1 perf-kvm(1) 1 perf-kvm(1)
2 =========== 2 ===========
3 3
4 NAME 4 NAME
5 ---- 5 ----
6 perf-kvm - Tool to trace/measure kvm guest os 6 perf-kvm - Tool to trace/measure kvm guest os
7 7
8 SYNOPSIS 8 SYNOPSIS
9 -------- 9 --------
10 [verse] 10 [verse]
11 'perf kvm' [--host] [--guest] [--guestmount=<p 11 'perf kvm' [--host] [--guest] [--guestmount=<path>
12 [--guestkallsyms=<path> --guestmodules 12 [--guestkallsyms=<path> --guestmodules=<path> | --guestvmlinux=<path>]]
13 {top|record|report|diff|buildid-list} !! 13 {top|record|report|diff|buildid-list}
14 'perf kvm' [--host] [--guest] [--guestkallsyms 14 'perf kvm' [--host] [--guest] [--guestkallsyms=<path> --guestmodules=<path>
15 | --guestvmlinux=<path>] {top|record|r !! 15 | --guestvmlinux=<path>] {top|record|report|diff|buildid-list|stat}
16 'perf kvm stat [record|report|live] [<options> <<
17 16
18 DESCRIPTION 17 DESCRIPTION
19 ----------- 18 -----------
20 There are a couple of variants of perf kvm: 19 There are a couple of variants of perf kvm:
21 20
22 'perf kvm [options] top <command>' to genera 21 'perf kvm [options] top <command>' to generates and displays
23 a performance counter profile of guest os in 22 a performance counter profile of guest os in realtime
24 of an arbitrary workload. 23 of an arbitrary workload.
25 24
26 'perf kvm record <command>' to record the pe 25 'perf kvm record <command>' to record the performance counter profile
27 of an arbitrary workload and save it into a !! 26 of an arbitrary workload and save it into a perf data file. If both
28 default behavior of perf kvm as --guest, so !! 27 --host and --guest are input, the perf data file name is perf.data.kvm.
29 is input, the perf data file name is perf.da !! 28 If there is no --host but --guest, the file name is perf.data.guest.
30 the perf data file name is perf.data.kvm. If !! 29 If there is no --guest but --host, the file name is perf.data.host.
31 perf.data.host, please input --host --no-gue <<
32 following: <<
33 Default('') -> perf.data.guest <<
34 --host -> perf.data.kvm <<
35 --guest -> perf.data.guest <<
36 --host --guest -> perf.data.kvm <<
37 --host --no-guest -> perf.data.host <<
38 30
39 'perf kvm report' to display the performance 31 'perf kvm report' to display the performance counter profile information
40 recorded via perf kvm record. 32 recorded via perf kvm record.
41 33
42 'perf kvm diff' to displays the performance 34 'perf kvm diff' to displays the performance difference amongst two perf.data
43 files captured via perf record. 35 files captured via perf record.
44 36
45 'perf kvm buildid-list' to display the buil 37 'perf kvm buildid-list' to display the buildids found in a perf data file,
46 so that other tools can be used to fetch pac 38 so that other tools can be used to fetch packages with matching symbol tables
47 for use by perf report. As buildid is read f !! 39 for use by perf report.
48 if you want to list the buildid for guest, p <<
49 was captured with --guestmount in perf kvm r <<
50 40
51 'perf kvm stat <command>' to run a command a 41 'perf kvm stat <command>' to run a command and gather performance counter
52 statistics. 42 statistics.
53 Especially, perf 'kvm stat record/report' ge 43 Especially, perf 'kvm stat record/report' generates a statistical analysis
54 of KVM events. Currently, vmexit, mmio (x86 !! 44 of KVM events. Currently, vmexit, mmio and ioport events are supported.
55 events are supported. 'perf kvm stat record !! 45 'perf kvm stat record <command>' records kvm events and the events between
56 and the events between start and end <comman !! 46 start and end <command>.
57 And this command produces a file which conta 47 And this command produces a file which contains tracing results of kvm
58 events. 48 events.
59 49
60 'perf kvm stat report' reports statistical d 50 'perf kvm stat report' reports statistical data which includes events
61 handled sample, percent_sample, time, percen !! 51 handled time, samples, and so on.
62 <<
63 'perf kvm stat live' reports statistical dat <<
64 record + report but with statistical data up <<
65 rate). <<
66 52
67 OPTIONS 53 OPTIONS
68 ------- 54 -------
69 -i:: 55 -i::
70 --input=<path>:: !! 56 --input=::
71 Input file name, for the 'report', 'di !! 57 Input file name.
72 -o:: 58 -o::
73 --output=<path>:: !! 59 --output::
74 Output file name, for the 'record' sub !! 60 Output file name.
75 just redirect the output to a file whe !! 61 --host=::
76 --host:: <<
77 Collect host side performance profile. 62 Collect host side performance profile.
78 --guest:: !! 63 --guest=::
79 Collect guest side performance profile 64 Collect guest side performance profile.
80 !! 65 --guestmount=<path>::
81 :GMEXAMPLECMD: kvm --host --guest !! 66 Guest os root file system mount directory. Users mounts guest os
82 :GMEXAMPLESUBCMD: top !! 67 root directories under <path> by a specific filesystem access method,
83 include::guest-files.txt[] !! 68 typically, sshfs. For example, start 2 guest os. The one's pid is 8888
84 !! 69 and the other's is 9999.
85 --stdio:: Use the stdio interface. !! 70 #mkdir ~/guestmount; cd ~/guestmount
86 !! 71 #sshfs -o allow_other,direct_io -p 5551 localhost:/ 8888/
87 -v:: !! 72 #sshfs -o allow_other,direct_io -p 5552 localhost:/ 9999/
88 --verbose:: !! 73 #perf kvm --host --guest --guestmount=~/guestmount top
89 Be more verbose (show counter open err !! 74 --guestkallsyms=<path>::
>> 75 Guest os /proc/kallsyms file copy. 'perf' kvm' reads it to get guest
>> 76 kernel symbols. Users copy it out from guest os.
>> 77 --guestmodules=<path>::
>> 78 Guest os /proc/modules file copy. 'perf' kvm' reads it to get guest
>> 79 kernel module information. Users copy it out from guest os.
>> 80 --guestvmlinux=<path>::
>> 81 Guest os kernel vmlinux.
90 82
91 STAT REPORT OPTIONS 83 STAT REPORT OPTIONS
92 ------------------- 84 -------------------
93 --vcpu=<value>:: 85 --vcpu=<value>::
94 analyze events which occur on this vcpu !! 86 analyze events which occures on this vcpu. (default: all vcpus)
95 <<
96 --event=<value>:: <<
97 event to be analyzed. Possible values: <<
98 ioport (x86 only). (default: vmexit) <<
99 -k:: <<
100 --key=<value>:: <<
101 Sorting key. Possible values: sample (d <<
102 number), percent_sample (sort by sample <<
103 (sort by average time), precent_time (s <<
104 max_t (sort by maximum time), min_t (so <<
105 (sort by mean time). <<
106 -p:: <<
107 --pid=:: <<
108 Analyze events only for given process ID(s <<
109 <<
110 STAT LIVE OPTIONS <<
111 ----------------- <<
112 -d:: <<
113 --display:: <<
114 Time in seconds between display update <<
115 <<
116 -m:: <<
117 --mmap-pages=:: <<
118 Number of mmap data pages (must be a p <<
119 specification in bytes with appended u <<
120 The size is rounded up to the nearest <<
121 <<
122 -a:: <<
123 --all-cpus:: <<
124 System-wide collection from all CPUs. <<
125 <<
126 -p:: <<
127 --pid=:: <<
128 Analyze events only for given process ID(s <<
129 87
130 --vcpu=<value>:: !! 88 --events=<value>::
131 analyze events which occur on this vcpu !! 89 events to be analyzed. Possible values: vmexit, mmio, ioport.
132 <<
133 <<
134 --event=<value>:: <<
135 event to be analyzed. Possible values: <<
136 mmio (x86 only), ioport (x86 only). <<
137 (default: vmexit) 90 (default: vmexit)
138 <<
139 -k:: 91 -k::
140 --key=<value>:: 92 --key=<value>::
141 Sorting key. Possible values: sample (d 93 Sorting key. Possible values: sample (default, sort by samples
142 number), time (sort by average time). 94 number), time (sort by average time).
143 <<
144 --duration=<value>:: <<
145 Show events other than HLT (x86 only) o <<
146 that take longer than duration usecs. <<
147 <<
148 --proc-map-timeout:: <<
149 When processing pre-existing threads / <<
150 a long time, because the file may be h <<
151 in such cases. <<
152 This option sets the time out limit. T <<
153 95
154 SEE ALSO 96 SEE ALSO
155 -------- 97 --------
156 linkperf:perf-top[1], linkperf:perf-record[1], 98 linkperf:perf-top[1], linkperf:perf-record[1], linkperf:perf-report[1],
157 linkperf:perf-diff[1], linkperf:perf-buildid-l 99 linkperf:perf-diff[1], linkperf:perf-buildid-list[1],
158 linkperf:perf-stat[1] 100 linkperf:perf-stat[1]
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.