1PERF-INJECT(1) perf Manual PERF-INJECT(1)
2
6 perf-inject - Filter to augment the events stream with additional
7 information
8
10 perf inject <options>
11
13 perf-inject reads a perf-record event stream and repipes it to stdout.
14 At any point the processing code can inject other events into the event
15 stream - in this case build-ids (-b option) are read and injected as
16 needed into the event stream.
17 Build-ids are just the first user of perf-inject - potentially anything
19 that needs userspace processing to augment the events stream with
20 additional information could make use of this facility.
21
23 -b, --build-ids
24 Inject build-ids into the output stream
25 --buildid-all: Inject build-ids of all DSOs into the output stream
27 -v, --verbose
29 Be more verbose.
30 -i, --input=
32 Input file name. (default: stdin)
33 -o, --output=
35 Output file name. (default: stdout)
36 -s, --sched-stat
38 Merge sched_stat and sched_switch for getting events where and how
39 long tasks slept. sched_switch contains a callchain where a task
40 slept and sched_stat contains a timeslice how long a task slept.
41 --kallsyms=<file>
43 kallsyms pathname
44 --itrace
46 Decode Instruction Tracing data, replacing it with synthesized
47 events. Options are:
48 i synthesize instructions events
50 b synthesize branches events (branch misses for Arm SPE)
51 c synthesize branches events (calls only)
52 r synthesize branches events (returns only)
53 x synthesize transactions events
54 w synthesize ptwrite events
55 p synthesize power events (incl. PSB events for Intel PT)
56 o synthesize other events recorded due to the use
57 of aux-output (refer to perf record)
58 e synthesize error events
59 d create a debug log
60 f synthesize first level cache events
61 m synthesize last level cache events
62 M synthesize memory events
63 t synthesize TLB events
64 a synthesize remote access events
65 g synthesize a call chain (use with i or x)
66 G synthesize a call chain on existing event records
67 l synthesize last branch entries (use with i or x)
68 L synthesize last branch entries on existing event records
69 s skip initial number of events
70 q quicker (less detailed) decoding
71 Z prefer to ignore timestamps (so-called "timeless" decoding)
72 The default is all events i.e. the same as --itrace=ibxwpe,
74 except for perf script where it is --itrace=ce
75 In addition, the period (default 100000, except for perf script where it is 1)
77 for instructions events can be specified in units of:
78 i instructions
80 t ticks
81 ms milliseconds
82 us microseconds
83 ns nanoseconds (default)
84 Also the call chain size (default 16, max. 1024) for instructions or
86 transactions events can be specified.
87 Also the number of last branch entries (default 64, max. 1024) for
89 instructions or transactions events can be specified.
90 Similar to options g and l, size may also be specified for options G and L.
92 On x86, note that G and L work poorly when data has been recorded with
93 large PEBS. Refer linkperf:perf-intel-pt[1] man page for details.
94 It is also possible to skip events generated (instructions, branches, transactions,
96 ptwrite, power) at the beginning. This is useful to ignore initialization code.
97 --itrace=i0nss1000000
99 skips the first million instructions.
101 The 'e' option may be followed by flags which affect what errors will or
103 will not be reported. Each flag must be preceded by either '+' or '-'.
104 The flags are:
105 o overflow
106 l trace data lost
107 If supported, the 'd' option may be followed by flags which affect what
109 debug messages will or will not be logged. Each flag must be preceded
110 by either '+' or '-'. The flags are:
111 a all perf events
112 If supported, the 'q' option may be repeated to increase the effect.
114 --strip
116 Use with --itrace to strip out non-synthesized events.
117 -j, --jit
119 Process jitdump files by injecting the mmap records corresponding
120 to jitted functions. This option also generates the ELF images for
121 each jitted function found in the jitdumps files captured in the
122 input perf.data file. Use this option if you are monitoring
123 environment using JIT runtimes, such as Java, DART or V8.
124 -f, --force
126 Don’t complain, do it.
127 --vm-time-correlation[=OPTIONS]
129 Some architectures may capture AUX area data which contains
130 timestamps affected by virtualization. This option will update
131 those timestamps in place, to correlate with host timestamps. The
132 in-place update means that an output file is not specified, and
133 instead the input file is modified. The options are architecture
134 specific, except that they may start with "dry-run" which will
135 cause the file to be processed but without updating it. Currently
136 this option is supported only by Intel PT, refer perf-intel-pt(1)
137
139 perf-record(1), perf-report(1), perf-archive(1), perf-intel-pt(1)
140perf 11/22/2021 PERF-INJECT(1)