PERF-TRACE(1)

📦
version 7.3.1 · 43 versions
, in arch rolling

okay bestie, we’re looking at perf-trace(1) from the perf Manual — and no cap, this is the strace-inspired glow-up you need in your life 💅✨

🎀 NAME 🎀

perf-trace — strace inspired tool 🦋

✨ SYNOPSIS ✨

perf trace
perf trace record

💖 DESCRIPTION 💖

okay so this command will show the events associated with the target — initially syscalls, but also other system events like pagefaults, task lifetime events, scheduling events, etc. literally the whole vibe check on your system 👑

This is a live mode tool in addition to working with perf.data files like the other perf tools. Files can be generated using the perf record command, but the session needs to include the raw_syscalls events (-e raw_syscalls:*). Alternatively, perf trace record can be used as a shortcut to automatically include the raw_syscalls events when writing events to a file. slay 🎀

The following options apply to perf trace; options to perf trace record are found in the perf record man page. iykyk 💅

🌸 OPTIONS 🌸

  • -a, --all-cpus — System-wide collection from all CPUs. main character energy fr fr 🫶
  • -e, --expr, --event — List of syscalls and other perf events (tracepoints, HW cache events, etc) to show. Globbing is supported, e.g.: "epoll_*", "msg", etc. See perf list for a complete list of events. Prefixing with ! shows all syscalls but the ones specified. You may need to escape it. no cap, that ! goes hard 😭
  • --filter=<filter> — Event filter. This option should follow an event selector (-e) which selects tracepoint event(s). lowkey important 💜
  • -D msecs, --delay msecs — After starting the program, wait msecs before measuring. This is useful to filter out the startup phase of the program, which is often very different. bye bye messy startup energy 🦋
  • -o, --output= — Output file name. put it somewhere cute 💖
  • -p, --pid= — Record events on existing process ID (comma separated list). bestie’s got a whole list? iconic 💅
  • -t, --tid= — Record events on existing thread ID (comma separated list). thread-specific tea ☕
  • -u, --uid= — Record events in threads owned by uid. Name or number. yes, name or number, flexible queen 🌈
  • -G, --cgroup — Record events in threads in a cgroup.

Look for cgroups to set at the /sys/fs/cgroup/perf_event directory, then remove the /sys/fs/cgroup/perf_event/ part and try:

perf trace -G A -e sched:*switch

Will set all raw_syscalls:sys_{enter,exit}, pgfault, vfs_getname, etc _and_ sched:sched_switch to the 'A' cgroup, while:

perf trace -e sched:*switch -G A

will only set the sched:sched_switch event to the 'A' cgroup, all the other events (raw_syscalls:sys_{enter,exit}, etc are left "without" a cgroup (on the root cgroup, sys wide, etc).

Multiple cgroups:

perf trace -G A -e sched:*switch -G B

the syscall ones go to the 'A' cgroup, the sched:sched_switch goes to the 'B' cgroup. cgroup assignments are giving organized bestie energy 🎀

  • --filter-pids= — Filter out events for these pids and for trace itself (comma separated list). not me filtering out the noise 🙈
  • -v, --verbose — Increase the verbosity level. more words, more lore 💀
  • --no-inherit — Child tasks do not inherit counters. boundaries, bestie 🫶
  • -m, --mmap-pages= — Number of mmap data pages (must be a power of two) or size specification in bytes with appended unit character - B/K/M/G. The size is rounded up to the nearest power-of-two page value. math girlies rise 🌸
  • -C, --cpu — Collect samples only on the list of CPUs provided. Multiple CPUs can be provided as a comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2. In per-thread mode with inheritance mode on (default), Events are captured only when the thread executes on the designated CPUs. Default is to monitor all CPUs. CPU roster, but make it fashion 💅
  • --duration — Show only events that had a duration greater than N.M ms. long events only, periodt ⏳
  • --sched — Accrue thread runtime and provide a summary at the end of the session. end-of-session recap, very wrap-up video energy 🎬
  • --failure — Show only syscalls that failed, i.e. that returned < 0. failure focus, but make it useful 😭
  • -i, --input — Process events from a given perf data file. reading old tea 🍵
  • -T, --time — Print full timestamp rather time relative to first sample. absolute time, no vibes-only timing ✨
  • --comm — Show process COMM right beside its ID, on by default, disable with --no-comm. name tag next to the ID 🏷️
  • -s, --summary — Show only a summary of syscalls by thread with min, max, and average times (in msec) and relative stddev. stats girl summer 📊
  • -S, --with-summary — Show all syscalls followed by a summary by thread with min, max, and average times (in msec) and relative stddev. all the events and the recap — ate 💅
  • --errno-summary — To be used with -s or -S, to show stats for the errnos experienced by syscalls, using only this option will trigger --summary. errno lore unlocked 🔓
  • --summary-mode=mode — To be used with -s or -S, to select how to show summary. By default it’ll show the syscall summary by thread. Possible values are: thread, total, cgroup. choose your summary aesthetic 💖
  • --tool_stats — Show tool stats such as number of times fd→pathname was discovered thru hooking the open syscall return + vfs_getname or via reading /proc/pid/fd, etc. behind-the-scenes receipts 🧾
  • -f, --force — Don’t complain, do it. no thoughts, just send it 🔥
  • -F=[all|min|maj], --pf=[all|min|maj] — Trace pagefaults. Optionally, you can specify whether you want minor, major or all pagefaults. Default value is maj. pagefaults but make it configurable 🦋
  • --syscalls — Trace system calls. This option is enabled by default, disable with --no-syscalls. syscalls are the default main characters 👑
  • --call-graph [mode,type,min[,limit],order[,key][,branch]] — Setup and enable call-graph (stack chain/backtrace) recording. See --call-graph section in perf-record and perf-report man pages for details. The ones that are most useful in perf trace are dwarf and lbr, where available, try: perf trace --call-graph dwarf. stack chain tea ☕
Using this will, for the root user, bump the value of --mmap-pages to 4 times the maximum for non-root users, based on the kernel.perf_event_mlock_kb sysctl. This is done only if the user doesn't specify a --mmap-pages value.
  • --kernel-syscall-graph — Show the kernel callchains on the syscall exit path. kernel-side backtrace moment 🧚
  • --max-events=N — Stop after processing N events. Note that strace-like events are considered only at exit time or when a syscall is interrupted, i.e. in those cases this option is equivalent to the number of lines printed. event limit, bestie 💀
  • --switch-on EVENT_NAME — Only consider events after this event is found. start recording after the chosen moment ✨
  • --switch-off EVENT_NAME — Stop considering events after this event is found. the cutoff 💅
  • --show-on-off-events — Show the --switch-on/off events too. don’t hide the switches 🎀
  • --show-cpu — Show cpu id. cpu name tag 🏷️
  • --max-stack — Set the stack depth limit when parsing the callchain, anything beyond the specified depth will be ignored. Note that at this point this is just about the presentation part, i.e. the kernel is still not limiting, the overhead of callchains needs to be set via the knobs in --call-graph dwarf.
Implies --call-graph dwarf when --call-graph not present on the command line, on systems where DWARF unwinding was built in.
Default: /proc/sys/kernel/perf_event_max_stack when present for live sessions (without --input/-i), 127 otherwise.
  • --min-stack — Set the stack depth limit when parsing the callchain, anything below the specified depth will be ignored. Disabled by default.
Implies --call-graph dwarf when --call-graph not present on the command line, on systems where DWARF unwinding was built in.
  • --print-sample — Print the PERF_RECORD_SAMPLE PERF_SAMPLE_ info for the raw_syscalls:sys_{enter,exit} tracepoints, for debugging. debug girlies, this one’s for you 🐛
  • --proc-map-timeout — When processing pre-existing threads /proc/XXX/mmap, it may take a long time, because the file may be huge. A time out is needed in such cases. This option sets the time out limit. The default value is 500 ms. timeout so we don’t wait forever 😭
  • --sort-events — Do sorting on batches of events, use when noticing out of order events that may happen, for instance, when a thread gets migrated to a different CPU while processing a syscall. sort the chaos 🌀
  • --libtraceevent_print — Use libtraceevent to print tracepoint arguments. By default perf trace uses the same beautifiers used in the strace-like enter+exit lines to augment the tracepoint arguments. library-based printing, fancy 💅
  • --force-btf — Use btf_dump to pretty print syscall argument data, instead of using hand-crafted pretty printers. This option is intended for testing BTF integration in perf trace. btf_dump-based pretty-printing serves as a fallback to hand-crafted pretty printers, as the latter can better pretty-print integer flags and struct pointers. BTF testing era 🧪
  • --bitmask-list — Show bitmasks as a human-readable, condensed list (e.g. "0,2-5,7") instead of the default hexadecimal representation. bitmask glow-up ⭐
  • --bpf-summary — Collect system call statistics in BPF. This is only for live mode and works well with -s/--summary option where no argument information is required. BPF girlies, rise 🌈
  • --max-summary=N — Maximum number of lines in the summary mode. Note that this applies to each entry (thread or cgroup). summary line limit, no endless scrolling 🙈

🦋 PAGEFAULTS 🦋

When tracing pagefaults, the format of the trace is as follows:

<min|maj>fault [<ip.symbol>+<ip.offset>] ⇒ <addr.dso@addr.offset[1]>
(<map type><addr level>).
  • min/maj indicates whether fault event is minor or major;
  • ip.symbol shows symbol for instruction pointer (the code that generated the fault); if no debug symbols available, perf trace will print raw IP;
  • addr.dso shows DSO for the faulted address;
  • map type is either d for non-executable maps or x for executable maps;
  • addr level is either k for kernel dso or . for user dso.

For symbols resolution you may need to install debugging symbols. lowkey essential for the full picture 🧚

Please be aware that duration is currently always 0 and doesn’t reflect actual time it took for fault to be handled! not me catching a fake duration 💀

When --verbose is specified, perf trace tries to print all available information for both IP and fault address in the form of dso@symbol[2]+offset. verbose mode really said “tell the whole story” 💖

✨ EXAMPLES ✨

Okay bestie, this is the part where it all clicks. Real commands, real output, real "ohhhh THAT's what it does" moments. Let's go 💖

🩺 Just the majors, m'am

Wanna see only major pagefaults? No syscall noise at all? Say less:

$ perf trace --no-syscalls -F

That's it. That's the whole vibe. The --no-syscalls keeps the syscall spam away and -F (pagefaults) brings the drama. 🎀

👀 Syscalls AND both kinds of pagefaults

Want the full package? Both major and minor pagefaults, plus syscalls:

$ perf trace -F all
1416.547 ( 0.000 ms): python/20235 majfault [CRYPTO_push_info_+0x0] => /lib/x86_64-linux-gnu/libcrypto.so.1.0.0@0x61be0 (x.)

So like, look at that line — there was a major pagefault in the python process, coming from the CRYPTO_push_info_ routine which faulted somewhere in libcrypto.so. Yeah. That's the tea. 🍵 The (x.) at the end is literally the kernel going "idk man" about the symbol info.

🚪 The first 4 opens, and nothing more

Trace the first 4 open, openat or open_by_handle_at syscalls — and yes, the manpage literally says (in the future more syscalls may match here), so don't come for me if that list grows 👑

$ perf trace -e open* --max-events 4
[root@jouet perf]# trace -e open* --max-events 4
2272.992 ( 0.037 ms): gnome-shell/1370 openat(dfd: CWD, filename: /proc/self/stat) = 31
2277.481 ( 0.139 ms): gnome-shell/3039 openat(dfd: CWD, filename: /proc/self/stat) = 65
3026.398 ( 0.076 ms): gnome-shell/3039 openat(dfd: CWD, filename: /proc/self/stat) = 65
4294.665 ( 0.015 ms): sed/15879 openat(dfd: CWD, filename: /etc/ld.so.cache, flags: CLOEXEC) = 3
$

Notice gnome-shell is out here opening /proc/self/stat over and over — main character energy fr fr. And then sed sneaks in with an /etc/ld.so.cache open with flags: CLOEXEC. Iconic. 💅

🐣 First minor pagefault while running a workload

Catch the very first minor page fault when you run something, with a lil stack trace (--max-stack=7), and stop after one event:

# perf trace -F min --max-stack=7 --max-events 1 sleep 1
   0.000 ( 0.000 ms): sleep/18006 minfault [__clear_user+0x1a] => 0x5626efa56080 (?k)
                                    __clear_user ([kernel.kallsyms])
                                    load_elf_binary ([kernel.kallsyms])
                                    search_binary_handler ([kernel.kallsyms])
                                    __do_execve_file.isra.33 ([kernel.kallsyms])
                                    __x64_sys_execve ([kernel.kallsyms])
                                    do_syscall_64 ([kernel.kallsyms])
                                    entry_SYSCALL_64 ([kernel.kallsyms])
#

Look at that stack 😭 it's literally the whole exec pipeline: __clear_user → load_elf_binary → search_binary_handler → __do_execve_file.isra.33 → __x64_sys_execve → do_syscall_64 → entry_SYSCALL_64, all from [kernel.kallsyms]. Chef's kiss. 🫶

🧚 Next min pagefault on CPU 0, with DWARF

Now with --call-graph=dwarf and pinned to --cpu 0:

# perf trace -F min --call-graph=dwarf --max-events 1 --cpu 0
   0.000 ( 0.000 ms): Web Content/17136 minfault [js::gc::Chunk::fetchNextDecommittedArena+0x4b] => 0x7fbe6181b000 (?.)
                                    js::gc::FreeSpan::initAsEmpty (inlined)
                                    js::gc::Arena::setAsNotAllocated (inlined)
                                    js::gc::Chunk::fetchNextDecommittedArena (/usr/lib64/firefox/libxul.so)
                                    js::gc::Chunk::allocateArena (/usr/lib64/firefox/libxul.so)
                                    js::gc::GCRuntime::allocateArena (/usr/lib64/firefox/libxul.so)
                                    js::gc::ArenaLists::allocateFromArena (/usr/lib64/firefox/libxul.so)
                                    js::gc::GCRuntime::tryNewTenuredThing<JSString, (js::AllowGC)1> (inlined)
                                    js::AllocateString<JSString, (js::AllowGC)1> (/usr/lib64/firefox/libxul.so)
                                    js::Allocate<JSThinInlineString, (js::AllowGC)1> (inlined)
                                    JSThinInlineString::new_<(js::AllowGC)1> (inlined)
                                    AllocateInlineString<(js::AllowGC)1, unsigned char> (inlined)
                                    js::ConcatStrings<(js::AllowGC)1> (/usr/lib64/firefox/libxul.so)
                                    [0x18b26e6bc2bd] (/tmp/perf-17136.map)
#

Firefox's Web Content process, deep in SpiderMonkey GC land, allocating a string arena. The (inlined) markers are DWARF doing its lil magic ✨ and that last frame [0x18b26e6bc2bd] (/tmp/perf-17136.map) is the JIT'd code being all mysterious about itself. So valid. 🌸

🎛️ The big one: mix and match events with per-event options

This is the flex example. Trace the next two sched:sched_switch events, four block:*_plug events, the next block:*_unplug and the next three net:*dev_queue events — that last one with a backtrace of at most 16 entries — system wide:

# perf trace -e sched:*switch/nr=2/,block:*_plug/nr=4/,block:*_unplug/nr=1/,net:*dev_queue/nr=3,max-stack=16/
   0.000 :0/0 sched:sched_switch:swapper/2:0 [120] S ==> rcu_sched:10 [120]
   0.015 rcu_sched/10 sched:sched_switch:rcu_sched:10 [120] R ==> swapper/2:0 [120]
 254.198 irq/50-iwlwifi/680 net:net_dev_queue:dev=wlp3s0 skbaddr=0xffff93498051f600 len=66
                                    __dev_queue_xmit ([kernel.kallsyms])
 273.977 :0/0 net:net_dev_queue:dev=wlp3s0 skbaddr=0xffff93498051f600 len=78
                                    __dev_queue_xmit ([kernel.kallsyms])
 274.007 :0/0 net:net_dev_queue:dev=wlp3s0 skbaddr=0xffff93498051ff00 len=78
                                    __dev_queue_xmit ([kernel.kallsyms])
2930.140 kworker/u16:58/2722 block:block_plug:[kworker/u16:58]
2930.162 kworker/u16:58/2722 block:block_unplug:[kworker/u16:58] 1
4466.094 jbd2/dm-2-8/748 block:block_plug:[jbd2/dm-2-8]
8050.123 kworker/u16:30/2694 block:block_plug:[kworker/u16:30]
8050.271 kworker/u16:30/2694 block:block_plug:[kworker/u16:30]
#

The way the /nr=N/ filter sits right there in the event spec, comma-separated, and net:*dev_queue gets its own max-stack=16 while everything else doesn't… yeah. That's the good stuff. 🔥 Count them: two switches ✅, three net_dev_queue ✅, three plugs shown before the unplug gives you the 1 at the end, and the unplug fires exactly once. The counts were respected. Periodt. 👑

🌈 SEE ALSO 🌈

Your homework, bestie — go read these next:

📮 NOTES 📮

These are the contact-y bits, kept exactly as they appear (yes, they look like email addresses, no I'm not making that up 🙈):

  1. addr.dso@addr.offset
mailto:addr.dso@addr.offset
  1. dso@symbol
mailto:dso@symbol

Okay that's the whole page! You made it to the end and I'm so proud of you 🥹💜 This whole thing lives in the perf source, dated 10/31/2026, and it's the PERF-TRACE(1) page — go trace something beautiful with it. Byeeee~ 🦋✨🎀

📦 where this page comes from