brintos

brintos / linux-shallow public Read only

0
0
Text · 13.2 KiB · 6e0cc50 Raw
348 lines · plain
1perf-trace(1)2=============3 4NAME5----6perf-trace - strace inspired tool7 8SYNOPSIS9--------10[verse]11'perf trace'12'perf trace record'13 14DESCRIPTION15-----------16This command will show the events associated with the target, initially17syscalls, but other system events like pagefaults, task lifetime events,18scheduling events, etc.19 20This is a live mode tool in addition to working with perf.data files like21the other perf tools. Files can be generated using the 'perf record' command22but the session needs to include the raw_syscalls events (-e 'raw_syscalls:*').23Alternatively, 'perf trace record' can be used as a shortcut to24automatically include the raw_syscalls events when writing events to a file.25 26The following options apply to perf trace; options to perf trace record are27found in the perf record man page.28 29OPTIONS30-------31 32-a::33--all-cpus::34        System-wide collection from all CPUs.35 36-e::37--expr::38--event::39	List of syscalls and other perf events (tracepoints, HW cache events,40	etc) to show. Globbing is supported, e.g.: "epoll_*", "*msg*", etc.41	See 'perf list' for a complete list of events.42	Prefixing with ! shows all syscalls but the ones specified.  You may43	need to escape it.44 45--filter=<filter>::46        Event filter. This option should follow an event selector (-e) which47	selects tracepoint event(s).48 49 50-D msecs::51--delay msecs::52After starting the program, wait msecs before measuring. This is useful to53filter out the startup phase of the program, which is often very different.54 55-o::56--output=::57	Output file name.58 59-p::60--pid=::61	Record events on existing process ID (comma separated list).62 63-t::64--tid=::65        Record events on existing thread ID (comma separated list).66 67-u::68--uid=::69        Record events in threads owned by uid. Name or number.70 71-G::72--cgroup::73	Record events in threads in a cgroup.74 75	Look for cgroups to set at the /sys/fs/cgroup/perf_event directory, then76	remove the /sys/fs/cgroup/perf_event/ part and try:77 78		perf trace -G A -e sched:*switch79 80	Will set all raw_syscalls:sys_{enter,exit}, pgfault, vfs_getname, etc81	_and_ sched:sched_switch to the 'A' cgroup, while:82 83		perf trace -e sched:*switch -G A84 85	will only set the sched:sched_switch event to the 'A' cgroup, all the86	other events (raw_syscalls:sys_{enter,exit}, etc are left "without"87	a cgroup (on the root cgroup, sys wide, etc).88 89	Multiple cgroups:90 91		perf trace -G A -e sched:*switch -G B92 93	the syscall ones go to the 'A' cgroup, the sched:sched_switch goes94	to the 'B' cgroup.95 96--filter-pids=::97	Filter out events for these pids and for 'trace' itself (comma separated list).98 99-v::100--verbose::101        Increase the verbosity level.102 103--no-inherit::104	Child tasks do not inherit counters.105 106-m::107--mmap-pages=::108	Number of mmap data pages (must be a power of two) or size109	specification in bytes with appended unit character - B/K/M/G.110	The size is rounded up to the nearest power-of-two page value.111 112-C::113--cpu::114Collect samples only on the list of CPUs provided. Multiple CPUs can be provided as a115comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2.116In per-thread mode with inheritance mode on (default), Events are captured only when117the thread executes on the designated CPUs. Default is to monitor all CPUs.118 119--duration::120	Show only events that had a duration greater than N.M ms.121 122--sched::123	Accrue thread runtime and provide a summary at the end of the session.124 125--failure::126	Show only syscalls that failed, i.e. that returned < 0.127 128-i::129--input::130	Process events from a given perf data file.131 132-T::133--time::134	Print full timestamp rather time relative to first sample.135 136--comm::137        Show process COMM right beside its ID, on by default, disable with --no-comm.138 139-s::140--summary::141	Show only a summary of syscalls by thread with min, max, and average times142    (in msec) and relative stddev.143 144-S::145--with-summary::146	Show all syscalls followed by a summary by thread with min, max, and147    average times (in msec) and relative stddev.148 149--errno-summary::150	To be used with -s or -S, to show stats for the errnos experienced by151	syscalls, using only this option will trigger --summary.152 153--tool_stats::154	Show tool stats such as number of times fd->pathname was discovered thru155	hooking the open syscall return + vfs_getname or via reading /proc/pid/fd, etc.156 157-f::158--force::159	Don't complain, do it.160 161-F=[all|min|maj]::162--pf=[all|min|maj]::163	Trace pagefaults. Optionally, you can specify whether you want minor,164	major or all pagefaults. Default value is maj.165 166--syscalls::167	Trace system calls. This options is enabled by default, disable with168	--no-syscalls.169 170--call-graph [mode,type,min[,limit],order[,key][,branch]]::171        Setup and enable call-graph (stack chain/backtrace) recording.172        See `--call-graph` section in perf-record and perf-report173        man pages for details. The ones that are most useful in 'perf trace'174        are 'dwarf' and 'lbr', where available, try: 'perf trace --call-graph dwarf'.175 176        Using this will, for the root user, bump the value of --mmap-pages to 4177        times the maximum for non-root users, based on the kernel.perf_event_mlock_kb178        sysctl. This is done only if the user doesn't specify a --mmap-pages value.179 180--kernel-syscall-graph::181	 Show the kernel callchains on the syscall exit path.182 183--max-events=N::184	Stop after processing N events. Note that strace-like events are considered185	only at exit time or when a syscall is interrupted, i.e. in those cases this186	option is equivalent to the number of lines printed.187 188--switch-on EVENT_NAME::189	Only consider events after this event is found.190 191--switch-off EVENT_NAME::192	Stop considering events after this event is found.193 194--show-on-off-events::195	Show the --switch-on/off events too.196 197--max-stack::198        Set the stack depth limit when parsing the callchain, anything199        beyond the specified depth will be ignored. Note that at this point200        this is just about the presentation part, i.e. the kernel is still201        not limiting, the overhead of callchains needs to be set via the202        knobs in --call-graph dwarf.203 204        Implies '--call-graph dwarf' when --call-graph not present on the205        command line, on systems where DWARF unwinding was built in.206 207        Default: /proc/sys/kernel/perf_event_max_stack when present for208                 live sessions (without --input/-i), 127 otherwise.209 210--min-stack::211        Set the stack depth limit when parsing the callchain, anything212        below the specified depth will be ignored. Disabled by default.213 214        Implies '--call-graph dwarf' when --call-graph not present on the215        command line, on systems where DWARF unwinding was built in.216 217--print-sample::218	Print the PERF_RECORD_SAMPLE PERF_SAMPLE_ info for the219	raw_syscalls:sys_{enter,exit} tracepoints, for debugging.220 221--proc-map-timeout::222	When processing pre-existing threads /proc/XXX/mmap, it may take a long time,223	because the file may be huge. A time out is needed in such cases.224	This option sets the time out limit. The default value is 500 ms.225 226--sort-events::227	Do sorting on batches of events, use when noticing out of order events that228	may happen, for instance, when a thread gets migrated to a different CPU229	while processing a syscall.230 231--libtraceevent_print::232	Use libtraceevent to print tracepoint arguments. By default 'perf trace' uses233	the same beautifiers used in the strace-like enter+exit lines to augment the234	tracepoint arguments.235 236--map-dump::237	Dump BPF maps setup by events passed via -e, for instance the augmented_raw_syscalls238	living in tools/perf/examples/bpf/augmented_raw_syscalls.c. For now this239	dumps just boolean map values and integer keys, in time this will print in hex240	by default and use BTF when available, as well as use functions to do pretty241	printing using the existing 'perf trace' syscall arg beautifiers to map integer242	arguments to strings (pid to comm, syscall id to syscall name, etc).243 244 245PAGEFAULTS246----------247 248When tracing pagefaults, the format of the trace is as follows:249 250<min|maj>fault [<ip.symbol>+<ip.offset>] => <addr.dso@addr.offset> (<map type><addr level>).251 252- min/maj indicates whether fault event is minor or major;253- ip.symbol shows symbol for instruction pointer (the code that generated the254  fault); if no debug symbols available, perf trace will print raw IP;255- addr.dso shows DSO for the faulted address;256- map type is either 'd' for non-executable maps or 'x' for executable maps;257- addr level is either 'k' for kernel dso or '.' for user dso.258 259For symbols resolution you may need to install debugging symbols.260 261Please be aware that duration is currently always 0 and doesn't reflect actual262time it took for fault to be handled!263 264When --verbose specified, perf trace tries to print all available information265for both IP and fault address in the form of dso@symbol+offset.266 267EXAMPLES268--------269 270Trace only major pagefaults:271 272 $ perf trace --no-syscalls -F273 274Trace syscalls, major and minor pagefaults:275 276 $ perf trace -F all277 278  1416.547 ( 0.000 ms): python/20235 majfault [CRYPTO_push_info_+0x0] => /lib/x86_64-linux-gnu/libcrypto.so.1.0.0@0x61be0 (x.)279 280  As you can see, there was major pagefault in python process, from281  CRYPTO_push_info_ routine which faulted somewhere in libcrypto.so.282 283Trace the first 4 open, openat or open_by_handle_at syscalls (in the future more syscalls may match here):284 285  $ perf trace -e open* --max-events 4286  [root@jouet perf]# trace -e open* --max-events 4287  2272.992 ( 0.037 ms): gnome-shell/1370 openat(dfd: CWD, filename: /proc/self/stat) = 31288  2277.481 ( 0.139 ms): gnome-shell/3039 openat(dfd: CWD, filename: /proc/self/stat) = 65289  3026.398 ( 0.076 ms): gnome-shell/3039 openat(dfd: CWD, filename: /proc/self/stat) = 65290  4294.665 ( 0.015 ms): sed/15879 openat(dfd: CWD, filename: /etc/ld.so.cache, flags: CLOEXEC) = 3291  $292 293Trace the first minor page fault when running a workload:294 295  # perf trace -F min --max-stack=7 --max-events 1 sleep 1296     0.000 ( 0.000 ms): sleep/18006 minfault [__clear_user+0x1a] => 0x5626efa56080 (?k)297                                       __clear_user ([kernel.kallsyms])298                                       load_elf_binary ([kernel.kallsyms])299                                       search_binary_handler ([kernel.kallsyms])300                                       __do_execve_file.isra.33 ([kernel.kallsyms])301                                       __x64_sys_execve ([kernel.kallsyms])302                                       do_syscall_64 ([kernel.kallsyms])303                                       entry_SYSCALL_64 ([kernel.kallsyms])304  #305 306Trace the next min page page fault to take place on the first CPU:307 308  # perf trace -F min --call-graph=dwarf --max-events 1 --cpu 0309     0.000 ( 0.000 ms): Web Content/17136 minfault [js::gc::Chunk::fetchNextDecommittedArena+0x4b] => 0x7fbe6181b000 (?.)310                                       js::gc::FreeSpan::initAsEmpty (inlined)311                                       js::gc::Arena::setAsNotAllocated (inlined)312                                       js::gc::Chunk::fetchNextDecommittedArena (/usr/lib64/firefox/libxul.so)313                                       js::gc::Chunk::allocateArena (/usr/lib64/firefox/libxul.so)314                                       js::gc::GCRuntime::allocateArena (/usr/lib64/firefox/libxul.so)315                                       js::gc::ArenaLists::allocateFromArena (/usr/lib64/firefox/libxul.so)316                                       js::gc::GCRuntime::tryNewTenuredThing<JSString, (js::AllowGC)1> (inlined)317                                       js::AllocateString<JSString, (js::AllowGC)1> (/usr/lib64/firefox/libxul.so)318                                       js::Allocate<JSThinInlineString, (js::AllowGC)1> (inlined)319                                       JSThinInlineString::new_<(js::AllowGC)1> (inlined)320                                       AllocateInlineString<(js::AllowGC)1, unsigned char> (inlined)321                                       js::ConcatStrings<(js::AllowGC)1> (/usr/lib64/firefox/libxul.so)322                                       [0x18b26e6bc2bd] (/tmp/perf-17136.map)323  #324 325Trace the next two sched:sched_switch events, four block:*_plug events, the326next block:*_unplug and the next three net:*dev_queue events, this last one327with a backtrace of at most 16 entries, system wide:328 329  # perf trace -e sched:*switch/nr=2/,block:*_plug/nr=4/,block:*_unplug/nr=1/,net:*dev_queue/nr=3,max-stack=16/330     0.000 :0/0 sched:sched_switch:swapper/2:0 [120] S ==> rcu_sched:10 [120]331     0.015 rcu_sched/10 sched:sched_switch:rcu_sched:10 [120] R ==> swapper/2:0 [120]332   254.198 irq/50-iwlwifi/680 net:net_dev_queue:dev=wlp3s0 skbaddr=0xffff93498051f600 len=66333                                       __dev_queue_xmit ([kernel.kallsyms])334   273.977 :0/0 net:net_dev_queue:dev=wlp3s0 skbaddr=0xffff93498051f600 len=78335                                       __dev_queue_xmit ([kernel.kallsyms])336   274.007 :0/0 net:net_dev_queue:dev=wlp3s0 skbaddr=0xffff93498051ff00 len=78337                                       __dev_queue_xmit ([kernel.kallsyms])338  2930.140 kworker/u16:58/2722 block:block_plug:[kworker/u16:58]339  2930.162 kworker/u16:58/2722 block:block_unplug:[kworker/u16:58] 1340  4466.094 jbd2/dm-2-8/748 block:block_plug:[jbd2/dm-2-8]341  8050.123 kworker/u16:30/2694 block:block_plug:[kworker/u16:30]342  8050.271 kworker/u16:30/2694 block:block_plug:[kworker/u16:30]343  #344 345SEE ALSO346--------347linkperf:perf-record[1], linkperf:perf-script[1]348