tools/perf/Documentation/perf-report.txt

   1 perf-report(1)
   2 ==============
   3
   4 NAME
   5 ----
   6 perf-report - Read perf.data (created by perf record) and display the profile
   7
   8 SYNOPSIS
   9 --------
  10 [verse]
  11 'perf report' [-i <file> | --input=file]
  12
  13 DESCRIPTION
  14 -----------
  15 This command displays the performance counter profile information recorded
  16 via perf record.
  17
  18 OPTIONS
  19 -------
  20 -i::
  21 --input=::
  22         Input file name. (default: perf.data unless stdin is a fifo)
  23
  24 -v::
  25 --verbose::
  26         Be more verbose. (show symbol address, etc)
  27
  28 -n::
  29 --show-nr-samples::
  30         Show the number of samples for each symbol
  31
  32 --showcpuutilization::
  33         Show sample percentage for different cpu modes.
  34
  35 -T::
  36 --threads::
  37         Show per-thread event counters
  38 -c::
  39 --comms=::
  40         Only consider symbols in these comms. CSV that understands
  41         file://filename entries.  This option will affect the percentage of
  42         the overhead column.  See --percentage for more info.
  43 --pid=::
  44         Only show events for given process ID (comma separated list).
  45
  46 --tid=::
  47         Only show events for given thread ID (comma separated list).
  48 -d::
  49 --dsos=::
  50         Only consider symbols in these dsos. CSV that understands
  51         file://filename entries.  This option will affect the percentage of
  52         the overhead column.  See --percentage for more info.
  53 -S::
  54 --symbols=::
  55         Only consider these symbols. CSV that understands
  56         file://filename entries.  This option will affect the percentage of
  57         the overhead column.  See --percentage for more info.
  58
  59 --symbol-filter=::
  60         Only show symbols that match (partially) with this filter.
  61
  62 -U::
  63 --hide-unresolved::
  64         Only display entries resolved to a symbol.
  65
  66 -s::
  67 --sort=::
  68         Sort histogram entries by given key(s) - multiple keys can be specified
  69         in CSV format.  Following sort keys are available:
  70         pid, comm, dso, symbol, parent, cpu, srcline, weight, local_weight.
  71
  72         Each key has following meaning:
  73
  74         - comm: command (name) of the task which can be read via /proc/<pid>/comm
  75         - pid: command and tid of the task
  76         - dso: name of library or module executed at the time of sample
  77         - symbol: name of function executed at the time of sample
  78         - parent: name of function matched to the parent regex filter. Unmatched
  79         entries are displayed as "[other]".
  80         - cpu: cpu number the task ran at the time of sample
  81         - srcline: filename and line number executed at the time of sample.  The
  82         DWARF debugging info must be provided.
  83         - weight: Event specific weight, e.g. memory latency or transaction
  84         abort cost. This is the global weight.
  85         - local_weight: Local weight version of the weight above.
  86         - transaction: Transaction abort flags.
  87         - overhead: Overhead percentage of sample
  88         - overhead_sys: Overhead percentage of sample running in system mode
  89         - overhead_us: Overhead percentage of sample running in user mode
  90         - overhead_guest_sys: Overhead percentage of sample running in system mode
  91         on guest machine
  92         - overhead_guest_us: Overhead percentage of sample running in user mode on
  93         guest machine
  94         - sample: Number of sample
  95         - period: Raw number of event count of sample
  96
  97         By default, comm, dso and symbol keys are used.
  98         (i.e. --sort comm,dso,symbol)
  99
 100         If --branch-stack option is used, following sort keys are also
 101         available:
 102         dso_from, dso_to, symbol_from, symbol_to, mispredict.
 103
 104         - dso_from: name of library or module branched from
 105         - dso_to: name of library or module branched to
 106         - symbol_from: name of function branched from
 107         - symbol_to: name of function branched to
 108         - mispredict: "N" for predicted branch, "Y" for mispredicted branch
 109         - in_tx: branch in TSX transaction
 110         - abort: TSX transaction abort.
 111
 112         And default sort keys are changed to comm, dso_from, symbol_from, dso_to
 113         and symbol_to, see '--branch-stack'.
 114
 115 -F::
 116 --fields=::
 117         Specify output field - multiple keys can be specified in CSV format.
 118         Following fields are available:
 119         overhead, overhead_sys, overhead_us, overhead_children, sample and period.
 120         Also it can contain any sort key(s).
 121
 122         By default, every sort keys not specified in -F will be appended
 123         automatically.
 124
 125         If --mem-mode option is used, following sort keys are also available
 126         (incompatible with --branch-stack):
 127         symbol_daddr, dso_daddr, locked, tlb, mem, snoop, dcacheline.
 128
 129         - symbol_daddr: name of data symbol being executed on at the time of sample
 130         - dso_daddr: name of library or module containing the data being executed
 131         on at the time of sample
 132         - locked: whether the bus was locked at the time of sample
 133         - tlb: type of tlb access for the data at the time of sample
 134         - mem: type of memory access for the data at the time of sample
 135         - snoop: type of snoop (if any) for the data at the time of sample
 136         - dcacheline: the cacheline the data address is on at the time of sample
 137
 138         And default sort keys are changed to local_weight, mem, sym, dso,
 139         symbol_daddr, dso_daddr, snoop, tlb, locked, see '--mem-mode'.
 140
 141 -p::
 142 --parent=<regex>::
 143         A regex filter to identify parent. The parent is a caller of this
 144         function and searched through the callchain, thus it requires callchain
 145         information recorded. The pattern is in the exteneded regex format and
 146         defaults to "\^sys_|^do_page_fault", see '--sort parent'.
 147
 148 -x::
 149 --exclude-other::
 150         Only display entries with parent-match.
 151
 152 -w::
 153 --column-widths=<width[,width...]>::
 154         Force each column width to the provided list, for large terminal
 155         readability.  0 means no limit (default behavior).
 156
 157 -t::
 158 --field-separator=::
 159         Use a special separator character and don't pad with spaces, replacing
 160         all occurrences of this separator in symbol names (and other output)
 161         with a '.' character, that thus it's the only non valid separator.
 162
 163 -D::
 164 --dump-raw-trace::
 165         Dump raw trace in ASCII.
 166
 167 -g [type,min[,limit],order[,key][,branch]]::
 168 --call-graph::
 169         Display call chains using type, min percent threshold, optional print
 170         limit and order.
 171         type can be either:
 172         - flat: single column, linear exposure of call chains.
 173         - graph: use a graph tree, displaying absolute overhead rates.
 174         - fractal: like graph, but displays relative rates. Each branch of
 175                  the tree is considered as a new profiled object. +
 176
 177         order can be either:
 178         - callee: callee based call graph.
 179         - caller: inverted caller based call graph.
 180
 181         key can be:
 182         - function: compare on functions
 183         - address: compare on individual code addresses
 184
 185         branch can be:
 186         - branch: include last branch information in callgraph
 187         when available. Usually more convenient to use --branch-history
 188         for this.
 189
 190         Default: fractal,0.5,callee,function.
 191
 192 --children::
 193         Accumulate callchain of children to parent entry so that then can
 194         show up in the output.  The output will have a new "Children" column
 195         and will be sorted on the data.  It requires callchains are recorded.
 196
 197 --max-stack::
 198         Set the stack depth limit when parsing the callchain, anything
 199         beyond the specified depth will be ignored. This is a trade-off
 200         between information loss and faster processing especially for
 201         workloads that can have a very long callchain stack.
 202
 203         Default: 127
 204
 205 -G::
 206 --inverted::
 207         alias for inverted caller based call graph.
 208
 209 --ignore-callees=<regex>::
 210         Ignore callees of the function(s) matching the given regex.
 211         This has the effect of collecting the callers of each such
 212         function into one place in the call-graph tree.
 213
 214 --pretty=<key>::
 215         Pretty printing style.  key: normal, raw
 216
 217 --stdio:: Use the stdio interface.
 218
 219 --tui:: Use the TUI interface, that is integrated with annotate and allows
 220         zooming into DSOs or threads, among other features. Use of --tui
 221         requires a tty, if one is not present, as when piping to other
 222         commands, the stdio interface is used.
 223
 224 --gtk:: Use the GTK2 interface.
 225
 226 -k::
 227 --vmlinux=<file>::
 228         vmlinux pathname
 229
 230 --kallsyms=<file>::
 231         kallsyms pathname
 232
 233 -m::
 234 --modules::
 235         Load module symbols. WARNING: This should only be used with -k and
 236         a LIVE kernel.
 237
 238 -f::
 239 --force::
 240         Don't complain, do it.
 241
 242 --symfs=<directory>::
 243         Look for files with symbols relative to this directory.
 244
 245 -C::
 246 --cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can
 247         be provided as a comma-separated list with no space: 0,1. Ranges of
 248         CPUs are specified with -: 0-2. Default is to report samples on all
 249         CPUs.
 250
 251 -M::
 252 --disassembler-style=:: Set disassembler style for objdump.
 253
 254 --source::
 255         Interleave source code with assembly code. Enabled by default,
 256         disable with --no-source.
 257
 258 --asm-raw::
 259         Show raw instruction encoding of assembly instructions.
 260
 261 --show-total-period:: Show a column with the sum of periods.
 262
 263 -I::
 264 --show-info::
 265         Display extended information about the perf.data file. This adds
 266         information which may be very large and thus may clutter the display.
 267         It currently includes: cpu and numa topology of the host system.
 268
 269 -b::
 270 --branch-stack::
 271         Use the addresses of sampled taken branches instead of the instruction
 272         address to build the histograms. To generate meaningful output, the
 273         perf.data file must have been obtained using perf record -b or
 274         perf record --branch-filter xxx where xxx is a branch filter option.
 275         perf report is able to auto-detect whether a perf.data file contains
 276         branch stacks and it will automatically switch to the branch view mode,
 277         unless --no-branch-stack is used.
 278
 279 --branch-history::
 280         Add the addresses of sampled taken branches to the callstack.
 281         This allows to examine the path the program took to each sample.
 282         The data collection must have used -b (or -j) and -g.
 283
 284 --objdump=<path>::
 285         Path to objdump binary.
 286
 287 --group::
 288         Show event group information together.
 289
 290 --demangle::
 291         Demangle symbol names to human readable form. It's enabled by default,
 292         disable with --no-demangle.
 293
 294 --demangle-kernel::
 295         Demangle kernel symbol names to human readable form (for C++ kernels).
 296
 297 --mem-mode::
 298         Use the data addresses of samples in addition to instruction addresses
 299         to build the histograms.  To generate meaningful output, the perf.data
 300         file must have been obtained using perf record -d -W and using a
 301         special event -e cpu/mem-loads/ or -e cpu/mem-stores/. See
 302         'perf mem' for simpler access.
 303
 304 --percent-limit::
 305         Do not show entries which have an overhead under that percent.
 306         (Default: 0).
 307
 308 --percentage::
 309         Determine how to display the overhead percentage of filtered entries.
 310         Filters can be applied by --comms, --dsos and/or --symbols options and
 311         Zoom operations on the TUI (thread, dso, etc).
 312
 313         "relative" means it's relative to filtered entries only so that the
 314         sum of shown entries will be always 100%.  "absolute" means it retains
 315         the original value before and after the filter is applied.
 316
 317 --header::
 318         Show header information in the perf.data file.  This includes
 319         various information like hostname, OS and perf version, cpu/mem
 320         info, perf command line, event list and so on.  Currently only
 321         --stdio output supports this feature.
 322
 323 --header-only::
 324         Show only perf.data header (forces --stdio).
 325
 326 SEE ALSO
 327 --------
 328 linkperf:perf-stat[1], linkperf:perf-annotate[1]