1perf-top(1) 2=========== 3 4NAME 5---- 6perf-top - System profiling tool. 7 8SYNOPSIS 9-------- 10[verse] 11'perf top' [-e <EVENT> | --event=EVENT] [<options>] 12 13DESCRIPTION 14----------- 15This command generates and displays a performance counter profile in real time. 16 17 18OPTIONS 19------- 20-a:: 21--all-cpus:: 22 System-wide collection. (default) 23 24-c <count>:: 25--count=<count>:: 26 Event period to sample. 27 28-C <cpu-list>:: 29--cpu=<cpu>:: 30Monitor only on the list of CPUs provided. Multiple CPUs can be provided as a 31comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2. 32Default is to monitor all CPUS. 33 34-d <seconds>:: 35--delay=<seconds>:: 36 Number of seconds to delay between refreshes. 37 38-e <event>:: 39--event=<event>:: 40 Select the PMU event. Selection can be a symbolic event name 41 (use 'perf list' to list all events) or a raw PMU 42 event (eventsel+umask) in the form of rNNN where NNN is a 43 hexadecimal event descriptor. 44 45-E <entries>:: 46--entries=<entries>:: 47 Display this many functions. 48 49-f <count>:: 50--count-filter=<count>:: 51 Only display functions with more events than this. 52 53--group:: 54 Put the counters into a counter group. 55 56--group-sort-idx:: 57 Sort the output by the event at the index n in group. If n is invalid, 58 sort by the first event. It can support multiple groups with different 59 amount of events. WARNING: This should be used on grouped events. 60 61-F <freq>:: 62--freq=<freq>:: 63 Profile at this frequency. Use 'max' to use the currently maximum 64 allowed frequency, i.e. the value in the kernel.perf_event_max_sample_rate 65 sysctl. 66 67-i:: 68--inherit:: 69 Child tasks do not inherit counters. 70 71-k <path>:: 72--vmlinux=<path>:: 73 Path to vmlinux. Required for annotation functionality. 74 75--ignore-vmlinux:: 76 Ignore vmlinux files. 77 78--kallsyms=<file>:: 79 kallsyms pathname 80 81-m <pages>:: 82--mmap-pages=<pages>:: 83 Number of mmap data pages (must be a power of two) or size 84 specification with appended unit character - B/K/M/G. The 85 size is rounded up to have nearest pages power of two value. 86 87-p <pid>:: 88--pid=<pid>:: 89 Profile events on existing Process ID (comma separated list). 90 91-t <tid>:: 92--tid=<tid>:: 93 Profile events on existing thread ID (comma separated list). 94 95-u:: 96--uid=:: 97 Record events in threads owned by uid. Name or number. 98 99-r <priority>:: 100--realtime=<priority>:: 101 Collect data with this RT SCHED_FIFO priority. 102 103--sym-annotate=<symbol>:: 104 Annotate this symbol. 105 106-K:: 107--hide_kernel_symbols:: 108 Hide kernel symbols. 109 110-U:: 111--hide_user_symbols:: 112 Hide user symbols. 113 114--demangle-kernel:: 115 Demangle kernel symbols. 116 117-D:: 118--dump-symtab:: 119 Dump the symbol table used for profiling. 120 121-v:: 122--verbose:: 123 Be more verbose (show counter open errors, etc). 124 125-z:: 126--zero:: 127 Zero history across display updates. 128 129-s:: 130--sort:: 131 Sort by key(s): pid, comm, dso, symbol, parent, srcline, weight, 132 local_weight, abort, in_tx, transaction, overhead, sample, period. 133 Please see description of --sort in the perf-report man page. 134 135--fields=:: 136 Specify output field - multiple keys can be specified in CSV format. 137 Following fields are available: 138 overhead, overhead_sys, overhead_us, overhead_children, sample and period. 139 Also it can contain any sort key(s). 140 141 By default, every sort keys not specified in --field will be appended 142 automatically. 143 144-n:: 145--show-nr-samples:: 146 Show a column with the number of samples. 147 148--show-total-period:: 149 Show a column with the sum of periods. 150 151--dsos:: 152 Only consider symbols in these dsos. This option will affect the 153 percentage of the overhead column. See --percentage for more info. 154 155--comms:: 156 Only consider symbols in these comms. This option will affect the 157 percentage of the overhead column. See --percentage for more info. 158 159--symbols:: 160 Only consider these symbols. This option will affect the 161 percentage of the overhead column. See --percentage for more info. 162 163-M:: 164--disassembler-style=:: Set disassembler style for objdump. 165 166--prefix=PREFIX:: 167--prefix-strip=N:: 168 Remove first N entries from source file path names in executables 169 and add PREFIX. This allows to display source code compiled on systems 170 with different file system layout. 171 172--source:: 173 Interleave source code with assembly code. Enabled by default, 174 disable with --no-source. 175 176--asm-raw:: 177 Show raw instruction encoding of assembly instructions. 178 179-g:: 180 Enables call-graph (stack chain/backtrace) recording. 181 182--call-graph [mode,type,min[,limit],order[,key][,branch]]:: 183 Setup and enable call-graph (stack chain/backtrace) recording, 184 implies -g. See `--call-graph` section in perf-record and 185 perf-report man pages for details. 186 187--children:: 188 Accumulate callchain of children to parent entry so that then can 189 show up in the output. The output will have a new "Children" column 190 and will be sorted on the data. It requires -g/--call-graph option 191 enabled. See the `overhead calculation' section for more details. 192 Enabled by default, disable with --no-children. 193 194--max-stack:: 195 Set the stack depth limit when parsing the callchain, anything 196 beyond the specified depth will be ignored. This is a trade-off 197 between information loss and faster processing especially for 198 workloads that can have a very long callchain stack. 199 200 Default: /proc/sys/kernel/perf_event_max_stack when present, 127 otherwise. 201 202--ignore-callees=<regex>:: 203 Ignore callees of the function(s) matching the given regex. 204 This has the effect of collecting the callers of each such 205 function into one place in the call-graph tree. 206 207--percent-limit:: 208 Do not show entries which have an overhead under that percent. 209 (Default: 0). 210 211--percentage:: 212 Determine how to display the overhead percentage of filtered entries. 213 Filters can be applied by --comms, --dsos and/or --symbols options and 214 Zoom operations on the TUI (thread, dso, etc). 215 216 "relative" means it's relative to filtered entries only so that the 217 sum of shown entries will be always 100%. "absolute" means it retains 218 the original value before and after the filter is applied. 219 220-w:: 221--column-widths=<width[,width...]>:: 222 Force each column width to the provided list, for large terminal 223 readability. 0 means no limit (default behavior). 224 225--proc-map-timeout:: 226 When processing pre-existing threads /proc/XXX/mmap, it may take 227 a long time, because the file may be huge. A time out is needed 228 in such cases. 229 This option sets the time out limit. The default value is 500 ms. 230 231 232-b:: 233--branch-any:: 234 Enable taken branch stack sampling. Any type of taken branch may be sampled. 235 This is a shortcut for --branch-filter any. See --branch-filter for more infos. 236 237-j:: 238--branch-filter:: 239 Enable taken branch stack sampling. Each sample captures a series of consecutive 240 taken branches. The number of branches captured with each sample depends on the 241 underlying hardware, the type of branches of interest, and the executed code. 242 It is possible to select the types of branches captured by enabling filters. 243 For a full list of modifiers please see the perf record manpage. 244 245 The option requires at least one branch type among any, any_call, any_ret, ind_call, cond. 246 The privilege levels may be omitted, in which case, the privilege levels of the associated 247 event are applied to the branch filter. Both kernel (k) and hypervisor (hv) privilege 248 levels are subject to permissions. When sampling on multiple events, branch stack sampling 249 is enabled for all the sampling events. The sampled branch type is the same for all events. 250 The various filters must be specified as a comma separated list: --branch-filter any_ret,u,k 251 Note that this feature may not be available on all processors. 252 253--raw-trace:: 254 When displaying traceevent output, do not use print fmt or plugins. 255 256--hierarchy:: 257 Enable hierarchy output. 258 259--overwrite:: 260 Enable this to use just the most recent records, which helps in high core count 261 machines such as Knights Landing/Mill, but right now is disabled by default as 262 the pausing used in this technique is leading to loss of metadata events such 263 as PERF_RECORD_MMAP which makes 'perf top' unable to resolve samples, leading 264 to lots of unknown samples appearing on the UI. Enable this if you are in such 265 machines and profiling a workload that doesn't creates short lived threads and/or 266 doesn't uses many executable mmap operations. Work is being planed to solve 267 this situation, till then, this will remain disabled by default. 268 269--force:: 270 Don't do ownership validation. 271 272--num-thread-synthesize:: 273 The number of threads to run when synthesizing events for existing processes. 274 By default, the number of threads equals to the number of online CPUs. 275 276--namespaces:: 277 Record events of type PERF_RECORD_NAMESPACES and display it with the 278 'cgroup_id' sort key. 279 280-G name:: 281--cgroup name:: 282monitor only in the container (cgroup) called "name". This option is available only 283in per-cpu mode. The cgroup filesystem must be mounted. All threads belonging to 284container "name" are monitored when they run on the monitored CPUs. Multiple cgroups 285can be provided. Each cgroup is applied to the corresponding event, i.e., first cgroup 286to first event, second cgroup to second event and so on. It is possible to provide 287an empty cgroup (monitor all the time) using, e.g., -G foo,,bar. Cgroups must have 288corresponding events, i.e., they always refer to events defined earlier on the command 289line. If the user wants to track multiple events for a specific cgroup, the user can 290use '-e e1 -e e2 -G foo,foo' or just use '-e e1 -e e2 -G foo'. 291 292--all-cgroups:: 293 Record events of type PERF_RECORD_CGROUP and display it with the 294 'cgroup' sort key. 295 296--switch-on EVENT_NAME:: 297 Only consider events after this event is found. 298 299 E.g.: 300 301 Find out where broadcast packets are handled 302 303 perf probe -L icmp_rcv 304 305 Insert a probe there: 306 307 perf probe icmp_rcv:59 308 309 Start perf top and ask it to only consider the cycles events when a 310 broadcast packet arrives This will show a menu with two entries and 311 will start counting when a broadcast packet arrives: 312 313 perf top -e cycles,probe:icmp_rcv --switch-on=probe:icmp_rcv 314 315 Alternatively one can ask for --group and then two overhead columns 316 will appear, the first for cycles and the second for the switch-on event. 317 318 perf top --group -e cycles,probe:icmp_rcv --switch-on=probe:icmp_rcv 319 320 This may be interesting to measure a workload only after some initialization 321 phase is over, i.e. insert a perf probe at that point and use the above 322 examples replacing probe:icmp_rcv with the just-after-init probe. 323 324--switch-off EVENT_NAME:: 325 Stop considering events after this event is found. 326 327--show-on-off-events:: 328 Show the --switch-on/off events too. This has no effect in 'perf top' now 329 but probably we'll make the default not to show the switch-on/off events 330 on the --group mode and if there is only one event besides the off/on ones, 331 go straight to the histogram browser, just like 'perf top' with no events 332 explicitly specified does. 333 334--stitch-lbr:: 335 Show callgraph with stitched LBRs, which may have more complete 336 callgraph. The option must be used with --call-graph lbr recording. 337 Disabled by default. In common cases with call stack overflows, 338 it can recreate better call stacks than the default lbr call stack 339 output. But this approach is not full proof. There can be cases 340 where it creates incorrect call stacks from incorrect matches. 341 The known limitations include exception handing such as 342 setjmp/longjmp will have calls/returns not match. 343 344ifdef::HAVE_LIBPFM[] 345--pfm-events events:: 346Select a PMU event using libpfm4 syntax (see http://perfmon2.sf.net) 347including support for event filters. For example '--pfm-events 348inst_retired:any_p:u:c=1:i'. More than one event can be passed to the 349option using the comma separator. Hardware events and generic hardware 350events cannot be mixed together. The latter must be used with the -e 351option. The -e option and this one can be mixed and matched. Events 352can be grouped using the {} notation. 353endif::HAVE_LIBPFM[] 354 355INTERACTIVE PROMPTING KEYS 356-------------------------- 357 358[d]:: 359 Display refresh delay. 360 361[e]:: 362 Number of entries to display. 363 364[E]:: 365 Event to display when multiple counters are active. 366 367[f]:: 368 Profile display filter (>= hit count). 369 370[F]:: 371 Annotation display filter (>= % of total). 372 373[s]:: 374 Annotate symbol. 375 376[S]:: 377 Stop annotation, return to full profile display. 378 379[K]:: 380 Hide kernel symbols. 381 382[U]:: 383 Hide user symbols. 384 385[z]:: 386 Toggle event count zeroing across display updates. 387 388[qQ]:: 389 Quit. 390 391Pressing any unmapped key displays a menu, and prompts for input. 392 393include::callchain-overhead-calculation.txt[] 394 395SEE ALSO 396-------- 397linkperf:perf-stat[1], linkperf:perf-list[1], linkperf:perf-report[1] 398