1 perf-lock(1) 1 perf-lock(1)
2 ============ 2 ============
3 3
4 NAME 4 NAME
5 ---- 5 ----
6 perf-lock - Analyze lock events 6 perf-lock - Analyze lock events
7 7
8 SYNOPSIS 8 SYNOPSIS
9 -------- 9 --------
10 [verse] 10 [verse]
11 'perf lock' {record|report|script|info|content 11 'perf lock' {record|report|script|info|contention}
12 12
13 DESCRIPTION 13 DESCRIPTION
14 ----------- 14 -----------
15 You can analyze various lock behaviours 15 You can analyze various lock behaviours
16 and statistics with this 'perf lock' command. 16 and statistics with this 'perf lock' command.
17 17
18 'perf lock record <command>' records lock ev 18 'perf lock record <command>' records lock events
19 between start and end <command>. And this co 19 between start and end <command>. And this command
20 produces the file "perf.data" which contains 20 produces the file "perf.data" which contains tracing
21 results of lock events. 21 results of lock events.
22 22
23 'perf lock report' reports statistical data. 23 'perf lock report' reports statistical data.
24 24
25 'perf lock script' shows raw lock events. 25 'perf lock script' shows raw lock events.
26 26
27 'perf lock info' shows metadata like threads 27 'perf lock info' shows metadata like threads or addresses
28 of lock instances. 28 of lock instances.
29 29
30 'perf lock contention' shows contention stat 30 'perf lock contention' shows contention statistics.
31 31
32 COMMON OPTIONS 32 COMMON OPTIONS
33 -------------- 33 --------------
34 34
35 -i:: 35 -i::
36 --input=<file>:: 36 --input=<file>::
37 Input file name. (default: perf.data u 37 Input file name. (default: perf.data unless stdin is a fifo)
38 38
39 --output=<file>:: <<
40 Output file name for perf lock content <<
41 <<
42 -v:: 39 -v::
43 --verbose:: 40 --verbose::
44 Be more verbose (show symbol address, 41 Be more verbose (show symbol address, etc).
45 42
46 -q:: <<
47 --quiet:: <<
48 Do not show any warnings or messages. <<
49 <<
50 -D:: 43 -D::
51 --dump-raw-trace:: 44 --dump-raw-trace::
52 Dump raw trace in ASCII. 45 Dump raw trace in ASCII.
53 46
54 -f:: 47 -f::
55 --force:: 48 --force::
56 Don't complain, do it. 49 Don't complain, do it.
57 50
58 --vmlinux=<file>:: 51 --vmlinux=<file>::
59 vmlinux pathname 52 vmlinux pathname
60 53
61 --kallsyms=<file>:: 54 --kallsyms=<file>::
62 kallsyms pathname 55 kallsyms pathname
63 56
64 57
65 REPORT OPTIONS 58 REPORT OPTIONS
66 -------------- 59 --------------
67 60
68 -k:: 61 -k::
69 --key=<value>:: 62 --key=<value>::
70 Sorting key. Possible values: acquired 63 Sorting key. Possible values: acquired (default), contended,
71 avg_wait, wait_total, wait_max, wait_m 64 avg_wait, wait_total, wait_max, wait_min.
72 65
73 -F:: 66 -F::
74 --field=<value>:: 67 --field=<value>::
75 Output fields. By default it shows all 68 Output fields. By default it shows all the fields but users can
76 customize that using this. Possible v 69 customize that using this. Possible values: acquired, contended,
77 avg_wait, wait_total, wait_max, wait_m 70 avg_wait, wait_total, wait_max, wait_min.
78 71
79 -c:: 72 -c::
80 --combine-locks:: 73 --combine-locks::
81 Merge lock instances in the same class 74 Merge lock instances in the same class (based on name).
82 75
83 -t:: 76 -t::
84 --threads:: 77 --threads::
85 The -t option is to show per-thread lock s 78 The -t option is to show per-thread lock stat like below:
86 79
87 $ perf lock report -t -F acquired,conten 80 $ perf lock report -t -F acquired,contended,avg_wait
88 81
89 Name acquired contended 82 Name acquired contended avg wait (ns)
90 83
91 perf 240569 9 84 perf 240569 9 5784
92 swapper 106610 19 85 swapper 106610 19 543
93 :15789 17370 2 86 :15789 17370 2 14538
94 ContainerMgr 8981 6 87 ContainerMgr 8981 6 874
95 sleep 5275 1 88 sleep 5275 1 11281
96 ContainerThread 4416 4 89 ContainerThread 4416 4 944
97 RootPressureThr 3215 5 90 RootPressureThr 3215 5 1215
98 rcu_preempt 2954 0 91 rcu_preempt 2954 0 0
99 ContainerMgr 2560 0 92 ContainerMgr 2560 0 0
100 unnamed 1873 0 93 unnamed 1873 0 0
101 EventManager_De 1845 1 94 EventManager_De 1845 1 636
102 futex-default-S 1609 0 95 futex-default-S 1609 0 0
103 96
104 -E:: <<
105 --entries=<value>:: <<
106 Display this many entries. <<
107 <<
108 <<
109 INFO OPTIONS 97 INFO OPTIONS
110 ------------ 98 ------------
111 99
112 -t:: 100 -t::
113 --threads:: 101 --threads::
114 dump only the thread list in perf.data !! 102 dump thread list in perf.data
115 103
116 -m:: 104 -m::
117 --map:: 105 --map::
118 dump only the map of lock instances (a !! 106 dump map of lock instances (address:name table)
119 <<
120 107
121 CONTENTION OPTIONS 108 CONTENTION OPTIONS
122 ------------------ !! 109 --------------
123 110
124 -k:: 111 -k::
125 --key=<value>:: 112 --key=<value>::
126 Sorting key. Possible values: contende 113 Sorting key. Possible values: contended, wait_total (default),
127 wait_max, wait_min, avg_wait. 114 wait_max, wait_min, avg_wait.
128 115
129 -F:: 116 -F::
130 --field=<value>:: 117 --field=<value>::
131 Output fields. By default it shows all 118 Output fields. By default it shows all but the wait_min fields
132 and users can customize that using thi 119 and users can customize that using this. Possible values:
133 contended, wait_total, wait_max, wait_ 120 contended, wait_total, wait_max, wait_min, avg_wait.
134 121
135 -t:: 122 -t::
136 --threads:: 123 --threads::
137 Show per-thread lock contention stat 124 Show per-thread lock contention stat
138 125
139 -b:: 126 -b::
140 --use-bpf:: 127 --use-bpf::
141 Use BPF program to collect lock conten 128 Use BPF program to collect lock contention stats instead of
142 using the input data. 129 using the input data.
143 130
144 -a:: 131 -a::
145 --all-cpus:: 132 --all-cpus::
146 System-wide collection from all CPUs. 133 System-wide collection from all CPUs.
147 134
148 -C:: 135 -C::
149 --cpu=<value>:: !! 136 --cpu::
150 Collect samples only on the list of CP 137 Collect samples only on the list of CPUs provided. Multiple CPUs can be
151 provided as a comma-separated list wit 138 provided as a comma-separated list with no space: 0,1. Ranges of CPUs
152 are specified with -: 0-2. Default is 139 are specified with -: 0-2. Default is to monitor all CPUs.
153 140
154 -p:: 141 -p::
155 --pid=<value>:: !! 142 --pid=::
156 Record events on existing process ID ( 143 Record events on existing process ID (comma separated list).
157 144
158 --tid=<value>:: !! 145 --tid=::
159 Record events on existing thread ID (c 146 Record events on existing thread ID (comma separated list).
160 147
161 -M:: !! 148 --map-nr-entries::
162 --map-nr-entries=<value>:: !! 149 Maximum number of BPF map entries (default: 10240).
163 Maximum number of BPF map entries (def <<
164 This will be aligned to a power of 2. <<
165 <<
166 --max-stack=<value>:: <<
167 Maximum stack depth when collecting lo <<
168 <<
169 --stack-skip=<value>:: <<
170 Number of stack depth to skip when fin <<
171 <<
172 -E:: <<
173 --entries=<value>:: <<
174 Display this many entries. <<
175 <<
176 -l:: <<
177 --lock-addr:: <<
178 Show lock contention stat by address <<
179 <<
180 -o:: <<
181 --lock-owner:: <<
182 Show lock contention stat by owners. <<
183 requires --use-bpf. <<
184 <<
185 -Y:: <<
186 --type-filter=<value>:: <<
187 Show lock contention only for given lo <<
188 Available values are: <<
189 semaphore, spinlock, rwlock, rwlock: <<
190 rtmutex, rwlock-rt, rwlock-rt:R, rwl <<
191 mutex <<
192 <<
193 Note that RW-variant of locks have :R <<
194 suffix are shortcuts for the both vari <<
195 <<
196 -L:: <<
197 --lock-filter=<value>:: <<
198 Show lock contention only for given lo <<
199 <<
200 -S:: <<
201 --callstack-filter=<value>:: <<
202 Show lock contention only if the calls <<
203 Note that it matches the substring so <<
204 and 'irq_enter_rcu'. <<
205 <<
206 -x:: <<
207 --field-separator=<SEP>:: <<
208 Show results using a CSV-style output <<
209 into spreadsheets. Columns are separat <<
210 <<
211 --lock-cgroup:: <<
212 Show lock contention stat by cgroup. <<
213 <<
214 -G:: <<
215 --cgroup-filter=<value>:: <<
216 Show lock contention only in the given <<
217 150
218 151
219 SEE ALSO 152 SEE ALSO
220 -------- 153 --------
221 linkperf:perf[1] 154 linkperf:perf[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.