]>
Commit | Line | Data |
---|---|---|
210ff7bf | 1 | SYSTEMTAP EXAMPLES INDEX BY KEYWORD |
09a98ae0 | 2 | (see also index.txt) |
210ff7bf FCE |
3 | |
4 | = BACKTRACE = | |
5 | ||
1cc8a4c7 WC |
6 | interrupt/scf.stp - Tally Backtraces for Inter-Processor Interrupt (IPI) |
7 | keywords: interrupt backtrace | |
8 | ||
9 | The Linux kernel function smp_call_function causes expensive | |
10 | inter-processor interrupts (IPIs). The scf.stp script tallies the | |
11 | processes and backtraces causing the interprocessor interrupts to | |
12 | identify the cause of the expensive IPI. On exit the script prints | |
13 | the tallies in descending frequency. | |
14 | ||
15 | ||
210ff7bf | 16 | io/io_submit.stp - Tally Reschedule Reason During AIO io_submit Call |
09a98ae0 | 17 | keywords: io backtrace |
210ff7bf FCE |
18 | |
19 | When a reschedule occurs during an AIO io_submit call, accumulate the | |
20 | traceback in a histogram. When the script exits prints out a sorted | |
21 | list from most common to least common backtrace. | |
22 | ||
23 | ||
cae71dd3 | 24 | process/sleepingBeauties.stp - Generating Backtraces of Threads Waiting for IO Operations |
09a98ae0 | 25 | keywords: io scheduler backtrace |
cae71dd3 FCE |
26 | |
27 | The script monitors the time that threads spend waiting for IO | |
28 | operations (in "D" state) in the wait_for_completion function. If a | |
29 | thread spends over 10ms, its name and backtrace is printed, and later | |
30 | so is the total delay. | |
31 | ||
32 | ||
210ff7bf FCE |
33 | = CALLGRAPH = |
34 | ||
aa15b9f0 | 35 | general/para-callgraph.stp - Callgraph tracing with arguments |
09a98ae0 | 36 | keywords: trace callgraph |
210ff7bf | 37 | |
aa15b9f0 FCE |
38 | Print a timed per-thread callgraph, complete with function parameters |
39 | and return values. The first parameter names the function probe | |
40 | points to trace. The optional second parameter names the probe | |
41 | points for trigger functions, which acts to enable tracing for only | |
42 | those functions that occur while the current thread is nested within | |
43 | the trigger. | |
210ff7bf FCE |
44 | |
45 | ||
46 | = CPU = | |
47 | ||
48 | general/graphs.stp - Graphing Disk and CPU Utilization | |
09a98ae0 | 49 | keywords: disk cpu use graph |
210ff7bf FCE |
50 | |
51 | The script tracks the disk and CPU utilization. The resulting output | |
52 | of the script can be piped into gnuplot to generate a graph of disk | |
53 | and CPU USE. | |
54 | ||
55 | ||
56 | = DISK = | |
57 | ||
58 | general/graphs.stp - Graphing Disk and CPU Utilization | |
09a98ae0 | 59 | keywords: disk cpu use graph |
210ff7bf FCE |
60 | |
61 | The script tracks the disk and CPU utilization. The resulting output | |
62 | of the script can be piped into gnuplot to generate a graph of disk | |
63 | and CPU USE. | |
64 | ||
65 | ||
66 | io/disktop.stp - Summarize Disk Read/Write Traffic | |
09a98ae0 | 67 | keywords: disk |
210ff7bf FCE |
68 | |
69 | Get the status of reading/writing disk every 5 seconds, output top | |
70 | ten entries during that period. | |
71 | ||
72 | ||
73 | = FUNCTIONS = | |
74 | ||
75 | profiling/functioncallcount.stp - Count Times Functions Called | |
09a98ae0 | 76 | keywords: profiling functions |
210ff7bf FCE |
77 | |
78 | The functioncallcount.stp script takes one argument, a list of | |
79 | functions to probe. The script will run and count the number of times | |
80 | that each of the functions on the list is called. On exit the script | |
81 | will print a sorted list from most frequently to least frequently | |
82 | called function. | |
83 | ||
84 | ||
85 | = FUTEX = | |
86 | ||
87 | process/futexes.stp - System-Wide Futex Contention | |
09a98ae0 | 88 | keywords: syscall locking futex |
210ff7bf FCE |
89 | |
90 | The script watches the futex syscall on the system. On exit the | |
91 | futexes address, the number of contentions, and the average time for | |
92 | each contention on the futex are printed from lowest pid number to | |
93 | highest. | |
94 | ||
95 | ||
96 | = GRAPH = | |
97 | ||
98 | general/graphs.stp - Graphing Disk and CPU Utilization | |
09a98ae0 | 99 | keywords: disk cpu use graph |
210ff7bf FCE |
100 | |
101 | The script tracks the disk and CPU utilization. The resulting output | |
102 | of the script can be piped into gnuplot to generate a graph of disk | |
103 | and CPU USE. | |
104 | ||
105 | ||
1cc8a4c7 WC |
106 | = INTERRUPT = |
107 | ||
108 | interrupt/scf.stp - Tally Backtraces for Inter-Processor Interrupt (IPI) | |
109 | keywords: interrupt backtrace | |
110 | ||
111 | The Linux kernel function smp_call_function causes expensive | |
112 | inter-processor interrupts (IPIs). The scf.stp script tallies the | |
113 | processes and backtraces causing the interprocessor interrupts to | |
114 | identify the cause of the expensive IPI. On exit the script prints | |
115 | the tallies in descending frequency. | |
116 | ||
117 | ||
210ff7bf FCE |
118 | = IO = |
119 | ||
120 | io/io_submit.stp - Tally Reschedule Reason During AIO io_submit Call | |
09a98ae0 | 121 | keywords: io backtrace |
210ff7bf FCE |
122 | |
123 | When a reschedule occurs during an AIO io_submit call, accumulate the | |
124 | traceback in a histogram. When the script exits prints out a sorted | |
125 | list from most common to least common backtrace. | |
126 | ||
127 | ||
128 | io/iotime.stp - Trace Time Spent in Read and Write for Files | |
09a98ae0 | 129 | keywords: syscall read write time io |
210ff7bf FCE |
130 | |
131 | The script watches each open, close, read, and write syscalls on the | |
132 | system. For each file the scripts observes opened it accumulates the | |
133 | amount of wall clock time spend in read and write operations and the | |
134 | number of bytes read and written. When a file is closed the script | |
135 | prints out a pair of lines for the file. Both lines begin with a | |
136 | timestamp in microseconds, the PID number, and the executable name in | |
137 | parenthesese. The first line with the "access" keyword lists the file | |
138 | name, the attempted number of bytes for the read and write | |
139 | operations. The second line with the "iotime" keyword list the file | |
140 | name and the number of microseconds accumulated in the read and write | |
141 | syscalls. | |
142 | ||
143 | ||
144 | io/iotop.stp - Periodically Print I/O Activity by Process Name | |
09a98ae0 | 145 | keywords: io |
210ff7bf FCE |
146 | |
147 | Every five seconds print out the top ten executables generating I/O | |
148 | traffic during that interval sorted in descending order. | |
149 | ||
150 | ||
151 | io/traceio.stp - Track Cumulative I/O Activity by Process Name | |
09a98ae0 | 152 | keywords: io |
210ff7bf FCE |
153 | |
154 | Every second print out the top ten executables sorted in descending | |
155 | order based on cumulative I/O traffic observed. | |
156 | ||
157 | ||
158 | io/traceio2.stp - Watch I/O Activity on a Particular Device | |
09a98ae0 | 159 | keywords: io |
210ff7bf FCE |
160 | |
161 | Print out the executable name and process number as reads and writes | |
162 | to the specified device occur. | |
163 | ||
164 | ||
165 | process/sleepingBeauties.stp - Generating Backtraces of Threads Waiting for IO Operations | |
09a98ae0 | 166 | keywords: io scheduler backtrace |
210ff7bf | 167 | |
cae71dd3 FCE |
168 | The script monitors the time that threads spend waiting for IO |
169 | operations (in "D" state) in the wait_for_completion function. If a | |
170 | thread spends over 10ms, its name and backtrace is printed, and later | |
171 | so is the total delay. | |
210ff7bf FCE |
172 | |
173 | ||
174 | = LOCKING = | |
175 | ||
176 | process/futexes.stp - System-Wide Futex Contention | |
09a98ae0 | 177 | keywords: syscall locking futex |
210ff7bf FCE |
178 | |
179 | The script watches the futex syscall on the system. On exit the | |
180 | futexes address, the number of contentions, and the average time for | |
181 | each contention on the futex are printed from lowest pid number to | |
182 | highest. | |
183 | ||
184 | ||
185 | = NETWORK = | |
186 | ||
187 | network/nettop.stp - Periodic Listing of Processes Using Network Interfaces | |
09a98ae0 | 188 | keywords: network traffic per-process |
210ff7bf FCE |
189 | |
190 | Every five seconds the nettop.stp script prints out a list of | |
191 | processed (PID and command) with the number of packets sent/received | |
192 | and the amount of data sent/received by the process during that | |
193 | interval. | |
194 | ||
195 | ||
196 | network/socket-trace.stp - Trace Functions called in Network Socket Code | |
09a98ae0 | 197 | keywords: network socket |
210ff7bf FCE |
198 | |
199 | The script instrument each of the functions inn the Linux kernel's | |
200 | net/socket.c file. The script prints out trace. The first element of | |
201 | a line is time delta in microseconds from the previous entry. This | |
202 | is followed by the command name and the PID. The "->" and "<-" | |
203 | indicates function entry and function exit, respectively. The last | |
204 | element of the line is the function name. | |
205 | ||
206 | ||
207 | = PER-PROCESS = | |
208 | ||
209 | network/nettop.stp - Periodic Listing of Processes Using Network Interfaces | |
09a98ae0 | 210 | keywords: network traffic per-process |
210ff7bf FCE |
211 | |
212 | Every five seconds the nettop.stp script prints out a list of | |
213 | processed (PID and command) with the number of packets sent/received | |
214 | and the amount of data sent/received by the process during that | |
215 | interval. | |
216 | ||
217 | ||
218 | = PROFILING = | |
219 | ||
220 | process/pf2.stp - Profile kernel functions | |
09a98ae0 | 221 | keywords: profiling |
210ff7bf FCE |
222 | |
223 | The pf2.stp script sets up time-based sampling. Every five seconds it | |
224 | prints out a sorted list with the top ten kernel functions with | |
225 | samples. | |
226 | ||
227 | ||
228 | profiling/functioncallcount.stp - Count Times Functions Called | |
09a98ae0 | 229 | keywords: profiling functions |
210ff7bf FCE |
230 | |
231 | The functioncallcount.stp script takes one argument, a list of | |
232 | functions to probe. The script will run and count the number of times | |
233 | that each of the functions on the list is called. On exit the script | |
234 | will print a sorted list from most frequently to least frequently | |
235 | called function. | |
236 | ||
237 | ||
238 | profiling/thread-times.stp - Profile kernel functions | |
09a98ae0 | 239 | keywords: profiling |
210ff7bf FCE |
240 | |
241 | The thread-times.stp script sets up time-based sampling. Every five | |
242 | seconds it prints out a sorted list with the top twenty processes | |
243 | with samples broken down into percentage total time spent in | |
244 | user-space and kernel-space. | |
245 | ||
246 | ||
ff90b297 WC |
247 | profiling/timeout.stp - Show Processes Doing Polling Operations |
248 | keywords: profiling | |
249 | ||
250 | The timeout.stp script is based on a blog entry | |
251 | (http://udrepper.livejournal.com/19041.html) mentioning a need for a | |
252 | tool to help developers find applications that are polling. The | |
253 | timeout.stp script monitors systemcall used for polling and records | |
254 | the systemcalls that timed out rather than returned because some | |
255 | action occurred. The script updates the screen once a second with the | |
256 | top twenty processes. | |
257 | ||
258 | ||
210ff7bf FCE |
259 | = READ = |
260 | ||
261 | io/iotime.stp - Trace Time Spent in Read and Write for Files | |
09a98ae0 | 262 | keywords: syscall read write time io |
210ff7bf FCE |
263 | |
264 | The script watches each open, close, read, and write syscalls on the | |
265 | system. For each file the scripts observes opened it accumulates the | |
266 | amount of wall clock time spend in read and write operations and the | |
267 | number of bytes read and written. When a file is closed the script | |
268 | prints out a pair of lines for the file. Both lines begin with a | |
269 | timestamp in microseconds, the PID number, and the executable name in | |
270 | parenthesese. The first line with the "access" keyword lists the file | |
271 | name, the attempted number of bytes for the read and write | |
272 | operations. The second line with the "iotime" keyword list the file | |
273 | name and the number of microseconds accumulated in the read and write | |
274 | syscalls. | |
275 | ||
276 | ||
277 | = SCHEDULER = | |
278 | ||
279 | process/sleepingBeauties.stp - Generating Backtraces of Threads Waiting for IO Operations | |
09a98ae0 | 280 | keywords: io scheduler backtrace |
cae71dd3 FCE |
281 | |
282 | The script monitors the time that threads spend waiting for IO | |
283 | operations (in "D" state) in the wait_for_completion function. If a | |
284 | thread spends over 10ms, its name and backtrace is printed, and later | |
285 | so is the total delay. | |
210ff7bf FCE |
286 | |
287 | ||
288 | = SIGNALS = | |
289 | ||
290 | process/sig_by_pid.stp - Signal Counts by Process ID | |
09a98ae0 | 291 | keywords: signals |
210ff7bf FCE |
292 | |
293 | Print signal counts by process ID in descending order. | |
294 | ||
295 | ||
296 | process/sig_by_proc.stp - Signal Counts by Process Name | |
09a98ae0 | 297 | keywords: signals |
210ff7bf FCE |
298 | |
299 | Print signal counts by process name in descending order. | |
300 | ||
301 | ||
302 | process/sigkill.stp - Track SIGKILL Signals | |
09a98ae0 | 303 | keywords: signals |
210ff7bf FCE |
304 | |
305 | The script traces any SIGKILL signals. When that SIGKILL signal is | |
306 | sent to a process, the script prints out the signal name, the | |
307 | desination executable and process ID, the executable name user ID | |
308 | that sent the signal. | |
309 | ||
310 | ||
311 | process/syscalls_by_pid.stp - System-Wide Count of Syscalls by PID | |
09a98ae0 | 312 | keywords: signals |
210ff7bf FCE |
313 | |
314 | The script watches for a particular signal sent to a specific | |
315 | process. When that signal is sent to the specified process, the | |
316 | script prints out the PID and executable of the process sending the | |
317 | signal, the PID and executable name of the process receiving the | |
318 | signal, and the signal number and name. | |
319 | ||
320 | ||
321 | = SIMPLE = | |
322 | ||
323 | general/helloworld.stp - SystemTap "Hello World" Program | |
09a98ae0 | 324 | keywords: simple |
210ff7bf FCE |
325 | |
326 | A basic "Hello World" program implemented in SystemTap script. It | |
327 | prints out "hello world" message and then immediately exits. | |
328 | ||
329 | ||
330 | = SLEEP = | |
331 | ||
332 | process/sleeptime.stp - Trace Time Spent in nanosleep Syscalls | |
09a98ae0 | 333 | keywords: syscall sleep |
210ff7bf FCE |
334 | |
335 | The script watches each nanosleep syscall on the system. At the end | |
336 | of each nanosleep syscall the script prints out a line with a | |
337 | timestamp in microseconds, the pid, the executable name in | |
338 | paretheses, the "nanosleep:" key, and the duration of the sleep in | |
339 | microseconds. | |
340 | ||
341 | ||
342 | = SOCKET = | |
343 | ||
344 | network/socket-trace.stp - Trace Functions called in Network Socket Code | |
09a98ae0 | 345 | keywords: network socket |
210ff7bf FCE |
346 | |
347 | The script instrument each of the functions inn the Linux kernel's | |
348 | net/socket.c file. The script prints out trace. The first element of | |
349 | a line is time delta in microseconds from the previous entry. This | |
350 | is followed by the command name and the PID. The "->" and "<-" | |
351 | indicates function entry and function exit, respectively. The last | |
352 | element of the line is the function name. | |
353 | ||
354 | ||
355 | = SYSCALL = | |
356 | ||
357 | io/iotime.stp - Trace Time Spent in Read and Write for Files | |
09a98ae0 | 358 | keywords: syscall read write time io |
210ff7bf FCE |
359 | |
360 | The script watches each open, close, read, and write syscalls on the | |
361 | system. For each file the scripts observes opened it accumulates the | |
362 | amount of wall clock time spend in read and write operations and the | |
363 | number of bytes read and written. When a file is closed the script | |
364 | prints out a pair of lines for the file. Both lines begin with a | |
365 | timestamp in microseconds, the PID number, and the executable name in | |
366 | parenthesese. The first line with the "access" keyword lists the file | |
367 | name, the attempted number of bytes for the read and write | |
368 | operations. The second line with the "iotime" keyword list the file | |
369 | name and the number of microseconds accumulated in the read and write | |
370 | syscalls. | |
371 | ||
372 | ||
373 | process/futexes.stp - System-Wide Futex Contention | |
09a98ae0 | 374 | keywords: syscall locking futex |
210ff7bf FCE |
375 | |
376 | The script watches the futex syscall on the system. On exit the | |
377 | futexes address, the number of contentions, and the average time for | |
378 | each contention on the futex are printed from lowest pid number to | |
379 | highest. | |
380 | ||
381 | ||
382 | process/sleeptime.stp - Trace Time Spent in nanosleep Syscalls | |
09a98ae0 | 383 | keywords: syscall sleep |
210ff7bf FCE |
384 | |
385 | The script watches each nanosleep syscall on the system. At the end | |
386 | of each nanosleep syscall the script prints out a line with a | |
387 | timestamp in microseconds, the pid, the executable name in | |
388 | paretheses, the "nanosleep:" key, and the duration of the sleep in | |
389 | microseconds. | |
390 | ||
391 | ||
392 | process/syscalls_by_pid.stp - System-Wide Count of Syscalls by PID | |
09a98ae0 | 393 | keywords: syscall |
210ff7bf FCE |
394 | |
395 | The script watches all syscall on the system. On exit the script | |
396 | prints a list showing the number of systemcalls executed by each PID | |
397 | ordered from greatest to least number of syscalls. | |
398 | ||
399 | ||
400 | process/syscalls_by_proc.stp - System-Wide Count of Syscalls by Executable | |
09a98ae0 | 401 | keywords: syscall |
210ff7bf FCE |
402 | |
403 | The script watches all syscall on the system. On exit the script | |
404 | prints a list showing the number of systemcalls executed by each | |
405 | executable ordered from greates to least number of syscalls. | |
406 | ||
407 | ||
408 | process/wait4time.stp - Trace Time Spent in wait4 Syscalls | |
09a98ae0 | 409 | keywords: syscall wait4 |
210ff7bf FCE |
410 | |
411 | The script watches each wait4 syscall on the system. At the end of | |
412 | each wait4 syscall the script prints out a line with a timestamp in | |
413 | microseconds, the pid, the executable name in paretheses, the | |
414 | "wait4:" key, the duration of the wait and the PID that the wait4 was | |
415 | waiting for. If the waited for PID is not specified , it is "-1". | |
416 | ||
417 | ||
418 | = TIME = | |
419 | ||
420 | io/iotime.stp - Trace Time Spent in Read and Write for Files | |
09a98ae0 | 421 | keywords: syscall read write time io |
210ff7bf FCE |
422 | |
423 | The script watches each open, close, read, and write syscalls on the | |
424 | system. For each file the scripts observes opened it accumulates the | |
425 | amount of wall clock time spend in read and write operations and the | |
426 | number of bytes read and written. When a file is closed the script | |
427 | prints out a pair of lines for the file. Both lines begin with a | |
428 | timestamp in microseconds, the PID number, and the executable name in | |
429 | parenthesese. The first line with the "access" keyword lists the file | |
430 | name, the attempted number of bytes for the read and write | |
431 | operations. The second line with the "iotime" keyword list the file | |
432 | name and the number of microseconds accumulated in the read and write | |
433 | syscalls. | |
434 | ||
435 | ||
436 | = TRACE = | |
437 | ||
aa15b9f0 | 438 | general/para-callgraph.stp - Callgraph tracing with arguments |
09a98ae0 | 439 | keywords: trace callgraph |
210ff7bf | 440 | |
aa15b9f0 FCE |
441 | Print a timed per-thread callgraph, complete with function parameters |
442 | and return values. The first parameter names the function probe | |
443 | points to trace. The optional second parameter names the probe | |
444 | points for trigger functions, which acts to enable tracing for only | |
445 | those functions that occur while the current thread is nested within | |
446 | the trigger. | |
210ff7bf FCE |
447 | |
448 | ||
449 | = TRAFFIC = | |
450 | ||
451 | network/nettop.stp - Periodic Listing of Processes Using Network Interfaces | |
09a98ae0 | 452 | keywords: network traffic per-process |
210ff7bf FCE |
453 | |
454 | Every five seconds the nettop.stp script prints out a list of | |
455 | processed (PID and command) with the number of packets sent/received | |
456 | and the amount of data sent/received by the process during that | |
457 | interval. | |
458 | ||
459 | ||
460 | = USE = | |
461 | ||
462 | general/graphs.stp - Graphing Disk and CPU Utilization | |
09a98ae0 | 463 | keywords: disk cpu use graph |
210ff7bf FCE |
464 | |
465 | The script tracks the disk and CPU utilization. The resulting output | |
466 | of the script can be piped into gnuplot to generate a graph of disk | |
467 | and CPU USE. | |
468 | ||
469 | ||
470 | = WAIT4 = | |
471 | ||
472 | process/wait4time.stp - Trace Time Spent in wait4 Syscalls | |
09a98ae0 | 473 | keywords: syscall wait4 |
210ff7bf FCE |
474 | |
475 | The script watches each wait4 syscall on the system. At the end of | |
476 | each wait4 syscall the script prints out a line with a timestamp in | |
477 | microseconds, the pid, the executable name in paretheses, the | |
478 | "wait4:" key, the duration of the wait and the PID that the wait4 was | |
479 | waiting for. If the waited for PID is not specified , it is "-1". | |
480 | ||
481 | ||
482 | = WRITE = | |
483 | ||
484 | io/iotime.stp - Trace Time Spent in Read and Write for Files | |
09a98ae0 | 485 | keywords: syscall read write time io |
210ff7bf FCE |
486 | |
487 | The script watches each open, close, read, and write syscalls on the | |
488 | system. For each file the scripts observes opened it accumulates the | |
489 | amount of wall clock time spend in read and write operations and the | |
490 | number of bytes read and written. When a file is closed the script | |
491 | prints out a pair of lines for the file. Both lines begin with a | |
492 | timestamp in microseconds, the PID number, and the executable name in | |
493 | parenthesese. The first line with the "access" keyword lists the file | |
494 | name, the attempted number of bytes for the read and write | |
495 | operations. The second line with the "iotime" keyword list the file | |
496 | name and the number of microseconds accumulated in the read and write | |
497 | syscalls. | |
498 | ||
499 |