1perf-inject(1) 2============== 3 4NAME 5---- 6perf-inject - Filter to augment the events stream with additional information 7 8SYNOPSIS 9-------- 10[verse] 11'perf inject <options>' 12 13DESCRIPTION 14----------- 15perf-inject reads a perf-record event stream and repipes it to stdout. At any 16point the processing code can inject other events into the event stream - in 17this case build-ids (-b option) are read and injected as needed into the event 18stream. 19 20Build-ids are just the first user of perf-inject - potentially anything that 21needs userspace processing to augment the events stream with additional 22information could make use of this facility. 23 24OPTIONS 25------- 26-b:: 27--build-ids:: 28 Inject build-ids into the output stream 29 30--buildid-all:: 31 Inject build-ids of all DSOs into the output stream 32 33--known-build-ids=:: 34 Override build-ids to inject using these comma-separated pairs of 35 build-id and path. Understands file://filename to read these pairs 36 from a file, which can be generated with perf buildid-list. 37 38-v:: 39--verbose:: 40 Be more verbose. 41-i:: 42--input=:: 43 Input file name. (default: stdin) 44-o:: 45--output=:: 46 Output file name. (default: stdout) 47-s:: 48--sched-stat:: 49 Merge sched_stat and sched_switch for getting events where and how long 50 tasks slept. sched_switch contains a callchain where a task slept and 51 sched_stat contains a timeslice how long a task slept. 52 53-k:: 54--vmlinux=<file>:: 55 vmlinux pathname 56 57--ignore-vmlinux:: 58 Ignore vmlinux files. 59 60--kallsyms=<file>:: 61 kallsyms pathname 62 63--itrace:: 64 Decode Instruction Tracing data, replacing it with synthesized events. 65 Options are: 66 67include::itrace.txt[] 68 69--strip:: 70 Use with --itrace to strip out non-synthesized events. 71 72-j:: 73--jit:: 74 Process jitdump files by injecting the mmap records corresponding to jitted 75 functions. This option also generates the ELF images for each jitted function 76 found in the jitdumps files captured in the input perf.data file. Use this option 77 if you are monitoring environment using JIT runtimes, such as Java, DART or V8. 78 79-f:: 80--force:: 81 Don't complain, do it. 82 83--vm-time-correlation[=OPTIONS]:: 84 Some architectures may capture AUX area data which contains timestamps 85 affected by virtualization. This option will update those timestamps 86 in place, to correlate with host timestamps. The in-place update means 87 that an output file is not specified, and instead the input file is 88 modified. The options are architecture specific, except that they may 89 start with "dry-run" which will cause the file to be processed but 90 without updating it. Currently this option is supported only by 91 Intel PT, refer linkperf:perf-intel-pt[1] 92 93--guest-data=<path>,<pid>[,<time offset>[,<time scale>]]:: 94 Insert events from a perf.data file recorded in a virtual machine at 95 the same time as the input perf.data file was recorded on the host. 96 The Process ID (PID) of the QEMU hypervisor process must be provided, 97 and the time offset and time scale (multiplier) will likely be needed 98 to convert guest time stamps into host time stamps. For example, for 99 x86 the TSC Offset and Multiplier could be provided for a virtual machine 100 using Linux command line option no-kvmclock. 101 Currently only mmap, mmap2, comm, task, context_switch, ksymbol, 102 and text_poke events are inserted, as well as build ID information. 103 The QEMU option -name debug-threads=on is needed so that thread names 104 can be used to determine which thread is running which VCPU. Note 105 libvirt seems to use this by default. 106 When using perf record in the guest, option --sample-identifier 107 should be used, and also --buildid-all and --switch-events may be 108 useful. 109 110:GMEXAMPLECMD: inject 111:GMEXAMPLESUBCMD: 112include::guestmount.txt[] 113 114SEE ALSO 115-------- 116linkperf:perf-record[1], linkperf:perf-report[1], linkperf:perf-archive[1], 117linkperf:perf-intel-pt[1] 118