1perf-script(1) 2============= 3 4NAME 5---- 6perf-script - Read perf.data (created by perf record) and display trace output 7 8SYNOPSIS 9-------- 10[verse] 11'perf script' [<options>] 12'perf script' [<options>] record <script> [<record-options>] <command> 13'perf script' [<options>] report <script> [script-args] 14'perf script' [<options>] <script> <required-script-args> [<record-options>] <command> 15'perf script' [<options>] <top-script> [script-args] 16 17DESCRIPTION 18----------- 19This command reads the input file and displays the trace recorded. 20 21There are several variants of perf script: 22 23 'perf script' to see a detailed trace of the workload that was 24 recorded. 25 26 You can also run a set of pre-canned scripts that aggregate and 27 summarize the raw trace data in various ways (the list of scripts is 28 available via 'perf script -l'). The following variants allow you to 29 record and run those scripts: 30 31 'perf script record <script> <command>' to record the events required 32 for 'perf script report'. <script> is the name displayed in the 33 output of 'perf script --list' i.e. the actual script name minus any 34 language extension. If <command> is not specified, the events are 35 recorded using the -a (system-wide) 'perf record' option. 36 37 'perf script report <script> [args]' to run and display the results 38 of <script>. <script> is the name displayed in the output of 'perf 39 script --list' i.e. the actual script name minus any language 40 extension. The perf.data output from a previous run of 'perf script 41 record <script>' is used and should be present for this command to 42 succeed. [args] refers to the (mainly optional) args expected by 43 the script. 44 45 'perf script <script> <required-script-args> <command>' to both 46 record the events required for <script> and to run the <script> 47 using 'live-mode' i.e. without writing anything to disk. <script> 48 is the name displayed in the output of 'perf script --list' i.e. the 49 actual script name minus any language extension. If <command> is 50 not specified, the events are recorded using the -a (system-wide) 51 'perf record' option. If <script> has any required args, they 52 should be specified before <command>. This mode doesn't allow for 53 optional script args to be specified; if optional script args are 54 desired, they can be specified using separate 'perf script record' 55 and 'perf script report' commands, with the stdout of the record step 56 piped to the stdin of the report script, using the '-o -' and '-i -' 57 options of the corresponding commands. 58 59 'perf script <top-script>' to both record the events required for 60 <top-script> and to run the <top-script> using 'live-mode' 61 i.e. without writing anything to disk. <top-script> is the name 62 displayed in the output of 'perf script --list' i.e. the actual 63 script name minus any language extension; a <top-script> is defined 64 as any script name ending with the string 'top'. 65 66 [<record-options>] can be passed to the record steps of 'perf script 67 record' and 'live-mode' variants; this isn't possible however for 68 <top-script> 'live-mode' or 'perf script report' variants. 69 70 See the 'SEE ALSO' section for links to language-specific 71 information on how to write and run your own trace scripts. 72 73OPTIONS 74------- 75<command>...:: 76 Any command you can specify in a shell. 77 78-D:: 79--dump-raw-trace=:: 80 Display verbose dump of the trace data. 81 82-L:: 83--Latency=:: 84 Show latency attributes (irqs/preemption disabled, etc). 85 86-l:: 87--list=:: 88 Display a list of available trace scripts. 89 90-s ['lang']:: 91--script=:: 92 Process trace data with the given script ([lang]:script[.ext]). 93 If the string 'lang' is specified in place of a script name, a 94 list of supported languages will be displayed instead. 95 96-g:: 97--gen-script=:: 98 Generate perf-script.[ext] starter script for given language, 99 using current perf.data. 100 101--dlfilter=<file>:: 102 Filter sample events using the given shared object file. 103 Refer linkperf:perf-dlfilter[1] 104 105--dlarg=<arg>:: 106 Pass 'arg' as an argument to the dlfilter. --dlarg may be repeated 107 to add more arguments. 108 109--list-dlfilters:: 110 Display a list of available dlfilters. Use with option -v (must come 111 before option --list-dlfilters) to show long descriptions. 112 113-a:: 114 Force system-wide collection. Scripts run without a <command> 115 normally use -a by default, while scripts run with a <command> 116 normally don't - this option allows the latter to be run in 117 system-wide mode. 118 119-i:: 120--input=:: 121 Input file name. (default: perf.data unless stdin is a fifo) 122 123-d:: 124--debug-mode:: 125 Do various checks like samples ordering and lost events. 126 127-F:: 128--fields:: 129 Comma separated list of fields to print. Options are: 130 comm, tid, pid, time, cpu, event, trace, ip, sym, dso, addr, symoff, 131 srcline, period, iregs, uregs, brstack, brstacksym, flags, bpf-output, 132 brstackinsn, brstackoff, callindent, insn, insnlen, synth, phys_addr, 133 metric, misc, srccode, ipc, data_page_size, code_page_size. 134 Field list can be prepended with the type, trace, sw or hw, 135 to indicate to which event type the field list applies. 136 e.g., -F sw:comm,tid,time,ip,sym and -F trace:time,cpu,trace 137 138 perf script -F <fields> 139 140 is equivalent to: 141 142 perf script -F trace:<fields> -F sw:<fields> -F hw:<fields> 143 144 i.e., the specified fields apply to all event types if the type string 145 is not given. 146 147 In addition to overriding fields, it is also possible to add or remove 148 fields from the defaults. For example 149 150 -F -cpu,+insn 151 152 removes the cpu field and adds the insn field. Adding/removing fields 153 cannot be mixed with normal overriding. 154 155 The arguments are processed in the order received. A later usage can 156 reset a prior request. e.g.: 157 158 -F trace: -F comm,tid,time,ip,sym 159 160 The first -F suppresses trace events (field list is ""), but then the 161 second invocation sets the fields to comm,tid,time,ip,sym. In this case a 162 warning is given to the user: 163 164 "Overriding previous field request for all events." 165 166 Alternatively, consider the order: 167 168 -F comm,tid,time,ip,sym -F trace: 169 170 The first -F sets the fields for all events and the second -F 171 suppresses trace events. The user is given a warning message about 172 the override, and the result of the above is that only S/W and H/W 173 events are displayed with the given fields. 174 175 It's possible tp add/remove fields only for specific event type: 176 177 -Fsw:-cpu,-period 178 179 removes cpu and period from software events. 180 181 For the 'wildcard' option if a user selected field is invalid for an 182 event type, a message is displayed to the user that the option is 183 ignored for that type. For example: 184 185 $ perf script -F comm,tid,trace 186 'trace' not valid for hardware events. Ignoring. 187 'trace' not valid for software events. Ignoring. 188 189 Alternatively, if the type is given an invalid field is specified it 190 is an error. For example: 191 192 perf script -v -F sw:comm,tid,trace 193 'trace' not valid for software events. 194 195 At this point usage is displayed, and perf-script exits. 196 197 The flags field is synthesized and may have a value when Instruction 198 Trace decoding. The flags are "bcrosyiABExgh" which stand for branch, 199 call, return, conditional, system, asynchronous, interrupt, 200 transaction abort, trace begin, trace end, in transaction, VM-Entry, and VM-Exit 201 respectively. Known combinations of flags are printed more nicely e.g. 202 "call" for "bc", "return" for "br", "jcc" for "bo", "jmp" for "b", 203 "int" for "bci", "iret" for "bri", "syscall" for "bcs", "sysret" for "brs", 204 "async" for "by", "hw int" for "bcyi", "tx abrt" for "bA", "tr strt" for "bB", 205 "tr end" for "bE", "vmentry" for "bcg", "vmexit" for "bch". 206 However the "x" flag will be displayed separately in those 207 cases e.g. "jcc (x)" for a condition branch within a transaction. 208 209 The callindent field is synthesized and may have a value when 210 Instruction Trace decoding. For calls and returns, it will display the 211 name of the symbol indented with spaces to reflect the stack depth. 212 213 When doing instruction trace decoding insn and insnlen give the 214 instruction bytes and the instruction length of the current 215 instruction. 216 217 The synth field is used by synthesized events which may be created when 218 Instruction Trace decoding. 219 220 The ipc (instructions per cycle) field is synthesized and may have a value when 221 Instruction Trace decoding. 222 223 Finally, a user may not set fields to none for all event types. 224 i.e., -F "" is not allowed. 225 226 The brstack output includes branch related information with raw addresses using the 227 /v/v/v/v/cycles syntax in the following order: 228 FROM: branch source instruction 229 TO : branch target instruction 230 M/P/-: M=branch target mispredicted or branch direction was mispredicted, P=target predicted or direction predicted, -=not supported 231 X/- : X=branch inside a transactional region, -=not in transaction region or not supported 232 A/- : A=TSX abort entry, -=not aborted region or not supported 233 cycles 234 235 The brstacksym is identical to brstack, except that the FROM and TO addresses are printed in a symbolic form if possible. 236 237 When brstackinsn is specified the full assembler sequences of branch sequences for each sample 238 is printed. This is the full execution path leading to the sample. This is only supported when the 239 sample was recorded with perf record -b or -j any. 240 241 The brstackoff field will print an offset into a specific dso/binary. 242 243 With the metric option perf script can compute metrics for 244 sampling periods, similar to perf stat. This requires 245 specifying a group with multiple events defining metrics with the :S option 246 for perf record. perf will sample on the first event, and 247 print computed metrics for all the events in the group. Please note 248 that the metric computed is averaged over the whole sampling 249 period (since the last sample), not just for the sample point. 250 251 For sample events it's possible to display misc field with -F +misc option, 252 following letters are displayed for each bit: 253 254 PERF_RECORD_MISC_KERNEL K 255 PERF_RECORD_MISC_USER U 256 PERF_RECORD_MISC_HYPERVISOR H 257 PERF_RECORD_MISC_GUEST_KERNEL G 258 PERF_RECORD_MISC_GUEST_USER g 259 PERF_RECORD_MISC_MMAP_DATA* M 260 PERF_RECORD_MISC_COMM_EXEC E 261 PERF_RECORD_MISC_SWITCH_OUT S 262 PERF_RECORD_MISC_SWITCH_OUT_PREEMPT Sp 263 264 $ perf script -F +misc ... 265 sched-messaging 1414 K 28690.636582: 4590 cycles ... 266 sched-messaging 1407 U 28690.636600: 325620 cycles ... 267 sched-messaging 1414 K 28690.636608: 19473 cycles ... 268 misc field ___________/ 269 270-k:: 271--vmlinux=<file>:: 272 vmlinux pathname 273 274--kallsyms=<file>:: 275 kallsyms pathname 276 277--symfs=<directory>:: 278 Look for files with symbols relative to this directory. 279 280-G:: 281--hide-call-graph:: 282 When printing symbols do not display call chain. 283 284--stop-bt:: 285 Stop display of callgraph at these symbols 286 287-C:: 288--cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can 289 be provided as a comma-separated list with no space: 0,1. Ranges of 290 CPUs are specified with -: 0-2. Default is to report samples on all 291 CPUs. 292 293-c:: 294--comms=:: 295 Only display events for these comms. CSV that understands 296 file://filename entries. 297 298--pid=:: 299 Only show events for given process ID (comma separated list). 300 301--tid=:: 302 Only show events for given thread ID (comma separated list). 303 304-I:: 305--show-info:: 306 Display extended information about the perf.data file. This adds 307 information which may be very large and thus may clutter the display. 308 It currently includes: cpu and numa topology of the host system. 309 It can only be used with the perf script report mode. 310 311--show-kernel-path:: 312 Try to resolve the path of [kernel.kallsyms] 313 314--show-task-events 315 Display task related events (e.g. FORK, COMM, EXIT). 316 317--show-mmap-events 318 Display mmap related events (e.g. MMAP, MMAP2). 319 320--show-namespace-events 321 Display namespace events i.e. events of type PERF_RECORD_NAMESPACES. 322 323--show-switch-events 324 Display context switch events i.e. events of type PERF_RECORD_SWITCH or 325 PERF_RECORD_SWITCH_CPU_WIDE. 326 327--show-lost-events 328 Display lost events i.e. events of type PERF_RECORD_LOST. 329 330--show-round-events 331 Display finished round events i.e. events of type PERF_RECORD_FINISHED_ROUND. 332 333--show-bpf-events 334 Display bpf events i.e. events of type PERF_RECORD_KSYMBOL and PERF_RECORD_BPF_EVENT. 335 336--show-cgroup-events 337 Display cgroup events i.e. events of type PERF_RECORD_CGROUP. 338 339--show-text-poke-events 340 Display text poke events i.e. events of type PERF_RECORD_TEXT_POKE and 341 PERF_RECORD_KSYMBOL. 342 343--demangle:: 344 Demangle symbol names to human readable form. It's enabled by default, 345 disable with --no-demangle. 346 347--demangle-kernel:: 348 Demangle kernel symbol names to human readable form (for C++ kernels). 349 350--header 351 Show perf.data header. 352 353--header-only 354 Show only perf.data header. 355 356--itrace:: 357 Options for decoding instruction tracing data. The options are: 358 359include::itrace.txt[] 360 361 To disable decoding entirely, use --no-itrace. 362 363--full-source-path:: 364 Show the full path for source files for srcline output. 365 366--max-stack:: 367 Set the stack depth limit when parsing the callchain, anything 368 beyond the specified depth will be ignored. This is a trade-off 369 between information loss and faster processing especially for 370 workloads that can have a very long callchain stack. 371 Note that when using the --itrace option the synthesized callchain size 372 will override this value if the synthesized callchain size is bigger. 373 374 Default: 127 375 376--ns:: 377 Use 9 decimal places when displaying time (i.e. show the nanoseconds) 378 379-f:: 380--force:: 381 Don't do ownership validation. 382 383--time:: 384 Only analyze samples within given time window: <start>,<stop>. Times 385 have the format seconds.nanoseconds. If start is not given (i.e. time 386 string is ',x.y') then analysis starts at the beginning of the file. If 387 stop time is not given (i.e. time string is 'x.y,') then analysis goes 388 to end of file. Multiple ranges can be separated by spaces, which 389 requires the argument to be quoted e.g. --time "1234.567,1234.789 1235," 390 391 Also support time percent with multiple time ranges. Time string is 392 'a%/n,b%/m,...' or 'a%-b%,c%-%d,...'. 393 394 For example: 395 Select the second 10% time slice: 396 perf script --time 10%/2 397 398 Select from 0% to 10% time slice: 399 perf script --time 0%-10% 400 401 Select the first and second 10% time slices: 402 perf script --time 10%/1,10%/2 403 404 Select from 0% to 10% and 30% to 40% slices: 405 perf script --time 0%-10%,30%-40% 406 407--max-blocks:: 408 Set the maximum number of program blocks to print with brstackinsn for 409 each sample. 410 411--reltime:: 412 Print time stamps relative to trace start. 413 414--deltatime:: 415 Print time stamps relative to previous event. 416 417--per-event-dump:: 418 Create per event files with a "perf.data.EVENT.dump" name instead of 419 printing to stdout, useful, for instance, for generating flamegraphs. 420 421--inline:: 422 If a callgraph address belongs to an inlined function, the inline stack 423 will be printed. Each entry has function name and file/line. Enabled by 424 default, disable with --no-inline. 425 426--insn-trace:: 427 Show instruction stream for intel_pt traces. Combine with --xed to 428 show disassembly. 429 430--xed:: 431 Run xed disassembler on output. Requires installing the xed disassembler. 432 433-S:: 434--symbols=symbol[,symbol...]:: 435 Only consider the listed symbols. Symbols are typically a name 436 but they may also be hexadecimal address. 437 438 The hexadecimal address may be the start address of a symbol or 439 any other address to filter the trace records 440 441 For example, to select the symbol noploop or the address 0x4007a0: 442 perf script --symbols=noploop,0x4007a0 443 444 Support filtering trace records by symbol name, start address of 445 symbol, any hexadecimal address and address range. 446 447 The comparison order is: 448 449 1. symbol name comparison 450 2. symbol start address comparison. 451 3. any hexadecimal address comparison. 452 4. address range comparison (see --addr-range). 453 454--addr-range:: 455 Use with -S or --symbols to list traced records within address range. 456 457 For example, to list the traced records within the address range 458 [0x4007a0, 0x0x4007a9]: 459 perf script -S 0x4007a0 --addr-range 10 460 461--dsos=:: 462 Only consider symbols in these DSOs. 463 464--call-trace:: 465 Show call stream for intel_pt traces. The CPUs are interleaved, but 466 can be filtered with -C. 467 468--call-ret-trace:: 469 Show call and return stream for intel_pt traces. 470 471--graph-function:: 472 For itrace only show specified functions and their callees for 473 itrace. Multiple functions can be separated by comma. 474 475--switch-on EVENT_NAME:: 476 Only consider events after this event is found. 477 478--switch-off EVENT_NAME:: 479 Stop considering events after this event is found. 480 481--show-on-off-events:: 482 Show the --switch-on/off events too. 483 484--stitch-lbr:: 485 Show callgraph with stitched LBRs, which may have more complete 486 callgraph. The perf.data file must have been obtained using 487 perf record --call-graph lbr. 488 Disabled by default. In common cases with call stack overflows, 489 it can recreate better call stacks than the default lbr call stack 490 output. But this approach is not full proof. There can be cases 491 where it creates incorrect call stacks from incorrect matches. 492 The known limitations include exception handing such as 493 setjmp/longjmp will have calls/returns not match. 494 495SEE ALSO 496-------- 497linkperf:perf-record[1], linkperf:perf-script-perl[1], 498linkperf:perf-script-python[1], linkperf:perf-intel-pt[1], 499linkperf:perf-dlfilter[1] 500