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} [<options>]
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} [<options>]
16 'perf kvm stat [record|report|live] [<options> 16 'perf kvm stat [record|report|live] [<options>]
17 17
18 DESCRIPTION 18 DESCRIPTION
19 ----------- 19 -----------
20 There are a couple of variants of perf kvm: 20 There are a couple of variants of perf kvm:
21 21
22 'perf kvm [options] top <command>' to genera 22 'perf kvm [options] top <command>' to generates and displays
23 a performance counter profile of guest os in 23 a performance counter profile of guest os in realtime
24 of an arbitrary workload. 24 of an arbitrary workload.
25 25
26 'perf kvm record <command>' to record the pe 26 'perf kvm record <command>' to record the performance counter profile
27 of an arbitrary workload and save it into a 27 of an arbitrary workload and save it into a perf data file. We set the
28 default behavior of perf kvm as --guest, so 28 default behavior of perf kvm as --guest, so if neither --host nor --guest
29 is input, the perf data file name is perf.da 29 is input, the perf data file name is perf.data.guest. If --host is input,
30 the perf data file name is perf.data.kvm. If 30 the perf data file name is perf.data.kvm. If you want to record data into
31 perf.data.host, please input --host --no-gue 31 perf.data.host, please input --host --no-guest. The behaviors are shown as
32 following: 32 following:
33 Default('') -> perf.data.guest 33 Default('') -> perf.data.guest
34 --host -> perf.data.kvm 34 --host -> perf.data.kvm
35 --guest -> perf.data.guest 35 --guest -> perf.data.guest
36 --host --guest -> perf.data.kvm 36 --host --guest -> perf.data.kvm
37 --host --no-guest -> perf.data.host 37 --host --no-guest -> perf.data.host
38 38
39 'perf kvm report' to display the performance 39 'perf kvm report' to display the performance counter profile information
40 recorded via perf kvm record. 40 recorded via perf kvm record.
41 41
42 'perf kvm diff' to displays the performance 42 'perf kvm diff' to displays the performance difference amongst two perf.data
43 files captured via perf record. 43 files captured via perf record.
44 44
45 'perf kvm buildid-list' to display the buil 45 '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 46 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 47 for use by perf report. As buildid is read from /sys/kernel/notes in os, then
48 if you want to list the buildid for guest, p 48 if you want to list the buildid for guest, please make sure your perf data file
49 was captured with --guestmount in perf kvm r 49 was captured with --guestmount in perf kvm record.
50 50
51 'perf kvm stat <command>' to run a command a 51 'perf kvm stat <command>' to run a command and gather performance counter
52 statistics. 52 statistics.
53 Especially, perf 'kvm stat record/report' ge 53 Especially, perf 'kvm stat record/report' generates a statistical analysis
54 of KVM events. Currently, vmexit, mmio (x86 54 of KVM events. Currently, vmexit, mmio (x86 only) and ioport (x86 only)
55 events are supported. 'perf kvm stat record 55 events are supported. 'perf kvm stat record <command>' records kvm events
56 and the events between start and end <comman 56 and the events between start and end <command>.
57 And this command produces a file which conta 57 And this command produces a file which contains tracing results of kvm
58 events. 58 events.
59 59
60 'perf kvm stat report' reports statistical d 60 'perf kvm stat report' reports statistical data which includes events
61 handled sample, percent_sample, time, percen !! 61 handled time, samples, and so on.
62 62
63 'perf kvm stat live' reports statistical dat 63 'perf kvm stat live' reports statistical data in a live mode (similar to
64 record + report but with statistical data up 64 record + report but with statistical data updated live at a given display
65 rate). 65 rate).
66 66
67 OPTIONS 67 OPTIONS
68 ------- 68 -------
69 -i:: 69 -i::
70 --input=<path>:: 70 --input=<path>::
71 Input file name, for the 'report', 'di 71 Input file name, for the 'report', 'diff' and 'buildid-list' subcommands.
72 -o:: 72 -o::
73 --output=<path>:: 73 --output=<path>::
74 Output file name, for the 'record' sub 74 Output file name, for the 'record' subcommand. Doesn't work with 'report',
75 just redirect the output to a file whe 75 just redirect the output to a file when using 'report'.
76 --host:: 76 --host::
77 Collect host side performance profile. 77 Collect host side performance profile.
78 --guest:: 78 --guest::
79 Collect guest side performance profile 79 Collect guest side performance profile.
80 !! 80 --guestmount=<path>::
81 :GMEXAMPLECMD: kvm --host --guest !! 81 Guest os root file system mount directory. Users mounts guest os
82 :GMEXAMPLESUBCMD: top !! 82 root directories under <path> by a specific filesystem access method,
83 include::guest-files.txt[] !! 83 typically, sshfs. For example, start 2 guest os. The one's pid is 8888
84 !! 84 and the other's is 9999.
85 --stdio:: Use the stdio interface. !! 85 #mkdir ~/guestmount; cd ~/guestmount
86 !! 86 #sshfs -o allow_other,direct_io -p 5551 localhost:/ 8888/
>> 87 #sshfs -o allow_other,direct_io -p 5552 localhost:/ 9999/
>> 88 #perf kvm --host --guest --guestmount=~/guestmount top
>> 89 --guestkallsyms=<path>::
>> 90 Guest os /proc/kallsyms file copy. 'perf' kvm' reads it to get guest
>> 91 kernel symbols. Users copy it out from guest os.
>> 92 --guestmodules=<path>::
>> 93 Guest os /proc/modules file copy. 'perf' kvm' reads it to get guest
>> 94 kernel module information. Users copy it out from guest os.
>> 95 --guestvmlinux=<path>::
>> 96 Guest os kernel vmlinux.
87 -v:: 97 -v::
88 --verbose:: 98 --verbose::
89 Be more verbose (show counter open err 99 Be more verbose (show counter open errors, etc).
90 100
91 STAT REPORT OPTIONS 101 STAT REPORT OPTIONS
92 ------------------- 102 -------------------
93 --vcpu=<value>:: 103 --vcpu=<value>::
94 analyze events which occur on this vcpu 104 analyze events which occur on this vcpu. (default: all vcpus)
95 105
96 --event=<value>:: 106 --event=<value>::
97 event to be analyzed. Possible values: 107 event to be analyzed. Possible values: vmexit, mmio (x86 only),
98 ioport (x86 only). (default: vmexit) 108 ioport (x86 only). (default: vmexit)
99 -k:: 109 -k::
100 --key=<value>:: 110 --key=<value>::
101 Sorting key. Possible values: sample (d 111 Sorting key. Possible values: sample (default, sort by samples
102 number), percent_sample (sort by sample !! 112 number), time (sort by average time).
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:: 113 -p::
107 --pid=:: 114 --pid=::
108 Analyze events only for given process ID(s 115 Analyze events only for given process ID(s) (comma separated list).
109 116
110 STAT LIVE OPTIONS 117 STAT LIVE OPTIONS
111 ----------------- 118 -----------------
112 -d:: 119 -d::
113 --display:: 120 --display::
114 Time in seconds between display update 121 Time in seconds between display updates
115 122
116 -m:: 123 -m::
117 --mmap-pages=:: 124 --mmap-pages=::
118 Number of mmap data pages (must be a p !! 125 Number of mmap data pages (must be a power of two) or size
119 specification in bytes with appended u !! 126 specification with appended unit character - B/K/M/G. The
120 The size is rounded up to the nearest !! 127 size is rounded up to have nearest pages power of two value.
121 128
122 -a:: 129 -a::
123 --all-cpus:: 130 --all-cpus::
124 System-wide collection from all CPUs. 131 System-wide collection from all CPUs.
125 132
126 -p:: 133 -p::
127 --pid=:: 134 --pid=::
128 Analyze events only for given process ID(s 135 Analyze events only for given process ID(s) (comma separated list).
129 136
130 --vcpu=<value>:: 137 --vcpu=<value>::
131 analyze events which occur on this vcpu 138 analyze events which occur on this vcpu. (default: all vcpus)
132 139
133 140
134 --event=<value>:: 141 --event=<value>::
135 event to be analyzed. Possible values: 142 event to be analyzed. Possible values: vmexit,
136 mmio (x86 only), ioport (x86 only). 143 mmio (x86 only), ioport (x86 only).
137 (default: vmexit) 144 (default: vmexit)
138 145
139 -k:: 146 -k::
140 --key=<value>:: 147 --key=<value>::
141 Sorting key. Possible values: sample (d 148 Sorting key. Possible values: sample (default, sort by samples
142 number), time (sort by average time). 149 number), time (sort by average time).
143 150
144 --duration=<value>:: 151 --duration=<value>::
145 Show events other than HLT (x86 only) o 152 Show events other than HLT (x86 only) or Wait state (s390 only)
146 that take longer than duration usecs. 153 that take longer than duration usecs.
147 154
148 --proc-map-timeout:: 155 --proc-map-timeout::
149 When processing pre-existing threads / 156 When processing pre-existing threads /proc/XXX/mmap, it may take
150 a long time, because the file may be h 157 a long time, because the file may be huge. A time out is needed
151 in such cases. 158 in such cases.
152 This option sets the time out limit. T 159 This option sets the time out limit. The default value is 500 ms.
153 160
154 SEE ALSO 161 SEE ALSO
155 -------- 162 --------
156 linkperf:perf-top[1], linkperf:perf-record[1], 163 linkperf:perf-top[1], linkperf:perf-record[1], linkperf:perf-report[1],
157 linkperf:perf-diff[1], linkperf:perf-buildid-l 164 linkperf:perf-diff[1], linkperf:perf-buildid-list[1],
158 linkperf:perf-stat[1] 165 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.