1PERF-SCRIPT(1) perf Manual PERF-SCRIPT(1)
2
6 perf-script - Read perf.data (created by perf record) and display trace
7 output
8
10 perf script [<options>]
11 perf script [<options>] record <script> [<record-options>] <command>
12 perf script [<options>] report <script> [script-args]
13 perf script [<options>] <script> <required-script-args> [<record-options>] <command>
14 perf script [<options>] <top-script> [script-args]
15
17 This command reads the input file and displays the trace recorded.
18 There are several variants of perf script:
20 'perf script' to see a detailed trace of the workload that was
22 recorded.
23 You can also run a set of pre-canned scripts that aggregate and
25 summarize the raw trace data in various ways (the list of scripts is
26 available via 'perf script -l'). The following variants allow you to
27 record and run those scripts:
28 'perf script record <script> <command>' to record the events required
30 for 'perf script report'. <script> is the name displayed in the
31 output of 'perf script --list' i.e. the actual script name minus any
32 language extension. If <command> is not specified, the events are
33 recorded using the -a (system-wide) 'perf record' option.
34 'perf script report <script> [args]' to run and display the results
36 of <script>. <script> is the name displayed in the output of 'perf
37 script --list' i.e. the actual script name minus any language
38 extension. The perf.data output from a previous run of 'perf script
39 record <script>' is used and should be present for this command to
40 succeed. [args] refers to the (mainly optional) args expected by
41 the script.
42 'perf script <script> <required-script-args> <command>' to both
44 record the events required for <script> and to run the <script>
45 using 'live-mode' i.e. without writing anything to disk. <script>
46 is the name displayed in the output of 'perf script --list' i.e. the
47 actual script name minus any language extension. If <command> is
48 not specified, the events are recorded using the -a (system-wide)
49 'perf record' option. If <script> has any required args, they
50 should be specified before <command>. This mode doesn't allow for
51 optional script args to be specified; if optional script args are
52 desired, they can be specified using separate 'perf script record'
53 and 'perf script report' commands, with the stdout of the record step
54 piped to the stdin of the report script, using the '-o -' and '-i -'
55 options of the corresponding commands.
56 'perf script <top-script>' to both record the events required for
58 <top-script> and to run the <top-script> using 'live-mode'
59 i.e. without writing anything to disk. <top-script> is the name
60 displayed in the output of 'perf script --list' i.e. the actual
61 script name minus any language extension; a <top-script> is defined
62 as any script name ending with the string 'top'.
63 [<record-options>] can be passed to the record steps of 'perf script
65 record' and 'live-mode' variants; this isn't possible however for
66 <top-script> 'live-mode' or 'perf script report' variants.
67 See the 'SEE ALSO' section for links to language-specific
69 information on how to write and run your own trace scripts.
70
72 <command>...
73 Any command you can specify in a shell.
74 -D, --dump-raw-trace=
76 Display verbose dump of the trace data.
77 -L, --Latency=
79 Show latency attributes (irqs/preemption disabled, etc).
80 -l, --list=
82 Display a list of available trace scripts.
83 -s [lang], --script=
85 Process trace data with the given script ([lang]:script[.ext]). If
86 the string lang is specified in place of a script name, a list of
87 supported languages will be displayed instead.
88 -g, --gen-script=
90 Generate perf-script.[ext] starter script for given language, using
91 current perf.data.
92 -a
94 Force system-wide collection. Scripts run without a <command>
95 normally use -a by default, while scripts run with a <command>
96 normally don’t - this option allows the latter to be run in
97 system-wide mode.
98 -i, --input=
100 Input file name. (default: perf.data unless stdin is a fifo)
101 -d, --debug-mode
103 Do various checks like samples ordering and lost events.
104 -F, --fields
106 Comma separated list of fields to print. Options are: comm, tid,
107 pid, time, cpu, event, trace, ip, sym, dso, addr, symoff, srcline,
108 period, iregs, uregs, brstack, brstacksym, flags, bpf-output,
109 brstackinsn, brstackoff, callindent, insn, insnlen, synth,
110 phys_addr, metric, misc. Field list can be prepended with the type,
111 trace, sw or hw, to indicate to which event type the field list
112 applies. e.g., -F sw:comm,tid,time,ip,sym and -F
113 trace:time,cpu,trace
114 perf script -F <fields>
116 is equivalent to:
118 perf script -F trace:<fields> -F sw:<fields> -F hw:<fields>
120 i.e., the specified fields apply to all event types if the type string
122 is not given.
123 In addition to overriding fields, it is also possible to add or remove
125 fields from the defaults. For example
126 -F -cpu,+insn
128 removes the cpu field and adds the insn field. Adding/removing fields
130 cannot be mixed with normal overriding.
131 The arguments are processed in the order received. A later usage can
133 reset a prior request. e.g.:
134 -F trace: -F comm,tid,time,ip,sym
136 The first -F suppresses trace events (field list is ""), but then the
138 second invocation sets the fields to comm,tid,time,ip,sym. In this case a
139 warning is given to the user:
140 "Overriding previous field request for all events."
142 Alternatively, consider the order:
144 -F comm,tid,time,ip,sym -F trace:
146 The first -F sets the fields for all events and the second -F
148 suppresses trace events. The user is given a warning message about
149 the override, and the result of the above is that only S/W and H/W
150 events are displayed with the given fields.
151 For the 'wildcard' option if a user selected field is invalid for an
153 event type, a message is displayed to the user that the option is
154 ignored for that type. For example:
155 $ perf script -F comm,tid,trace
157 'trace' not valid for hardware events. Ignoring.
158 'trace' not valid for software events. Ignoring.
159 Alternatively, if the type is given an invalid field is specified it
161 is an error. For example:
162 perf script -v -F sw:comm,tid,trace
164 'trace' not valid for software events.
165 At this point usage is displayed, and perf-script exits.
167 The flags field is synthesized and may have a value when Instruction
169 Trace decoding. The flags are "bcrosyiABEx" which stand for branch,
170 call, return, conditional, system, asynchronous, interrupt,
171 transaction abort, trace begin, trace end, and in transaction,
172 respectively. Known combinations of flags are printed more nicely e.g.
173 "call" for "bc", "return" for "br", "jcc" for "bo", "jmp" for "b",
174 "int" for "bci", "iret" for "bri", "syscall" for "bcs", "sysret" for "brs",
175 "async" for "by", "hw int" for "bcyi", "tx abrt" for "bA", "tr strt" for "bB",
176 "tr end" for "bE". However the "x" flag will be display separately in those
177 cases e.g. "jcc (x)" for a condition branch within a transaction.
178 The callindent field is synthesized and may have a value when
180 Instruction Trace decoding. For calls and returns, it will display the
181 name of the symbol indented with spaces to reflect the stack depth.
182 When doing instruction trace decoding insn and insnlen give the
184 instruction bytes and the instruction length of the current
185 instruction.
186 The synth field is used by synthesized events which may be created when
188 Instruction Trace decoding.
189 Finally, a user may not set fields to none for all event types.
191 i.e., -F "" is not allowed.
192 The brstack output includes branch related information with raw addresses using the
194 /v/v/v/v/cycles syntax in the following order:
195 FROM: branch source instruction
196 TO : branch target instruction
197 M/P/-: M=branch target mispredicted or branch direction was mispredicted, P=target predicted or direction predicted, -=not supported
198 X/- : X=branch inside a transactional region, -=not in transaction region or not supported
199 A/- : A=TSX abort entry, -=not aborted region or not supported
200 cycles
201 The brstacksym is identical to brstack, except that the FROM and TO addresses are printed in a symbolic form if possible.
203 When brstackinsn is specified the full assembler sequences of branch sequences for each sample
205 is printed. This is the full execution path leading to the sample. This is only supported when the
206 sample was recorded with perf record -b or -j any.
207 The brstackoff field will print an offset into a specific dso/binary.
209 With the metric option perf script can compute metrics for
211 sampling periods, similar to perf stat. This requires
212 specifying a group with multiple metrics with the :S option
213 for perf record. perf will sample on the first event, and
214 compute metrics for all the events in the group. Please note
215 that the metric computed is averaged over the whole sampling
216 period, not just for the sample point.
217 For sample events it's possible to display misc field with -F +misc option,
219 following letters are displayed for each bit:
220 PERF_RECORD_MISC_KERNEL K
222 PERF_RECORD_MISC_USER U
223 PERF_RECORD_MISC_HYPERVISOR H
224 PERF_RECORD_MISC_GUEST_KERNEL G
225 PERF_RECORD_MISC_GUEST_USER g
226 PERF_RECORD_MISC_MMAP_DATA* M
227 PERF_RECORD_MISC_COMM_EXEC E
228 PERF_RECORD_MISC_SWITCH_OUT S
229 PERF_RECORD_MISC_SWITCH_OUT_PREEMPT Sp
230 $ perf script -F +misc ...
232 sched-messaging 1414 K 28690.636582: 4590 cycles ...
233 sched-messaging 1407 U 28690.636600: 325620 cycles ...
234 sched-messaging 1414 K 28690.636608: 19473 cycles ...
235 misc field ___________/
236 -k, --vmlinux=<file>
238 vmlinux pathname
239 --kallsyms=<file>
241 kallsyms pathname
242 --symfs=<directory>
244 Look for files with symbols relative to this directory.
245 -G, --hide-call-graph
247 When printing symbols do not display call chain.
248 --stop-bt
250 Stop display of callgraph at these symbols
251 -C, --cpu
253 Only report samples for the list of CPUs provided. Multiple CPUs
254 can be provided as a comma-separated list with no space: 0,1.
255 Ranges of CPUs are specified with -: 0-2. Default is to report
256 samples on all CPUs.
257 -c, --comms=
259 Only display events for these comms. CSV that understands
260 file://filename entries.
261 --pid=
263 Only show events for given process ID (comma separated list).
264 --tid=
266 Only show events for given thread ID (comma separated list).
267 -I, --show-info
269 Display extended information about the perf.data file. This adds
270 information which may be very large and thus may clutter the
271 display. It currently includes: cpu and numa topology of the host
272 system. It can only be used with the perf script report mode.
273 --show-kernel-path
275 Try to resolve the path of [kernel.kallsyms]
276 --show-task-events Display task related events (e.g. FORK, COMM, EXIT).
278 --show-mmap-events Display mmap related events (e.g. MMAP, MMAP2).
280 --show-namespace-events Display namespace events i.e. events of type
282 PERF_RECORD_NAMESPACES.
283 --show-switch-events Display context switch events i.e. events of type
285 PERF_RECORD_SWITCH or PERF_RECORD_SWITCH_CPU_WIDE.
286 --show-lost-events Display lost events i.e. events of type
288 PERF_RECORD_LOST.
289 --show-round-events Display finished round events i.e. events of type
291 PERF_RECORD_FINISHED_ROUND.
292 --demangle
294 Demangle symbol names to human readable form. It’s enabled by
295 default, disable with --no-demangle.
296 --demangle-kernel
298 Demangle kernel symbol names to human readable form (for C++
299 kernels).
300 --header Show perf.data header.
302 --header-only Show only perf.data header.
304 --itrace
306 Options for decoding instruction tracing data. The options are:
307 i synthesize instructions events
309 b synthesize branches events
310 c synthesize branches events (calls only)
311 r synthesize branches events (returns only)
312 x synthesize transactions events
313 w synthesize ptwrite events
314 p synthesize power events
315 e synthesize error events
316 d create a debug log
317 g synthesize a call chain (use with i or x)
318 l synthesize last branch entries (use with i or x)
319 s skip initial number of events
320 The default is all events i.e. the same as --itrace=ibxwpe
322 In addition, the period (default 100000) for instructions events
324 can be specified in units of:
325 i instructions
327 t ticks
328 ms milliseconds
329 us microseconds
330 ns nanoseconds (default)
331 Also the call chain size (default 16, max. 1024) for instructions or
333 transactions events can be specified.
334 Also the number of last branch entries (default 64, max. 1024) for
336 instructions or transactions events can be specified.
337 It is also possible to skip events generated (instructions, branches, transactions,
339 ptwrite, power) at the beginning. This is useful to ignore initialization code.
340 --itrace=i0nss1000000
342 skips the first million instructions.
344 To disable decoding entirely, use --no-itrace.
346 --full-source-path
348 Show the full path for source files for srcline output.
349 --max-stack
351 Set the stack depth limit when parsing the callchain, anything
352 beyond the specified depth will be ignored. This is a trade-off
353 between information loss and faster processing especially for
354 workloads that can have a very long callchain stack. Note that when
355 using the --itrace option the synthesized callchain size will
356 override this value if the synthesized callchain size is bigger.
357 Default: 127
359 --ns
361 Use 9 decimal places when displaying time (i.e. show the
362 nanoseconds)
363 -f, --force
365 Don’t do ownership validation.
366 --time
368 Only analyze samples within given time window: <start>,<stop>.
369 Times have the format seconds.microseconds. If start is not given
370 (i.e., time string is ,x.y) then analysis starts at the beginning
371 of the file. If stop time is not given (i.e, time string is x.y,)
372 then analysis goes to end of file.
373 Also support time percent with multipe time range. Time string is
375 'a%/n,b%/m,...' or 'a%-b%,c%-%d,...'.
376 For example:
378 Select the second 10% time slice:
379 perf script --time 10%/2
380 Select from 0% to 10% time slice:
382 perf script --time 0%-10%
383 Select the first and second 10% time slices:
385 perf script --time 10%/1,10%/2
386 Select from 0% to 10% and 30% to 40% slices:
388 perf script --time 0%-10%,30%-40%
389 --max-blocks
391 Set the maximum number of program blocks to print with brstackasm
392 for each sample.
393 --per-event-dump
395 Create per event files with a "perf.data.EVENT.dump" name instead
396 of printing to stdout, useful, for instance, for generating
397 flamegraphs.
398 --inline
400 If a callgraph address belongs to an inlined function, the inline
401 stack will be printed. Each entry has function name and file/line.
402 Enabled by default, disable with --no-inline.
403
405 perf-record(1), perf-script-perl(1), perf-script-python(1)
406perf 09/24/2019 PERF-SCRIPT(1)