6 perf-config - Get and set variables in a configuration file.
11 'perf config' [<file-option>] [section.name ...]
13 'perf config' [<file-option>] -l | --list
17 You can manage variables in a configuration file with this command.
24 Show current config variables, name and value, for all sections.
27 For writing and reading options: write to user
28 '$HOME/.perfconfig' file or read it.
31 For writing and reading options: write to system-wide
32 '$(sysconfdir)/perfconfig' or read it.
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 When reading or writing, the values are read from the system and user
44 configuration files by default, and options '--system' and '--user'
45 can be used to tell the command to read from or write to only that location.
50 The file consist of sections. A section starts with its name
51 surrounded by square brackets and continues till the next section
52 begins. Each variable must be in a section, and have the form
53 'name = value', for example:
59 Section names are case sensitive and can contain any characters except
60 newline (double quote `"` and backslash have to be escaped as `\"` and `\\`,
61 respectively). Section headers can't span multiple lines.
66 Given a $HOME/.perfconfig like this:
69 # This is the config file, and
70 # a '#' and ';' character indicates a comment
76 medium = green, default
77 normal = lightgray, default
78 selected = white, lightgray
79 jump_arrows = blue, default
80 addr = magenta, default
84 # Defaults if linked with libslang
90 # Default, disable using /dev/null
101 # Format can be man, info, web or html
109 # fp (framepointer), dwarf
117 sort-order = comm,dso,symbol
123 To query the record mode of call graph, do
125 % perf config call-graph.record-mode
127 If you want to know multiple config key/value pairs, you can do like
129 % perf config report.queue-size call-graph.order report.children
131 To query the config value of sort order of call graph in user config file (i.e. `~/.perfconfig`), do
133 % perf config --user call-graph.sort-order
135 To query the config value of buildid directory in system config file (i.e. `$(sysconf)/perfconfig`), do
137 % perf config --system buildid.dir
143 The variables for customizing the colors used in the output for the
144 'report', 'top' and 'annotate' in the TUI. They should specify the
145 foreground and background colors, separated by a comma, for example:
147 medium = green, lightgray
149 If you want to use the color configured for you terminal, just leave it
150 as 'default', for example:
152 medium = default, lightgray
155 red, yellow, green, cyan, gray, black, blue,
156 white, default, magenta, lightgray
159 'top' means a overhead percentage which is more than 5%.
160 And values of this variable specify percentage colors.
161 Basic key values are foreground-color 'red' and
162 background-color 'default'.
164 'medium' means a overhead percentage which has more than 0.5%.
165 Default values are 'green' and 'default'.
167 'normal' means the rest of overhead percentages
168 except 'top', 'medium', 'selected'.
169 Default values are 'lightgray' and 'default'.
171 This selects the colors for the current entry in a list of entries
172 from sub-commands (top, report, annotate).
173 Default values are 'black' and 'lightgray'.
175 Colors for jump arrows on assembly code listings
176 such as 'jns', 'jmp', 'jane', etc.
177 Default values are 'blue', 'default'.
179 This selects colors for addresses from 'annotate'.
180 Default values are 'magenta', 'default'.
182 Colors for headers in the output of a sub-commands (top, report).
183 Default values are 'white', 'blue'.
186 Subcommands that can be configured here are 'top', 'report' and 'annotate'.
187 These values are booleans, for example:
192 will make the TUI be the default for the 'top' subcommand. Those will be
193 available if the required libs were detected at tool build time.
197 Each executable and shared library in modern distributions comes with a
198 content based identifier that, if available, will be inserted in a
199 'perf.data' file header to, at analysis time find what is needed to do
200 symbol resolution, code annotation, etc.
202 The recording tools also stores a hard link or copy in a per-user
203 directory, $HOME/.debug/, of binaries, shared libraries, /proc/kallsyms
204 and /proc/kcore files to be used at analysis time.
206 The buildid.dir variable can be used to either change this directory
207 cache location, or to disable it altogether. If you want to disable it,
208 set buildid.dir to /dev/null. The default is $HOME/.debug
211 These options work only for TUI.
212 These are in control of addresses, jump function, source code
213 in lines of assembly code from a specific program.
215 annotate.hide_src_code::
216 If a program which is analyzed has source code,
217 this option lets 'annotate' print a list of assembly code with the source code.
218 For example, let's see a part of a program. There're four lines.
219 If this option is 'true', they can be printed
220 without source code from a program as below.
227 But if this option is 'false', source code of the part
228 can be also printed as below. Default is 'false'.
230 │ struct rb_node *rb_next(const struct rb_node *node)
235 │ struct rb_node *parent;
237 │ if (RB_EMPTY_NODE(node))
241 annotate.use_offset::
242 Basing on a first address of a loaded function, offset can be used.
243 Instead of using original addresses of assembly code,
244 addresses subtracted from a base address can be printed.
245 Let's illustrate an example.
246 If a base address is 0XFFFFFFFF81624d50 as below,
248 ffffffff81624d50 <load0>
250 an address on assembly code has a specific absolute address as below
252 ffffffff816250b8:│ mov 0x8(%r14),%rdi
254 but if use_offset is 'true', an address subtracted from a base address is printed.
255 Default is true. This option is only applied to TUI.
257 368:│ mov 0x8(%r14),%rdi
259 annotate.jump_arrows::
260 There can be jump instruction among assembly code.
261 Depending on a boolean value of jump_arrows,
262 arrows can be printed or not which represent
263 where do the instruction jump into as below.
267 │1330:│ mov %r15,%r10
268 │1333:└─→cmp %r15,%r14
270 If jump_arrow is 'false', the arrows isn't printed as below.
278 annotate.show_linenr::
279 When showing source code if this option is 'true',
280 line numbers are printed as below.
282 │1628 if (type & PERF_SAMPLE_IDENTIFIER) {
284 │1628 data->id = *array;
288 However if this option is 'false', they aren't printed as below.
291 │ if (type & PERF_SAMPLE_IDENTIFIER) {
297 annotate.show_nr_jumps::
298 Let's see a part of assembly code.
300 │1382: movb $0x1,-0x270(%rbp)
302 If use this, the number of branches jumping to that address can be printed as below.
305 │1 1382: movb $0x1,-0x270(%rbp)
307 annotate.show_total_period::
308 To compare two records on an instruction base, with this option
309 provided, display total number of samples that belong to a line
310 in assembly code. If this option is 'true', total periods are printed
311 instead of percent values as below.
315 But if this option is 'false', percent values for overhead are printed i.e.
318 99.93 │ mov %eax,%eax
322 This option control the way to calculate overhead of filtered entries -
323 that means the value of this option is effective only if there's a
324 filter (by comm, dso or symbol name). Suppose a following example:
332 This is an original overhead and we'll filter out the first 'foo'
333 entry. The value of 'relative' would increase the overhead of 'bar'
334 and 'baz' to 50.00% for each, while 'absolute' would show their
335 current overhead (33.33%).
339 This option controls display of column headers (like 'Overhead' and 'Symbol')
340 in 'report' and 'top'. If this option is false, they are hidden.
341 This option is only applied to TUI.
344 When sub-commands 'top' and 'report' work with -g/—-children
345 there're options in control of call-graph.
347 call-graph.record-mode::
348 The record-mode can be 'fp' (frame pointer), 'dwarf' and 'lbr'.
349 The value of 'dwarf' is effective only if perf detect needed library
350 (libunwind or a recent version of libdw).
351 'lbr' only work for cpus that support it.
353 call-graph.dump-size::
354 The size of stack to dump in order to do post-unwinding. Default is 8192 (byte).
355 When using dwarf into record-mode, the default size will be used if omitted.
357 call-graph.print-type::
358 The print-types can be graph (graph absolute), fractal (graph relative),
359 flat and folded. This option controls a way to show overhead for each callchain
360 entry. Suppose a following example.
374 This output is a 'fractal' format. The 'foo' came from 'bar' and 'baz' exactly
375 half and half so 'fractal' shows 50.00% for each
376 (meaning that it assumes 100% total overhead of 'foo').
378 The 'graph' uses absolute overhead value of 'foo' as total so each of
379 'bar' and 'baz' callchain will have 20.00% of overhead.
380 If 'flat' is used, single column and linear exposure of call chains.
381 'folded' mean call chains are displayed in a line, separated by semicolons.
384 This option controls print order of callchains. The default is
385 'callee' which means callee is printed at top and then followed by its
386 caller and so on. The 'caller' prints it in reverse order.
388 If this option is not set and report.children or top.children is
389 set to true (or the equivalent command line option is given),
390 the default value of this option is changed to 'caller' for the
391 execution of 'perf report' or 'perf top'. Other commands will
392 still default to 'callee'.
394 call-graph.sort-key::
395 The callchains are merged if they contain same information.
396 The sort-key option determines a way to compare the callchains.
397 A value of 'sort-key' can be 'function' or 'address'.
398 The default is 'function'.
400 call-graph.threshold::
401 When there're many callchains it'd print tons of lines. So perf omits
402 small callchains under a certain overhead (threshold) and this option
403 control the threshold. Default is 0.5 (%). The overhead is calculated
404 by value depends on call-graph.print-type.
406 call-graph.print-limit::
407 This is a maximum number of lines of callchain printed for a single
408 histogram entry. Default is 0 which means no limitation.
412 Allows changing the default sort order from "comm,dso,symbol" to
413 some other default, for instance "sym,dso" may be more fitting for
415 report.percent-limit::
416 This one is mostly the same as call-graph.threshold but works for
417 histogram entries. Entries having an overhead lower than this
418 percentage will not be printed. Default is '0'. If percent-limit
419 is '10', only entries which have more than 10% of overhead will be
423 This option sets up the maximum allocation size of the internal
424 event queue for ordering events. Default is 0, meaning no limit.
427 'Children' means functions called from another function.
428 If this option is true, 'perf report' cumulates callchains of children
429 and show (accumulated) total overhead as well as 'Self' overhead.
430 Please refer to the 'perf report' manual. The default is 'true'.
433 This option is to show event group information together.
434 Example output with this turned on, notice that there is one column
435 per event in the group, ref-cycles and cycles:
437 # group: {ref-cycles,cycles}
440 # Samples: 7K of event 'anon group { ref-cycles, cycles }'
441 # Event count (approx.): 6876107743
443 # Overhead Command Shared Object Symbol
444 # ................ ....... ................. ...................
446 99.84% 99.76% noploop noploop [.] main
447 0.07% 0.00% noploop ld-2.15.so [.] strcmp
448 0.03% 0.00% noploop [kernel.kallsyms] [k] timerqueue_del
452 Same as 'report.children'. So if it is enabled, the output of 'top'
453 command will have 'Children' overhead column as well as 'Self' overhead
455 The default is 'true'.
459 This option can assign a tool to view manual pages when 'help'
460 subcommand was invoked. Supported tools are 'man', 'woman'
461 (with emacs client) and 'konqueror'. Default is 'man'.
463 New man viewer tool can be also added using 'man.<tool>.cmd'
464 or use different path using 'man.<tool>.path' config option.
468 When the subcommand is run on stdio, determine whether it uses
469 pager or not based on this value. Default is 'unspecified'.
473 This option decides which allocator is to be analyzed if neither
474 '--slab' nor '--page' option is used. Default is 'slab'.
478 This option can be 'cache', 'no-cache' or 'skip'.
479 'cache' is to post-process data and save/update the binaries into
480 the build-id cache (in ~/.debug). This is the default.
481 But if this option is 'no-cache', it will not update the build-id cache.
482 'skip' skips post-processing and does not update the cache.