(2),\ngetxuid(2), kernfeatures(2), and metaggettls(2) syscalls.\n\nThe -c option is useful for determining which system calls might be useful to\ntrace. For example, trace=open,close,read,write means to only trace those four\nsystem calls. Be careful when making inferences about the user/kernel boundary\nif only a subset of system calls are being monitored. The default is trace=all.\n", "flag": "-e" }, { "name": "-e", "content": "--signal=set\nTrace only the specified subset of signals. The default is signal=all. For ex‐\nample, signal=!SIGIO (or signal=!io) causes SIGIO signals not to be traced.\n", "flag": "-e" }, { "name": "-e", "content": "--status=set\nPrint only system calls with the specified return status. The default is sta‐‐\ntus=all. When using the status qualifier, because strace waits for system calls\nto return before deciding whether they should be printed or not, the traditional\norder of events may not be preserved anymore. If two system calls are executed\nby concurrent threads, strace will first print both the entry and exit of the\nfirst system call to exit, regardless of their respective entry time. The entry\nand exit of the second system call to exit will be printed afterwards. Here is\nan example when select(2) is called, but a different thread calls clockget‐‐\ntime(2) before select(2) finishes:\n\n[pid 28779] 1130322148.939977 clockgettime(CLOCKREALTIME, {1130322148, 939977000}) = 0\n[pid 28772] 1130322148.438139 select(4, [3], NULL, NULL, NULL) = 1 (in [3])\n\nset can include the following elements:\n\nsuccessful Trace system calls that returned without an error code. The -z op‐\ntion has the effect of status=successful.\nfailed Trace system calls that returned with an error code. The -Z option\nhas the effect of status=failed.\nunfinished Trace system calls that did not return. This might happen, for ex‐\nample, due to an execve call in a neighbour thread.\nunavailable Trace system calls that returned but strace failed to fetch the er‐\nror status.\ndetached Trace system calls for which strace detached before the return.\n", "flag": "-e" }, { "name": "-P", "content": "--trace-path=path\nTrace only system calls accessing path. Multiple -P options can be used to spec‐\nify several paths.\n", "flag": "-P" }, { "name": "-z", "content": "", "flag": "-z" }, { "name": "--successful-only", "content": "Print only syscalls that returned without an error code.\n", "long": "--successful-only" }, { "name": "-Z", "content": "", "flag": "-Z" }, { "name": "--failed-only", "content": "Print only syscalls that returned with an error code.\n", "long": "--failed-only" }, { "name": "Output format", "content": "" }, { "name": "-a", "content": "--columns=column\nAlign return values in a specific column (default column 40).\n", "flag": "-a" }, { "name": "-e", "content": "--abbrev=syscallset\nAbbreviate the output from printing each member of large structures. The syntax\nof the syscallset specification is the same as in the -e trace option. The de‐\nfault is abbrev=all. The -v option has the effect of abbrev=none.\n", "flag": "-e" }, { "name": "-e", "content": "--verbose=syscallset\nDereference structures for the specified set of system calls. The syntax of the\nsyscallset specification is the same as in the -e trace option. The default is\nverbose=all.\n", "flag": "-e" }, { "name": "-e", "content": "--raw=syscallset\nPrint raw, undecoded arguments for the specified set of system calls. The syntax\nof the syscallset specification is the same as in the -e trace option. This op‐\ntion has the effect of causing all arguments to be printed in hexadecimal. This\nis mostly useful if you don't trust the decoding or you need to know the actual\nnumeric value of an argument. See also -X raw option.\n", "flag": "-e" }, { "name": "-e", "content": "--read=set Perform a full hexadecimal and ASCII dump of all the data read from file descrip‐\ntors listed in the specified set. For example, to see all input activity on file\ndescriptors 3 and 5 use -e read=3,5. Note that this is independent from the nor‐\nmal tracing of the read(2) system call which is controlled by the option\n-e trace=read.\n", "flag": "-e" }, { "name": "-e", "content": "--write=set Perform a full hexadecimal and ASCII dump of all the data written to file de‐\nscriptors listed in the specified set. For example, to see all output activity\non file descriptors 3 and 5 use -e write=3,5. Note that this is independent from\nthe normal tracing of the write(2) system call which is controlled by the option\n-e trace=write.\n", "flag": "-e" }, { "name": "-e", "content": "--quiet=set\n--silent=set\n--silence=set\nSuppress various information messages. The default is quiet=none. set can in‐\nclude the following elements:\n\nattach Suppress messages about attaching and detaching (\"[ Process NNNN\nattached ]\", \"[ Process NNNN detached ]\").\nexit Suppress messages about process exits (\"+++ exited with SSS\n+++\").\npath-resolution Suppress messages about resolution of paths provided via the -P\noption (\"Requested path \"...\" resolved into \"...\"\").\npersonality Suppress messages about process personality changes (\"[ Process\nPID=NNNN runs in PPP mode. ]\").\nthread-execve\nsuperseded Suppress messages about process being superseded by execve(2) in\nanother thread (\"+++ superseded by execve in pid NNNN +++\").\n", "flag": "-e" }, { "name": "-e", "content": "--decode-fds=set\nDecode various information associated with file descriptors. The default is de‐‐\ncode-fds=none. set can include the following elements:\n\npath Print file paths. Also enables printing of tracee's current working di‐\nrectory when ATFDCWD constant is used.\nsocket Print socket protocol-specific information,\ndev Print character/block device numbers.\npidfd Print PIDs associated with pidfd file descriptors.\n", "flag": "-e" }, { "name": "-e", "content": "--decode-pids=set\nDecode various information associated with process IDs (and also thread IDs,\nprocess group IDs, and session IDs). The default is decode-pids=none. set can\ninclude the following elements:\n\ncomm Print command names associated with thread or process IDs.\npidns Print thread, process, process group, and session IDs in strace's PID\nnamespace if the tracee is in a different PID namespace.\n", "flag": "-e" }, { "name": "-e", "content": "--kvm=vcpu Print the exit reason of kvm vcpu. Requires Linux kernel version 4.16.0 or\nhigher.\n", "flag": "-e" }, { "name": "-i", "content": "", "flag": "-i" }, { "name": "--instruction-pointer", "content": "Print the instruction pointer at the time of the system call.\n", "long": "--instruction-pointer" }, { "name": "-n", "content": "", "flag": "-n" }, { "name": "--syscall-number", "content": "Print the syscall number.\n", "long": "--syscall-number" }, { "name": "-k", "content": "", "flag": "-k" }, { "name": "--stack-traces", "content": "Print the execution stack trace of the traced processes after each system call.\n", "long": "--stack-traces" }, { "name": "-o", "content": "--output=filename\nWrite the trace output to the file filename rather than to stderr. filename.pid\nform is used if -ff option is supplied. If the argument begins with '|' or '!',\nthe rest of the argument is treated as a command and all output is piped to it.\nThis is convenient for piping the debugging output to a program without affecting\nthe redirections of executed programs. The latter is not compatible with -ff op‐\ntion currently.\n", "flag": "-o" }, { "name": "-A", "content": "", "flag": "-A" }, { "name": "--output-append-mode", "content": "Open the file provided in the -o option in append mode.\n", "long": "--output-append-mode" }, { "name": "-q", "content": "", "flag": "-q" }, { "name": "--quiet", "content": "--quiet=attach,personality\nSuppress messages about attaching, detaching, and personality changes. This hap‐\npens automatically when output is redirected to a file and the command is run di‐\nrectly instead of attaching.\n", "long": "--quiet" }, { "name": "-qq", "content": "--quiet=attach,personality,exit\nSuppress messages attaching, detaching, personality changes, and about process\nexit status.\n" }, { "name": "-qqq", "content": "--quiet=all Suppress all suppressible messages (please refer to the -e quiet option descrip‐\ntion for the full list of suppressible messages).\n" }, { "name": "-r", "content": "--relative-timestamps[=precision]\nPrint a relative timestamp upon entry to each system call. This records the time\ndifference between the beginning of successive system calls. precision can be\none of s (for seconds), ms (milliseconds), us (microseconds), or ns (nanosec‐\nonds), and allows setting the precision of time value being printed. Default is", "flag": "-r" }, { "name": "us (microseconds)", "content": "measuring time difference and not the wall clock time, its measurements can dif‐\nfer from the difference in time reported by the -t option.\n" }, { "name": "-s", "content": "--string-limit=strsize\nSpecify the maximum string size to print (the default is 32). Note that file‐\nnames are not considered strings and are always printed in full.\n\n--absolute-timestamps[=[[format:]format],[[precision:]precision]]\n--timestamps[=[[format:]format],[[precision:]precision]]\nPrefix each line of the trace with the wall clock time in the specified format\nwith the specified precision. format can be one of the following:\n\nnone No time stamp is printed. Can be used to override the previous\nsetting.\ntime Wall clock time (strftime(3) format string is %T).\nunix Number of seconds since the epoch (strftime(3) format string is\n%s).\n\nprecision can be one of s (for seconds), ms (milliseconds), us (microseconds), or", "flag": "-s" }, { "name": "ns (nanoseconds)", "content": "" }, { "name": "-t", "content": "", "flag": "-t" }, { "name": "--absolute-timestamps", "content": "Prefix each line of the trace with the wall clock time.\n", "long": "--absolute-timestamps" }, { "name": "-tt", "content": "--absolute-timestamps=precision:us\nIf given twice, the time printed will include the microseconds.\n" }, { "name": "-ttt", "content": "--absolute-timestamps=format:unix,precision:us\nIf given thrice, the time printed will include the microseconds and the leading\nportion will be printed as the number of seconds since the epoch.\n" }, { "name": "-T", "content": "--syscall-times[=precision]\nShow the time spent in system calls. This records the time difference between\nthe beginning and the end of each system call. precision can be one of s (for\nseconds), ms (milliseconds), us (microseconds), or ns (nanoseconds), and allows\nsetting the precision of time value being printed. Default is us (microseconds).\n", "flag": "-T" }, { "name": "-v", "content": "--no-abbrev Print unabbreviated versions of environment, stat, termios, etc. calls. These\nstructures are very common in calls and so the default behavior displays a rea‐\nsonable subset of structure members. Use this option to get all of the gory de‐\ntails.\n\n--strings-in-hex[=option]\nControl usage of escape sequences with hexadecimal numbers in the printed\nstrings. Normally (when no --strings-in-hex or -x option is supplied), escape\nsequences are used to print non-printable and non-ASCII characters (that is,\ncharacters with a character code less than 32 or greater than 127), or to disam‐\nbiguate the output (so, for quotes and other characters that encase the printed\nstring, for example, angle brackets, in case of file descriptor path output); for\nthe former use case, unless it is a white space character that has a symbolic es‐\ncape sequence defined in the C standard (that is, “\\t” for a horizontal tab, “\\n”\nfor a newline, “\\v” for a vertical tab, “\\f” for a form feed page break, and “\\r”\nfor a carriage return) are printed using escape sequences with numbers that cor‐\nrespond to their byte values, with octal number format being the default. option\ncan be one of the following:\n\nnone Hexadecimal numbers are not used in the output at all. When\nthere is a need to emit an escape sequence, octal numbers are\nused.\nnon-ascii-chars Hexadecimal numbers are used instead of octal in the escape se‐\nquences.\nnon-ascii Strings that contain non-ASCII characters are printed using es‐\ncape sequences with hexadecimal numbers.\nall All strings are printed using escape sequences with hexadecimal\nnumbers.\n\nWhen the option is supplied without an argument, all is assumed.\n", "flag": "-v" }, { "name": "-x", "content": "--strings-in-hex=non-ascii\nPrint all non-ASCII strings in hexadecimal string format.\n", "flag": "-x" }, { "name": "-xx", "content": "--strings-in-hex[=all]\nPrint all strings in hexadecimal string format.\n" }, { "name": "-X", "content": "--const-print-style=format\nSet the format for printing of named constants and flags. Supported format val‐\nues are:\n\nraw Raw number output, without decoding.\nabbrev Output a named constant or a set of flags instead of the raw number if\nthey are found. This is the default strace behaviour.\nverbose Output both the raw value and the decoded string (as a comment).\n", "flag": "-X" }, { "name": "-y", "content": "", "flag": "-y" }, { "name": "--decode-fds", "content": "--decode-fds=path\nPrint paths associated with file descriptor arguments and with the ATFDCWD con‐\nstant.\n", "long": "--decode-fds" }, { "name": "-yy", "content": "--decode-fds=all\nPrint all available information associated with file descriptors: protocol-spe‐\ncific information associated with socket file descriptors, block/character device\nnumber associated with device file descriptors, and PIDs associated with pidfd\nfile descriptors.\n" }, { "name": "--pidns-translation", "content": "--decode-pids=pidns\nIf strace and tracee are in different PID namespaces, print PIDs in strace's\nnamespace, too.\n", "long": "--pidns-translation" }, { "name": "-Y", "content": "--decode-pids=comm\nPrint command names for PIDs.\n", "flag": "-Y" }, { "name": "Statistics", "content": "" }, { "name": "-c", "content": "", "flag": "-c" }, { "name": "--summary-only", "content": "Count time, calls, and errors for each system call and report a summary on pro‐\ngram exit, suppressing the regular output. This attempts to show system time\n(CPU time spent running in the kernel) independent of wall clock time. If -c is\nused with -f, only aggregate totals for all traced processes are kept.\n", "long": "--summary-only" }, { "name": "-C", "content": "--summary Like -c but also print regular output while processes are running.\n", "flag": "-C" }, { "name": "-O", "content": "--summary-syscall-overhead=overhead\nSet the overhead for tracing system calls to overhead. This is useful for over‐\nriding the default heuristic for guessing how much time is spent in mere measur‐\ning when timing system calls using the -c option. The accuracy of the heuristic\ncan be gauged by timing a given program run without tracing (using time(1)) and\ncomparing the accumulated system call time to the total produced using -c.\n\nThe format of overhead specification is described in section Time specification\nformat description.\n", "flag": "-O" }, { "name": "-S", "content": "--summary-sort-by=sortby\nSort the output of the histogram printed by the -c option by the specified crite‐\nrion. Legal values are time (or time-percent or time-total or total-time),\nmin-time (or shortest or time-min), max-time (or longest or time-max), avg-time\n(or time-avg), calls (or count), errors (or error), name (or syscall or\nsyscall-name), and nothing (or none); default is time.\n", "flag": "-S" }, { "name": "-U", "content": "--summary-columns=columns\nConfigure a set (and order) of columns being shown in the call summary. The col‐\numns argument is a comma-separated list with items being one of the following:\n\ntime-percent (or time) Percentage of cumulative time consumed by a\nspecific system call.\ntotal-time (or time-total) Total system (or wall clock, if -w option is\nprovided) time consumed by a specific system\ncall.\nmin-time (or shortest or time-min) Minimum observed call duration.\nmax-time (or longest or time-max) Maximum observed call duration.\navg-time (or time-avg) Average call duration.\ncalls (or count) Call count.\nerrors (or error) Error count.\nname (or syscall or syscall-name) Syscall name.\n\nThe default value is time-percent,total-time,avg-time,calls,errors,name. If the\nname field is not supplied explicitly, it is added as the last column.\n", "flag": "-U" }, { "name": "-w", "content": "", "flag": "-w" }, { "name": "--summary-wall-clock", "content": "Summarise the time difference between the beginning and end of each system call.\nThe default is to summarise the system time.\n", "long": "--summary-wall-clock" }, { "name": "Tampering", "content": "" }, { "name": "-e", "content": "ter=delay][:delayexit=delay][:pokeen‐‐\nter=@argN=DATAN,@argM=DATAM...][:pokeexit=@argN=DATAN,@argM=DATAM...][:when=expr]\n--inject=syscallset[:error=errno|:retval=value][:signal=sig][:syscall=syscall][:delayen‐‐\nter=delay][:de‐‐\nlayexit=delay][:pokeen‐‐\nter=@argN=DATAN,@argM=DATAM...][:pokeexit=@argN=DATAN,@argM=DATAM...][:when=expr]\nPerform syscall tampering for the specified set of syscalls. The syntax of the\nsyscallset specification is the same as in the -e trace option.\n\nAt least one of error, retval, signal, delayenter, delayexit, pokeenter, or\npokeexit options has to be specified. error and retval are mutually exclusive.\n\nIf :error=errno option is specified, a fault is injected into a syscall invoca‐\ntion: the syscall number is replaced by -1 which corresponds to an invalid\nsyscall (unless a syscall is specified with :syscall= option), and the error code\nis specified using a symbolic errno value like ENOSYS or a numeric value within\n1..4095 range.\n\nIf :retval=value option is specified, success injection is performed: the syscall\nnumber is replaced by -1, but a bogus success value is returned to the callee.\n\nIf :signal=sig option is specified with either a symbolic value like SIGSEGV or a\nnumeric value within 1..SIGRTMAX range, that signal is delivered on entering ev‐\nery syscall specified by the set.\n\nIf :delayenter=delay or :delayexit=delay options are specified, delay injection\nis performed: the tracee is delayed by time period specified by delay on entering\nor exiting the syscall, respectively. The format of delay specification is de‐\nscribed in section Time specification format description.\n\nIf :pokeenter=@argN=DATAN,@argM=DATAM... or\n:pokeexit=@argN=DATAN,@argM=DATAM... options are specified, tracee's memory at\nlocations, pointed to by system call arguments argN and argM (going from arg1 to\narg7) is overwritten by data DATAN and DATAM (specified in hexadecimal format;\nfor example :pokeenter=@arg1=0000DEAD0000BEEF). :pokeenter modifies memory on\nsyscall enter, and :pokeexit - on exit.\n\nIf :signal=sig option is specified without :error=errno, :retval=value or :de‐‐\nlay{enter,exit}=usecs options, then only a signal sig is delivered without a\nsyscall fault or delay injection. Conversely, :error=errno or :retval=value op‐\ntion without :delayenter=delay, :delayexit=delay or :signal=sig options injects\na fault without delivering a signal or injecting a delay, etc.\n\nIf :signal=sig option is specified together with :error=errno or :retval=value,\nthen both injection of a fault or success and signal delivery are performed.\n\nif :syscall=syscall option is specified, the corresponding syscall with no side\neffects is injected instead of -1. Currently, only \"pure\" (see -e trace=%pure\ndescription) syscalls can be specified there.\n\nUnless a :when=expr subexpression is specified, an injection is being made into\nevery invocation of each syscall from the set.\n\nThe format of the subexpression is:\n\nfirst[..last][+[step]]\n\nNumber first stands for the first invocation number in the range, number last\nstands for the last invocation number in the range, and step stands for the step\nbetween two consecutive invocations. The following combinations are useful:\n\nfirst For every syscall from the set, perform an injection for the\nsyscall invocation number first only.\nfirst..last For every syscall from the set, perform an injection for the\nsyscall invocation number first and all subsequent invocations\nuntil the invocation number last (inclusive).\nfirst+ For every syscall from the set, perform injections for the\nsyscall invocation number first and all subsequent invocations.\nfirst..last+ For every syscall from the set, perform injections for the\nsyscall invocation number first and all subsequent invocations\nuntil the invocation number last (inclusive).\nfirst+step For every syscall from the set, perform injections for syscall\ninvocations number first, first+step, first+step+step, and so\non.\nfirst..last+step Same as the previous, but consider only syscall invocations\nwith numbers up to last (inclusive).\n\nFor example, to fail each third and subsequent chdir syscalls with ENOENT, use\n-e inject=chdir:error=ENOENT:when=3+.\n\nThe valid range for numbers first and step is 1..65535, and for number last is\n1..65534.\n\nAn injection expression can contain only one error= or retval= specification, and\nonly one signal= specification. If an injection expression contains multiple\nwhen= specifications, the last one takes precedence.\n\nAccounting of syscalls that are subject to injection is done per syscall and per\ntracee.\n\nSpecification of syscall injection can be combined with other syscall filtering\noptions, for example, -P /dev/urandom -e inject=file:error=ENOENT.\n", "flag": "-e" }, { "name": "-e", "content": "--fault=syscallset[:error=errno][:when=expr]\nPerform syscall fault injection for the specified set of syscalls.\n\nThis is equivalent to more generic -e inject= expression with default value of\nerrno option set to ENOSYS.\n", "flag": "-e" }, { "name": "Miscellaneous", "content": "" }, { "name": "-d", "content": "--debug Show some debugging output of strace itself on the standard error.\n", "flag": "-d" }, { "name": "-F", "content": "may be removed in future releases. Usage of multiple instances of -F option is\nstill equivalent to a single -f, and it is ignored at all if used along with one\nor more instances of -f option.\n", "flag": "-F" }, { "name": "-h", "content": "--help Print the help summary.\n", "flag": "-h" }, { "name": "--seccomp-bpf", "content": "Try to enable use of seccomp-bpf (see seccomp(2)) to have ptrace(2)-stops only\nwhen system calls that are being traced occur in the traced processes. This op‐\ntion has no effect unless -f/--follow-forks is also specified. --seccomp-bpf is\nalso not applicable to processes attached using -p/--attach option. An attempt\nto enable system calls filtering using seccomp-bpf may fail for various reasons,\ne.g. there are too many system calls to filter, the seccomp API is not available,\nor strace itself is being traced. In cases when seccomp-bpf filter setup failed,\nstrace proceeds as usual and stops traced processes on every system call.\n", "long": "--seccomp-bpf" }, { "name": "-V", "content": "--version Print the version number of strace.\n", "flag": "-V" }, { "name": "Time specification format description", "content": "Time values can be specified as a decimal floating point number (in a format accepted by str‐‐\ntod(3)), optionally followed by one of the following suffices that specify the unit of time:" }, { "name": "s (seconds), ms (milliseconds), us (microseconds),", "content": "specified, the value is interpreted as microseconds.\n\nThe described format is used for -O, -e inject=delayenter, and -e inject=delayexit options.\n" } ] }, "DIAGNOSTICS": { "content": "When command exits, strace exits with the same exit status. If command is terminated by a\nsignal, strace terminates itself with the same signal, so that strace can be used as a wrap‐\nper process transparent to the invoking parent process. Note that parent-child relationship\n(signal stop notifications, getppid(2) value, etc) between traced process and its parent are\nnot preserved unless -D is used.\n\nWhen using -p without a command, the exit status of strace is zero unless no processes has\nbeen attached or there was an unexpected error in doing the tracing.\n", "subsections": [] }, "SETUID INSTALLATION": { "content": "If strace is installed setuid to root then the invoking user will be able to attach to and\ntrace processes owned by any user. In addition setuid and setgid programs will be executed\nand traced with the correct effective privileges. Since only users trusted with full root\nprivileges should be allowed to do these things, it only makes sense to install strace as se‐\ntuid to root when the users who can execute it are restricted to those users who have this\ntrust. For example, it makes sense to install a special version of strace with mode 'rwsr-\nxr--', user root and group trace, where members of the trace group are trusted users. If you\ndo use this feature, please remember to install a regular non-setuid version of strace for\nordinary users to use.\n", "subsections": [] }, "MULTIPLE PERSONALITIES SUPPORT": { "content": "On some architectures, strace supports decoding of syscalls for processes that use different\nABI rather than the one strace uses. Specifically, in addition to decoding native ABI,\nstrace can decode the following ABIs on the following architectures:\n\n┌───────────────────┬─────────────────────────┐\n│Architecture │ ABIs supported │\n├───────────────────┼─────────────────────────┤\n│x8664 │ i386, x32 [1]; i386 [2] │\n├───────────────────┼─────────────────────────┤\n│AArch64 │ ARM 32-bit EABI │\n├───────────────────┼─────────────────────────┤\n│PowerPC 64-bit [3] │ PowerPC 32-bit │\n├───────────────────┼─────────────────────────┤\n│s390x │ s390 │\n├───────────────────┼─────────────────────────┤\n│SPARC 64-bit │ SPARC 32-bit │\n├───────────────────┼─────────────────────────┤\n│TILE 64-bit │ TILE 32-bit │\n└───────────────────┴─────────────────────────┘\n[1] When strace is built as an x8664 application\n[2] When strace is built as an x32 application\n[3] Big endian only\n\nThis support is optional and relies on ability to generate and parse structure definitions\nduring the build time. Please refer to the output of the strace -V command in order to fig‐\nure out what support is available in your strace build (\"non-native\" refers to an ABI that\ndiffers from the ABI strace has):\n\nm32-mpers strace can trace and properly decode non-native 32-bit binaries.\nno-m32-mpers strace can trace, but cannot properly decode non-native 32-bit binaries.\nmx32-mpers strace can trace and properly decode non-native 32-on-64-bit binaries.\nno-mx32-mpers strace can trace, but cannot properly decode non-native 32-on-64-bit binaries.\n\nIf the output contains neither m32-mpers nor no-m32-mpers, then decoding of non-native 32-bit\nbinaries is not implemented at all or not applicable.\n\nLikewise, if the output contains neither mx32-mpers nor no-mx32-mpers, then decoding of non-\nnative 32-on-64-bit binaries is not implemented at all or not applicable.\n", "subsections": [] }, "NOTES": { "content": "It is a pity that so much tracing clutter is produced by systems employing shared libraries.\n\nIt is instructive to think about system call inputs and outputs as data-flow across the\nuser/kernel boundary. Because user-space and kernel-space are separate and address-pro‐\ntected, it is sometimes possible to make deductive inferences about process behavior using\ninputs and outputs as propositions.\n\nIn some cases, a system call will differ from the documented behavior or have a different\nname. For example, the faccessat(2) system call does not have flags argument, and the setr‐‐\nlimit(2) library function uses prlimit64(2) system call on modern (2.6.38+) kernels. These\ndiscrepancies are normal but idiosyncratic characteristics of the system call interface and\nare accounted for by C library wrapper functions.\n\nSome system calls have different names in different architectures and personalities. In\nthese cases, system call filtering and printing uses the names that match corresponding\nNR* kernel macros of the tracee's architecture and personality. There are two exceptions\nfrom this general rule: armfadvise6464(2) ARM syscall and xtensafadvise6464(2) Xtensa\nsyscall are filtered and printed as fadvise6464(2).\n\nOn x32, syscalls that are intended to be used by 64-bit processes and not x32 ones (for exam‐\nple, readv(2), that has syscall number 19 on x8664, with its x32 counterpart has syscall\nnumber 515), but called with X32SYSCALLBIT flag being set, are designated with #64 suf‐\nfix.\n\nOn some platforms a process that is attached to with the -p option may observe a spurious\nEINTR return from the current system call that is not restartable. (Ideally, all system\ncalls should be restarted on strace attach, making the attach invisible to the traced\nprocess, but a few system calls aren't. Arguably, every instance of such behavior is a ker‐\nnel bug.) This may have an unpredictable effect on the process if the process takes no ac‐\ntion to restart the system call.\n\nAs strace executes the specified command directly and does not employ a shell for that,\nscripts without shebang that usually run just fine when invoked by shell fail to execute with\nENOEXEC error. It is advisable to manually supply a shell as a command with the script as\nits argument.\n", "subsections": [] }, "BUGS": { "content": "Programs that use the setuid bit do not have effective user ID privileges while being traced.\n\nA traced process runs slowly (but check out the --seccomp-bpf option).\n\nTraced processes which are descended from command may be left running after an interrupt sig‐\nnal (CTRL-C).\n", "subsections": [] }, "HISTORY": { "content": "The original strace was written by Paul Kranenburg for SunOS and was inspired by its trace\nutility. The SunOS version of strace was ported to Linux and enhanced by Branko Lankester,\nwho also wrote the Linux kernel support. Even though Paul released strace 2.5 in 1992,\nBranko's work was based on Paul's strace 1.5 release from 1991. In 1993, Rick Sladkey merged\nstrace 2.5 for SunOS and the second release of strace for Linux, added many of the features\nof truss(1) from SVR4, and produced an strace that worked on both platforms. In 1994 Rick\nported strace to SVR4 and Solaris and wrote the automatic configuration support. In 1995 he\nported strace to Irix and tired of writing about himself in the third person.\n\nBeginning with 1996, strace was maintained by Wichert Akkerman. During his tenure, strace\ndevelopment migrated to CVS; ports to FreeBSD and many architectures on Linux (including ARM,\nIA-64, MIPS, PA-RISC, PowerPC, s390, SPARC) were introduced. In 2002, the burden of strace\nmaintainership was transferred to Roland McGrath. Since then, strace gained support for sev‐\neral new Linux architectures (AMD64, s390x, SuperH), bi-architecture support for some of\nthem, and received numerous additions and improvements in syscalls decoders on Linux; strace\ndevelopment migrated to git during that period. Since 2009, strace is actively maintained by\nDmitry Levin. strace gained support for AArch64, ARC, AVR32, Blackfin, Meta, Nios II, Open‐\nRISC 1000, RISC-V, Tile/TileGx, Xtensa architectures since that time. In 2012, unmaintained\nand apparently broken support for non-Linux operating systems was removed. Also, in 2012\nstrace gained support for path tracing and file descriptor path decoding. In 2014, support\nfor stack traces printing was added. In 2016, syscall fault injection was implemented.\n\nFor the additional information, please refer to the NEWS file and strace repository commit\nlog.\n", "subsections": [] }, "REPORTING BUGS": { "content": "Problems with strace should be reported to the strace mailing list\n⟨mailto:strace-devel@lists.strace.io⟩.\n", "subsections": [] }, "SEE ALSO": { "content": "strace-log-merge(1), ltrace(1), perf-trace(1), trace-cmd(1), time(1), ptrace(2), proc(5)\n\nstrace Home Page ⟨https://strace.io/⟩\n", "subsections": [] }, "AUTHORS": { "content": "The complete list of strace contributors can be found in the CREDITS file.\n\n\n\nstrace 5.16 2022-01-04 STRACE(1)", "subsections": [] } } } }