6 perf-lock - Analyze lock events
11 'perf lock' {record|report|script|info|contention}
15 You can analyze various lock behaviours
16 and statistics with this 'perf lock' command.
18 'perf lock record <command>' records lock events
19 between start and end <command>. And this command
20 produces the file "perf.data" which contains tracing
21 results of lock events.
23 'perf lock report' reports statistical data.
25 'perf lock script' shows raw lock events.
27 'perf lock info' shows metadata like threads or addresses
30 'perf lock contention' shows contention statistics.
37 Input file name. (default: perf.data unless stdin is a fifo)
41 Be more verbose (show symbol address, etc).
45 Do not show any warnings or messages. (Suppress -v)
49 Dump raw trace in ASCII.
53 Don't complain, do it.
67 Sorting key. Possible values: acquired (default), contended,
68 avg_wait, wait_total, wait_max, wait_min.
72 Output fields. By default it shows all the fields but users can
73 customize that using this. Possible values: acquired, contended,
74 avg_wait, wait_total, wait_max, wait_min.
78 Merge lock instances in the same class (based on name).
82 The -t option is to show per-thread lock stat like below:
84 $ perf lock report -t -F acquired,contended,avg_wait
86 Name acquired contended avg wait (ns)
91 ContainerMgr 8981 6 874
93 ContainerThread 4416 4 944
94 RootPressureThr 3215 5 1215
98 EventManager_De 1845 1 636
99 futex-default-S 1609 0 0
103 Display this many entries.
111 dump thread list in perf.data
115 dump map of lock instances (address:name table)
123 Sorting key. Possible values: contended, wait_total (default),
124 wait_max, wait_min, avg_wait.
128 Output fields. By default it shows all but the wait_min fields
129 and users can customize that using this. Possible values:
130 contended, wait_total, wait_max, wait_min, avg_wait.
134 Show per-thread lock contention stat
138 Use BPF program to collect lock contention stats instead of
139 using the input data.
143 System-wide collection from all CPUs.
147 Collect samples only on the list of CPUs provided. Multiple CPUs can be
148 provided as a comma-separated list with no space: 0,1. Ranges of CPUs
149 are specified with -: 0-2. Default is to monitor all CPUs.
153 Record events on existing process ID (comma separated list).
156 Record events on existing thread ID (comma separated list).
159 --map-nr-entries=<value>::
160 Maximum number of BPF map entries (default: 16384).
161 This will be aligned to a power of 2.
163 --max-stack=<value>::
164 Maximum stack depth when collecting lock contention (default: 8).
166 --stack-skip=<value>::
167 Number of stack depth to skip when finding a lock caller (default: 3).
171 Display this many entries.
175 Show lock contention stat by address
179 Show lock contention stat by owners. Implies --threads and
183 --type-filter=<value>::
184 Show lock contention only for given lock types (comma separated list).
185 Available values are:
186 semaphore, spinlock, rwlock, rwlock:R, rwlock:W, rwsem, rwsem:R, rwsem:W,
187 rtmutex, rwlock-rt, rwlock-rt:R, rwlock-rt:W, pcpu-sem, pcpu-sem:R, pcpu-sem:W,
190 Note that RW-variant of locks have :R and :W suffix. Names without the
191 suffix are shortcuts for the both variants. Ex) rwsem = rwsem:R + rwsem:W.
194 --lock-filter=<value>::
195 Show lock contention only for given lock addresses or names (comma separated list).
198 --callstack-filter=<value>::
199 Show lock contention only if the callstack contains the given string.
200 Note that it matches the substring so 'rq' would match both 'raw_spin_rq_lock'
204 --field-separator=<SEP>::
205 Show results using a CSV-style output to make it easy to import directly
206 into spreadsheets. Columns are separated by the string specified in SEP.