1perf-config(1) 2============== 3 4NAME 5---- 6perf-config - Get and set variables in a configuration file. 7 8SYNOPSIS 9-------- 10[verse] 11'perf config' [<file-option>] [section.name[=value] ...] 12or 13'perf config' [<file-option>] -l | --list 14 15DESCRIPTION 16----------- 17You can manage variables in a configuration file with this command. 18 19OPTIONS 20------- 21 22-l:: 23--list:: 24 Show current config variables, name and value, for all sections. 25 26--user:: 27 For writing and reading options: write to user 28 '$HOME/.perfconfig' file or read it. 29 30--system:: 31 For writing and reading options: write to system-wide 32 '$(sysconfdir)/perfconfig' or read it. 33 34CONFIGURATION FILE 35------------------ 36 37The perf configuration file contains many variables to change various 38aspects of each of its tools, including output, disk usage, etc. 39The '$HOME/.perfconfig' file is used to store a per-user configuration. 40The file '$(sysconfdir)/perfconfig' can be used to 41store a system-wide default configuration. 42 43One an disable reading config files by setting the PERF_CONFIG environment 44variable to /dev/null, or provide an alternate config file by setting that 45variable. 46 47When reading or writing, the values are read from the system and user 48configuration files by default, and options '--system' and '--user' 49can be used to tell the command to read from or write to only that location. 50 51Syntax 52~~~~~~ 53 54The file consist of sections. A section starts with its name 55surrounded by square brackets and continues till the next section 56begins. Each variable must be in a section, and have the form 57'name = value', for example: 58 59 [section] 60 name1 = value1 61 name2 = value2 62 63Section names are case sensitive and can contain any characters except 64newline (double quote `"` and backslash have to be escaped as `\"` and `\\`, 65respectively). Section headers can't span multiple lines. 66 67Example 68~~~~~~~ 69 70Given a $HOME/.perfconfig like this: 71 72# 73# This is the config file, and 74# a '#' and ';' character indicates a comment 75# 76 77 [colors] 78 # Color variables 79 top = red, default 80 medium = green, default 81 normal = lightgray, default 82 selected = white, lightgray 83 jump_arrows = blue, default 84 addr = magenta, default 85 root = white, blue 86 87 [tui] 88 # Defaults if linked with libslang 89 report = on 90 annotate = on 91 top = on 92 93 [buildid] 94 # Default, disable using /dev/null 95 dir = ~/.debug 96 97 [annotate] 98 # Defaults 99 hide_src_code = false 100 use_offset = true 101 jump_arrows = true 102 show_nr_jumps = false 103 104 [help] 105 # Format can be man, info, web or html 106 format = man 107 autocorrect = 0 108 109 [ui] 110 show-headers = true 111 112 [call-graph] 113 # fp (framepointer), dwarf 114 record-mode = fp 115 print-type = graph 116 order = caller 117 sort-key = function 118 119 [report] 120 # Defaults 121 sort_order = comm,dso,symbol 122 percent-limit = 0 123 queue-size = 0 124 children = true 125 group = true 126 skip-empty = true 127 128 [llvm] 129 dump-obj = true 130 clang-opt = -g 131 132You can hide source code of annotate feature setting the config to false with 133 134 % perf config annotate.hide_src_code=true 135 136If you want to add or modify several config items, you can do like 137 138 % perf config ui.show-headers=false kmem.default=slab 139 140To modify the sort order of report functionality in user config file(i.e. `~/.perfconfig`), do 141 142 % perf config --user report.sort-order=srcline 143 144To change colors of selected line to other foreground and background colors 145in system config file (i.e. `$(sysconf)/perfconfig`), do 146 147 % perf config --system colors.selected=yellow,green 148 149To query the record mode of call graph, do 150 151 % perf config call-graph.record-mode 152 153If you want to know multiple config key/value pairs, you can do like 154 155 % perf config report.queue-size call-graph.order report.children 156 157To query the config value of sort order of call graph in user config file (i.e. `~/.perfconfig`), do 158 159 % perf config --user call-graph.sort-order 160 161To query the config value of buildid directory in system config file (i.e. `$(sysconf)/perfconfig`), do 162 163 % perf config --system buildid.dir 164 165Variables 166~~~~~~~~~ 167 168colors.*:: 169 The variables for customizing the colors used in the output for the 170 'report', 'top' and 'annotate' in the TUI. They should specify the 171 foreground and background colors, separated by a comma, for example: 172 173 medium = green, lightgray 174 175 If you want to use the color configured for you terminal, just leave it 176 as 'default', for example: 177 178 medium = default, lightgray 179 180 Available colors: 181 red, yellow, green, cyan, gray, black, blue, 182 white, default, magenta, lightgray 183 184 colors.top:: 185 'top' means a overhead percentage which is more than 5%. 186 And values of this variable specify percentage colors. 187 Basic key values are foreground-color 'red' and 188 background-color 'default'. 189 colors.medium:: 190 'medium' means a overhead percentage which has more than 0.5%. 191 Default values are 'green' and 'default'. 192 colors.normal:: 193 'normal' means the rest of overhead percentages 194 except 'top', 'medium', 'selected'. 195 Default values are 'lightgray' and 'default'. 196 colors.selected:: 197 This selects the colors for the current entry in a list of entries 198 from sub-commands (top, report, annotate). 199 Default values are 'black' and 'lightgray'. 200 colors.jump_arrows:: 201 Colors for jump arrows on assembly code listings 202 such as 'jns', 'jmp', 'jane', etc. 203 Default values are 'blue', 'default'. 204 colors.addr:: 205 This selects colors for addresses from 'annotate'. 206 Default values are 'magenta', 'default'. 207 colors.root:: 208 Colors for headers in the output of a sub-commands (top, report). 209 Default values are 'white', 'blue'. 210 211core.*:: 212 core.proc-map-timeout:: 213 Sets a timeout (in milliseconds) for parsing /proc/<pid>/maps files. 214 Can be overridden by the --proc-map-timeout option on supported 215 subcommands. The default timeout is 500ms. 216 217tui.*, gtk.*:: 218 Subcommands that can be configured here are 'top', 'report' and 'annotate'. 219 These values are booleans, for example: 220 221 [tui] 222 top = true 223 224 will make the TUI be the default for the 'top' subcommand. Those will be 225 available if the required libs were detected at tool build time. 226 227buildid.*:: 228 buildid.dir:: 229 Each executable and shared library in modern distributions comes with a 230 content based identifier that, if available, will be inserted in a 231 'perf.data' file header to, at analysis time find what is needed to do 232 symbol resolution, code annotation, etc. 233 234 The recording tools also stores a hard link or copy in a per-user 235 directory, $HOME/.debug/, of binaries, shared libraries, /proc/kallsyms 236 and /proc/kcore files to be used at analysis time. 237 238 The buildid.dir variable can be used to either change this directory 239 cache location, or to disable it altogether. If you want to disable it, 240 set buildid.dir to /dev/null. The default is $HOME/.debug 241 242buildid-cache.*:: 243 buildid-cache.debuginfod=URLs 244 Specify debuginfod URLs to be used when retrieving perf.data binaries, 245 it follows the same syntax as the DEBUGINFOD_URLS variable, like: 246 247 buildid-cache.debuginfod=http://192.168.122.174:8002 248 249annotate.*:: 250 These are in control of addresses, jump function, source code 251 in lines of assembly code from a specific program. 252 253 annotate.addr2line:: 254 addr2line binary to use for file names and line numbers. 255 256 annotate.objdump:: 257 objdump binary to use for disassembly and annotations. 258 259 annotate.disassembler_style:: 260 Use this to change the default disassembler style to some other value 261 supported by binutils, such as "intel", see the '-M' option help in the 262 'objdump' man page. 263 264 annotate.hide_src_code:: 265 If a program which is analyzed has source code, 266 this option lets 'annotate' print a list of assembly code with the source code. 267 For example, let's see a part of a program. There're four lines. 268 If this option is 'true', they can be printed 269 without source code from a program as below. 270 271 │ push %rbp 272 │ mov %rsp,%rbp 273 │ sub $0x10,%rsp 274 │ mov (%rdi),%rdx 275 276 But if this option is 'false', source code of the part 277 can be also printed as below. Default is 'false'. 278 279 │ struct rb_node *rb_next(const struct rb_node *node) 280 │ { 281 │ push %rbp 282 │ mov %rsp,%rbp 283 │ sub $0x10,%rsp 284 │ struct rb_node *parent; 285 │ 286 │ if (RB_EMPTY_NODE(node)) 287 │ mov (%rdi),%rdx 288 │ return n; 289 290 This option works with tui, stdio2 browsers. 291 292 annotate.use_offset:: 293 Basing on a first address of a loaded function, offset can be used. 294 Instead of using original addresses of assembly code, 295 addresses subtracted from a base address can be printed. 296 Let's illustrate an example. 297 If a base address is 0XFFFFFFFF81624d50 as below, 298 299 ffffffff81624d50 <load0> 300 301 an address on assembly code has a specific absolute address as below 302 303 ffffffff816250b8:│ mov 0x8(%r14),%rdi 304 305 but if use_offset is 'true', an address subtracted from a base address is printed. 306 Default is true. This option is only applied to TUI. 307 308 368:│ mov 0x8(%r14),%rdi 309 310 This option works with tui, stdio2 browsers. 311 312 annotate.jump_arrows:: 313 There can be jump instruction among assembly code. 314 Depending on a boolean value of jump_arrows, 315 arrows can be printed or not which represent 316 where do the instruction jump into as below. 317 318 │ ┌──jmp 1333 319 │ │ xchg %ax,%ax 320 │1330:│ mov %r15,%r10 321 │1333:└─→cmp %r15,%r14 322 323 If jump_arrow is 'false', the arrows isn't printed as below. 324 Default is 'false'. 325 326 │ ↓ jmp 1333 327 │ xchg %ax,%ax 328 │1330: mov %r15,%r10 329 │1333: cmp %r15,%r14 330 331 This option works with tui browser. 332 333 annotate.show_linenr:: 334 When showing source code if this option is 'true', 335 line numbers are printed as below. 336 337 │1628 if (type & PERF_SAMPLE_IDENTIFIER) { 338 │ ↓ jne 508 339 │1628 data->id = *array; 340 │1629 array++; 341 │1630 } 342 343 However if this option is 'false', they aren't printed as below. 344 Default is 'false'. 345 346 │ if (type & PERF_SAMPLE_IDENTIFIER) { 347 │ ↓ jne 508 348 │ data->id = *array; 349 │ array++; 350 │ } 351 352 This option works with tui, stdio2 browsers. 353 354 annotate.show_nr_jumps:: 355 Let's see a part of assembly code. 356 357 │1382: movb $0x1,-0x270(%rbp) 358 359 If use this, the number of branches jumping to that address can be printed as below. 360 Default is 'false'. 361 362 │1 1382: movb $0x1,-0x270(%rbp) 363 364 This option works with tui, stdio2 browsers. 365 366 annotate.show_total_period:: 367 To compare two records on an instruction base, with this option 368 provided, display total number of samples that belong to a line 369 in assembly code. If this option is 'true', total periods are printed 370 instead of percent values as below. 371 372 302 │ mov %eax,%eax 373 374 But if this option is 'false', percent values for overhead are printed i.e. 375 Default is 'false'. 376 377 99.93 │ mov %eax,%eax 378 379 This option works with tui, stdio2, stdio browsers. 380 381 annotate.show_nr_samples:: 382 By default perf annotate shows percentage of samples. This option 383 can be used to print absolute number of samples. Ex, when set as 384 false: 385 386 Percent│ 387 74.03 │ mov %fs:0x28,%rax 388 389 When set as true: 390 391 Samples│ 392 6 │ mov %fs:0x28,%rax 393 394 This option works with tui, stdio2, stdio browsers. 395 396 annotate.offset_level:: 397 Default is '1', meaning just jump targets will have offsets show right beside 398 the instruction. When set to '2' 'call' instructions will also have its offsets 399 shown, 3 or higher will show offsets for all instructions. 400 401 This option works with tui, stdio2 browsers. 402 403 annotate.demangle:: 404 Demangle symbol names to human readable form. Default is 'true'. 405 406 annotate.demangle_kernel:: 407 Demangle kernel symbol names to human readable form. Default is 'true'. 408 409hist.*:: 410 hist.percentage:: 411 This option control the way to calculate overhead of filtered entries - 412 that means the value of this option is effective only if there's a 413 filter (by comm, dso or symbol name). Suppose a following example: 414 415 Overhead Symbols 416 ........ ....... 417 33.33% foo 418 33.33% bar 419 33.33% baz 420 421 This is an original overhead and we'll filter out the first 'foo' 422 entry. The value of 'relative' would increase the overhead of 'bar' 423 and 'baz' to 50.00% for each, while 'absolute' would show their 424 current overhead (33.33%). 425 426ui.*:: 427 ui.show-headers:: 428 This option controls display of column headers (like 'Overhead' and 'Symbol') 429 in 'report' and 'top'. If this option is false, they are hidden. 430 This option is only applied to TUI. 431 432call-graph.*:: 433 The following controls the handling of call-graphs (obtained via the 434 -g/--call-graph options). 435 436 call-graph.record-mode:: 437 The mode for user space can be 'fp' (frame pointer), 'dwarf' 438 and 'lbr'. The value 'dwarf' is effective only if libunwind 439 (or a recent version of libdw) is present on the system; 440 the value 'lbr' only works for certain cpus. The method for 441 kernel space is controlled not by this option but by the 442 kernel config (CONFIG_UNWINDER_*). 443 444 call-graph.dump-size:: 445 The size of stack to dump in order to do post-unwinding. Default is 8192 (byte). 446 When using dwarf into record-mode, the default size will be used if omitted. 447 448 call-graph.print-type:: 449 The print-types can be graph (graph absolute), fractal (graph relative), 450 flat and folded. This option controls a way to show overhead for each callchain 451 entry. Suppose a following example. 452 453 Overhead Symbols 454 ........ ....... 455 40.00% foo 456 | 457 ---foo 458 | 459 |--50.00%--bar 460 | main 461 | 462 --50.00%--baz 463 main 464 465 This output is a 'fractal' format. The 'foo' came from 'bar' and 'baz' exactly 466 half and half so 'fractal' shows 50.00% for each 467 (meaning that it assumes 100% total overhead of 'foo'). 468 469 The 'graph' uses absolute overhead value of 'foo' as total so each of 470 'bar' and 'baz' callchain will have 20.00% of overhead. 471 If 'flat' is used, single column and linear exposure of call chains. 472 'folded' mean call chains are displayed in a line, separated by semicolons. 473 474 call-graph.order:: 475 This option controls print order of callchains. The default is 476 'callee' which means callee is printed at top and then followed by its 477 caller and so on. The 'caller' prints it in reverse order. 478 479 If this option is not set and report.children or top.children is 480 set to true (or the equivalent command line option is given), 481 the default value of this option is changed to 'caller' for the 482 execution of 'perf report' or 'perf top'. Other commands will 483 still default to 'callee'. 484 485 call-graph.sort-key:: 486 The callchains are merged if they contain same information. 487 The sort-key option determines a way to compare the callchains. 488 A value of 'sort-key' can be 'function' or 'address'. 489 The default is 'function'. 490 491 call-graph.threshold:: 492 When there're many callchains it'd print tons of lines. So perf omits 493 small callchains under a certain overhead (threshold) and this option 494 control the threshold. Default is 0.5 (%). The overhead is calculated 495 by value depends on call-graph.print-type. 496 497 call-graph.print-limit:: 498 This is a maximum number of lines of callchain printed for a single 499 histogram entry. Default is 0 which means no limitation. 500 501report.*:: 502 report.sort_order:: 503 Allows changing the default sort order from "comm,dso,symbol" to 504 some other default, for instance "sym,dso" may be more fitting for 505 kernel developers. 506 report.percent-limit:: 507 This one is mostly the same as call-graph.threshold but works for 508 histogram entries. Entries having an overhead lower than this 509 percentage will not be printed. Default is '0'. If percent-limit 510 is '10', only entries which have more than 10% of overhead will be 511 printed. 512 513 report.queue-size:: 514 This option sets up the maximum allocation size of the internal 515 event queue for ordering events. Default is 0, meaning no limit. 516 517 report.children:: 518 'Children' means functions called from another function. 519 If this option is true, 'perf report' cumulates callchains of children 520 and show (accumulated) total overhead as well as 'Self' overhead. 521 Please refer to the 'perf report' manual. The default is 'true'. 522 523 report.group:: 524 This option is to show event group information together. 525 Example output with this turned on, notice that there is one column 526 per event in the group, ref-cycles and cycles: 527 528 # group: {ref-cycles,cycles} 529 # ======== 530 # 531 # Samples: 7K of event 'anon group { ref-cycles, cycles }' 532 # Event count (approx.): 6876107743 533 # 534 # Overhead Command Shared Object Symbol 535 # ................ ....... ................. ................... 536 # 537 99.84% 99.76% noploop noploop [.] main 538 0.07% 0.00% noploop ld-2.15.so [.] strcmp 539 0.03% 0.00% noploop [kernel.kallsyms] [k] timerqueue_del 540 541 report.skip-empty:: 542 This option can change default stat behavior with empty results. 543 If it's set true, 'perf report --stat' will not show 0 stats. 544 545top.*:: 546 top.children:: 547 Same as 'report.children'. So if it is enabled, the output of 'top' 548 command will have 'Children' overhead column as well as 'Self' overhead 549 column by default. 550 The default is 'true'. 551 552 top.call-graph:: 553 This is identical to 'call-graph.record-mode', except it is 554 applicable only for 'top' subcommand. This option ONLY setup 555 the unwind method. To enable 'perf top' to actually use it, 556 the command line option -g must be specified. 557 558man.*:: 559 man.viewer:: 560 This option can assign a tool to view manual pages when 'help' 561 subcommand was invoked. Supported tools are 'man', 'woman' 562 (with emacs client) and 'konqueror'. Default is 'man'. 563 564 New man viewer tool can be also added using 'man.<tool>.cmd' 565 or use different path using 'man.<tool>.path' config option. 566 567pager.*:: 568 pager.<subcommand>:: 569 When the subcommand is run on stdio, determine whether it uses 570 pager or not based on this value. Default is 'unspecified'. 571 572kmem.*:: 573 kmem.default:: 574 This option decides which allocator is to be analyzed if neither 575 '--slab' nor '--page' option is used. Default is 'slab'. 576 577record.*:: 578 record.build-id:: 579 This option can be 'cache', 'no-cache', 'skip' or 'mmap'. 580 'cache' is to post-process data and save/update the binaries into 581 the build-id cache (in ~/.debug). This is the default. 582 But if this option is 'no-cache', it will not update the build-id cache. 583 'skip' skips post-processing and does not update the cache. 584 'mmap' skips post-processing and reads build-ids from MMAP events. 585 586 record.call-graph:: 587 This is identical to 'call-graph.record-mode', except it is 588 applicable only for 'record' subcommand. This option ONLY setup 589 the unwind method. To enable 'perf record' to actually use it, 590 the command line option -g must be specified. 591 592 record.aio:: 593 Use 'n' control blocks in asynchronous (Posix AIO) trace writing 594 mode ('n' default: 1, max: 4). 595 596 record.debuginfod:: 597 Specify debuginfod URL to be used when cacheing perf.data binaries, 598 it follows the same syntax as the DEBUGINFOD_URLS variable, like: 599 600 http://192.168.122.174:8002 601 602 If the URLs is 'system', the value of DEBUGINFOD_URLS system environment 603 variable is used. 604 605diff.*:: 606 diff.order:: 607 This option sets the number of columns to sort the result. 608 The default is 0, which means sorting by baseline. 609 Setting it to 1 will sort the result by delta (or other 610 compute method selected). 611 612 diff.compute:: 613 This options sets the method for computing the diff result. 614 Possible values are 'delta', 'delta-abs', 'ratio' and 615 'wdiff'. Default is 'delta'. 616 617trace.*:: 618 trace.add_events:: 619 Allows adding a set of events to add to the ones specified 620 by the user, or use as a default one if none was specified. 621 The initial use case is to add augmented_raw_syscalls.o to 622 activate the 'perf trace' logic that looks for syscall 623 pointer contents after the normal tracepoint payload. 624 625 trace.args_alignment:: 626 Number of columns to align the argument list, default is 70, 627 use 40 for the strace default, zero to no alignment. 628 629 trace.no_inherit:: 630 Do not follow children threads. 631 632 trace.show_arg_names:: 633 Should syscall argument names be printed? If not then trace.show_zeros 634 will be set. 635 636 trace.show_duration:: 637 Show syscall duration. 638 639 trace.show_prefix:: 640 If set to 'yes' will show common string prefixes in tables. The default 641 is to remove the common prefix in things like "MAP_SHARED", showing just "SHARED". 642 643 trace.show_timestamp:: 644 Show syscall start timestamp. 645 646 trace.show_zeros:: 647 Do not suppress syscall arguments that are equal to zero. 648 649 trace.tracepoint_beautifiers:: 650 Use "libtraceevent" to use that library to augment the tracepoint arguments, 651 "libbeauty", the default, to use the same argument beautifiers used in the 652 strace-like sys_enter+sys_exit lines. 653 654ftrace.*:: 655 ftrace.tracer:: 656 Can be used to select the default tracer when neither -G nor 657 -F option is not specified. Possible values are 'function' and 658 'function_graph'. 659 660llvm.*:: 661 llvm.clang-path:: 662 Path to clang. If omit, search it from $PATH. 663 664 llvm.clang-bpf-cmd-template:: 665 Cmdline template. Below lines show its default value. Environment 666 variable is used to pass options. 667 "$CLANG_EXEC -D__KERNEL__ -D__NR_CPUS__=$NR_CPUS "\ 668 "-DLINUX_VERSION_CODE=$LINUX_VERSION_CODE " \ 669 "$CLANG_OPTIONS $PERF_BPF_INC_OPTIONS $KERNEL_INC_OPTIONS " \ 670 "-Wno-unused-value -Wno-pointer-sign " \ 671 "-working-directory $WORKING_DIR " \ 672 "-c \"$CLANG_SOURCE\" -target bpf $CLANG_EMIT_LLVM -O2 -o - $LLVM_OPTIONS_PIPE" 673 674 llvm.clang-opt:: 675 Options passed to clang. 676 677 llvm.kbuild-dir:: 678 kbuild directory. If not set, use /lib/modules/`uname -r`/build. 679 If set to "" deliberately, skip kernel header auto-detector. 680 681 llvm.kbuild-opts:: 682 Options passed to 'make' when detecting kernel header options. 683 684 llvm.dump-obj:: 685 Enable perf dump BPF object files compiled by LLVM. 686 687 llvm.opts:: 688 Options passed to llc. 689 690samples.*:: 691 692 samples.context:: 693 Define how many ns worth of time to show 694 around samples in perf report sample context browser. 695 696scripts.*:: 697 698 Any option defines a script that is added to the scripts menu 699 in the interactive perf browser and whose output is displayed. 700 The name of the option is the name, the value is a script command line. 701 The script gets the same options passed as a full perf script, 702 in particular -i perfdata file, --cpu, --tid 703 704convert.*:: 705 706 convert.queue-size:: 707 Limit the size of ordered_events queue, so we could control 708 allocation size of perf data files without proper finished 709 round events. 710stat.*:: 711 712 stat.big-num:: 713 (boolean) Change the default for "--big-num". To make 714 "--no-big-num" the default, set "stat.big-num=false". 715 716intel-pt.*:: 717 718 intel-pt.cache-divisor:: 719 720 intel-pt.mispred-all:: 721 If set, Intel PT decoder will set the mispred flag on all 722 branches. 723 724 intel-pt.max-loops:: 725 If set and non-zero, the maximum number of unconditional 726 branches decoded without consuming any trace packets. If 727 the maximum is exceeded there will be a "Never-ending loop" 728 error. The default is 100000. 729 730auxtrace.*:: 731 732 auxtrace.dumpdir:: 733 s390 only. The directory to save the auxiliary trace buffer 734 can be changed using this option. Ex, auxtrace.dumpdir=/tmp. 735 If the directory does not exist or has the wrong file type, 736 the current directory is used. 737 738itrace.*:: 739 740 debug-log-buffer-size:: 741 Log size in bytes to output when using the option --itrace=d+e 742 Refer 'itrace' option of linkperf:perf-script[1] or 743 linkperf:perf-report[1]. The default is 16384. 744 745daemon.*:: 746 747 daemon.base:: 748 Base path for daemon data. All sessions data are stored under 749 this path. 750 751session-<NAME>.*:: 752 753 session-<NAME>.run:: 754 755 Defines new record session for daemon. The value is record's 756 command line without the 'record' keyword. 757 758 759SEE ALSO 760-------- 761linkperf:perf[1] 762