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