xref: /linux/tools/perf/Documentation/perf-record.txt (revision bd628c1bed7902ec1f24ba0fe70758949146abbe)
1perf-record(1)
2==============
3
4NAME
5----
6perf-record - Run a command and record its profile into perf.data
7
8SYNOPSIS
9--------
10[verse]
11'perf record' [-e <EVENT> | --event=EVENT] [-a] <command>
12'perf record' [-e <EVENT> | --event=EVENT] [-a] -- <command> [<options>]
13
14DESCRIPTION
15-----------
16This command runs a command and gathers a performance counter profile
17from it, into perf.data - without displaying anything.
18
19This file can then be inspected later on, using 'perf report'.
20
21
22OPTIONS
23-------
24<command>...::
25	Any command you can specify in a shell.
26
27-e::
28--event=::
29	Select the PMU event. Selection can be:
30
31        - a symbolic event name	(use 'perf list' to list all events)
32
33        - a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a
34	  hexadecimal event descriptor.
35
36	- a symbolically formed PMU event like 'pmu/param1=0x3,param2/' where
37	  'param1', 'param2', etc are defined as formats for the PMU in
38	  /sys/bus/event_source/devices/<pmu>/format/*.
39
40	- a symbolically formed event like 'pmu/config=M,config1=N,config3=K/'
41
42          where M, N, K are numbers (in decimal, hex, octal format). Acceptable
43          values for each of 'config', 'config1' and 'config2' are defined by
44          corresponding entries in /sys/bus/event_source/devices/<pmu>/format/*
45          param1 and param2 are defined as formats for the PMU in:
46          /sys/bus/event_source/devices/<pmu>/format/*
47
48	  There are also some parameters which are not defined in .../<pmu>/format/*.
49	  These params can be used to overload default config values per event.
50	  Here are some common parameters:
51	  - 'period': Set event sampling period
52	  - 'freq': Set event sampling frequency
53	  - 'time': Disable/enable time stamping. Acceptable values are 1 for
54		    enabling time stamping. 0 for disabling time stamping.
55		    The default is 1.
56	  - 'call-graph': Disable/enable callgraph. Acceptable str are "fp" for
57			 FP mode, "dwarf" for DWARF mode, "lbr" for LBR mode and
58			 "no" for disable callgraph.
59	  - 'stack-size': user stack size for dwarf mode
60	  - 'name' : User defined event name. Single quotes (') may be used to
61		    escape symbols in the name from parsing by shell and tool
62		    like this: name=\'CPU_CLK_UNHALTED.THREAD:cmask=0x1\'.
63
64          See the linkperf:perf-list[1] man page for more parameters.
65
66	  Note: If user explicitly sets options which conflict with the params,
67	  the value set by the parameters will be overridden.
68
69	  Also not defined in .../<pmu>/format/* are PMU driver specific
70	  configuration parameters.  Any configuration parameter preceded by
71	  the letter '@' is not interpreted in user space and sent down directly
72	  to the PMU driver.  For example:
73
74	  perf record -e some_event/@cfg1,@cfg2=config/ ...
75
76	  will see 'cfg1' and 'cfg2=config' pushed to the PMU driver associated
77	  with the event for further processing.  There is no restriction on
78	  what the configuration parameters are, as long as their semantic is
79	  understood and supported by the PMU driver.
80
81        - a hardware breakpoint event in the form of '\mem:addr[/len][:access]'
82          where addr is the address in memory you want to break in.
83          Access is the memory access type (read, write, execute) it can
84          be passed as follows: '\mem:addr[:[r][w][x]]'. len is the range,
85          number of bytes from specified addr, which the breakpoint will cover.
86          If you want to profile read-write accesses in 0x1000, just set
87          'mem:0x1000:rw'.
88          If you want to profile write accesses in [0x1000~1008), just set
89          'mem:0x1000/8:w'.
90
91	- a group of events surrounded by a pair of brace ("{event1,event2,...}").
92	  Each event is separated by commas and the group should be quoted to
93	  prevent the shell interpretation.  You also need to use --group on
94	  "perf report" to view group events together.
95
96--filter=<filter>::
97        Event filter. This option should follow an event selector (-e) which
98	selects either tracepoint event(s) or a hardware trace PMU
99	(e.g. Intel PT or CoreSight).
100
101	- tracepoint filters
102
103	In the case of tracepoints, multiple '--filter' options are combined
104	using '&&'.
105
106	- address filters
107
108	A hardware trace PMU advertises its ability to accept a number of
109	address filters	by specifying a non-zero value in
110	/sys/bus/event_source/devices/<pmu>/nr_addr_filters.
111
112	Address filters have the format:
113
114	filter|start|stop|tracestop <start> [/ <size>] [@<file name>]
115
116	Where:
117	- 'filter': defines a region that will be traced.
118	- 'start': defines an address at which tracing will begin.
119	- 'stop': defines an address at which tracing will stop.
120	- 'tracestop': defines a region in which tracing will stop.
121
122	<file name> is the name of the object file, <start> is the offset to the
123	code to trace in that file, and <size> is the size of the region to
124	trace. 'start' and 'stop' filters need not specify a <size>.
125
126	If no object file is specified then the kernel is assumed, in which case
127	the start address must be a current kernel memory address.
128
129	<start> can also be specified by providing the name of a symbol. If the
130	symbol name is not unique, it can be disambiguated by inserting #n where
131	'n' selects the n'th symbol in address order. Alternately #0, #g or #G
132	select only a global symbol. <size> can also be specified by providing
133	the name of a symbol, in which case the size is calculated to the end
134	of that symbol. For 'filter' and 'tracestop' filters, if <size> is
135	omitted and <start> is a symbol, then the size is calculated to the end
136	of that symbol.
137
138	If <size> is omitted and <start> is '*', then the start and size will
139	be calculated from the first and last symbols, i.e. to trace the whole
140	file.
141
142	If symbol names (or '*') are provided, they must be surrounded by white
143	space.
144
145	The filter passed to the kernel is not necessarily the same as entered.
146	To see the filter that is passed, use the -v option.
147
148	The kernel may not be able to configure a trace region if it is not
149	within a single mapping.  MMAP events (or /proc/<pid>/maps) can be
150	examined to determine if that is a possibility.
151
152	Multiple filters can be separated with space or comma.
153
154--exclude-perf::
155	Don't record events issued by perf itself. This option should follow
156	an event selector (-e) which selects tracepoint event(s). It adds a
157	filter expression 'common_pid != $PERFPID' to filters. If other
158	'--filter' exists, the new filter expression will be combined with
159	them by '&&'.
160
161-a::
162--all-cpus::
163        System-wide collection from all CPUs (default if no target is specified).
164
165-p::
166--pid=::
167	Record events on existing process ID (comma separated list).
168
169-t::
170--tid=::
171        Record events on existing thread ID (comma separated list).
172        This option also disables inheritance by default.  Enable it by adding
173        --inherit.
174
175-u::
176--uid=::
177        Record events in threads owned by uid. Name or number.
178
179-r::
180--realtime=::
181	Collect data with this RT SCHED_FIFO priority.
182
183--no-buffering::
184	Collect data without buffering.
185
186-c::
187--count=::
188	Event period to sample.
189
190-o::
191--output=::
192	Output file name.
193
194-i::
195--no-inherit::
196	Child tasks do not inherit counters.
197
198-F::
199--freq=::
200	Profile at this frequency. Use 'max' to use the currently maximum
201	allowed frequency, i.e. the value in the kernel.perf_event_max_sample_rate
202	sysctl. Will throttle down to the currently maximum allowed frequency.
203	See --strict-freq.
204
205--strict-freq::
206	Fail if the specified frequency can't be used.
207
208-m::
209--mmap-pages=::
210	Number of mmap data pages (must be a power of two) or size
211	specification with appended unit character - B/K/M/G. The
212	size is rounded up to have nearest pages power of two value.
213	Also, by adding a comma, the number of mmap pages for AUX
214	area tracing can be specified.
215
216--group::
217	Put all events in a single event group.  This precedes the --event
218	option and remains only for backward compatibility.  See --event.
219
220-g::
221	Enables call-graph (stack chain/backtrace) recording.
222
223--call-graph::
224	Setup and enable call-graph (stack chain/backtrace) recording,
225	implies -g.  Default is "fp".
226
227	Allows specifying "fp" (frame pointer) or "dwarf"
228	(DWARF's CFI - Call Frame Information) or "lbr"
229	(Hardware Last Branch Record facility) as the method to collect
230	the information used to show the call graphs.
231
232	In some systems, where binaries are build with gcc
233	--fomit-frame-pointer, using the "fp" method will produce bogus
234	call graphs, using "dwarf", if available (perf tools linked to
235	the libunwind or libdw library) should be used instead.
236	Using the "lbr" method doesn't require any compiler options. It
237	will produce call graphs from the hardware LBR registers. The
238	main limitation is that it is only available on new Intel
239	platforms, such as Haswell. It can only get user call chain. It
240	doesn't work with branch stack sampling at the same time.
241
242	When "dwarf" recording is used, perf also records (user) stack dump
243	when sampled.  Default size of the stack dump is 8192 (bytes).
244	User can change the size by passing the size after comma like
245	"--call-graph dwarf,4096".
246
247-q::
248--quiet::
249	Don't print any message, useful for scripting.
250
251-v::
252--verbose::
253	Be more verbose (show counter open errors, etc).
254
255-s::
256--stat::
257	Record per-thread event counts.  Use it with 'perf report -T' to see
258	the values.
259
260-d::
261--data::
262	Record the sample virtual addresses.
263
264--phys-data::
265	Record the sample physical addresses.
266
267-T::
268--timestamp::
269	Record the sample timestamps. Use it with 'perf report -D' to see the
270	timestamps, for instance.
271
272-P::
273--period::
274	Record the sample period.
275
276--sample-cpu::
277	Record the sample cpu.
278
279-n::
280--no-samples::
281	Don't sample.
282
283-R::
284--raw-samples::
285Collect raw sample records from all opened counters (default for tracepoint counters).
286
287-C::
288--cpu::
289Collect samples only on the list of CPUs provided. Multiple CPUs can be provided as a
290comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2.
291In per-thread mode with inheritance mode on (default), samples are captured only when
292the thread executes on the designated CPUs. Default is to monitor all CPUs.
293
294-B::
295--no-buildid::
296Do not save the build ids of binaries in the perf.data files. This skips
297post processing after recording, which sometimes makes the final step in
298the recording process to take a long time, as it needs to process all
299events looking for mmap records. The downside is that it can misresolve
300symbols if the workload binaries used when recording get locally rebuilt
301or upgraded, because the only key available in this case is the
302pathname. You can also set the "record.build-id" config variable to
303'skip to have this behaviour permanently.
304
305-N::
306--no-buildid-cache::
307Do not update the buildid cache. This saves some overhead in situations
308where the information in the perf.data file (which includes buildids)
309is sufficient.  You can also set the "record.build-id" config variable to
310'no-cache' to have the same effect.
311
312-G name,...::
313--cgroup name,...::
314monitor only in the container (cgroup) called "name". This option is available only
315in per-cpu mode. The cgroup filesystem must be mounted. All threads belonging to
316container "name" are monitored when they run on the monitored CPUs. Multiple cgroups
317can be provided. Each cgroup is applied to the corresponding event, i.e., first cgroup
318to first event, second cgroup to second event and so on. It is possible to provide
319an empty cgroup (monitor all the time) using, e.g., -G foo,,bar. Cgroups must have
320corresponding events, i.e., they always refer to events defined earlier on the command
321line. If the user wants to track multiple events for a specific cgroup, the user can
322use '-e e1 -e e2 -G foo,foo' or just use '-e e1 -e e2 -G foo'.
323
324If wanting to monitor, say, 'cycles' for a cgroup and also for system wide, this
325command line can be used: 'perf stat -e cycles -G cgroup_name -a -e cycles'.
326
327-b::
328--branch-any::
329Enable taken branch stack sampling. Any type of taken branch may be sampled.
330This is a shortcut for --branch-filter any. See --branch-filter for more infos.
331
332-j::
333--branch-filter::
334Enable taken branch stack sampling. Each sample captures a series of consecutive
335taken branches. The number of branches captured with each sample depends on the
336underlying hardware, the type of branches of interest, and the executed code.
337It is possible to select the types of branches captured by enabling filters. The
338following filters are defined:
339
340        - any:  any type of branches
341        - any_call: any function call or system call
342        - any_ret: any function return or system call return
343        - ind_call: any indirect branch
344        - call: direct calls, including far (to/from kernel) calls
345        - u:  only when the branch target is at the user level
346        - k: only when the branch target is in the kernel
347        - hv: only when the target is at the hypervisor level
348	- in_tx: only when the target is in a hardware transaction
349	- no_tx: only when the target is not in a hardware transaction
350	- abort_tx: only when the target is a hardware transaction abort
351	- cond: conditional branches
352	- save_type: save branch type during sampling in case binary is not available later
353
354+
355The option requires at least one branch type among any, any_call, any_ret, ind_call, cond.
356The privilege levels may be omitted, in which case, the privilege levels of the associated
357event are applied to the branch filter. Both kernel (k) and hypervisor (hv) privilege
358levels are subject to permissions.  When sampling on multiple events, branch stack sampling
359is enabled for all the sampling events. The sampled branch type is the same for all events.
360The various filters must be specified as a comma separated list: --branch-filter any_ret,u,k
361Note that this feature may not be available on all processors.
362
363--weight::
364Enable weightened sampling. An additional weight is recorded per sample and can be
365displayed with the weight and local_weight sort keys.  This currently works for TSX
366abort events and some memory events in precise mode on modern Intel CPUs.
367
368--namespaces::
369Record events of type PERF_RECORD_NAMESPACES.
370
371--transaction::
372Record transaction flags for transaction related events.
373
374--per-thread::
375Use per-thread mmaps.  By default per-cpu mmaps are created.  This option
376overrides that and uses per-thread mmaps.  A side-effect of that is that
377inheritance is automatically disabled.  --per-thread is ignored with a warning
378if combined with -a or -C options.
379
380-D::
381--delay=::
382After starting the program, wait msecs before measuring. This is useful to
383filter out the startup phase of the program, which is often very different.
384
385-I::
386--intr-regs::
387Capture machine state (registers) at interrupt, i.e., on counter overflows for
388each sample. List of captured registers depends on the architecture. This option
389is off by default. It is possible to select the registers to sample using their
390symbolic names, e.g. on x86, ax, si. To list the available registers use
391--intr-regs=\?. To name registers, pass a comma separated list such as
392--intr-regs=ax,bx. The list of register is architecture dependent.
393
394--user-regs::
395Capture user registers at sample time. Same arguments as -I.
396
397--running-time::
398Record running and enabled time for read events (:S)
399
400-k::
401--clockid::
402Sets the clock id to use for the various time fields in the perf_event_type
403records. See clock_gettime(). In particular CLOCK_MONOTONIC and
404CLOCK_MONOTONIC_RAW are supported, some events might also allow
405CLOCK_BOOTTIME, CLOCK_REALTIME and CLOCK_TAI.
406
407-S::
408--snapshot::
409Select AUX area tracing Snapshot Mode. This option is valid only with an
410AUX area tracing event. Optionally the number of bytes to capture per
411snapshot can be specified. In Snapshot Mode, trace data is captured only when
412signal SIGUSR2 is received.
413
414--proc-map-timeout::
415When processing pre-existing threads /proc/XXX/mmap, it may take a long time,
416because the file may be huge. A time out is needed in such cases.
417This option sets the time out limit. The default value is 500 ms.
418
419--switch-events::
420Record context switch events i.e. events of type PERF_RECORD_SWITCH or
421PERF_RECORD_SWITCH_CPU_WIDE.
422
423--clang-path=PATH::
424Path to clang binary to use for compiling BPF scriptlets.
425(enabled when BPF support is on)
426
427--clang-opt=OPTIONS::
428Options passed to clang when compiling BPF scriptlets.
429(enabled when BPF support is on)
430
431--vmlinux=PATH::
432Specify vmlinux path which has debuginfo.
433(enabled when BPF prologue is on)
434
435--buildid-all::
436Record build-id of all DSOs regardless whether it's actually hit or not.
437
438--aio[=n]::
439Use <n> control blocks in asynchronous (Posix AIO) trace writing mode (default: 1, max: 4).
440Asynchronous mode is supported only when linking Perf tool with libc library
441providing implementation for Posix AIO API.
442
443--all-kernel::
444Configure all used events to run in kernel space.
445
446--all-user::
447Configure all used events to run in user space.
448
449--timestamp-filename
450Append timestamp to output file name.
451
452--timestamp-boundary::
453Record timestamp boundary (time of first/last samples).
454
455--switch-output[=mode]::
456Generate multiple perf.data files, timestamp prefixed, switching to a new one
457based on 'mode' value:
458  "signal" - when receiving a SIGUSR2 (default value) or
459  <size>   - when reaching the size threshold, size is expected to
460             be a number with appended unit character - B/K/M/G
461  <time>   - when reaching the time threshold, size is expected to
462             be a number with appended unit character - s/m/h/d
463
464             Note: the precision of  the size  threshold  hugely depends
465             on your configuration  - the number and size of  your  ring
466             buffers (-m). It is generally more precise for higher sizes
467             (like >5M), for lower values expect different sizes.
468
469A possible use case is to, given an external event, slice the perf.data file
470that gets then processed, possibly via a perf script, to decide if that
471particular perf.data snapshot should be kept or not.
472
473Implies --timestamp-filename, --no-buildid and --no-buildid-cache.
474The reason for the latter two is to reduce the data file switching
475overhead. You can still switch them on with:
476
477  --switch-output --no-no-buildid  --no-no-buildid-cache
478
479--dry-run::
480Parse options then exit. --dry-run can be used to detect errors in cmdline
481options.
482
483'perf record --dry-run -e' can act as a BPF script compiler if llvm.dump-obj
484in config file is set to true.
485
486--tail-synthesize::
487Instead of collecting non-sample events (for example, fork, comm, mmap) at
488the beginning of record, collect them during finalizing an output file.
489The collected non-sample events reflects the status of the system when
490record is finished.
491
492--overwrite::
493Makes all events use an overwritable ring buffer. An overwritable ring
494buffer works like a flight recorder: when it gets full, the kernel will
495overwrite the oldest records, that thus will never make it to the
496perf.data file.
497
498When '--overwrite' and '--switch-output' are used perf records and drops
499events until it receives a signal, meaning that something unusual was
500detected that warrants taking a snapshot of the most current events,
501those fitting in the ring buffer at that moment.
502
503'overwrite' attribute can also be set or canceled for an event using
504config terms. For example: 'cycles/overwrite/' and 'instructions/no-overwrite/'.
505
506Implies --tail-synthesize.
507
508SEE ALSO
509--------
510linkperf:perf-stat[1], linkperf:perf-list[1]
511