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)
40 Output file name for perf lock contention and report.
44 Be more verbose (show symbol address, etc).
48 Do not show any warnings or messages. (Suppress -v)
52 Dump raw trace in ASCII.
56 Don't complain, do it.
70 Sorting key. Possible values: acquired (default), contended,
71 avg_wait, wait_total, wait_max, wait_min.
75 Output fields. By default it shows all the fields but users can
76 customize that using this. Possible values: acquired, contended,
77 avg_wait, wait_total, wait_max, wait_min.
81 Merge lock instances in the same class (based on name).
85 The -t option is to show per-thread lock stat like below:
87 $ perf lock report -t -F acquired,contended,avg_wait
89 Name acquired contended avg wait (ns)
94 ContainerMgr 8981 6 874
96 ContainerThread 4416 4 944
97 RootPressureThr 3215 5 1215
101 EventManager_De 1845 1 636
102 futex-default-S 1609 0 0
106 Display this many entries.
114 dump only the thread list in perf.data
118 dump only the map of lock instances (address:name table)
126 Sorting key. Possible values: contended, wait_total (default),
127 wait_max, wait_min, avg_wait.
131 Output fields. By default it shows all but the wait_min fields
132 and users can customize that using this. Possible values:
133 contended, wait_total, wait_max, wait_min, avg_wait.
137 Show per-thread lock contention stat
141 Use BPF program to collect lock contention stats instead of
142 using the input data.
146 System-wide collection from all CPUs.
150 Collect samples only on the list of CPUs provided. Multiple CPUs can be
151 provided as a comma-separated list with no space: 0,1. Ranges of CPUs
152 are specified with -: 0-2. Default is to monitor all CPUs.
156 Record events on existing process ID (comma separated list).
159 Record events on existing thread ID (comma separated list).
162 --map-nr-entries=<value>::
163 Maximum number of BPF map entries (default: 16384).
164 This will be aligned to a power of 2.
166 --max-stack=<value>::
167 Maximum stack depth when collecting lock contention (default: 8).
169 --stack-skip=<value>::
170 Number of stack depth to skip when finding a lock caller (default: 3).
174 Display this many entries.
178 Show lock contention stat by address
182 Show lock contention stat by owners. Implies --threads and
186 --type-filter=<value>::
187 Show lock contention only for given lock types (comma separated list).
188 Available values are:
189 semaphore, spinlock, rwlock, rwlock:R, rwlock:W, rwsem, rwsem:R, rwsem:W,
190 rtmutex, rwlock-rt, rwlock-rt:R, rwlock-rt:W, pcpu-sem, pcpu-sem:R, pcpu-sem:W,
193 Note that RW-variant of locks have :R and :W suffix. Names without the
194 suffix are shortcuts for the both variants. Ex) rwsem = rwsem:R + rwsem:W.
197 --lock-filter=<value>::
198 Show lock contention only for given lock addresses or names (comma separated list).
201 --callstack-filter=<value>::
202 Show lock contention only if the callstack contains the given string.
203 Note that it matches the substring so 'rq' would match both 'raw_spin_rq_lock'
207 --field-separator=<SEP>::
208 Show results using a CSV-style output to make it easy to import directly
209 into spreadsheets. Columns are separated by the string specified in SEP.
212 Show lock contention stat by cgroup. Requires --use-bpf.
215 --cgroup-filter=<value>::
216 Show lock contention only in the given cgroups (comma separated list).