1 perf-report(1) 1 perf-report(1)
2 ============== 2 ==============
3 3
4 NAME 4 NAME
5 ---- 5 ----
6 perf-report - Read perf.data (created by perf 6 perf-report - Read perf.data (created by perf record) and display the profile
7 7
8 SYNOPSIS 8 SYNOPSIS
9 -------- 9 --------
10 [verse] 10 [verse]
11 'perf report' [-i <file> | --input=file] 11 'perf report' [-i <file> | --input=file]
12 12
13 DESCRIPTION 13 DESCRIPTION
14 ----------- 14 -----------
15 This command displays the performance counter 15 This command displays the performance counter profile information recorded
16 via perf record. 16 via perf record.
17 17
18 OPTIONS 18 OPTIONS
19 ------- 19 -------
20 -i:: 20 -i::
21 --input=:: 21 --input=::
22 Input file name. (default: perf.data u 22 Input file name. (default: perf.data unless stdin is a fifo)
23 23
24 -v:: 24 -v::
25 --verbose:: 25 --verbose::
26 Be more verbose. (show symbol address, 26 Be more verbose. (show symbol address, etc)
27 27
28 -q:: 28 -q::
29 --quiet:: 29 --quiet::
30 Do not show any warnings or messages. 30 Do not show any warnings or messages. (Suppress -v)
31 31
32 -n:: 32 -n::
33 --show-nr-samples:: 33 --show-nr-samples::
34 Show the number of samples for each sy 34 Show the number of samples for each symbol
35 35
36 --show-cpu-utilization:: 36 --show-cpu-utilization::
37 Show sample percentage for different c 37 Show sample percentage for different cpu modes.
38 38
39 -T:: 39 -T::
40 --threads:: 40 --threads::
41 Show per-thread event counters. The i 41 Show per-thread event counters. The input data file should be recorded
42 with -s option. 42 with -s option.
43 -c:: 43 -c::
44 --comms=:: 44 --comms=::
45 Only consider symbols in these comms. 45 Only consider symbols in these comms. CSV that understands
46 file://filename entries. This option 46 file://filename entries. This option will affect the percentage of
47 the overhead column. See --percentage 47 the overhead column. See --percentage for more info.
48 --pid=:: 48 --pid=::
49 Only show events for given process ID 49 Only show events for given process ID (comma separated list).
50 50
51 --tid=:: 51 --tid=::
52 Only show events for given thread ID ( 52 Only show events for given thread ID (comma separated list).
53 -d:: 53 -d::
54 --dsos=:: 54 --dsos=::
55 Only consider symbols in these dsos. C 55 Only consider symbols in these dsos. CSV that understands
56 file://filename entries. This option 56 file://filename entries. This option will affect the percentage of
57 the overhead column. See --percentage 57 the overhead column. See --percentage for more info.
58 -S:: 58 -S::
59 --symbols=:: 59 --symbols=::
60 Only consider these symbols. CSV that 60 Only consider these symbols. CSV that understands
61 file://filename entries. This option 61 file://filename entries. This option will affect the percentage of
62 the overhead column. See --percentage 62 the overhead column. See --percentage for more info.
63 63
64 --symbol-filter=:: 64 --symbol-filter=::
65 Only show symbols that match (partiall 65 Only show symbols that match (partially) with this filter.
66 66
67 -U:: 67 -U::
68 --hide-unresolved:: 68 --hide-unresolved::
69 Only display entries resolved to a sym 69 Only display entries resolved to a symbol.
70 70
71 -s:: 71 -s::
72 --sort=:: 72 --sort=::
73 Sort histogram entries by given key(s) 73 Sort histogram entries by given key(s) - multiple keys can be specified
74 in CSV format. Following sort keys ar 74 in CSV format. Following sort keys are available:
75 pid, comm, dso, symbol, parent, cpu, s 75 pid, comm, dso, symbol, parent, cpu, socket, srcline, weight,
76 local_weight, cgroup_id, addr. 76 local_weight, cgroup_id, addr.
77 77
78 Each key has following meaning: 78 Each key has following meaning:
79 79
80 - comm: command (name) of the task whi 80 - comm: command (name) of the task which can be read via /proc/<pid>/comm
81 - pid: command and tid of the task 81 - pid: command and tid of the task
82 - dso: name of library or module execu 82 - dso: name of library or module executed at the time of sample
83 - dso_size: size of library or module 83 - dso_size: size of library or module executed at the time of sample
84 - symbol: name of function executed at 84 - symbol: name of function executed at the time of sample
85 - symbol_size: size of function execut 85 - symbol_size: size of function executed at the time of sample
86 - parent: name of function matched to 86 - parent: name of function matched to the parent regex filter. Unmatched
87 entries are displayed as "[other]". 87 entries are displayed as "[other]".
88 - cpu: cpu number the task ran at the 88 - cpu: cpu number the task ran at the time of sample
89 - socket: processor socket number the 89 - socket: processor socket number the task ran at the time of sample
90 - srcline: filename and line number ex 90 - srcline: filename and line number executed at the time of sample. The
91 DWARF debugging info must be provided. 91 DWARF debugging info must be provided.
92 - srcfile: file name of the source fil 92 - srcfile: file name of the source file of the samples. Requires dwarf
93 information. 93 information.
94 - weight: Event specific weight, e.g. 94 - weight: Event specific weight, e.g. memory latency or transaction
95 abort cost. This is the global weight. 95 abort cost. This is the global weight.
96 - local_weight: Local weight version o 96 - local_weight: Local weight version of the weight above.
97 - cgroup_id: ID derived from cgroup na 97 - cgroup_id: ID derived from cgroup namespace device and inode numbers.
98 - cgroup: cgroup pathname in the cgrou 98 - cgroup: cgroup pathname in the cgroupfs.
99 - transaction: Transaction abort flags 99 - transaction: Transaction abort flags.
100 - overhead: Overhead percentage of sam 100 - overhead: Overhead percentage of sample
101 - overhead_sys: Overhead percentage of 101 - overhead_sys: Overhead percentage of sample running in system mode
102 - overhead_us: Overhead percentage of 102 - overhead_us: Overhead percentage of sample running in user mode
103 - overhead_guest_sys: Overhead percent 103 - overhead_guest_sys: Overhead percentage of sample running in system mode
104 on guest machine 104 on guest machine
105 - overhead_guest_us: Overhead percenta 105 - overhead_guest_us: Overhead percentage of sample running in user mode on
106 guest machine 106 guest machine
107 - sample: Number of sample 107 - sample: Number of sample
108 - period: Raw number of event count of 108 - period: Raw number of event count of sample
109 - time: Separate the samples by time s 109 - time: Separate the samples by time stamp with the resolution specified by
110 --time-quantum (default 100ms). Specif 110 --time-quantum (default 100ms). Specify with overhead and before it.
111 - code_page_size: the code page size o 111 - code_page_size: the code page size of sampled code address (ip)
112 - ins_lat: Instruction latency in core 112 - ins_lat: Instruction latency in core cycles. This is the global instruction
113 latency 113 latency
114 - local_ins_lat: Local instruction lat 114 - local_ins_lat: Local instruction latency version
115 - p_stage_cyc: On powerpc, this presen 115 - p_stage_cyc: On powerpc, this presents the number of cycles spent in a
116 pipeline stage. And currently suppor 116 pipeline stage. And currently supported only on powerpc.
117 - addr: (Full) virtual address of the 117 - addr: (Full) virtual address of the sampled instruction
118 - retire_lat: On X86, this reports pip 118 - retire_lat: On X86, this reports pipeline stall of this instruction compared
119 to the previous instruction in cycle 119 to the previous instruction in cycles. And currently supported only on X86
120 - simd: Flags describing a SIMD operat 120 - simd: Flags describing a SIMD operation. "e" for empty Arm SVE predicate. "p" for partial Arm SVE predicate
121 - type: Data type of sample memory acc 121 - type: Data type of sample memory access.
122 - typeoff: Offset in the data type of 122 - typeoff: Offset in the data type of sample memory access.
123 - symoff: Offset in the symbol. 123 - symoff: Offset in the symbol.
124 - weight1: Average value of event spec 124 - weight1: Average value of event specific weight (1st field of weight_struct).
125 - weight2: Average value of event spec 125 - weight2: Average value of event specific weight (2nd field of weight_struct).
126 - weight3: Average value of event spec 126 - weight3: Average value of event specific weight (3rd field of weight_struct).
127 127
128 By default, comm, dso and symbol keys 128 By default, comm, dso and symbol keys are used.
129 (i.e. --sort comm,dso,symbol) 129 (i.e. --sort comm,dso,symbol)
130 130
131 If --branch-stack option is used, foll 131 If --branch-stack option is used, following sort keys are also
132 available: 132 available:
133 133
134 - dso_from: name of library or module 134 - dso_from: name of library or module branched from
135 - dso_to: name of library or module br 135 - dso_to: name of library or module branched to
136 - symbol_from: name of function branch 136 - symbol_from: name of function branched from
137 - symbol_to: name of function branched 137 - symbol_to: name of function branched to
138 - srcline_from: source file and line b 138 - srcline_from: source file and line branched from
139 - srcline_to: source file and line bra 139 - srcline_to: source file and line branched to
140 - mispredict: "N" for predicted branch 140 - mispredict: "N" for predicted branch, "Y" for mispredicted branch
141 - in_tx: branch in TSX transaction 141 - in_tx: branch in TSX transaction
142 - abort: TSX transaction abort. 142 - abort: TSX transaction abort.
143 - cycles: Cycles in basic block 143 - cycles: Cycles in basic block
144 144
145 And default sort keys are changed to c 145 And default sort keys are changed to comm, dso_from, symbol_from, dso_to
146 and symbol_to, see '--branch-stack'. 146 and symbol_to, see '--branch-stack'.
147 147
148 When the sort key symbol is specified, 148 When the sort key symbol is specified, columns "IPC" and "IPC Coverage"
149 are enabled automatically. Column "IPC 149 are enabled automatically. Column "IPC" reports the average IPC per function
150 and column "IPC coverage" reports the 150 and column "IPC coverage" reports the percentage of instructions with
151 sampled IPC in this function. IPC mean 151 sampled IPC in this function. IPC means Instruction Per Cycle. If it's low,
152 it indicates there may be a performanc 152 it indicates there may be a performance bottleneck when the function is
153 executed, such as a memory access bott 153 executed, such as a memory access bottleneck. If a function has high overhead
154 and low IPC, it's worth further analyz 154 and low IPC, it's worth further analyzing it to optimize its performance.
155 155
156 If the --mem-mode option is used, the 156 If the --mem-mode option is used, the following sort keys are also available
157 (incompatible with --branch-stack): 157 (incompatible with --branch-stack):
158 symbol_daddr, dso_daddr, locked, tlb, 158 symbol_daddr, dso_daddr, locked, tlb, mem, snoop, dcacheline, blocked.
159 159
160 - symbol_daddr: name of data symbol be 160 - symbol_daddr: name of data symbol being executed on at the time of sample
161 - dso_daddr: name of library or module 161 - dso_daddr: name of library or module containing the data being executed
162 on at the time of the sample 162 on at the time of the sample
163 - locked: whether the bus was locked a 163 - locked: whether the bus was locked at the time of the sample
164 - tlb: type of tlb access for the data 164 - tlb: type of tlb access for the data at the time of the sample
165 - mem: type of memory access for the d 165 - mem: type of memory access for the data at the time of the sample
166 - snoop: type of snoop (if any) for th 166 - snoop: type of snoop (if any) for the data at the time of the sample
167 - dcacheline: the cacheline the data a 167 - dcacheline: the cacheline the data address is on at the time of the sample
168 - phys_daddr: physical address of data 168 - phys_daddr: physical address of data being executed on at the time of sample
169 - data_page_size: the data page size o 169 - data_page_size: the data page size of data being executed on at the time of sample
170 - blocked: reason of blocked load acce 170 - blocked: reason of blocked load access for the data at the time of the sample
171 171
172 And the default sort keys are changed 172 And the default sort keys are changed to local_weight, mem, sym, dso,
173 symbol_daddr, dso_daddr, snoop, tlb, l 173 symbol_daddr, dso_daddr, snoop, tlb, locked, blocked, local_ins_lat,
174 see '--mem-mode'. 174 see '--mem-mode'.
175 175
176 If the data file has tracepoint event( 176 If the data file has tracepoint event(s), following (dynamic) sort keys
177 are also available: 177 are also available:
178 trace, trace_fields, [<event>.]<field> 178 trace, trace_fields, [<event>.]<field>[/raw]
179 179
180 - trace: pretty printed trace output i 180 - trace: pretty printed trace output in a single column
181 - trace_fields: fields in tracepoints 181 - trace_fields: fields in tracepoints in separate columns
182 - <field name>: optional event and fie 182 - <field name>: optional event and field name for a specific field
183 183
184 The last form consists of event and fi 184 The last form consists of event and field names. If event name is
185 omitted, it searches all events for ma 185 omitted, it searches all events for matching field name. The matched
186 field will be shown only for the event 186 field will be shown only for the event has the field. The event name
187 supports substring match so user doesn 187 supports substring match so user doesn't need to specify full subsystem
188 and event name everytime. For example 188 and event name everytime. For example, 'sched:sched_switch' event can
189 be shortened to 'switch' as long as it 189 be shortened to 'switch' as long as it's not ambiguous. Also event can
190 be specified by its index (starting fr 190 be specified by its index (starting from 1) preceded by the '%'.
191 So '%1' is the first event, '%2' is th 191 So '%1' is the first event, '%2' is the second, and so on.
192 192
193 The field name can have '/raw' suffix 193 The field name can have '/raw' suffix which disables pretty printing
194 and shows raw field value like hex num 194 and shows raw field value like hex numbers. The --raw-trace option
195 has the same effect for all dynamic so 195 has the same effect for all dynamic sort keys.
196 196
197 The default sort keys are changed to ' 197 The default sort keys are changed to 'trace' if all events in the data
198 file are tracepoint. 198 file are tracepoint.
199 199
200 -F:: 200 -F::
201 --fields=:: 201 --fields=::
202 Specify output field - multiple keys c 202 Specify output field - multiple keys can be specified in CSV format.
203 Following fields are available: 203 Following fields are available:
204 overhead, overhead_sys, overhead_us, o 204 overhead, overhead_sys, overhead_us, overhead_children, sample, period,
205 weight1, weight2, weight3, ins_lat, p_ 205 weight1, weight2, weight3, ins_lat, p_stage_cyc and retire_lat. The
206 last 3 names are alias for the corresp 206 last 3 names are alias for the corresponding weights. When the weight
207 fields are used, they will show the av 207 fields are used, they will show the average value of the weight.
208 208
209 Also it can contain any sort key(s). 209 Also it can contain any sort key(s).
210 210
211 By default, every sort keys not specif 211 By default, every sort keys not specified in -F will be appended
212 automatically. 212 automatically.
213 213
214 If the keys starts with a prefix '+', 214 If the keys starts with a prefix '+', then it will append the specified
215 field(s) to the default field order. F 215 field(s) to the default field order. For example: perf report -F +period,sample.
216 216
217 -p:: 217 -p::
218 --parent=<regex>:: 218 --parent=<regex>::
219 A regex filter to identify parent. The 219 A regex filter to identify parent. The parent is a caller of this
220 function and searched through the call 220 function and searched through the callchain, thus it requires callchain
221 information recorded. The pattern is i 221 information recorded. The pattern is in the extended regex format and
222 defaults to "\^sys_|^do_page_fault", s 222 defaults to "\^sys_|^do_page_fault", see '--sort parent'.
223 223
224 -x:: 224 -x::
225 --exclude-other:: 225 --exclude-other::
226 Only display entries with parent-match 226 Only display entries with parent-match.
227 227
228 -w:: 228 -w::
229 --column-widths=<width[,width...]>:: 229 --column-widths=<width[,width...]>::
230 Force each column width to the provide 230 Force each column width to the provided list, for large terminal
231 readability. 0 means no limit (defaul 231 readability. 0 means no limit (default behavior).
232 232
233 -t:: 233 -t::
234 --field-separator=:: 234 --field-separator=::
235 Use a special separator character and 235 Use a special separator character and don't pad with spaces, replacing
236 all occurrences of this separator in s 236 all occurrences of this separator in symbol names (and other output)
237 with a '.' character, that thus it's t 237 with a '.' character, that thus it's the only non valid separator.
238 238
239 -D:: 239 -D::
240 --dump-raw-trace:: 240 --dump-raw-trace::
241 Dump raw trace in ASCII. 241 Dump raw trace in ASCII.
242 242
243 --disable-order:: 243 --disable-order::
244 Disable raw trace ordering. 244 Disable raw trace ordering.
245 245
246 -g:: 246 -g::
247 --call-graph=<print_type,threshold[,print_limi 247 --call-graph=<print_type,threshold[,print_limit],order,sort_key[,branch],value>::
248 Display call chains using type, min pe 248 Display call chains using type, min percent threshold, print limit,
249 call order, sort key, optional branch 249 call order, sort key, optional branch and value. Note that ordering
250 is not fixed so any parameter can be g 250 is not fixed so any parameter can be given in an arbitrary order.
251 One exception is the print_limit which 251 One exception is the print_limit which should be preceded by threshold.
252 252
253 print_type can be either: 253 print_type can be either:
254 - flat: single column, linear exposure 254 - flat: single column, linear exposure of call chains.
255 - graph: use a graph tree, displaying 255 - graph: use a graph tree, displaying absolute overhead rates. (default)
256 - fractal: like graph, but displays re 256 - fractal: like graph, but displays relative rates. Each branch of
257 the tree is considered as a n 257 the tree is considered as a new profiled object.
258 - folded: call chains are displayed in 258 - folded: call chains are displayed in a line, separated by semicolons
259 - none: disable call chain display. 259 - none: disable call chain display.
260 260
261 threshold is a percentage value which 261 threshold is a percentage value which specifies a minimum percent to be
262 included in the output call graph. De 262 included in the output call graph. Default is 0.5 (%).
263 263
264 print_limit is only applied when stdio 264 print_limit is only applied when stdio interface is used. It's to limit
265 number of call graph entries in a sing 265 number of call graph entries in a single hist entry. Note that it needs
266 to be given after threshold (but not n 266 to be given after threshold (but not necessarily consecutive).
267 Default is 0 (unlimited). 267 Default is 0 (unlimited).
268 268
269 order can be either: 269 order can be either:
270 - callee: callee based call graph. 270 - callee: callee based call graph.
271 - caller: inverted caller based call g 271 - caller: inverted caller based call graph.
272 Default is 'caller' when --children is 272 Default is 'caller' when --children is used, otherwise 'callee'.
273 273
274 sort_key can be: 274 sort_key can be:
275 - function: compare on functions (defa 275 - function: compare on functions (default)
276 - address: compare on individual code 276 - address: compare on individual code addresses
277 - srcline: compare on source filename 277 - srcline: compare on source filename and line number
278 278
279 branch can be: 279 branch can be:
280 - branch: include last branch informat 280 - branch: include last branch information in callgraph when available.
281 Usually more convenient to u 281 Usually more convenient to use --branch-history for this.
282 282
283 value can be: 283 value can be:
284 - percent: display overhead percent (d 284 - percent: display overhead percent (default)
285 - period: display event period 285 - period: display event period
286 - count: display event count 286 - count: display event count
287 287
288 --children:: 288 --children::
289 Accumulate callchain of children to pa 289 Accumulate callchain of children to parent entry so that then can
290 show up in the output. The output wil 290 show up in the output. The output will have a new "Children" column
291 and will be sorted on the data. It re 291 and will be sorted on the data. It requires callchains are recorded.
292 See the `overhead calculation' section 292 See the `overhead calculation' section for more details. Enabled by
293 default, disable with --no-children. 293 default, disable with --no-children.
294 294
295 --max-stack:: 295 --max-stack::
296 Set the stack depth limit when parsing 296 Set the stack depth limit when parsing the callchain, anything
297 beyond the specified depth will be ign 297 beyond the specified depth will be ignored. This is a trade-off
298 between information loss and faster pr 298 between information loss and faster processing especially for
299 workloads that can have a very long ca 299 workloads that can have a very long callchain stack.
300 Note that when using the --itrace opti 300 Note that when using the --itrace option the synthesized callchain size
301 will override this value if the synthe 301 will override this value if the synthesized callchain size is bigger.
302 302
303 Default: 127 303 Default: 127
304 304
305 -G:: 305 -G::
306 --inverted:: 306 --inverted::
307 alias for inverted caller based call g 307 alias for inverted caller based call graph.
308 308
309 --ignore-callees=<regex>:: 309 --ignore-callees=<regex>::
310 Ignore callees of the function(s) matc 310 Ignore callees of the function(s) matching the given regex.
311 This has the effect of collecting the 311 This has the effect of collecting the callers of each such
312 function into one place in the call-gr 312 function into one place in the call-graph tree.
313 313
314 --pretty=<key>:: 314 --pretty=<key>::
315 Pretty printing style. key: normal, r 315 Pretty printing style. key: normal, raw
316 316
317 --stdio:: Use the stdio interface. 317 --stdio:: Use the stdio interface.
318 318
319 --stdio-color:: 319 --stdio-color::
320 'always', 'never' or 'auto', allowing 320 'always', 'never' or 'auto', allowing configuring color output
321 via the command line, in addition to v 321 via the command line, in addition to via "color.ui" .perfconfig.
322 Use '--stdio-color always' to generate 322 Use '--stdio-color always' to generate color even when redirecting
323 to a pipe or file. Using just '--stdio 323 to a pipe or file. Using just '--stdio-color' is equivalent to
324 using 'always'. 324 using 'always'.
325 325
326 --tui:: Use the TUI interface, that is integra 326 --tui:: Use the TUI interface, that is integrated with annotate and allows
327 zooming into DSOs or threads, among ot 327 zooming into DSOs or threads, among other features. Use of --tui
328 requires a tty, if one is not present, 328 requires a tty, if one is not present, as when piping to other
329 commands, the stdio interface is used. 329 commands, the stdio interface is used.
330 330
331 --gtk:: Use the GTK2 interface. 331 --gtk:: Use the GTK2 interface.
332 332
333 -k:: 333 -k::
334 --vmlinux=<file>:: 334 --vmlinux=<file>::
335 vmlinux pathname 335 vmlinux pathname
336 336
337 --ignore-vmlinux:: 337 --ignore-vmlinux::
338 Ignore vmlinux files. 338 Ignore vmlinux files.
339 339
340 --kallsyms=<file>:: 340 --kallsyms=<file>::
341 kallsyms pathname 341 kallsyms pathname
342 342
343 -m:: 343 -m::
344 --modules:: 344 --modules::
345 Load module symbols. WARNING: This sho 345 Load module symbols. WARNING: This should only be used with -k and
346 a LIVE kernel. 346 a LIVE kernel.
347 347
348 -f:: 348 -f::
349 --force:: 349 --force::
350 Don't do ownership validation. 350 Don't do ownership validation.
351 351
352 --symfs=<directory>:: 352 --symfs=<directory>::
353 Look for files with symbols relative t 353 Look for files with symbols relative to this directory.
354 354
355 -C:: 355 -C::
356 --cpu:: Only report samples for the list of CP 356 --cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can
357 be provided as a comma-separated list 357 be provided as a comma-separated list with no space: 0,1. Ranges of
358 CPUs are specified with -: 0-2. Defaul 358 CPUs are specified with -: 0-2. Default is to report samples on all
359 CPUs. 359 CPUs.
360 360
361 -M:: 361 -M::
362 --disassembler-style=:: Set disassembler style 362 --disassembler-style=:: Set disassembler style for objdump.
363 363
364 --source:: 364 --source::
365 Interleave source code with assembly c 365 Interleave source code with assembly code. Enabled by default,
366 disable with --no-source. 366 disable with --no-source.
367 367
368 --asm-raw:: 368 --asm-raw::
369 Show raw instruction encoding of assem 369 Show raw instruction encoding of assembly instructions.
370 370
371 --show-total-period:: Show a column with the s 371 --show-total-period:: Show a column with the sum of periods.
372 372
373 -I:: 373 -I::
374 --show-info:: 374 --show-info::
375 Display extended information about the 375 Display extended information about the perf.data file. This adds
376 information which may be very large an 376 information which may be very large and thus may clutter the display.
377 It currently includes: cpu and numa to 377 It currently includes: cpu and numa topology of the host system.
378 378
379 -b:: 379 -b::
380 --branch-stack:: 380 --branch-stack::
381 Use the addresses of sampled taken bra 381 Use the addresses of sampled taken branches instead of the instruction
382 address to build the histograms. To ge 382 address to build the histograms. To generate meaningful output, the
383 perf.data file must have been obtained 383 perf.data file must have been obtained using perf record -b or
384 perf record --branch-filter xxx where 384 perf record --branch-filter xxx where xxx is a branch filter option.
385 perf report is able to auto-detect whe 385 perf report is able to auto-detect whether a perf.data file contains
386 branch stacks and it will automaticall 386 branch stacks and it will automatically switch to the branch view mode,
387 unless --no-branch-stack is used. 387 unless --no-branch-stack is used.
388 388
389 --branch-history:: 389 --branch-history::
390 Add the addresses of sampled taken bra 390 Add the addresses of sampled taken branches to the callstack.
391 This allows to examine the path the pr 391 This allows to examine the path the program took to each sample.
392 The data collection must have used -b 392 The data collection must have used -b (or -j) and -g.
393 393
394 --addr2line=<path>:: 394 --addr2line=<path>::
395 Path to addr2line binary. 395 Path to addr2line binary.
396 396
397 --objdump=<path>:: 397 --objdump=<path>::
398 Path to objdump binary. 398 Path to objdump binary.
399 399
400 --prefix=PREFIX:: 400 --prefix=PREFIX::
401 --prefix-strip=N:: 401 --prefix-strip=N::
402 Remove first N entries from source fil 402 Remove first N entries from source file path names in executables
403 and add PREFIX. This allows to display 403 and add PREFIX. This allows to display source code compiled on systems
404 with different file system layout. 404 with different file system layout.
405 405
406 --group:: 406 --group::
407 Show event group information together. 407 Show event group information together. It forces group output also
408 if there are no groups defined in data 408 if there are no groups defined in data file.
409 409
410 --group-sort-idx:: 410 --group-sort-idx::
411 Sort the output by the event at the in 411 Sort the output by the event at the index n in group. If n is invalid,
412 sort by the first event. It can suppor 412 sort by the first event. It can support multiple groups with different
413 amount of events. WARNING: This should 413 amount of events. WARNING: This should be used on grouped events.
414 414
415 --demangle:: 415 --demangle::
416 Demangle symbol names to human readabl 416 Demangle symbol names to human readable form. It's enabled by default,
417 disable with --no-demangle. 417 disable with --no-demangle.
418 418
419 --demangle-kernel:: 419 --demangle-kernel::
420 Demangle kernel symbol names to human 420 Demangle kernel symbol names to human readable form (for C++ kernels).
421 421
422 --mem-mode:: 422 --mem-mode::
423 Use the data addresses of samples in a 423 Use the data addresses of samples in addition to instruction addresses
424 to build the histograms. To generate 424 to build the histograms. To generate meaningful output, the perf.data
425 file must have been obtained using per 425 file must have been obtained using perf record -d -W and using a
426 special event -e cpu/mem-loads/p or -e 426 special event -e cpu/mem-loads/p or -e cpu/mem-stores/p. See
427 'perf mem' for simpler access. 427 'perf mem' for simpler access.
428 428
429 --percent-limit:: 429 --percent-limit::
430 Do not show entries which have an over 430 Do not show entries which have an overhead under that percent.
431 (Default: 0). Note that this option a 431 (Default: 0). Note that this option also sets the percent limit (threshold)
432 of callchains. However the default va 432 of callchains. However the default value of callchain threshold is
433 different than the default value of hi 433 different than the default value of hist entries. Please see the
434 --call-graph option for details. 434 --call-graph option for details.
435 435
436 --percentage:: 436 --percentage::
437 Determine how to display the overhead 437 Determine how to display the overhead percentage of filtered entries.
438 Filters can be applied by --comms, --d 438 Filters can be applied by --comms, --dsos and/or --symbols options and
439 Zoom operations on the TUI (thread, ds 439 Zoom operations on the TUI (thread, dso, etc).
440 440
441 "relative" means it's relative to filt 441 "relative" means it's relative to filtered entries only so that the
442 sum of shown entries will be always 10 442 sum of shown entries will be always 100%. "absolute" means it retains
443 the original value before and after th 443 the original value before and after the filter is applied.
444 444
445 --header:: 445 --header::
446 Show header information in the perf.da 446 Show header information in the perf.data file. This includes
447 various information like hostname, OS 447 various information like hostname, OS and perf version, cpu/mem
448 info, perf command line, event list an 448 info, perf command line, event list and so on. Currently only
449 --stdio output supports this feature. 449 --stdio output supports this feature.
450 450
451 --header-only:: 451 --header-only::
452 Show only perf.data header (forces --s 452 Show only perf.data header (forces --stdio).
453 453
454 --time:: 454 --time::
455 Only analyze samples within given time 455 Only analyze samples within given time window: <start>,<stop>. Times
456 have the format seconds.nanoseconds. I 456 have the format seconds.nanoseconds. If start is not given (i.e. time
457 string is ',x.y') then analysis starts 457 string is ',x.y') then analysis starts at the beginning of the file. If
458 stop time is not given (i.e. time stri 458 stop time is not given (i.e. time string is 'x.y,') then analysis goes
459 to end of file. Multiple ranges can be 459 to end of file. Multiple ranges can be separated by spaces, which
460 requires the argument to be quoted e.g 460 requires the argument to be quoted e.g. --time "1234.567,1234.789 1235,"
461 461
462 Also support time percent with multipl 462 Also support time percent with multiple time ranges. Time string is
463 'a%/n,b%/m,...' or 'a%-b%,c%-%d,...'. 463 'a%/n,b%/m,...' or 'a%-b%,c%-%d,...'.
464 464
465 For example: 465 For example:
466 Select the second 10% time slice: 466 Select the second 10% time slice:
467 467
468 perf report --time 10%/2 468 perf report --time 10%/2
469 469
470 Select from 0% to 10% time slice: 470 Select from 0% to 10% time slice:
471 471
472 perf report --time 0%-10% 472 perf report --time 0%-10%
473 473
474 Select the first and second 10% time s 474 Select the first and second 10% time slices:
475 475
476 perf report --time 10%/1,10%/2 476 perf report --time 10%/1,10%/2
477 477
478 Select from 0% to 10% and 30% to 40% s 478 Select from 0% to 10% and 30% to 40% slices:
479 479
480 perf report --time 0%-10%,30%-40% 480 perf report --time 0%-10%,30%-40%
481 481
482 --switch-on EVENT_NAME:: 482 --switch-on EVENT_NAME::
483 Only consider events after this event 483 Only consider events after this event is found.
484 484
485 This may be interesting to measure a w 485 This may be interesting to measure a workload only after some initialization
486 phase is over, i.e. insert a perf prob 486 phase is over, i.e. insert a perf probe at that point and then using this
487 option with that probe. 487 option with that probe.
488 488
489 --switch-off EVENT_NAME:: 489 --switch-off EVENT_NAME::
490 Stop considering events after this eve 490 Stop considering events after this event is found.
491 491
492 --show-on-off-events:: 492 --show-on-off-events::
493 Show the --switch-on/off events too. T 493 Show the --switch-on/off events too. This has no effect in 'perf report' now
494 but probably we'll make the default no 494 but probably we'll make the default not to show the switch-on/off events
495 on the --group mode and if there is on 495 on the --group mode and if there is only one event besides the off/on ones,
496 go straight to the histogram browser, 496 go straight to the histogram browser, just like 'perf report' with no events
497 explicitly specified does. 497 explicitly specified does.
498 498
499 --itrace:: 499 --itrace::
500 Options for decoding instruction traci 500 Options for decoding instruction tracing data. The options are:
501 501
502 include::itrace.txt[] 502 include::itrace.txt[]
503 503
504 To disable decoding entirely, use --no 504 To disable decoding entirely, use --no-itrace.
505 505
506 --full-source-path:: 506 --full-source-path::
507 Show the full path for source files fo 507 Show the full path for source files for srcline output.
508 508
509 --show-ref-call-graph:: 509 --show-ref-call-graph::
510 When multiple events are sampled, it m 510 When multiple events are sampled, it may not be needed to collect
511 callgraphs for all of them. The sample 511 callgraphs for all of them. The sample sites are usually nearby,
512 and it's enough to collect the callgra 512 and it's enough to collect the callgraphs on a reference event.
513 So user can use "call-graph=no" event 513 So user can use "call-graph=no" event modifier to disable callgraph
514 for other events to reduce the overhea 514 for other events to reduce the overhead.
515 However, perf report cannot show callg 515 However, perf report cannot show callgraphs for the event which
516 disable the callgraph. 516 disable the callgraph.
517 This option extends the perf report to 517 This option extends the perf report to show reference callgraphs,
518 which collected by reference event, in 518 which collected by reference event, in no callgraph event.
519 519
520 --stitch-lbr:: 520 --stitch-lbr::
521 Show callgraph with stitched LBRs, whi 521 Show callgraph with stitched LBRs, which may have more complete
522 callgraph. The perf.data file must hav 522 callgraph. The perf.data file must have been obtained using
523 perf record --call-graph lbr. 523 perf record --call-graph lbr.
524 Disabled by default. In common cases w 524 Disabled by default. In common cases with call stack overflows,
525 it can recreate better call stacks tha 525 it can recreate better call stacks than the default lbr call stack
526 output. But this approach is not foolp 526 output. But this approach is not foolproof. There can be cases
527 where it creates incorrect call stacks 527 where it creates incorrect call stacks from incorrect matches.
528 The known limitations include exceptio 528 The known limitations include exception handing such as
529 setjmp/longjmp will have calls/returns 529 setjmp/longjmp will have calls/returns not match.
530 530
531 --socket-filter:: 531 --socket-filter::
532 Only report the samples on the process 532 Only report the samples on the processor socket that match with this filter
533 533
534 --samples=N:: 534 --samples=N::
535 Save N individual samples for each his 535 Save N individual samples for each histogram entry to show context in perf
536 report tui browser. 536 report tui browser.
537 537
538 --raw-trace:: 538 --raw-trace::
539 When displaying traceevent output, do 539 When displaying traceevent output, do not use print fmt or plugins.
540 540
541 -H:: 541 -H::
542 --hierarchy:: 542 --hierarchy::
543 Enable hierarchical output. In the hi 543 Enable hierarchical output. In the hierarchy mode, each sort key groups
544 samples based on the criteria and then 544 samples based on the criteria and then sub-divide it using the lower
545 level sort key. 545 level sort key.
546 546
547 For example: 547 For example:
548 In normal output: 548 In normal output:
549 549
550 perf report -s dso,sym 550 perf report -s dso,sym
551 # Overhead Shared Object Symbo 551 # Overhead Shared Object Symbol
552 50.00% [kernel.kallsyms] [k] k 552 50.00% [kernel.kallsyms] [k] kfunc1
553 20.00% perf [.] f 553 20.00% perf [.] foo
554 15.00% [kernel.kallsyms] [k] k 554 15.00% [kernel.kallsyms] [k] kfunc2
555 10.00% perf [.] b 555 10.00% perf [.] bar
556 5.00% libc.so [.] l 556 5.00% libc.so [.] libcall
557 557
558 In hierarchy output: 558 In hierarchy output:
559 559
560 perf report -s dso,sym --hierarchy 560 perf report -s dso,sym --hierarchy
561 # Overhead Shared Object / Symbol 561 # Overhead Shared Object / Symbol
562 65.00% [kernel.kallsyms] 562 65.00% [kernel.kallsyms]
563 50.00% [k] kfunc1 563 50.00% [k] kfunc1
564 15.00% [k] kfunc2 564 15.00% [k] kfunc2
565 30.00% perf 565 30.00% perf
566 20.00% [.] foo 566 20.00% [.] foo
567 10.00% [.] bar 567 10.00% [.] bar
568 5.00% libc.so 568 5.00% libc.so
569 5.00% [.] libcall 569 5.00% [.] libcall
570 570
571 --inline:: 571 --inline::
572 If a callgraph address belongs to an i 572 If a callgraph address belongs to an inlined function, the inline stack
573 will be printed. Each entry is functio 573 will be printed. Each entry is function name or file/line. Enabled by
574 default, disable with --no-inline. 574 default, disable with --no-inline.
575 575
576 --mmaps:: 576 --mmaps::
577 Show --tasks output plus mmap informat 577 Show --tasks output plus mmap information in a format similar to
578 /proc/<PID>/maps. 578 /proc/<PID>/maps.
579 579
580 Please note that not all mmaps are sto 580 Please note that not all mmaps are stored, options affecting which ones
581 are include 'perf record --data', for 581 are include 'perf record --data', for instance.
582 582
583 --ns:: 583 --ns::
584 Show time stamps in nanoseconds. 584 Show time stamps in nanoseconds.
585 585
586 --stats:: 586 --stats::
587 Display overall events statistics with 587 Display overall events statistics without any further processing.
588 (like the one at the end of the perf r 588 (like the one at the end of the perf report -D command)
589 589
590 --tasks:: 590 --tasks::
591 Display monitored tasks stored in perf 591 Display monitored tasks stored in perf data. Displaying pid/tid/ppid
592 plus the command string aligned to dis 592 plus the command string aligned to distinguish parent and child tasks.
593 593
594 --percent-type:: 594 --percent-type::
595 Set annotation percent type from follo 595 Set annotation percent type from following choices:
596 global-period, local-period, global- 596 global-period, local-period, global-hits, local-hits
597 597
598 The local/global keywords set if the p 598 The local/global keywords set if the percentage is computed
599 in the scope of the function (local) o 599 in the scope of the function (local) or the whole data (global).
600 The period/hits keywords set the base 600 The period/hits keywords set the base the percentage is computed
601 on - the samples period or the number 601 on - the samples period or the number of samples (hits).
602 602
603 --time-quantum:: 603 --time-quantum::
604 Configure time quantum for time sort k 604 Configure time quantum for time sort key. Default 100ms.
605 Accepts s, us, ms, ns units. 605 Accepts s, us, ms, ns units.
606 606
607 --total-cycles:: 607 --total-cycles::
608 When --total-cycles is specified, it s 608 When --total-cycles is specified, it supports sorting for all blocks by
609 'Sampled Cycles%'. This is useful to c 609 'Sampled Cycles%'. This is useful to concentrate on the globally hottest
610 blocks. In output, there are some new 610 blocks. In output, there are some new columns:
611 611
612 'Sampled Cycles%' - block sampled cycl 612 'Sampled Cycles%' - block sampled cycles aggregation / total sampled cycles
613 'Sampled Cycles' - block sampled cycl 613 'Sampled Cycles' - block sampled cycles aggregation
614 'Avg Cycles%' - block average samp 614 'Avg Cycles%' - block average sampled cycles / sum of total block average
615 sampled cycles 615 sampled cycles
616 'Avg Cycles' - block average samp 616 'Avg Cycles' - block average sampled cycles
617 'Branch Counter' - block branch count 617 'Branch Counter' - block branch counter histogram (with -v showing the number)
618 618
619 --skip-empty:: 619 --skip-empty::
620 Do not print 0 results in the --stat o 620 Do not print 0 results in the --stat output.
621 621
622 include::callchain-overhead-calculation.txt[] 622 include::callchain-overhead-calculation.txt[]
623 623
624 SEE ALSO 624 SEE ALSO
625 -------- 625 --------
626 linkperf:perf-stat[1], linkperf:perf-annotate[ 626 linkperf:perf-stat[1], linkperf:perf-annotate[1], linkperf:perf-record[1],
627 linkperf:perf-intel-pt[1] 627 linkperf:perf-intel-pt[1]
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.