| blob | 5c379adf83043d28a257310deee55e0e0c5f422a |
1 perf-config(1)
2 ==============
4 NAME
5 ----
6 perf-config - Get and set variables in a configuration file.
8 SYNOPSIS
9 --------
10 [verse]
11 'perf config' [<file-option>] [section.name[=value] ...]
12 or
13 'perf config' [<file-option>] -l | --list
15 DESCRIPTION
16 -----------
17 You can manage variables in a configuration file with this command.
19 OPTIONS
20 -------
22 -l::
23 --list::
24 Show current config variables, name and value, for all sections.
26 --user::
27 For writing and reading options: write to user
28 '$HOME/.perfconfig' file or read it.
30 --system::
31 For writing and reading options: write to system-wide
32 '$(sysconfdir)/perfconfig' or read it.
34 CONFIGURATION FILE
35 ------------------
37 The perf configuration file contains many variables to change various
38 aspects of each of its tools, including output, disk usage, etc.
39 The '$HOME/.perfconfig' file is used to store a per-user configuration.
40 The file '$(sysconfdir)/perfconfig' can be used to
41 store a system-wide default configuration.
43 One an disable reading config files by setting the PERF_CONFIG environment
44 variable to /dev/null, or provide an alternate config file by setting that
45 variable.
47 When reading or writing, the values are read from the system and user
48 configuration files by default, and options '--system' and '--user'
49 can be used to tell the command to read from or write to only that location.
51 Syntax
52 ~~~~~~
54 The file consist of sections. A section starts with its name
55 surrounded by square brackets and continues till the next section
56 begins. Each variable must be in a section, and have the form
57 'name = value', for example:
59 [section]
60 name1 = value1
61 name2 = value2
63 Section names are case sensitive and can contain any characters except
64 newline (double quote `"` and backslash have to be escaped as `\"` and `\\`,
65 respectively). Section headers can't span multiple lines.
67 Example
68 ~~~~~~~
70 Given a $HOME/.perfconfig like this:
72 #
73 # This is the config file, and
74 # a '#' and ';' character indicates a comment
75 #
77 [colors]
78 # Color variables
79 top = red, default
80 medium = green, default
81 normal = lightgray, default
82 selected = white, lightgray
83 jump_arrows = blue, default
84 addr = magenta, default
85 root = white, blue
87 [tui]
88 # Defaults if linked with libslang
89 report = on
90 annotate = on
91 top = on
93 [buildid]
94 # Default, disable using /dev/null
95 dir = ~/.debug
97 [annotate]
98 # Defaults
99 hide_src_code = false
100 use_offset = true
101 jump_arrows = true
102 show_nr_jumps = false
104 [help]
105 # Format can be man, info, web or html
106 format = man
107 autocorrect = 0
109 [ui]
110 show-headers = true
112 [call-graph]
113 # fp (framepointer), dwarf
114 record-mode = fp
115 print-type = graph
116 order = caller
117 sort-key = function
119 [report]
120 # Defaults
121 sort_order = comm,dso,symbol
122 percent-limit = 0
123 queue-size = 0
124 children = true
125 group = true
127 [llvm]
128 dump-obj = true
129 clang-opt = -g
131 You can hide source code of annotate feature setting the config to false with
133 % perf config annotate.hide_src_code=true
135 If you want to add or modify several config items, you can do like
137 % perf config ui.show-headers=false kmem.default=slab
139 To modify the sort order of report functionality in user config file(i.e. `~/.perfconfig`), do
141 % perf config --user report.sort-order=srcline
143 To change colors of selected line to other foreground and background colors
144 in system config file (i.e. `$(sysconf)/perfconfig`), do
146 % perf config --system colors.selected=yellow,green
148 To query the record mode of call graph, do
150 % perf config call-graph.record-mode
152 If you want to know multiple config key/value pairs, you can do like
154 % perf config report.queue-size call-graph.order report.children
156 To query the config value of sort order of call graph in user config file (i.e. `~/.perfconfig`), do
158 % perf config --user call-graph.sort-order
160 To query the config value of buildid directory in system config file (i.e. `$(sysconf)/perfconfig`), do
162 % perf config --system buildid.dir
164 Variables
165 ~~~~~~~~~
167 colors.*::
168 The variables for customizing the colors used in the output for the
169 'report', 'top' and 'annotate' in the TUI. They should specify the
170 foreground and background colors, separated by a comma, for example:
172 medium = green, lightgray
174 If you want to use the color configured for you terminal, just leave it
175 as 'default', for example:
177 medium = default, lightgray
179 Available colors:
180 red, yellow, green, cyan, gray, black, blue,
181 white, default, magenta, lightgray
183 colors.top::
184 'top' means a overhead percentage which is more than 5%.
185 And values of this variable specify percentage colors.
186 Basic key values are foreground-color 'red' and
187 background-color 'default'.
188 colors.medium::
189 'medium' means a overhead percentage which has more than 0.5%.
190 Default values are 'green' and 'default'.
191 colors.normal::
192 'normal' means the rest of overhead percentages
193 except 'top', 'medium', 'selected'.
194 Default values are 'lightgray' and 'default'.
195 colors.selected::
196 This selects the colors for the current entry in a list of entries
197 from sub-commands (top, report, annotate).
198 Default values are 'black' and 'lightgray'.
199 colors.jump_arrows::
200 Colors for jump arrows on assembly code listings
201 such as 'jns', 'jmp', 'jane', etc.
202 Default values are 'blue', 'default'.
203 colors.addr::
204 This selects colors for addresses from 'annotate'.
205 Default values are 'magenta', 'default'.
206 colors.root::
207 Colors for headers in the output of a sub-commands (top, report).
208 Default values are 'white', 'blue'.
210 core.*::
211 core.proc-map-timeout::
212 Sets a timeout (in milliseconds) for parsing /proc/<pid>/maps files.
213 Can be overridden by the --proc-map-timeout option on supported
214 subcommands. The default timeout is 500ms.
216 tui.*, gtk.*::
217 Subcommands that can be configured here are 'top', 'report' and 'annotate'.
218 These values are booleans, for example:
220 [tui]
221 top = true
223 will make the TUI be the default for the 'top' subcommand. Those will be
224 available if the required libs were detected at tool build time.
226 buildid.*::
227 buildid.dir::
228 Each executable and shared library in modern distributions comes with a
229 content based identifier that, if available, will be inserted in a
230 'perf.data' file header to, at analysis time find what is needed to do
231 symbol resolution, code annotation, etc.
233 The recording tools also stores a hard link or copy in a per-user
234 directory, $HOME/.debug/, of binaries, shared libraries, /proc/kallsyms
235 and /proc/kcore files to be used at analysis time.
237 The buildid.dir variable can be used to either change this directory
238 cache location, or to disable it altogether. If you want to disable it,
239 set buildid.dir to /dev/null. The default is $HOME/.debug
241 annotate.*::
242 These are in control of addresses, jump function, source code
243 in lines of assembly code from a specific program.
245 annotate.disassembler_style:
246 Use this to change the default disassembler style to some other value
247 supported by binutils, such as "intel", see the '-M' option help in the
248 'objdump' man page.
250 annotate.hide_src_code::
251 If a program which is analyzed has source code,
252 this option lets 'annotate' print a list of assembly code with the source code.
253 For example, let's see a part of a program. There're four lines.
254 If this option is 'true', they can be printed
255 without source code from a program as below.
257 │ push %rbp
258 │ mov %rsp,%rbp
259 │ sub $0x10,%rsp
260 │ mov (%rdi),%rdx
262 But if this option is 'false', source code of the part
263 can be also printed as below. Default is 'false'.
265 │ struct rb_node *rb_next(const struct rb_node *node)
266 │ {
267 │ push %rbp
268 │ mov %rsp,%rbp
269 │ sub $0x10,%rsp
270 │ struct rb_node *parent;
271 │
272 │ if (RB_EMPTY_NODE(node))
273 │ mov (%rdi),%rdx
274 │ return n;
276 This option works with tui, stdio2 browsers.
278 annotate.use_offset::
279 Basing on a first address of a loaded function, offset can be used.
280 Instead of using original addresses of assembly code,
281 addresses subtracted from a base address can be printed.
282 Let's illustrate an example.
283 If a base address is 0XFFFFFFFF81624d50 as below,
285 ffffffff81624d50 <load0>
287 an address on assembly code has a specific absolute address as below
289 ffffffff816250b8:│ mov 0x8(%r14),%rdi
291 but if use_offset is 'true', an address subtracted from a base address is printed.
292 Default is true. This option is only applied to TUI.
294 368:│ mov 0x8(%r14),%rdi
296 This option works with tui, stdio2 browsers.
298 annotate.jump_arrows::
299 There can be jump instruction among assembly code.
300 Depending on a boolean value of jump_arrows,
301 arrows can be printed or not which represent
302 where do the instruction jump into as below.
304 │ ┌──jmp 1333
305 │ │ xchg %ax,%ax
306 │1330:│ mov %r15,%r10
307 │1333:└─→cmp %r15,%r14
309 If jump_arrow is 'false', the arrows isn't printed as below.
310 Default is 'false'.
312 │ ↓ jmp 1333
313 │ xchg %ax,%ax
314 │1330: mov %r15,%r10
315 │1333: cmp %r15,%r14
317 This option works with tui browser.
319 annotate.show_linenr::
320 When showing source code if this option is 'true',
321 line numbers are printed as below.
323 │1628 if (type & PERF_SAMPLE_IDENTIFIER) {
324 │ ↓ jne 508
325 │1628 data->id = *array;
326 │1629 array++;
327 │1630 }
329 However if this option is 'false', they aren't printed as below.
330 Default is 'false'.
332 │ if (type & PERF_SAMPLE_IDENTIFIER) {
333 │ ↓ jne 508
334 │ data->id = *array;
335 │ array++;
336 │ }
338 This option works with tui, stdio2 browsers.
340 annotate.show_nr_jumps::
341 Let's see a part of assembly code.
343 │1382: movb $0x1,-0x270(%rbp)
345 If use this, the number of branches jumping to that address can be printed as below.
346 Default is 'false'.
348 │1 1382: movb $0x1,-0x270(%rbp)
350 This option works with tui, stdio2 browsers.
352 annotate.show_total_period::
353 To compare two records on an instruction base, with this option
354 provided, display total number of samples that belong to a line
355 in assembly code. If this option is 'true', total periods are printed
356 instead of percent values as below.
358 302 │ mov %eax,%eax
360 But if this option is 'false', percent values for overhead are printed i.e.
361 Default is 'false'.
363 99.93 │ mov %eax,%eax
365 This option works with tui, stdio2, stdio browsers.
367 annotate.show_nr_samples::
368 By default perf annotate shows percentage of samples. This option
369 can be used to print absolute number of samples. Ex, when set as
370 false:
372 Percent│
373 74.03 │ mov %fs:0x28,%rax
375 When set as true:
377 Samples│
378 6 │ mov %fs:0x28,%rax
380 This option works with tui, stdio2, stdio browsers.
382 annotate.offset_level::
383 Default is '1', meaning just jump targets will have offsets show right beside
384 the instruction. When set to '2' 'call' instructions will also have its offsets
385 shown, 3 or higher will show offsets for all instructions.
387 This option works with tui, stdio2 browsers.
389 hist.*::
390 hist.percentage::
391 This option control the way to calculate overhead of filtered entries -
392 that means the value of this option is effective only if there's a
393 filter (by comm, dso or symbol name). Suppose a following example:
395 Overhead Symbols
396 ........ .......
397 33.33% foo
398 33.33% bar
399 33.33% baz
401 This is an original overhead and we'll filter out the first 'foo'
402 entry. The value of 'relative' would increase the overhead of 'bar'
403 and 'baz' to 50.00% for each, while 'absolute' would show their
404 current overhead (33.33%).
406 ui.*::
407 ui.show-headers::
408 This option controls display of column headers (like 'Overhead' and 'Symbol')
409 in 'report' and 'top'. If this option is false, they are hidden.
410 This option is only applied to TUI.
412 call-graph.*::
413 The following controls the handling of call-graphs (obtained via the
414 -g/--call-graph options).
416 call-graph.record-mode::
417 The mode for user space can be 'fp' (frame pointer), 'dwarf'
418 and 'lbr'. The value 'dwarf' is effective only if libunwind
419 (or a recent version of libdw) is present on the system;
420 the value 'lbr' only works for certain cpus. The method for
421 kernel space is controlled not by this option but by the
422 kernel config (CONFIG_UNWINDER_*).
424 call-graph.dump-size::
425 The size of stack to dump in order to do post-unwinding. Default is 8192 (byte).
426 When using dwarf into record-mode, the default size will be used if omitted.
428 call-graph.print-type::
429 The print-types can be graph (graph absolute), fractal (graph relative),
430 flat and folded. This option controls a way to show overhead for each callchain
431 entry. Suppose a following example.
433 Overhead Symbols
434 ........ .......
435 40.00% foo
436 |
437 ---foo
438 |
439 |--50.00%--bar
440 | main
441 |
442 --50.00%--baz
443 main
445 This output is a 'fractal' format. The 'foo' came from 'bar' and 'baz' exactly
446 half and half so 'fractal' shows 50.00% for each
447 (meaning that it assumes 100% total overhead of 'foo').
449 The 'graph' uses absolute overhead value of 'foo' as total so each of
450 'bar' and 'baz' callchain will have 20.00% of overhead.
451 If 'flat' is used, single column and linear exposure of call chains.
452 'folded' mean call chains are displayed in a line, separated by semicolons.
454 call-graph.order::
455 This option controls print order of callchains. The default is
456 'callee' which means callee is printed at top and then followed by its
457 caller and so on. The 'caller' prints it in reverse order.
459 If this option is not set and report.children or top.children is
460 set to true (or the equivalent command line option is given),
461 the default value of this option is changed to 'caller' for the
462 execution of 'perf report' or 'perf top'. Other commands will
463 still default to 'callee'.
465 call-graph.sort-key::
466 The callchains are merged if they contain same information.
467 The sort-key option determines a way to compare the callchains.
468 A value of 'sort-key' can be 'function' or 'address'.
469 The default is 'function'.
471 call-graph.threshold::
472 When there're many callchains it'd print tons of lines. So perf omits
473 small callchains under a certain overhead (threshold) and this option
474 control the threshold. Default is 0.5 (%). The overhead is calculated
475 by value depends on call-graph.print-type.
477 call-graph.print-limit::
478 This is a maximum number of lines of callchain printed for a single
479 histogram entry. Default is 0 which means no limitation.
481 report.*::
482 report.sort_order::
483 Allows changing the default sort order from "comm,dso,symbol" to
484 some other default, for instance "sym,dso" may be more fitting for
485 kernel developers.
486 report.percent-limit::
487 This one is mostly the same as call-graph.threshold but works for
488 histogram entries. Entries having an overhead lower than this
489 percentage will not be printed. Default is '0'. If percent-limit
490 is '10', only entries which have more than 10% of overhead will be
491 printed.
493 report.queue-size::
494 This option sets up the maximum allocation size of the internal
495 event queue for ordering events. Default is 0, meaning no limit.
497 report.children::
498 'Children' means functions called from another function.
499 If this option is true, 'perf report' cumulates callchains of children
500 and show (accumulated) total overhead as well as 'Self' overhead.
501 Please refer to the 'perf report' manual. The default is 'true'.
503 report.group::
504 This option is to show event group information together.
505 Example output with this turned on, notice that there is one column
506 per event in the group, ref-cycles and cycles:
508 # group: {ref-cycles,cycles}
509 # ========
510 #
511 # Samples: 7K of event 'anon group { ref-cycles, cycles }'
512 # Event count (approx.): 6876107743
513 #
514 # Overhead Command Shared Object Symbol
515 # ................ ....... ................. ...................
516 #
517 99.84% 99.76% noploop noploop [.] main
518 0.07% 0.00% noploop ld-2.15.so [.] strcmp
519 0.03% 0.00% noploop [kernel.kallsyms] [k] timerqueue_del
521 top.*::
522 top.children::
523 Same as 'report.children'. So if it is enabled, the output of 'top'
524 command will have 'Children' overhead column as well as 'Self' overhead
525 column by default.
526 The default is 'true'.
528 top.call-graph::
529 This is identical to 'call-graph.record-mode', except it is
530 applicable only for 'top' subcommand. This option ONLY setup
531 the unwind method. To enable 'perf top' to actually use it,
532 the command line option -g must be specified.
534 man.*::
535 man.viewer::
536 This option can assign a tool to view manual pages when 'help'
537 subcommand was invoked. Supported tools are 'man', 'woman'
538 (with emacs client) and 'konqueror'. Default is 'man'.
540 New man viewer tool can be also added using 'man.<tool>.cmd'
541 or use different path using 'man.<tool>.path' config option.
543 pager.*::
544 pager.<subcommand>::
545 When the subcommand is run on stdio, determine whether it uses
546 pager or not based on this value. Default is 'unspecified'.
548 kmem.*::
549 kmem.default::
550 This option decides which allocator is to be analyzed if neither
551 '--slab' nor '--page' option is used. Default is 'slab'.
553 record.*::
554 record.build-id::
555 This option can be 'cache', 'no-cache' or 'skip'.
556 'cache' is to post-process data and save/update the binaries into
557 the build-id cache (in ~/.debug). This is the default.
558 But if this option is 'no-cache', it will not update the build-id cache.
559 'skip' skips post-processing and does not update the cache.
561 record.call-graph::
562 This is identical to 'call-graph.record-mode', except it is
563 applicable only for 'record' subcommand. This option ONLY setup
564 the unwind method. To enable 'perf record' to actually use it,
565 the command line option -g must be specified.
567 record.aio::
568 Use 'n' control blocks in asynchronous (Posix AIO) trace writing
569 mode ('n' default: 1, max: 4).
571 diff.*::
572 diff.order::
573 This option sets the number of columns to sort the result.
574 The default is 0, which means sorting by baseline.
575 Setting it to 1 will sort the result by delta (or other
576 compute method selected).
578 diff.compute::
579 This options sets the method for computing the diff result.
580 Possible values are 'delta', 'delta-abs', 'ratio' and
581 'wdiff'. Default is 'delta'.
583 trace.*::
584 trace.add_events::
585 Allows adding a set of events to add to the ones specified
586 by the user, or use as a default one if none was specified.
587 The initial use case is to add augmented_raw_syscalls.o to
588 activate the 'perf trace' logic that looks for syscall
589 pointer contents after the normal tracepoint payload.
591 trace.args_alignment::
592 Number of columns to align the argument list, default is 70,
593 use 40 for the strace default, zero to no alignment.
595 trace.no_inherit::
596 Do not follow children threads.
598 trace.show_arg_names::
599 Should syscall argument names be printed? If not then trace.show_zeros
600 will be set.
602 trace.show_duration::
603 Show syscall duration.
605 trace.show_prefix::
606 If set to 'yes' will show common string prefixes in tables. The default
607 is to remove the common prefix in things like "MAP_SHARED", showing just "SHARED".
609 trace.show_timestamp::
610 Show syscall start timestamp.
612 trace.show_zeros::
613 Do not suppress syscall arguments that are equal to zero.
615 trace.tracepoint_beautifiers::
616 Use "libtraceevent" to use that library to augment the tracepoint arguments,
617 "libbeauty", the default, to use the same argument beautifiers used in the
618 strace-like sys_enter+sys_exit lines.
620 ftrace.*::
621 ftrace.tracer::
622 Can be used to select the default tracer when neither -G nor
623 -F option is not specified. Possible values are 'function' and
624 'function_graph'.
626 llvm.*::
627 llvm.clang-path::
628 Path to clang. If omit, search it from $PATH.
630 llvm.clang-bpf-cmd-template::
631 Cmdline template. Below lines show its default value. Environment
632 variable is used to pass options.
633 "$CLANG_EXEC -D__KERNEL__ -D__NR_CPUS__=$NR_CPUS "\
634 "-DLINUX_VERSION_CODE=$LINUX_VERSION_CODE " \
635 "$CLANG_OPTIONS $PERF_BPF_INC_OPTIONS $KERNEL_INC_OPTIONS " \
636 "-Wno-unused-value -Wno-pointer-sign " \
637 "-working-directory $WORKING_DIR " \
638 "-c \"$CLANG_SOURCE\" -target bpf $CLANG_EMIT_LLVM -O2 -o - $LLVM_OPTIONS_PIPE"
640 llvm.clang-opt::
641 Options passed to clang.
643 llvm.kbuild-dir::
644 kbuild directory. If not set, use /lib/modules/`uname -r`/build.
645 If set to "" deliberately, skip kernel header auto-detector.
647 llvm.kbuild-opts::
648 Options passed to 'make' when detecting kernel header options.
650 llvm.dump-obj::
651 Enable perf dump BPF object files compiled by LLVM.
653 llvm.opts::
654 Options passed to llc.
656 samples.*::
658 samples.context::
659 Define how many ns worth of time to show
660 around samples in perf report sample context browser.
662 scripts.*::
664 Any option defines a script that is added to the scripts menu
665 in the interactive perf browser and whose output is displayed.
666 The name of the option is the name, the value is a script command line.
667 The script gets the same options passed as a full perf script,
668 in particular -i perfdata file, --cpu, --tid
670 convert.*::
672 convert.queue-size::
673 Limit the size of ordered_events queue, so we could control
674 allocation size of perf data files without proper finished
675 round events.
676 stat.*::
678 stat.big-num::
679 (boolean) Change the default for "--big-num". To make
680 "--no-big-num" the default, set "stat.big-num=false".
682 intel-pt.*::
684 intel-pt.cache-divisor::
686 intel-pt.mispred-all::
687 If set, Intel PT decoder will set the mispred flag on all
688 branches.
690 auxtrace.*::
692 auxtrace.dumpdir::
693 s390 only. The directory to save the auxiliary trace buffer
694 can be changed using this option. Ex, auxtrace.dumpdir=/tmp.
695 If the directory does not exist or has the wrong file type,
696 the current directory is used.
698 SEE ALSO
699 --------
700 linkperf:perf[1]