1 perf-intel-pt(1) 1 perf-intel-pt(1)
2 ================ 2 ================
3 3
4 NAME 4 NAME
5 ---- 5 ----
6 perf-intel-pt - Support for Intel Processor Tr 6 perf-intel-pt - Support for Intel Processor Trace within perf tools
7 7
8 SYNOPSIS 8 SYNOPSIS
9 -------- 9 --------
10 [verse] 10 [verse]
11 'perf record' -e intel_pt// 11 'perf record' -e intel_pt//
12 12
13 DESCRIPTION 13 DESCRIPTION
14 ----------- 14 -----------
15 15
16 Intel Processor Trace (Intel PT) is an extensi 16 Intel Processor Trace (Intel PT) is an extension of Intel Architecture that
17 collects information about software execution 17 collects information about software execution such as control flow, execution
18 modes and timings and formats it into highly c 18 modes and timings and formats it into highly compressed binary packets.
19 Technical details are documented in the Intel 19 Technical details are documented in the Intel 64 and IA-32 Architectures
20 Software Developer Manuals, Chapter 36 Intel P 20 Software Developer Manuals, Chapter 36 Intel Processor Trace.
21 21
22 Intel PT is first supported in Intel Core M an 22 Intel PT is first supported in Intel Core M and 5th generation Intel Core
23 processors that are based on the Intel micro-a 23 processors that are based on the Intel micro-architecture code name Broadwell.
24 24
25 Trace data is collected by 'perf record' and s 25 Trace data is collected by 'perf record' and stored within the perf.data file.
26 See below for options to 'perf record'. 26 See below for options to 'perf record'.
27 27
28 Trace data must be 'decoded' which involves wa 28 Trace data must be 'decoded' which involves walking the object code and matching
29 the trace data packets. For example a TNT pack 29 the trace data packets. For example a TNT packet only tells whether a
30 conditional branch was taken or not taken, so 30 conditional branch was taken or not taken, so to make use of that packet the
31 decoder must know precisely which instruction 31 decoder must know precisely which instruction was being executed.
32 32
33 Decoding is done on-the-fly. The decoder outp 33 Decoding is done on-the-fly. The decoder outputs samples in the same format as
34 samples output by perf hardware events, for ex 34 samples output by perf hardware events, for example as though the "instructions"
35 or "branches" events had been recorded. Prese 35 or "branches" events had been recorded. Presently 3 tools support this:
36 'perf script', 'perf report' and 'perf inject' 36 'perf script', 'perf report' and 'perf inject'. See below for more information
37 on using those tools. 37 on using those tools.
38 38
39 The main distinguishing feature of Intel PT is 39 The main distinguishing feature of Intel PT is that the decoder can determine
40 the exact flow of software execution. Intel P 40 the exact flow of software execution. Intel PT can be used to understand why
41 and how did software get to a certain point, o 41 and how did software get to a certain point, or behave a certain way. The
42 software does not have to be recompiled, so In 42 software does not have to be recompiled, so Intel PT works with debug or release
43 builds, however the executed images are needed 43 builds, however the executed images are needed - which makes use in JIT-compiled
44 environments, or with self-modified code, a ch 44 environments, or with self-modified code, a challenge. Also symbols need to be
45 provided to make sense of addresses. 45 provided to make sense of addresses.
46 46
47 A limitation of Intel PT is that it produces h 47 A limitation of Intel PT is that it produces huge amounts of trace data
48 (hundreds of megabytes per second per core) wh 48 (hundreds of megabytes per second per core) which takes a long time to decode,
49 for example two or three orders of magnitude l 49 for example two or three orders of magnitude longer than it took to collect.
50 Another limitation is the performance impact o 50 Another limitation is the performance impact of tracing, something that will
51 vary depending on the use-case and architectur 51 vary depending on the use-case and architecture.
52 52
53 53
54 Quickstart 54 Quickstart
55 ---------- 55 ----------
56 56
57 It is important to start small. That is becau 57 It is important to start small. That is because it is easy to capture vastly
58 more data than can possibly be processed. 58 more data than can possibly be processed.
59 59
60 The simplest thing to do with Intel PT is user 60 The simplest thing to do with Intel PT is userspace profiling of small programs.
61 Data is captured with 'perf record' e.g. to tr 61 Data is captured with 'perf record' e.g. to trace 'ls' userspace-only:
62 62
63 perf record -e intel_pt//u ls 63 perf record -e intel_pt//u ls
64 64
65 And profiled with 'perf report' e.g. 65 And profiled with 'perf report' e.g.
66 66
67 perf report 67 perf report
68 68
69 To also trace kernel space presents a problem, 69 To also trace kernel space presents a problem, namely kernel self-modifying
70 code. A fairly good kernel image is available 70 code. A fairly good kernel image is available in /proc/kcore but to get an
71 accurate image a copy of /proc/kcore needs to 71 accurate image a copy of /proc/kcore needs to be made under the same conditions
72 as the data capture. 'perf record' can make a 72 as the data capture. 'perf record' can make a copy of /proc/kcore if the option
73 --kcore is used, but access to /proc/kcore is 73 --kcore is used, but access to /proc/kcore is restricted e.g.
74 74
75 sudo perf record -o pt_ls --kcore -e i 75 sudo perf record -o pt_ls --kcore -e intel_pt// -- ls
76 76
77 which will create a directory named 'pt_ls' an 77 which will create a directory named 'pt_ls' and put the perf.data file (named
78 simply 'data') and copies of /proc/kcore, /pro 78 simply 'data') and copies of /proc/kcore, /proc/kallsyms and /proc/modules into
79 it. The other tools understand the directory 79 it. The other tools understand the directory format, so to use 'perf report'
80 becomes: 80 becomes:
81 81
82 sudo perf report -i pt_ls 82 sudo perf report -i pt_ls
83 83
84 Because samples are synthesized after-the-fact 84 Because samples are synthesized after-the-fact, the sampling period can be
85 selected for reporting. e.g. sample every micr 85 selected for reporting. e.g. sample every microsecond
86 86
87 sudo perf report pt_ls --itrace=i1usge 87 sudo perf report pt_ls --itrace=i1usge
88 88
89 See the sections below for more information ab 89 See the sections below for more information about the --itrace option.
90 90
91 Beware the smaller the period, the more sample 91 Beware the smaller the period, the more samples that are produced, and the
92 longer it takes to process them. 92 longer it takes to process them.
93 93
94 Also note that the coarseness of Intel PT timi 94 Also note that the coarseness of Intel PT timing information will start to
95 distort the statistical value of the sampling 95 distort the statistical value of the sampling as the sampling period becomes
96 smaller. 96 smaller.
97 97
98 To represent software control flow, "branches" 98 To represent software control flow, "branches" samples are produced. By default
99 a branch sample is synthesized for every singl 99 a branch sample is synthesized for every single branch. To get an idea what
100 data is available you can use the 'perf script 100 data is available you can use the 'perf script' tool with all itrace sampling
101 options, which will list all the samples. 101 options, which will list all the samples.
102 102
103 perf record -e intel_pt//u ls 103 perf record -e intel_pt//u ls
104 perf script --itrace=iybxwpe !! 104 perf script --itrace=ibxwpe
105 105
106 An interesting field that is not printed by de 106 An interesting field that is not printed by default is 'flags' which can be
107 displayed as follows: 107 displayed as follows:
108 108
109 perf script --itrace=iybxwpe -F+flags !! 109 perf script --itrace=ibxwpe -F+flags
110 110
111 The flags are "bcrosyiABExghDt" which stand fo 111 The flags are "bcrosyiABExghDt" which stand for branch, call, return, conditional,
112 system, asynchronous, interrupt, transaction a 112 system, asynchronous, interrupt, transaction abort, trace begin, trace end,
113 in transaction, VM-entry, VM-exit, interrupt d 113 in transaction, VM-entry, VM-exit, interrupt disabled, and interrupt disable
114 toggle respectively. 114 toggle respectively.
115 115
116 perf script also supports higher level ways to 116 perf script also supports higher level ways to dump instruction traces:
117 117
118 perf script --insn-trace=disasm <<
119 <<
120 or to use the xed disassembler, which requires <<
121 (see XED below): <<
122 <<
123 perf script --insn-trace --xed 118 perf script --insn-trace --xed
124 119
>> 120 Dump all instructions. This requires installing the xed tool (see XED below)
125 Dumping all instructions in a long trace can b 121 Dumping all instructions in a long trace can be fairly slow. It is usually better
126 to start with higher level decoding, like 122 to start with higher level decoding, like
127 123
128 perf script --call-trace 124 perf script --call-trace
129 125
130 or 126 or
131 127
132 perf script --call-ret-trace 128 perf script --call-ret-trace
133 129
134 and then select a time range of interest. The 130 and then select a time range of interest. The time range can then be examined
135 in detail with 131 in detail with
136 132
137 perf script --time starttime,stoptime !! 133 perf script --time starttime,stoptime --insn-trace --xed
138 134
139 While examining the trace it's also useful to 135 While examining the trace it's also useful to filter on specific CPUs using
140 the -C option 136 the -C option
141 137
142 perf script --time starttime,stoptime !! 138 perf script --time starttime,stoptime --insn-trace --xed -C 1
143 139
144 Dump all instructions in time range on CPU 1. 140 Dump all instructions in time range on CPU 1.
145 141
146 Another interesting field that is not printed 142 Another interesting field that is not printed by default is 'ipc' which can be
147 displayed as follows: 143 displayed as follows:
148 144
149 perf script --itrace=be -F+ipc 145 perf script --itrace=be -F+ipc
150 146
151 There are two ways that instructions-per-cycle 147 There are two ways that instructions-per-cycle (IPC) can be calculated depending
152 on the recording. 148 on the recording.
153 149
154 If the 'cyc' config term (see config terms sec !! 150 If the 'cyc' config term (see config terms section below) was used, then IPC is
155 and cycle events are calculated using the cycl !! 151 calculated using the cycle count from CYC packets, otherwise MTC packets are
156 MTC packets are used - refer to the 'mtc' conf !! 152 used - refer to the 'mtc' config term. When MTC is used, however, the values
157 the values are less accurate because the timin !! 153 are less accurate because the timing is less accurate.
158 154
159 Because Intel PT does not update the cycle cou 155 Because Intel PT does not update the cycle count on every branch or instruction,
160 the values will often be zero. When there are 156 the values will often be zero. When there are values, they will be the number
161 of instructions and number of cycles since the 157 of instructions and number of cycles since the last update, and thus represent
162 the average IPC cycle count since the last IPC !! 158 the average IPC since the last IPC for that event type. Note IPC for "branches"
163 Note IPC for "branches" events is calculated s !! 159 events is calculated separately from IPC for "instructions" events.
164 events. <<
165 160
166 Even with the 'cyc' config term, it is possibl 161 Even with the 'cyc' config term, it is possible to produce IPC information for
167 every change of timestamp, but at the expense 162 every change of timestamp, but at the expense of accuracy. That is selected by
168 specifying the itrace 'A' option. Due to the 163 specifying the itrace 'A' option. Due to the granularity of timestamps, the
169 actual number of cycles increases even though 164 actual number of cycles increases even though the cycles reported does not.
170 The number of instructions is known, but if IP 165 The number of instructions is known, but if IPC is reported, cycles can be too
171 low and so IPC is too high. Note that inaccur 166 low and so IPC is too high. Note that inaccuracy decreases as the period of
172 sampling increases i.e. if the number of cycle 167 sampling increases i.e. if the number of cycles is too low by a small amount,
173 that becomes less significant if the number of 168 that becomes less significant if the number of cycles is large. It may also be
174 useful to use the 'A' option in conjunction wi 169 useful to use the 'A' option in conjunction with dlfilter-show-cycles.so to
175 provide higher granularity cycle information. 170 provide higher granularity cycle information.
176 171
177 Also note that the IPC instruction count may o 172 Also note that the IPC instruction count may or may not include the current
178 instruction. If the cycle count is associated 173 instruction. If the cycle count is associated with an asynchronous branch
179 (e.g. page fault or interrupt), then the instr 174 (e.g. page fault or interrupt), then the instruction count does not include the
180 current instruction, otherwise it does. That 175 current instruction, otherwise it does. That is consistent with whether or not
181 that instruction has retired when the cycle co 176 that instruction has retired when the cycle count is updated.
182 177
183 Another note, in the case of "branches" events 178 Another note, in the case of "branches" events, non-taken branches are not
184 presently sampled, so IPC values for them do n 179 presently sampled, so IPC values for them do not appear e.g. a CYC packet with a
185 TNT packet that starts with a non-taken branch 180 TNT packet that starts with a non-taken branch. To see every possible IPC
186 value, "instructions" events can be used e.g. 181 value, "instructions" events can be used e.g. --itrace=i0ns
187 182
188 While it is possible to create scripts to anal 183 While it is possible to create scripts to analyze the data, an alternative
189 approach is available to export the data to a 184 approach is available to export the data to a sqlite or postgresql database.
190 Refer to script export-to-sqlite.py or export- 185 Refer to script export-to-sqlite.py or export-to-postgresql.py for more details,
191 and to script exported-sql-viewer.py for an ex 186 and to script exported-sql-viewer.py for an example of using the database.
192 187
193 There is also script intel-pt-events.py which 188 There is also script intel-pt-events.py which provides an example of how to
194 unpack the raw data for power events and PTWRI 189 unpack the raw data for power events and PTWRITE. The script also displays
195 branches, and supports 2 additional modes sele 190 branches, and supports 2 additional modes selected by option:
196 191
197 - --insn-trace - instruction trace !! 192 --insn-trace - instruction trace
198 - --src-trace - source trace !! 193 --src-trace - source trace
199 <<
200 The intel-pt-events.py script also has options <<
201 <<
202 - --all-switch-events - display all switch ev <<
203 - --interleave [<n>] - interleave sample outp <<
204 no more than n samples for a CPU are displaye <<
205 Note this only affects the order of output, a <<
206 same. <<
207 194
208 As mentioned above, it is easy to capture too 195 As mentioned above, it is easy to capture too much data. One way to limit the
209 data captured is to use 'snapshot' mode which 196 data captured is to use 'snapshot' mode which is explained further below.
210 Refer to 'new snapshot option' and 'Intel PT m 197 Refer to 'new snapshot option' and 'Intel PT modes of operation' further below.
211 198
212 Another problem that will be experienced is de 199 Another problem that will be experienced is decoder errors. They can be caused
213 by inability to access the executed image, sel 200 by inability to access the executed image, self-modified or JIT-ed code, or the
214 inability to match side-band information (such 201 inability to match side-band information (such as context switches and mmaps)
215 which results in the decoder not knowing what 202 which results in the decoder not knowing what code was executed.
216 203
217 There is also the problem of perf not being ab 204 There is also the problem of perf not being able to copy the data fast enough,
218 resulting in data lost because the buffer was 205 resulting in data lost because the buffer was full. See 'Buffer handling' below
219 for more details. 206 for more details.
220 207
221 208
222 perf record 209 perf record
223 ----------- 210 -----------
224 211
225 new event 212 new event
226 ~~~~~~~~~ 213 ~~~~~~~~~
227 214
228 The Intel PT kernel driver creates a new PMU f 215 The Intel PT kernel driver creates a new PMU for Intel PT. PMU events are
229 selected by providing the PMU name followed by 216 selected by providing the PMU name followed by the "config" separated by slashes.
230 An enhancement has been made to allow default 217 An enhancement has been made to allow default "config" e.g. the option
231 218
232 -e intel_pt// 219 -e intel_pt//
233 220
234 will use a default config value. Currently th 221 will use a default config value. Currently that is the same as
235 222
236 -e intel_pt/tsc,noretcomp=0/ 223 -e intel_pt/tsc,noretcomp=0/
237 224
238 which is the same as 225 which is the same as
239 226
240 -e intel_pt/tsc=1,noretcomp=0/ 227 -e intel_pt/tsc=1,noretcomp=0/
241 228
242 Note there are now new config terms - see sect 229 Note there are now new config terms - see section 'config terms' further below.
243 230
244 The config terms are listed in /sys/devices/in 231 The config terms are listed in /sys/devices/intel_pt/format. They are bit
245 fields within the config member of the struct 232 fields within the config member of the struct perf_event_attr which is
246 passed to the kernel by the perf_event_open sy 233 passed to the kernel by the perf_event_open system call. They correspond to bit
247 fields in the IA32_RTIT_CTL MSR. Here is a li 234 fields in the IA32_RTIT_CTL MSR. Here is a list of them and their definitions:
248 235
249 $ grep -H . /sys/bus/event_source/devi 236 $ grep -H . /sys/bus/event_source/devices/intel_pt/format/*
250 /sys/bus/event_source/devices/intel_pt 237 /sys/bus/event_source/devices/intel_pt/format/cyc:config:1
251 /sys/bus/event_source/devices/intel_pt 238 /sys/bus/event_source/devices/intel_pt/format/cyc_thresh:config:19-22
252 /sys/bus/event_source/devices/intel_pt 239 /sys/bus/event_source/devices/intel_pt/format/mtc:config:9
253 /sys/bus/event_source/devices/intel_pt 240 /sys/bus/event_source/devices/intel_pt/format/mtc_period:config:14-17
254 /sys/bus/event_source/devices/intel_pt 241 /sys/bus/event_source/devices/intel_pt/format/noretcomp:config:11
255 /sys/bus/event_source/devices/intel_pt 242 /sys/bus/event_source/devices/intel_pt/format/psb_period:config:24-27
256 /sys/bus/event_source/devices/intel_pt 243 /sys/bus/event_source/devices/intel_pt/format/tsc:config:10
257 244
258 Note that the default config must be overridde 245 Note that the default config must be overridden for each term i.e.
259 246
260 -e intel_pt/noretcomp=0/ 247 -e intel_pt/noretcomp=0/
261 248
262 is the same as: 249 is the same as:
263 250
264 -e intel_pt/tsc=1,noretcomp=0/ 251 -e intel_pt/tsc=1,noretcomp=0/
265 252
266 So, to disable TSC packets use: 253 So, to disable TSC packets use:
267 254
268 -e intel_pt/tsc=0/ 255 -e intel_pt/tsc=0/
269 256
270 It is also possible to specify the config valu 257 It is also possible to specify the config value explicitly:
271 258
272 -e intel_pt/config=0x400/ 259 -e intel_pt/config=0x400/
273 260
274 Note that, as with all events, the event is su 261 Note that, as with all events, the event is suffixed with event modifiers:
275 262
276 u userspace 263 u userspace
277 k kernel 264 k kernel
278 h hypervisor 265 h hypervisor
279 G guest 266 G guest
280 H host 267 H host
281 p precise ip 268 p precise ip
282 269
283 'h', 'G' and 'H' are for virtualization which 270 'h', 'G' and 'H' are for virtualization which are not used by Intel PT.
284 'p' is also not relevant to Intel PT. So only 271 'p' is also not relevant to Intel PT. So only options 'u' and 'k' are
285 meaningful for Intel PT. 272 meaningful for Intel PT.
286 273
287 perf_event_attr is displayed if the -vv option 274 perf_event_attr is displayed if the -vv option is used e.g.
288 275
289 -------------------------------------- 276 ------------------------------------------------------------
290 perf_event_attr: 277 perf_event_attr:
291 type 6 278 type 6
292 size 112 279 size 112
293 config 0x400 280 config 0x400
294 { sample_period, sample_freq } 1 281 { sample_period, sample_freq } 1
295 sample_type IP|TI 282 sample_type IP|TID|TIME|CPU|IDENTIFIER
296 read_format ID 283 read_format ID
297 disabled 1 284 disabled 1
298 inherit 1 285 inherit 1
299 exclude_kernel 1 286 exclude_kernel 1
300 exclude_hv 1 287 exclude_hv 1
301 enable_on_exec 1 288 enable_on_exec 1
302 sample_id_all 1 289 sample_id_all 1
303 -------------------------------------- 290 ------------------------------------------------------------
304 sys_perf_event_open: pid 31104 cpu 0 291 sys_perf_event_open: pid 31104 cpu 0 group_fd -1 flags 0x8
305 sys_perf_event_open: pid 31104 cpu 1 292 sys_perf_event_open: pid 31104 cpu 1 group_fd -1 flags 0x8
306 sys_perf_event_open: pid 31104 cpu 2 293 sys_perf_event_open: pid 31104 cpu 2 group_fd -1 flags 0x8
307 sys_perf_event_open: pid 31104 cpu 3 294 sys_perf_event_open: pid 31104 cpu 3 group_fd -1 flags 0x8
308 -------------------------------------- 295 ------------------------------------------------------------
309 296
310 297
311 config terms 298 config terms
312 ~~~~~~~~~~~~ 299 ~~~~~~~~~~~~
313 300
314 The June 2015 version of Intel 64 and IA-32 Ar 301 The June 2015 version of Intel 64 and IA-32 Architectures Software Developer
315 Manuals, Chapter 36 Intel Processor Trace, def 302 Manuals, Chapter 36 Intel Processor Trace, defined new Intel PT features.
316 Some of the features are reflect in new config 303 Some of the features are reflect in new config terms. All the config terms are
317 described below. 304 described below.
318 305
319 tsc Always supported. Produces TS 306 tsc Always supported. Produces TSC timestamp packets to provide
320 timing information. In some c 307 timing information. In some cases it is possible to decode
321 without timing information, fo 308 without timing information, for example a per-thread context
322 that does not overlap executab 309 that does not overlap executable memory maps.
323 310
324 The default config selects tsc 311 The default config selects tsc (i.e. tsc=1).
325 312
326 noretcomp Always supported. Disables "r 313 noretcomp Always supported. Disables "return compression" so a TIP packet
327 is produced when a function re 314 is produced when a function returns. Causes more packets to be
328 produced but might make decodi 315 produced but might make decoding more reliable.
329 316
330 The default config does not se 317 The default config does not select noretcomp (i.e. noretcomp=0).
331 318
332 psb_period Allows the frequency of PSB pa 319 psb_period Allows the frequency of PSB packets to be specified.
333 320
334 The PSB packet is a synchroniz 321 The PSB packet is a synchronization packet that provides a
335 starting point for decoding or 322 starting point for decoding or recovery from errors.
336 323
337 Support for psb_period is indi 324 Support for psb_period is indicated by:
338 325
339 /sys/bus/event_source/ 326 /sys/bus/event_source/devices/intel_pt/caps/psb_cyc
340 327
341 which contains "1" if the feat 328 which contains "1" if the feature is supported and "0"
342 otherwise. 329 otherwise.
343 330
344 Valid values are given by: 331 Valid values are given by:
345 332
346 /sys/bus/event_source/ 333 /sys/bus/event_source/devices/intel_pt/caps/psb_periods
347 334
348 which contains a hexadecimal v 335 which contains a hexadecimal value, the bits of which represent
349 valid values e.g. bit 2 set me 336 valid values e.g. bit 2 set means value 2 is valid.
350 337
351 The psb_period value is conver 338 The psb_period value is converted to the approximate number of
352 trace bytes between PSB packet 339 trace bytes between PSB packets as:
353 340
354 2 ^ (value + 11) 341 2 ^ (value + 11)
355 342
356 e.g. value 3 means 16KiB bytes 343 e.g. value 3 means 16KiB bytes between PSBs
357 344
358 If an invalid value is entered 345 If an invalid value is entered, the error message
359 will give a list of valid valu 346 will give a list of valid values e.g.
360 347
361 $ perf record -e intel 348 $ perf record -e intel_pt/psb_period=15/u uname
362 Invalid psb_period for 349 Invalid psb_period for intel_pt. Valid values are: 0-5
363 350
364 If MTC packets are selected, t 351 If MTC packets are selected, the default config selects a value
365 of 3 (i.e. psb_period=3) or th 352 of 3 (i.e. psb_period=3) or the nearest lower value that is
366 supported (0 is always support 353 supported (0 is always supported). Otherwise the default is 0.
367 354
368 If decoding is expected to be 355 If decoding is expected to be reliable and the buffer is large
369 then a large PSB period can be 356 then a large PSB period can be used.
370 357
371 Because a TSC packet is produc 358 Because a TSC packet is produced with PSB, the PSB period can
372 also affect the granularity to 359 also affect the granularity to timing information in the absence
373 of MTC or CYC. 360 of MTC or CYC.
374 361
375 mtc Produces MTC timing packets. 362 mtc Produces MTC timing packets.
376 363
377 MTC packets provide finer grai 364 MTC packets provide finer grain timestamp information than TSC
378 packets. MTC packets record t 365 packets. MTC packets record time using the hardware crystal
379 clock (CTC) which is related t 366 clock (CTC) which is related to TSC packets using a TMA packet.
380 367
381 Support for this feature is in 368 Support for this feature is indicated by:
382 369
383 /sys/bus/event_source/ 370 /sys/bus/event_source/devices/intel_pt/caps/mtc
384 371
385 which contains "1" if the feat 372 which contains "1" if the feature is supported and
386 "0" otherwise. 373 "0" otherwise.
387 374
388 The frequency of MTC packets c 375 The frequency of MTC packets can also be specified - see
389 mtc_period below. 376 mtc_period below.
390 377
391 mtc_period Specifies how frequently MTC p 378 mtc_period Specifies how frequently MTC packets are produced - see mtc
392 above for how to determine if 379 above for how to determine if MTC packets are supported.
393 380
394 Valid values are given by: 381 Valid values are given by:
395 382
396 /sys/bus/event_source/ 383 /sys/bus/event_source/devices/intel_pt/caps/mtc_periods
397 384
398 which contains a hexadecimal v 385 which contains a hexadecimal value, the bits of which represent
399 valid values e.g. bit 2 set me 386 valid values e.g. bit 2 set means value 2 is valid.
400 387
401 The mtc_period value is conver 388 The mtc_period value is converted to the MTC frequency as:
402 389
403 CTC-frequency / (2 ^ v 390 CTC-frequency / (2 ^ value)
404 391
405 e.g. value 3 means one eighth 392 e.g. value 3 means one eighth of CTC-frequency
406 393
407 Where CTC is the hardware crys 394 Where CTC is the hardware crystal clock, the frequency of which
408 can be related to TSC via valu 395 can be related to TSC via values provided in cpuid leaf 0x15.
409 396
410 If an invalid value is entered 397 If an invalid value is entered, the error message
411 will give a list of valid valu 398 will give a list of valid values e.g.
412 399
413 $ perf record -e intel 400 $ perf record -e intel_pt/mtc_period=15/u uname
414 Invalid mtc_period for 401 Invalid mtc_period for intel_pt. Valid values are: 0,3,6,9
415 402
416 The default value is 3 or the 403 The default value is 3 or the nearest lower value
417 that is supported (0 is always 404 that is supported (0 is always supported).
418 405
419 cyc Produces CYC timing packets. 406 cyc Produces CYC timing packets.
420 407
421 CYC packets provide even finer 408 CYC packets provide even finer grain timestamp information than
422 MTC and TSC packets. A CYC pa 409 MTC and TSC packets. A CYC packet contains the number of CPU
423 cycles since the last CYC pack 410 cycles since the last CYC packet. Unlike MTC and TSC packets,
424 CYC packets are only sent when 411 CYC packets are only sent when another packet is also sent.
425 412
426 Support for this feature is in 413 Support for this feature is indicated by:
427 414
428 /sys/bus/event_source/ 415 /sys/bus/event_source/devices/intel_pt/caps/psb_cyc
429 416
430 which contains "1" if the feat 417 which contains "1" if the feature is supported and
431 "0" otherwise. 418 "0" otherwise.
432 419
433 The number of CYC packets prod 420 The number of CYC packets produced can be reduced by specifying
434 a threshold - see cyc_thresh b 421 a threshold - see cyc_thresh below.
435 422
436 cyc_thresh Specifies how frequently CYC p 423 cyc_thresh Specifies how frequently CYC packets are produced - see cyc
437 above for how to determine if 424 above for how to determine if CYC packets are supported.
438 425
439 Valid cyc_thresh values are gi 426 Valid cyc_thresh values are given by:
440 427
441 /sys/bus/event_source/ 428 /sys/bus/event_source/devices/intel_pt/caps/cycle_thresholds
442 429
443 which contains a hexadecimal v 430 which contains a hexadecimal value, the bits of which represent
444 valid values e.g. bit 2 set me 431 valid values e.g. bit 2 set means value 2 is valid.
445 432
446 The cyc_thresh value represent 433 The cyc_thresh value represents the minimum number of CPU cycles
447 that must have passed before a 434 that must have passed before a CYC packet can be sent. The
448 number of CPU cycles is: 435 number of CPU cycles is:
449 436
450 2 ^ (value - 1) 437 2 ^ (value - 1)
451 438
452 e.g. value 4 means 8 CPU cycle 439 e.g. value 4 means 8 CPU cycles must pass before a CYC packet
453 can be sent. Note a CYC packe 440 can be sent. Note a CYC packet is still only sent when another
454 packet is sent, not at, e.g. e 441 packet is sent, not at, e.g. every 8 CPU cycles.
455 442
456 If an invalid value is entered 443 If an invalid value is entered, the error message
457 will give a list of valid valu 444 will give a list of valid values e.g.
458 445
459 $ perf record -e intel 446 $ perf record -e intel_pt/cyc,cyc_thresh=15/u uname
460 Invalid cyc_thresh for 447 Invalid cyc_thresh for intel_pt. Valid values are: 0-12
461 448
462 CYC packets are not requested 449 CYC packets are not requested by default.
463 450
464 pt Specifies pass-through which e 451 pt Specifies pass-through which enables the 'branch' config term.
465 452
466 The default config selects 'pt 453 The default config selects 'pt' if it is available, so a user will
467 never need to specify this ter 454 never need to specify this term.
468 455
469 branch Enable branch tracing. Branch 456 branch Enable branch tracing. Branch tracing is enabled by default so to
470 disable branch tracing use 'br 457 disable branch tracing use 'branch=0'.
471 458
472 The default config selects 'br 459 The default config selects 'branch' if it is available.
473 460
474 ptw Enable PTWRITE packets which a 461 ptw Enable PTWRITE packets which are produced when a ptwrite instruction
475 is executed. 462 is executed.
476 463
477 Support for this feature is in 464 Support for this feature is indicated by:
478 465
479 /sys/bus/event_source/ 466 /sys/bus/event_source/devices/intel_pt/caps/ptwrite
480 467
481 which contains "1" if the feat 468 which contains "1" if the feature is supported and
482 "0" otherwise. 469 "0" otherwise.
483 470
484 As an alternative, refer to "E 471 As an alternative, refer to "Emulated PTWRITE" further below.
485 472
486 fup_on_ptw Enable a FUP packet to follow 473 fup_on_ptw Enable a FUP packet to follow the PTWRITE packet. The FUP packet
487 provides the address of the pt 474 provides the address of the ptwrite instruction. In the absence of
488 fup_on_ptw, the decoder will u 475 fup_on_ptw, the decoder will use the address of the previous branch
489 if branch tracing is enabled, 476 if branch tracing is enabled, otherwise the address will be zero.
490 Note that fup_on_ptw will work 477 Note that fup_on_ptw will work even when branch tracing is disabled.
491 478
492 pwr_evt Enable power events. The powe 479 pwr_evt Enable power events. The power events provide information about
493 changes to the CPU C-state. 480 changes to the CPU C-state.
494 481
495 Support for this feature is in 482 Support for this feature is indicated by:
496 483
497 /sys/bus/event_source/ 484 /sys/bus/event_source/devices/intel_pt/caps/power_event_trace
498 485
499 which contains "1" if the feat 486 which contains "1" if the feature is supported and
500 "0" otherwise. 487 "0" otherwise.
501 488
502 event Enable Event Trace. The event 489 event Enable Event Trace. The events provide information about asynchronous
503 events. 490 events.
504 491
505 Support for this feature is in 492 Support for this feature is indicated by:
506 493
507 /sys/bus/event_source/ 494 /sys/bus/event_source/devices/intel_pt/caps/event_trace
508 495
509 which contains "1" if the feat 496 which contains "1" if the feature is supported and
510 "0" otherwise. 497 "0" otherwise.
511 498
512 notnt Disable TNT packets. Without 499 notnt Disable TNT packets. Without TNT packets, it is not possible to walk
513 executable code to reconstruct 500 executable code to reconstruct control flow, however FUP, TIP, TIP.PGE
514 and TIP.PGD packets still indi 501 and TIP.PGD packets still indicate asynchronous control flow, and (if
515 return compression is disabled 502 return compression is disabled - see noretcomp) return statements.
516 The advantage of eliminating T 503 The advantage of eliminating TNT packets is reducing the size of the
517 trace and corresponding tracin 504 trace and corresponding tracing overhead.
518 505
519 Support for this feature is in 506 Support for this feature is indicated by:
520 507
521 /sys/bus/event_source/ 508 /sys/bus/event_source/devices/intel_pt/caps/tnt_disable
522 509
523 which contains "1" if the feat 510 which contains "1" if the feature is supported and
524 "0" otherwise. 511 "0" otherwise.
525 512
526 513
527 AUX area sampling option 514 AUX area sampling option
528 ~~~~~~~~~~~~~~~~~~~~~~~~ 515 ~~~~~~~~~~~~~~~~~~~~~~~~
529 516
530 To select Intel PT "sampling" the AUX area sam 517 To select Intel PT "sampling" the AUX area sampling option can be used:
531 518
532 --aux-sample 519 --aux-sample
533 520
534 Optionally it can be followed by the sample si 521 Optionally it can be followed by the sample size in bytes e.g.
535 522
536 --aux-sample=8192 523 --aux-sample=8192
537 524
538 In addition, the Intel PT event to sample must 525 In addition, the Intel PT event to sample must be defined e.g.
539 526
540 -e intel_pt//u 527 -e intel_pt//u
541 528
542 Samples on other events will be created contai 529 Samples on other events will be created containing Intel PT data e.g. the
543 following will create Intel PT samples on the 530 following will create Intel PT samples on the branch-misses event, note the
544 events must be grouped using {}: 531 events must be grouped using {}:
545 532
546 perf record --aux-sample -e '{intel_pt 533 perf record --aux-sample -e '{intel_pt//u,branch-misses:u}'
547 534
548 An alternative to '--aux-sample' is to add the 535 An alternative to '--aux-sample' is to add the config term 'aux-sample-size' to
549 events. In this case, the grouping is implied 536 events. In this case, the grouping is implied e.g.
550 537
551 perf record -e intel_pt//u -e branch-m 538 perf record -e intel_pt//u -e branch-misses/aux-sample-size=8192/u
552 539
553 is the same as: 540 is the same as:
554 541
555 perf record -e '{intel_pt//u,branch-mi 542 perf record -e '{intel_pt//u,branch-misses/aux-sample-size=8192/u}'
556 543
557 but allows for also using an address filter e. 544 but allows for also using an address filter e.g.:
558 545
559 perf record -e intel_pt//u --filter 'f 546 perf record -e intel_pt//u --filter 'filter * @/bin/ls' -e branch-misses/aux-sample-size=8192/u -- ls
560 547
561 It is important to select a sample size that i 548 It is important to select a sample size that is big enough to contain at least
562 one PSB packet. If not a warning will be disp 549 one PSB packet. If not a warning will be displayed:
563 550
564 Intel PT sample size (%zu) may be too 551 Intel PT sample size (%zu) may be too small for PSB period (%zu)
565 552
566 The calculation used for that is: if sample_si 553 The calculation used for that is: if sample_size <= psb_period + 256 display the
567 warning. When sampling is used, psb_period de 554 warning. When sampling is used, psb_period defaults to 0 (2KiB).
568 555
569 The default sample size is 4KiB. 556 The default sample size is 4KiB.
570 557
571 The sample size is passed in aux_sample_size i 558 The sample size is passed in aux_sample_size in struct perf_event_attr. The
572 sample size is limited by the maximum event si 559 sample size is limited by the maximum event size which is 64KiB. It is
573 difficult to know how big the event might be w 560 difficult to know how big the event might be without the trace sample attached,
574 but the tool validates that the sample size is 561 but the tool validates that the sample size is not greater than 60KiB.
575 562
576 563
577 new snapshot option 564 new snapshot option
578 ~~~~~~~~~~~~~~~~~~~ 565 ~~~~~~~~~~~~~~~~~~~
579 566
580 The difference between full trace and snapshot 567 The difference between full trace and snapshot from the kernel's perspective is
581 that in full trace we don't overwrite trace da 568 that in full trace we don't overwrite trace data that the user hasn't collected
582 yet (and indicated that by advancing aux_tail) 569 yet (and indicated that by advancing aux_tail), whereas in snapshot mode we let
583 the trace run and overwrite older data in the 570 the trace run and overwrite older data in the buffer so that whenever something
584 interesting happens, we can stop it and grab a 571 interesting happens, we can stop it and grab a snapshot of what was going on
585 around that interesting moment. 572 around that interesting moment.
586 573
587 To select snapshot mode a new option has been 574 To select snapshot mode a new option has been added:
588 575
589 -S 576 -S
590 577
591 Optionally it can be followed by the snapshot 578 Optionally it can be followed by the snapshot size e.g.
592 579
593 -S0x100000 580 -S0x100000
594 581
595 The default snapshot size is the auxtrace mmap 582 The default snapshot size is the auxtrace mmap size. If neither auxtrace mmap size
596 nor snapshot size is specified, then the defau 583 nor snapshot size is specified, then the default is 4MiB for privileged users
597 (or if /proc/sys/kernel/perf_event_paranoid < 584 (or if /proc/sys/kernel/perf_event_paranoid < 0), 128KiB for unprivileged users.
598 If an unprivileged user does not specify mmap 585 If an unprivileged user does not specify mmap pages, the mmap pages will be
599 reduced as described in the 'new auxtrace mmap 586 reduced as described in the 'new auxtrace mmap size option' section below.
600 587
601 The snapshot size is displayed if the option - 588 The snapshot size is displayed if the option -vv is used e.g.
602 589
603 Intel PT snapshot size: %zu 590 Intel PT snapshot size: %zu
604 591
605 592
606 new auxtrace mmap size option 593 new auxtrace mmap size option
607 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 594 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
608 595
609 Intel PT buffer size is specified by an additi 596 Intel PT buffer size is specified by an addition to the -m option e.g.
610 597
611 -m,16 598 -m,16
612 599
613 selects a buffer size of 16 pages i.e. 64KiB. 600 selects a buffer size of 16 pages i.e. 64KiB.
614 601
615 Note that the existing functionality of -m is 602 Note that the existing functionality of -m is unchanged. The auxtrace mmap size
616 is specified by the optional addition of a com 603 is specified by the optional addition of a comma and the value.
617 604
618 The default auxtrace mmap size for Intel PT is 605 The default auxtrace mmap size for Intel PT is 4MiB/page_size for privileged users
619 (or if /proc/sys/kernel/perf_event_paranoid < 606 (or if /proc/sys/kernel/perf_event_paranoid < 0), 128KiB for unprivileged users.
620 If an unprivileged user does not specify mmap 607 If an unprivileged user does not specify mmap pages, the mmap pages will be
621 reduced from the default 512KiB/page_size to 2 608 reduced from the default 512KiB/page_size to 256KiB/page_size, otherwise the
622 user is likely to get an error as they exceed 609 user is likely to get an error as they exceed their mlock limit (Max locked
623 memory as shown in /proc/self/limits). Note t 610 memory as shown in /proc/self/limits). Note that perf does not count the first
624 512KiB (actually /proc/sys/kernel/perf_event_m 611 512KiB (actually /proc/sys/kernel/perf_event_mlock_kb minus 1 page) per cpu
625 against the mlock limit so an unprivileged use 612 against the mlock limit so an unprivileged user is allowed 512KiB per cpu plus
626 their mlock limit (which defaults to 64KiB but 613 their mlock limit (which defaults to 64KiB but is not multiplied by the number
627 of cpus). 614 of cpus).
628 615
629 In full-trace mode, powers of two are allowed 616 In full-trace mode, powers of two are allowed for buffer size, with a minimum
630 size of 2 pages. In snapshot mode or sampling 617 size of 2 pages. In snapshot mode or sampling mode, it is the same but the
631 minimum size is 1 page. 618 minimum size is 1 page.
632 619
633 The mmap size and auxtrace mmap size are displ 620 The mmap size and auxtrace mmap size are displayed if the -vv option is used e.g.
634 621
635 mmap length 528384 622 mmap length 528384
636 auxtrace mmap length 4198400 623 auxtrace mmap length 4198400
637 624
638 625
639 Intel PT modes of operation 626 Intel PT modes of operation
640 ~~~~~~~~~~~~~~~~~~~~~~~~~~~ 627 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
641 628
642 Intel PT can be used in 3 modes: 629 Intel PT can be used in 3 modes:
643 full-trace mode 630 full-trace mode
644 sample mode 631 sample mode
645 snapshot mode 632 snapshot mode
646 633
647 Full-trace mode traces continuously e.g. 634 Full-trace mode traces continuously e.g.
648 635
649 perf record -e intel_pt//u uname 636 perf record -e intel_pt//u uname
650 637
651 Sample mode attaches a Intel PT sample to othe 638 Sample mode attaches a Intel PT sample to other events e.g.
652 639
653 perf record --aux-sample -e intel_pt// 640 perf record --aux-sample -e intel_pt//u -e branch-misses:u
654 641
655 Snapshot mode captures the available data when 642 Snapshot mode captures the available data when a signal is sent or "snapshot"
656 control command is issued. e.g. using a signal 643 control command is issued. e.g. using a signal
657 644
658 perf record -v -e intel_pt//u -S ./loo 645 perf record -v -e intel_pt//u -S ./loopy 1000000000 &
659 [1] 11435 646 [1] 11435
660 kill -USR2 11435 647 kill -USR2 11435
661 Recording AUX area tracing snapshot 648 Recording AUX area tracing snapshot
662 649
663 Note that the signal sent is SIGUSR2. 650 Note that the signal sent is SIGUSR2.
664 Note that "Recording AUX area tracing snapshot 651 Note that "Recording AUX area tracing snapshot" is displayed because the -v
665 option is used. 652 option is used.
666 653
667 The advantage of using "snapshot" control comm 654 The advantage of using "snapshot" control command is that the access is
668 controlled by access to a FIFO e.g. 655 controlled by access to a FIFO e.g.
669 656
670 $ mkfifo perf.control 657 $ mkfifo perf.control
671 $ mkfifo perf.ack 658 $ mkfifo perf.ack
672 $ cat perf.ack & 659 $ cat perf.ack &
673 [1] 15235 660 [1] 15235
674 $ sudo ~/bin/perf record --control fif 661 $ sudo ~/bin/perf record --control fifo:perf.control,perf.ack -S -e intel_pt//u -- sleep 60 &
675 [2] 15243 662 [2] 15243
676 $ ps -e | grep perf 663 $ ps -e | grep perf
677 15244 pts/1 00:00:00 perf 664 15244 pts/1 00:00:00 perf
678 $ kill -USR2 15244 665 $ kill -USR2 15244
679 bash: kill: (15244) - Operation not pe 666 bash: kill: (15244) - Operation not permitted
680 $ echo snapshot > perf.control 667 $ echo snapshot > perf.control
681 ack 668 ack
682 669
683 The 3 Intel PT modes of operation cannot be us 670 The 3 Intel PT modes of operation cannot be used together.
684 671
685 672
686 Buffer handling 673 Buffer handling
687 ~~~~~~~~~~~~~~~ 674 ~~~~~~~~~~~~~~~
688 675
689 There may be buffer limitations (i.e. single T 676 There may be buffer limitations (i.e. single ToPa entry) which means that actual
690 buffer sizes are limited to powers of 2 up to !! 677 buffer sizes are limited to powers of 2 up to 4MiB (MAX_ORDER). In order to
691 provide other sizes, and in particular an arbi 678 provide other sizes, and in particular an arbitrarily large size, multiple
692 buffers are logically concatenated. However a 679 buffers are logically concatenated. However an interrupt must be used to switch
693 between buffers. That has two potential probl 680 between buffers. That has two potential problems:
694 a) the interrupt may not be handled in 681 a) the interrupt may not be handled in time so that the current buffer
695 becomes full and some trace data is lo 682 becomes full and some trace data is lost.
696 b) the interrupts may slow the system 683 b) the interrupts may slow the system and affect the performance
697 results. 684 results.
698 685
699 If trace data is lost, the driver sets 'trunca 686 If trace data is lost, the driver sets 'truncated' in the PERF_RECORD_AUX event
700 which the tools report as an error. 687 which the tools report as an error.
701 688
702 In full-trace mode, the driver waits for data 689 In full-trace mode, the driver waits for data to be copied out before allowing
703 the (logical) buffer to wrap-around. If data 690 the (logical) buffer to wrap-around. If data is not copied out quickly enough,
704 again 'truncated' is set in the PERF_RECORD_AU 691 again 'truncated' is set in the PERF_RECORD_AUX event. If the driver has to
705 wait, the intel_pt event gets disabled. Becau 692 wait, the intel_pt event gets disabled. Because it is difficult to know when
706 that happens, perf tools always re-enable the 693 that happens, perf tools always re-enable the intel_pt event after copying out
707 data. 694 data.
708 695
709 696
710 Intel PT and build ids 697 Intel PT and build ids
711 ~~~~~~~~~~~~~~~~~~~~~~ 698 ~~~~~~~~~~~~~~~~~~~~~~
712 699
713 By default "perf record" post-processes the ev 700 By default "perf record" post-processes the event stream to find all build ids
714 for executables for all addresses sampled. De 701 for executables for all addresses sampled. Deliberately, Intel PT is not
715 decoded for that purpose (it would take too lo 702 decoded for that purpose (it would take too long). Instead the build ids for
716 all executables encountered (due to mmap, comm 703 all executables encountered (due to mmap, comm or task events) are included
717 in the perf.data file. 704 in the perf.data file.
718 705
719 To see buildids included in the perf.data file 706 To see buildids included in the perf.data file use the command:
720 707
721 perf buildid-list 708 perf buildid-list
722 709
723 If the perf.data file contains Intel PT data, 710 If the perf.data file contains Intel PT data, that is the same as:
724 711
725 perf buildid-list --with-hits 712 perf buildid-list --with-hits
726 713
727 714
728 Snapshot mode and event disabling 715 Snapshot mode and event disabling
729 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 716 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
730 717
731 In order to make a snapshot, the intel_pt even 718 In order to make a snapshot, the intel_pt event is disabled using an IOCTL,
732 namely PERF_EVENT_IOC_DISABLE. However doing 719 namely PERF_EVENT_IOC_DISABLE. However doing that can also disable the
733 collection of side-band information. In order 720 collection of side-band information. In order to prevent that, a dummy
734 software event has been introduced that permit 721 software event has been introduced that permits tracking events (like mmaps) to
735 continue to be recorded while intel_pt is disa 722 continue to be recorded while intel_pt is disabled. That is important to ensure
736 there is complete side-band information to all 723 there is complete side-band information to allow the decoding of subsequent
737 snapshots. 724 snapshots.
738 725
739 A test has been created for that. To find the 726 A test has been created for that. To find the test:
740 727
741 perf test list 728 perf test list
742 ... 729 ...
743 23: Test using a dummy software event 730 23: Test using a dummy software event to keep tracking
744 731
745 To run the test: 732 To run the test:
746 733
747 perf test 23 734 perf test 23
748 23: Test using a dummy software event 735 23: Test using a dummy software event to keep tracking : Ok
749 736
750 737
751 perf record modes (nothing new here) 738 perf record modes (nothing new here)
752 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 739 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
753 740
754 perf record essentially operates in one of thr 741 perf record essentially operates in one of three modes:
755 per thread 742 per thread
756 per cpu 743 per cpu
757 workload only 744 workload only
758 745
759 "per thread" mode is selected by -t or by --pe 746 "per thread" mode is selected by -t or by --per-thread (with -p or -u or just a
760 workload). 747 workload).
761 "per cpu" is selected by -C or -a. 748 "per cpu" is selected by -C or -a.
762 "workload only" mode is selected by not using 749 "workload only" mode is selected by not using the other options but providing a
763 command to run (i.e. the workload). 750 command to run (i.e. the workload).
764 751
765 In per-thread mode an exact list of threads is 752 In per-thread mode an exact list of threads is traced. There is no inheritance.
766 Each thread has its own event buffer. 753 Each thread has its own event buffer.
767 754
768 In per-cpu mode all processes (or processes fr 755 In per-cpu mode all processes (or processes from the selected cgroup i.e. -G
769 option, or processes selected with -p or -u) a 756 option, or processes selected with -p or -u) are traced. Each cpu has its own
770 buffer. Inheritance is allowed. 757 buffer. Inheritance is allowed.
771 758
772 In workload-only mode, the workload is traced 759 In workload-only mode, the workload is traced but with per-cpu buffers.
773 Inheritance is allowed. Note that you can now 760 Inheritance is allowed. Note that you can now trace a workload in per-thread
774 mode by using the --per-thread option. 761 mode by using the --per-thread option.
775 762
776 763
777 Privileged vs non-privileged users 764 Privileged vs non-privileged users
778 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 765 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
779 766
780 Unless /proc/sys/kernel/perf_event_paranoid is 767 Unless /proc/sys/kernel/perf_event_paranoid is set to -1, unprivileged users
781 have memory limits imposed upon them. That af 768 have memory limits imposed upon them. That affects what buffer sizes they can
782 have as outlined above. 769 have as outlined above.
783 770
784 The v4.2 kernel introduced support for a conte 771 The v4.2 kernel introduced support for a context switch metadata event,
785 PERF_RECORD_SWITCH, which allows unprivileged 772 PERF_RECORD_SWITCH, which allows unprivileged users to see when their processes
786 are scheduled out and in, just not by whom, wh 773 are scheduled out and in, just not by whom, which is left for the
787 PERF_RECORD_SWITCH_CPU_WIDE, that is only acce 774 PERF_RECORD_SWITCH_CPU_WIDE, that is only accessible in system wide context,
788 which in turn requires CAP_PERFMON or CAP_SYS_ 775 which in turn requires CAP_PERFMON or CAP_SYS_ADMIN.
789 776
790 Please see the 45ac1403f564 ("perf: Add PERF_R 777 Please see the 45ac1403f564 ("perf: Add PERF_RECORD_SWITCH to indicate context
791 switches") commit, that introduces these metad 778 switches") commit, that introduces these metadata events for further info.
792 779
793 When working with kernels < v4.2, the followin 780 When working with kernels < v4.2, the following considerations must be taken,
794 as the sched:sched_switch tracepoints will be 781 as the sched:sched_switch tracepoints will be used to receive such information:
795 782
796 Unless /proc/sys/kernel/perf_event_paranoid is 783 Unless /proc/sys/kernel/perf_event_paranoid is set to -1, unprivileged users are
797 not permitted to use tracepoints which means t 784 not permitted to use tracepoints which means there is insufficient side-band
798 information to decode Intel PT in per-cpu mode 785 information to decode Intel PT in per-cpu mode, and potentially workload-only
799 mode too if the workload creates new processes 786 mode too if the workload creates new processes.
800 787
801 Note also, that to use tracepoints, read-acces 788 Note also, that to use tracepoints, read-access to debugfs is required. So if
802 debugfs is not mounted or the user does not ha 789 debugfs is not mounted or the user does not have read-access, it will again not
803 be possible to decode Intel PT in per-cpu mode 790 be possible to decode Intel PT in per-cpu mode.
804 791
805 792
806 sched_switch tracepoint 793 sched_switch tracepoint
807 ~~~~~~~~~~~~~~~~~~~~~~~ 794 ~~~~~~~~~~~~~~~~~~~~~~~
808 795
809 The sched_switch tracepoint is used to provide 796 The sched_switch tracepoint is used to provide side-band data for Intel PT
810 decoding in kernels where the PERF_RECORD_SWIT 797 decoding in kernels where the PERF_RECORD_SWITCH metadata event isn't
811 available. 798 available.
812 799
813 The sched_switch events are automatically adde 800 The sched_switch events are automatically added. e.g. the second event shown
814 below: 801 below:
815 802
816 $ perf record -vv -e intel_pt//u uname 803 $ perf record -vv -e intel_pt//u uname
817 -------------------------------------- 804 ------------------------------------------------------------
818 perf_event_attr: 805 perf_event_attr:
819 type 6 806 type 6
820 size 112 807 size 112
821 config 0x400 808 config 0x400
822 { sample_period, sample_freq } 1 809 { sample_period, sample_freq } 1
823 sample_type IP|TI 810 sample_type IP|TID|TIME|CPU|IDENTIFIER
824 read_format ID 811 read_format ID
825 disabled 1 812 disabled 1
826 inherit 1 813 inherit 1
827 exclude_kernel 1 814 exclude_kernel 1
828 exclude_hv 1 815 exclude_hv 1
829 enable_on_exec 1 816 enable_on_exec 1
830 sample_id_all 1 817 sample_id_all 1
831 -------------------------------------- 818 ------------------------------------------------------------
832 sys_perf_event_open: pid 31104 cpu 0 819 sys_perf_event_open: pid 31104 cpu 0 group_fd -1 flags 0x8
833 sys_perf_event_open: pid 31104 cpu 1 820 sys_perf_event_open: pid 31104 cpu 1 group_fd -1 flags 0x8
834 sys_perf_event_open: pid 31104 cpu 2 821 sys_perf_event_open: pid 31104 cpu 2 group_fd -1 flags 0x8
835 sys_perf_event_open: pid 31104 cpu 3 822 sys_perf_event_open: pid 31104 cpu 3 group_fd -1 flags 0x8
836 -------------------------------------- 823 ------------------------------------------------------------
837 perf_event_attr: 824 perf_event_attr:
838 type 2 825 type 2
839 size 112 826 size 112
840 config 0x108 827 config 0x108
841 { sample_period, sample_freq } 1 828 { sample_period, sample_freq } 1
842 sample_type IP|TI 829 sample_type IP|TID|TIME|CPU|PERIOD|RAW|IDENTIFIER
843 read_format ID 830 read_format ID
844 inherit 1 831 inherit 1
845 sample_id_all 1 832 sample_id_all 1
846 exclude_guest 1 833 exclude_guest 1
847 -------------------------------------- 834 ------------------------------------------------------------
848 sys_perf_event_open: pid -1 cpu 0 gr 835 sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8
849 sys_perf_event_open: pid -1 cpu 1 gr 836 sys_perf_event_open: pid -1 cpu 1 group_fd -1 flags 0x8
850 sys_perf_event_open: pid -1 cpu 2 gr 837 sys_perf_event_open: pid -1 cpu 2 group_fd -1 flags 0x8
851 sys_perf_event_open: pid -1 cpu 3 gr 838 sys_perf_event_open: pid -1 cpu 3 group_fd -1 flags 0x8
852 -------------------------------------- 839 ------------------------------------------------------------
853 perf_event_attr: 840 perf_event_attr:
854 type 1 841 type 1
855 size 112 842 size 112
856 config 0x9 843 config 0x9
857 { sample_period, sample_freq } 1 844 { sample_period, sample_freq } 1
858 sample_type IP|TI 845 sample_type IP|TID|TIME|IDENTIFIER
859 read_format ID 846 read_format ID
860 disabled 1 847 disabled 1
861 inherit 1 848 inherit 1
862 exclude_kernel 1 849 exclude_kernel 1
863 exclude_hv 1 850 exclude_hv 1
864 mmap 1 851 mmap 1
865 comm 1 852 comm 1
866 enable_on_exec 1 853 enable_on_exec 1
867 task 1 854 task 1
868 sample_id_all 1 855 sample_id_all 1
869 mmap2 1 856 mmap2 1
870 comm_exec 1 857 comm_exec 1
871 -------------------------------------- 858 ------------------------------------------------------------
872 sys_perf_event_open: pid 31104 cpu 0 859 sys_perf_event_open: pid 31104 cpu 0 group_fd -1 flags 0x8
873 sys_perf_event_open: pid 31104 cpu 1 860 sys_perf_event_open: pid 31104 cpu 1 group_fd -1 flags 0x8
874 sys_perf_event_open: pid 31104 cpu 2 861 sys_perf_event_open: pid 31104 cpu 2 group_fd -1 flags 0x8
875 sys_perf_event_open: pid 31104 cpu 3 862 sys_perf_event_open: pid 31104 cpu 3 group_fd -1 flags 0x8
876 mmap size 528384B 863 mmap size 528384B
877 AUX area mmap length 4194304 864 AUX area mmap length 4194304
878 perf event ring buffer mmapped per cpu 865 perf event ring buffer mmapped per cpu
879 Synthesizing auxtrace information 866 Synthesizing auxtrace information
880 Linux 867 Linux
881 [ perf record: Woken up 1 times to wri 868 [ perf record: Woken up 1 times to write data ]
882 [ perf record: Captured and wrote 0.04 869 [ perf record: Captured and wrote 0.042 MB perf.data ]
883 870
884 Note, the sched_switch event is only added if 871 Note, the sched_switch event is only added if the user is permitted to use it
885 and only in per-cpu mode. 872 and only in per-cpu mode.
886 873
887 Note also, the sched_switch event is only adde 874 Note also, the sched_switch event is only added if TSC packets are requested.
888 That is because, in the absence of timing info 875 That is because, in the absence of timing information, the sched_switch events
889 cannot be matched against the Intel PT trace. 876 cannot be matched against the Intel PT trace.
890 877
891 878
892 perf script 879 perf script
893 ----------- 880 -----------
894 881
895 By default, perf script will decode trace data 882 By default, perf script will decode trace data found in the perf.data file.
896 This can be further controlled by new option - 883 This can be further controlled by new option --itrace.
897 884
898 885
899 New --itrace option 886 New --itrace option
900 ~~~~~~~~~~~~~~~~~~~ 887 ~~~~~~~~~~~~~~~~~~~
901 888
902 Having no option is the same as 889 Having no option is the same as
903 890
904 --itrace 891 --itrace
905 892
906 which, in turn, is the same as 893 which, in turn, is the same as
907 894
908 --itrace=cepwxy !! 895 --itrace=cepwx
909 896
910 The letters are: 897 The letters are:
911 898
912 i synthesize "instructions" even 899 i synthesize "instructions" events
913 y synthesize "cycles" events <<
914 b synthesize "branches" events 900 b synthesize "branches" events
915 x synthesize "transactions" even 901 x synthesize "transactions" events
916 w synthesize "ptwrite" events 902 w synthesize "ptwrite" events
917 p synthesize "power" events (inc 903 p synthesize "power" events (incl. PSB events)
918 c synthesize branches events (ca 904 c synthesize branches events (calls only)
919 r synthesize branches events (re 905 r synthesize branches events (returns only)
920 o synthesize PEBS-via-PT events 906 o synthesize PEBS-via-PT events
921 I synthesize Event Trace events 907 I synthesize Event Trace events
922 e synthesize tracing error event 908 e synthesize tracing error events
923 d create a debug log 909 d create a debug log
924 g synthesize a call chain (use w 910 g synthesize a call chain (use with i or x)
925 G synthesize a call chain on exi 911 G synthesize a call chain on existing event records
926 l synthesize last branch entries 912 l synthesize last branch entries (use with i or x)
927 L synthesize last branch entries 913 L synthesize last branch entries on existing event records
928 s skip initial number of events 914 s skip initial number of events
929 q quicker (less detailed) decodi 915 q quicker (less detailed) decoding
930 A approximate IPC 916 A approximate IPC
931 Z prefer to ignore timestamps (s 917 Z prefer to ignore timestamps (so-called "timeless" decoding)
932 918
933 "Instructions" events look like they were reco 919 "Instructions" events look like they were recorded by "perf record -e
934 instructions". 920 instructions".
935 921
936 "Cycles" events look like they were recorded b <<
937 (ie., the default). Note that even with CYC pa <<
938 these are not fully accurate, since CYC packet <<
939 instruction, only when some other event (like <<
940 TNT packet representing multiple branches) hap <<
941 be emitted. Thus, it is more effective for att <<
942 (and possibly basic blocks) than to individual <<
943 is not even perfect for functions (although it <<
944 option is active). <<
945 <<
946 "Branches" events look like they were recorded 922 "Branches" events look like they were recorded by "perf record -e branches". "c"
947 and "r" can be combined to get calls and retur 923 and "r" can be combined to get calls and returns.
948 924
949 "Transactions" events correspond to the start 925 "Transactions" events correspond to the start or end of transactions. The
950 'flags' field can be used in perf script to de 926 'flags' field can be used in perf script to determine whether the event is a
951 transaction start, commit or abort. 927 transaction start, commit or abort.
952 928
953 Note that "instructions", "cycles", "branches" !! 929 Note that "instructions", "branches" and "transactions" events depend on code
954 depend on code flow packets which can be disab !! 930 flow packets which can be disabled by using the config term "branch=0". Refer
955 "branch=0". Refer to the config terms section !! 931 to the config terms section above.
956 932
957 "ptwrite" events record the payload of the ptw 933 "ptwrite" events record the payload of the ptwrite instruction and whether
958 "fup_on_ptw" was used. "ptwrite" events depen 934 "fup_on_ptw" was used. "ptwrite" events depend on PTWRITE packets which are
959 recorded only if the "ptw" config term was use 935 recorded only if the "ptw" config term was used. Refer to the config terms
960 section above. perf script "synth" field disp 936 section above. perf script "synth" field displays "ptwrite" information like
961 this: "ip: 0 payload: 0x123456789abcdef0" whe 937 this: "ip: 0 payload: 0x123456789abcdef0" where "ip" is 1 if "fup_on_ptw" was
962 used. 938 used.
963 939
964 "Power" events correspond to power event packe 940 "Power" events correspond to power event packets and CBR (core-to-bus ratio)
965 packets. While CBR packets are always recorde 941 packets. While CBR packets are always recorded when tracing is enabled, power
966 event packets are recorded only if the "pwr_ev 942 event packets are recorded only if the "pwr_evt" config term was used. Refer to
967 the config terms section above. The power eve 943 the config terms section above. The power events record information about
968 C-state changes, whereas CBR is indicative of 944 C-state changes, whereas CBR is indicative of CPU frequency. perf script
969 "event,synth" fields display information like 945 "event,synth" fields display information like this:
970 946
971 cbr: cbr: 22 freq: 2189 MHz (200%) 947 cbr: cbr: 22 freq: 2189 MHz (200%)
972 mwait: hints: 0x60 extensions: 0x1 948 mwait: hints: 0x60 extensions: 0x1
973 pwre: hw: 0 cstate: 2 sub-cstate: 0 949 pwre: hw: 0 cstate: 2 sub-cstate: 0
974 exstop: ip: 1 950 exstop: ip: 1
975 pwrx: deepest cstate: 2 last cstate: 951 pwrx: deepest cstate: 2 last cstate: 2 wake reason: 0x4
976 952
977 Where: 953 Where:
978 954
979 "cbr" includes the frequency and the p 955 "cbr" includes the frequency and the percentage of maximum non-turbo
980 "mwait" shows mwait hints and extensio 956 "mwait" shows mwait hints and extensions
981 "pwre" shows C-state transitions (to a 957 "pwre" shows C-state transitions (to a C-state deeper than C0) and
982 whether initiated by hardware 958 whether initiated by hardware
983 "exstop" indicates execution stopped a 959 "exstop" indicates execution stopped and whether the IP was recorded
984 exactly, 960 exactly,
985 "pwrx" indicates return to C0 961 "pwrx" indicates return to C0
986 962
987 For more details refer to the Intel 64 and IA- 963 For more details refer to the Intel 64 and IA-32 Architectures Software
988 Developer Manuals. 964 Developer Manuals.
989 965
990 PSB events show when a PSB+ occurred and also 966 PSB events show when a PSB+ occurred and also the byte-offset in the trace.
991 Emitting a PSB+ can cause a CPU a slight delay 967 Emitting a PSB+ can cause a CPU a slight delay. When doing timing analysis
992 of code with Intel PT, it is useful to know if 968 of code with Intel PT, it is useful to know if a timing bubble was caused
993 by Intel PT or not. 969 by Intel PT or not.
994 970
995 Error events show where the decoder lost the t 971 Error events show where the decoder lost the trace. Error events
996 are quite important. Users must know if what 972 are quite important. Users must know if what they are seeing is a complete
997 picture or not. The "e" option may be followed 973 picture or not. The "e" option may be followed by flags which affect what errors
998 will or will not be reported. Each flag must 974 will or will not be reported. Each flag must be preceded by either '+' or '-'.
999 The flags supported by Intel PT are: 975 The flags supported by Intel PT are:
1000 976
1001 -o Suppress overflow err 977 -o Suppress overflow errors
1002 -l Suppress trace data l 978 -l Suppress trace data lost errors
1003 979
1004 For example, for errors but not overflow or d 980 For example, for errors but not overflow or data lost errors:
1005 981
1006 --itrace=e-o-l 982 --itrace=e-o-l
1007 983
1008 The "d" option will cause the creation of a f 984 The "d" option will cause the creation of a file "intel_pt.log" containing all
1009 decoded packets and instructions. Note that 985 decoded packets and instructions. Note that this option slows down the decoder
1010 and that the resulting file may be very large 986 and that the resulting file may be very large. The "d" option may be followed
1011 by flags which affect what debug messages wil 987 by flags which affect what debug messages will or will not be logged. Each flag
1012 must be preceded by either '+' or '-'. The fl 988 must be preceded by either '+' or '-'. The flags support by Intel PT are:
1013 989
1014 -a Suppress logging of p 990 -a Suppress logging of perf events
1015 +a Log all perf events 991 +a Log all perf events
1016 +e Output only on decodi 992 +e Output only on decoding errors (size configurable)
1017 +o Output to stdout inst 993 +o Output to stdout instead of "intel_pt.log"
1018 994
1019 By default, logged perf events are filtered b 995 By default, logged perf events are filtered by any specified time ranges, but
1020 flag +a overrides that. The +e flag can be u 996 flag +a overrides that. The +e flag can be useful for analyzing errors. By
1021 default, the log size in that case is 16384 b 997 default, the log size in that case is 16384 bytes, but can be altered by
1022 linkperf:perf-config[1] e.g. perf config itra 998 linkperf:perf-config[1] e.g. perf config itrace.debug-log-buffer-size=30000
1023 999
1024 In addition, the period of the "instructions" 1000 In addition, the period of the "instructions" event can be specified. e.g.
1025 1001
1026 --itrace=i10us 1002 --itrace=i10us
1027 1003
1028 sets the period to 10us i.e. one instruction 1004 sets the period to 10us i.e. one instruction sample is synthesized for each 10
1029 microseconds of trace. Alternatives to "us" 1005 microseconds of trace. Alternatives to "us" are "ms" (milliseconds),
1030 "ns" (nanoseconds), "t" (TSC ticks) or "i" (i 1006 "ns" (nanoseconds), "t" (TSC ticks) or "i" (instructions).
1031 1007
1032 "ms", "us" and "ns" are converted to TSC tick 1008 "ms", "us" and "ns" are converted to TSC ticks.
1033 1009
1034 The timing information included with Intel PT 1010 The timing information included with Intel PT does not give the time of every
1035 instruction. Consequently, for the purpose o 1011 instruction. Consequently, for the purpose of sampling, the decoder estimates
1036 the time since the last timing packet based o 1012 the time since the last timing packet based on 1 tick per instruction. The time
1037 on the sample is *not* adjusted and reflects 1013 on the sample is *not* adjusted and reflects the last known value of TSC.
1038 1014
1039 For Intel PT, the default period is 100us. 1015 For Intel PT, the default period is 100us.
1040 1016
1041 Setting it to a zero period means "as often a 1017 Setting it to a zero period means "as often as possible".
1042 1018
1043 In the case of Intel PT that is the same as a 1019 In the case of Intel PT that is the same as a period of 1 and a unit of
1044 'instructions' (i.e. --itrace=i1i). 1020 'instructions' (i.e. --itrace=i1i).
1045 1021
1046 Also the call chain size (default 16, max. 10 1022 Also the call chain size (default 16, max. 1024) for instructions or
1047 transactions events can be specified. e.g. 1023 transactions events can be specified. e.g.
1048 1024
1049 --itrace=ig32 1025 --itrace=ig32
1050 --itrace=xg32 1026 --itrace=xg32
1051 1027
1052 Also the number of last branch entries (defau 1028 Also the number of last branch entries (default 64, max. 1024) for instructions or
1053 transactions events can be specified. e.g. 1029 transactions events can be specified. e.g.
1054 1030
1055 --itrace=il10 1031 --itrace=il10
1056 --itrace=xl10 1032 --itrace=xl10
1057 1033
1058 Note that last branch entries are cleared for 1034 Note that last branch entries are cleared for each sample, so there is no overlap
1059 from one sample to the next. 1035 from one sample to the next.
1060 1036
1061 The G and L options are designed in particula 1037 The G and L options are designed in particular for sample mode, and work much
1062 like g and l but add call chain and branch st 1038 like g and l but add call chain and branch stack to the other selected events
1063 instead of synthesized events. For example, t 1039 instead of synthesized events. For example, to record branch-misses events for
1064 'ls' and then add a call chain derived from t 1040 'ls' and then add a call chain derived from the Intel PT trace:
1065 1041
1066 perf record --aux-sample -e '{intel_p 1042 perf record --aux-sample -e '{intel_pt//u,branch-misses:u}' -- ls
1067 perf report --itrace=Ge 1043 perf report --itrace=Ge
1068 1044
1069 Although in fact G is a default for perf repo 1045 Although in fact G is a default for perf report, so that is the same as just:
1070 1046
1071 perf report 1047 perf report
1072 1048
1073 One caveat with the G and L options is that t 1049 One caveat with the G and L options is that they work poorly with "Large PEBS".
1074 Large PEBS means PEBS records will be accumul 1050 Large PEBS means PEBS records will be accumulated by hardware and the written
1075 into the event buffer in one go. That reduce 1051 into the event buffer in one go. That reduces interrupts, but can give very
1076 late timestamps. Because the Intel PT trace 1052 late timestamps. Because the Intel PT trace is synchronized by timestamps,
1077 the PEBS events do not match the trace. Curr 1053 the PEBS events do not match the trace. Currently, Large PEBS is used only in
1078 certain circumstances: 1054 certain circumstances:
1079 - hardware supports it 1055 - hardware supports it
1080 - PEBS is used 1056 - PEBS is used
1081 - event period is specified, instead 1057 - event period is specified, instead of frequency
1082 - the sample type is limited to the f 1058 - the sample type is limited to the following flags:
1083 PERF_SAMPLE_IP | PERF_SAMPLE_ 1059 PERF_SAMPLE_IP | PERF_SAMPLE_TID | PERF_SAMPLE_ADDR |
1084 PERF_SAMPLE_ID | PERF_SAMPLE_ 1060 PERF_SAMPLE_ID | PERF_SAMPLE_CPU | PERF_SAMPLE_STREAM_ID |
1085 PERF_SAMPLE_DATA_SRC | PERF_S 1061 PERF_SAMPLE_DATA_SRC | PERF_SAMPLE_IDENTIFIER |
1086 PERF_SAMPLE_TRANSACTION | PER 1062 PERF_SAMPLE_TRANSACTION | PERF_SAMPLE_PHYS_ADDR |
1087 PERF_SAMPLE_REGS_INTR | PERF_ 1063 PERF_SAMPLE_REGS_INTR | PERF_SAMPLE_REGS_USER |
1088 PERF_SAMPLE_PERIOD (and somet 1064 PERF_SAMPLE_PERIOD (and sometimes) | PERF_SAMPLE_TIME
1089 Because Intel PT sample mode uses a different 1065 Because Intel PT sample mode uses a different sample type to the list above,
1090 Large PEBS is not used with Intel PT sample m 1066 Large PEBS is not used with Intel PT sample mode. To avoid Large PEBS in other
1091 cases, avoid specifying the event period i.e. 1067 cases, avoid specifying the event period i.e. avoid the 'perf record' -c option,
1092 --count option, or 'period' config term. 1068 --count option, or 'period' config term.
1093 1069
1094 To disable trace decoding entirely, use the o 1070 To disable trace decoding entirely, use the option --no-itrace.
1095 1071
1096 It is also possible to skip events generated 1072 It is also possible to skip events generated (instructions, branches, transactions)
1097 at the beginning. This is useful to ignore in 1073 at the beginning. This is useful to ignore initialization code.
1098 1074
1099 --itrace=i0nss1000000 1075 --itrace=i0nss1000000
1100 1076
1101 skips the first million instructions. 1077 skips the first million instructions.
1102 1078
1103 The q option changes the way the trace is dec 1079 The q option changes the way the trace is decoded. The decoding is much faster
1104 but much less detailed. Specifically, with t 1080 but much less detailed. Specifically, with the q option, the decoder does not
1105 decode TNT packets, and does not walk object 1081 decode TNT packets, and does not walk object code, but gets the ip from FUP and
1106 TIP packets. The q option can be used with t 1082 TIP packets. The q option can be used with the b and i options but the period
1107 is not used. The q option decodes more quick 1083 is not used. The q option decodes more quickly, but is useful only if the
1108 control flow of interest is represented or in 1084 control flow of interest is represented or indicated by FUP, TIP, TIP.PGE, or
1109 TIP.PGD packets (refer below). However the q 1085 TIP.PGD packets (refer below). However the q option could be used to find time
1110 ranges that could then be decoded fully using 1086 ranges that could then be decoded fully using the --time option.
1111 1087
1112 What will *not* be decoded with the (single) 1088 What will *not* be decoded with the (single) q option:
1113 1089
1114 - direct calls and jmps 1090 - direct calls and jmps
1115 - conditional branches 1091 - conditional branches
1116 - non-branch instructions 1092 - non-branch instructions
1117 1093
1118 What *will* be decoded with the (single) q op 1094 What *will* be decoded with the (single) q option:
1119 1095
1120 - asynchronous branches such as inter 1096 - asynchronous branches such as interrupts
1121 - indirect branches 1097 - indirect branches
1122 - function return target address *if* 1098 - function return target address *if* the noretcomp config term (refer
1123 config terms section) was used 1099 config terms section) was used
1124 - start of (control-flow) tracing 1100 - start of (control-flow) tracing
1125 - end of (control-flow) tracing, if i 1101 - end of (control-flow) tracing, if it is not out of context
1126 - power events, ptwrite, transaction 1102 - power events, ptwrite, transaction start and abort
1127 - instruction pointer associated with 1103 - instruction pointer associated with PSB packets
1128 1104
1129 Note the q option does not specify what event 1105 Note the q option does not specify what events will be synthesized e.g. the p
1130 option must be used also to show power events 1106 option must be used also to show power events.
1131 1107
1132 Repeating the q option (double-q i.e. qq) res 1108 Repeating the q option (double-q i.e. qq) results in even faster decoding and even
1133 less detail. The decoder decodes only extend 1109 less detail. The decoder decodes only extended PSB (PSB+) packets, getting the
1134 instruction pointer if there is a FUP packet 1110 instruction pointer if there is a FUP packet within PSB+ (i.e. between PSB and
1135 PSBEND). Note PSB packets occur regularly in 1111 PSBEND). Note PSB packets occur regularly in the trace based on the psb_period
1136 config term (refer config terms section). Th 1112 config term (refer config terms section). There will be a FUP packet if the
1137 PSB+ occurs while control flow is being trace 1113 PSB+ occurs while control flow is being traced.
1138 1114
1139 What will *not* be decoded with the qq option 1115 What will *not* be decoded with the qq option:
1140 1116
1141 - everything except instruction point 1117 - everything except instruction pointer associated with PSB packets
1142 1118
1143 What *will* be decoded with the qq option: 1119 What *will* be decoded with the qq option:
1144 1120
1145 - instruction pointer associated with 1121 - instruction pointer associated with PSB packets
1146 1122
1147 The Z option is equivalent to having recorded 1123 The Z option is equivalent to having recorded a trace without TSC
1148 (i.e. config term tsc=0). It can be useful to 1124 (i.e. config term tsc=0). It can be useful to avoid timestamp issues when
1149 decoding a trace of a virtual machine. 1125 decoding a trace of a virtual machine.
1150 1126
1151 1127
1152 dlfilter-show-cycles.so 1128 dlfilter-show-cycles.so
1153 ~~~~~~~~~~~~~~~~~~~~~~~ 1129 ~~~~~~~~~~~~~~~~~~~~~~~
1154 1130
1155 Cycles can be displayed using dlfilter-show-c 1131 Cycles can be displayed using dlfilter-show-cycles.so in which case the itrace A
1156 option can be useful to provide higher granul 1132 option can be useful to provide higher granularity cycle information:
1157 1133
1158 perf script --itrace=A --call-trace - 1134 perf script --itrace=A --call-trace --dlfilter dlfilter-show-cycles.so
1159 1135
1160 To see a list of dlfilters: 1136 To see a list of dlfilters:
1161 1137
1162 perf script -v --list-dlfilters 1138 perf script -v --list-dlfilters
1163 1139
1164 See also linkperf:perf-dlfilters[1] 1140 See also linkperf:perf-dlfilters[1]
1165 1141
1166 1142
1167 dump option 1143 dump option
1168 ~~~~~~~~~~~ 1144 ~~~~~~~~~~~
1169 1145
1170 perf script has an option (-D) to "dump" the 1146 perf script has an option (-D) to "dump" the events i.e. display the binary
1171 data. 1147 data.
1172 1148
1173 When -D is used, Intel PT packets are display 1149 When -D is used, Intel PT packets are displayed. The packet decoder does not
1174 pay attention to PSB packets, but just decode 1150 pay attention to PSB packets, but just decodes the bytes - so the packets seen
1175 by the actual decoder may not be identical in 1151 by the actual decoder may not be identical in places where the data is corrupt.
1176 One example of that would be when the buffer- 1152 One example of that would be when the buffer-switching interrupt has been too
1177 slow, and the buffer has been filled complete 1153 slow, and the buffer has been filled completely. In that case, the last packet
1178 in the buffer might be truncated and immediat 1154 in the buffer might be truncated and immediately followed by a PSB as the trace
1179 continues in the next buffer. 1155 continues in the next buffer.
1180 1156
1181 To disable the display of Intel PT packets, c 1157 To disable the display of Intel PT packets, combine the -D option with
1182 --no-itrace. 1158 --no-itrace.
1183 1159
1184 1160
1185 perf report 1161 perf report
1186 ----------- 1162 -----------
1187 1163
1188 By default, perf report will decode trace dat 1164 By default, perf report will decode trace data found in the perf.data file.
1189 This can be further controlled by new option 1165 This can be further controlled by new option --itrace exactly the same as
1190 perf script, with the exception that the defa 1166 perf script, with the exception that the default is --itrace=igxe.
1191 1167
1192 1168
1193 perf inject 1169 perf inject
1194 ----------- 1170 -----------
1195 1171
1196 perf inject also accepts the --itrace option 1172 perf inject also accepts the --itrace option in which case tracing data is
1197 removed and replaced with the synthesized eve 1173 removed and replaced with the synthesized events. e.g.
1198 1174
1199 perf inject --itrace -i perf.data -o 1175 perf inject --itrace -i perf.data -o perf.data.new
1200 1176
1201 Below is an example of using Intel PT with au 1177 Below is an example of using Intel PT with autofdo. It requires autofdo
1202 (https://github.com/google/autofdo) and gcc v 1178 (https://github.com/google/autofdo) and gcc version 5. The bubble
1203 sort example is from the AutoFDO tutorial (ht 1179 sort example is from the AutoFDO tutorial (https://gcc.gnu.org/wiki/AutoFDO/Tutorial)
1204 amended to take the number of elements as a p 1180 amended to take the number of elements as a parameter.
1205 1181
1206 $ gcc-5 -O3 sort.c -o sort_optimized 1182 $ gcc-5 -O3 sort.c -o sort_optimized
1207 $ ./sort_optimized 30000 1183 $ ./sort_optimized 30000
1208 Bubble sorting array of 30000 element 1184 Bubble sorting array of 30000 elements
1209 2254 ms 1185 2254 ms
1210 1186
1211 $ cat ~/.perfconfig 1187 $ cat ~/.perfconfig
1212 [intel-pt] 1188 [intel-pt]
1213 mispred-all = on 1189 mispred-all = on
1214 1190
1215 $ perf record -e intel_pt//u ./sort 3 1191 $ perf record -e intel_pt//u ./sort 3000
1216 Bubble sorting array of 3000 elements 1192 Bubble sorting array of 3000 elements
1217 58 ms 1193 58 ms
1218 [ perf record: Woken up 2 times to wr 1194 [ perf record: Woken up 2 times to write data ]
1219 [ perf record: Captured and wrote 3.9 1195 [ perf record: Captured and wrote 3.939 MB perf.data ]
1220 $ perf inject -i perf.data -o inj --i 1196 $ perf inject -i perf.data -o inj --itrace=i100usle --strip
1221 $ ./create_gcov --binary=./sort --pro 1197 $ ./create_gcov --binary=./sort --profile=inj --gcov=sort.gcov -gcov_version=1
1222 $ gcc-5 -O3 -fauto-profile=sort.gcov 1198 $ gcc-5 -O3 -fauto-profile=sort.gcov sort.c -o sort_autofdo
1223 $ ./sort_autofdo 30000 1199 $ ./sort_autofdo 30000
1224 Bubble sorting array of 30000 element 1200 Bubble sorting array of 30000 elements
1225 2155 ms 1201 2155 ms
1226 1202
1227 Note there is currently no advantage to using 1203 Note there is currently no advantage to using Intel PT instead of LBR, but
1228 that may change in the future if greater use 1204 that may change in the future if greater use is made of the data.
1229 1205
1230 1206
1231 PEBS via Intel PT 1207 PEBS via Intel PT
1232 ----------------- 1208 -----------------
1233 1209
1234 Some hardware has the feature to redirect PEB 1210 Some hardware has the feature to redirect PEBS records to the Intel PT trace.
1235 Recording is selected by using the aux-output 1211 Recording is selected by using the aux-output config term e.g.
1236 1212
1237 perf record -c 10000 -e '{intel_pt/br 1213 perf record -c 10000 -e '{intel_pt/branch=0/,cycles/aux-output/ppp}' uname
1238 1214
1239 Originally, software only supported redirecti 1215 Originally, software only supported redirecting at most one PEBS event because it
1240 was not able to differentiate one event from 1216 was not able to differentiate one event from another. To overcome that, more recent
1241 kernels and perf tools add support for the PE 1217 kernels and perf tools add support for the PERF_RECORD_AUX_OUTPUT_HW_ID side-band event.
1242 To check for the presence of that event in a 1218 To check for the presence of that event in a PEBS-via-PT trace:
1243 1219
1244 perf script -D --no-itrace | grep PER 1220 perf script -D --no-itrace | grep PERF_RECORD_AUX_OUTPUT_HW_ID
1245 1221
1246 To display PEBS events from the Intel PT trac 1222 To display PEBS events from the Intel PT trace, use the itrace 'o' option e.g.
1247 1223
1248 perf script --itrace=oe 1224 perf script --itrace=oe
1249 1225
1250 XED 1226 XED
1251 --- 1227 ---
1252 1228
1253 include::build-xed.txt[] 1229 include::build-xed.txt[]
1254 1230
1255 1231
1256 Tracing Virtual Machines (kernel only) 1232 Tracing Virtual Machines (kernel only)
1257 -------------------------------------- 1233 --------------------------------------
1258 1234
1259 Currently, kernel tracing is supported with e 1235 Currently, kernel tracing is supported with either "timeless" decoding
1260 (i.e. no TSC timestamps) or VM Time Correlati 1236 (i.e. no TSC timestamps) or VM Time Correlation. VM Time Correlation is an extra step
1261 using 'perf inject' and requires unchanging V 1237 using 'perf inject' and requires unchanging VMX TSC Offset and no VMX TSC Scaling.
1262 1238
1263 Other limitations and caveats 1239 Other limitations and caveats
1264 1240
1265 VMX controls may suppress packets needed for 1241 VMX controls may suppress packets needed for decoding resulting in decoding errors
1266 VMX controls may block the perf NMI to the h 1242 VMX controls may block the perf NMI to the host potentially resulting in lost trace data
1267 Guest kernel self-modifying code (e.g. jump 1243 Guest kernel self-modifying code (e.g. jump labels or JIT-compiled eBPF) will result in decoding errors
1268 Guest thread information is unknown 1244 Guest thread information is unknown
1269 Guest VCPU is unknown but may be able to be 1245 Guest VCPU is unknown but may be able to be inferred from the host thread
1270 Callchains are not supported 1246 Callchains are not supported
1271 1247
1272 Example using "timeless" decoding 1248 Example using "timeless" decoding
1273 1249
1274 Start VM 1250 Start VM
1275 1251
1276 $ sudo virsh start kubuntu20.04 1252 $ sudo virsh start kubuntu20.04
1277 Domain kubuntu20.04 started 1253 Domain kubuntu20.04 started
1278 1254
1279 Mount the guest file system. Note sshfs need 1255 Mount the guest file system. Note sshfs needs -o direct_io to enable reading of proc files. root access is needed to read /proc/kcore.
1280 1256
1281 $ mkdir vm0 1257 $ mkdir vm0
1282 $ sshfs -o direct_io root@vm0:/ vm0 1258 $ sshfs -o direct_io root@vm0:/ vm0
1283 1259
1284 Copy the guest /proc/kallsyms, /proc/modules 1260 Copy the guest /proc/kallsyms, /proc/modules and /proc/kcore
1285 1261
1286 $ perf buildid-cache -v --kcore vm0/proc/kco 1262 $ perf buildid-cache -v --kcore vm0/proc/kcore
1287 kcore added to build-id cache directory /hom 1263 kcore added to build-id cache directory /home/user/.debug/[kernel.kcore]/9600f316a53a0f54278885e8d9710538ec5f6a08/2021021807494306
1288 $ KALLSYMS=/home/user/.debug/[kernel.kcore]/ 1264 $ KALLSYMS=/home/user/.debug/[kernel.kcore]/9600f316a53a0f54278885e8d9710538ec5f6a08/2021021807494306/kallsyms
1289 1265
1290 Find the VM process 1266 Find the VM process
1291 1267
1292 $ ps -eLl | grep 'KVM\|PID' 1268 $ ps -eLl | grep 'KVM\|PID'
1293 F S UID PID PPID LWP C PRI NI 1269 F S UID PID PPID LWP C PRI NI ADDR SZ WCHAN TTY TIME CMD
1294 3 S 64055 1430 1 1440 1 80 0 1270 3 S 64055 1430 1 1440 1 80 0 - 1921718 - ? 00:02:47 CPU 0/KVM
1295 3 S 64055 1430 1 1441 1 80 0 1271 3 S 64055 1430 1 1441 1 80 0 - 1921718 - ? 00:02:41 CPU 1/KVM
1296 3 S 64055 1430 1 1442 1 80 0 1272 3 S 64055 1430 1 1442 1 80 0 - 1921718 - ? 00:02:38 CPU 2/KVM
1297 3 S 64055 1430 1 1443 2 80 0 1273 3 S 64055 1430 1 1443 2 80 0 - 1921718 - ? 00:03:18 CPU 3/KVM
1298 1274
1299 Start an open-ended perf record, tracing the 1275 Start an open-ended perf record, tracing the VM process, do something on the VM, and then ctrl-C to stop.
1300 TSC is not supported and tsc=0 must be specif 1276 TSC is not supported and tsc=0 must be specified. That means mtc is useless, so add mtc=0.
1301 However, IPC can still be determined, hence c 1277 However, IPC can still be determined, hence cyc=1 can be added.
1302 Only kernel decoding is supported, so 'k' mus 1278 Only kernel decoding is supported, so 'k' must be specified.
1303 Intel PT traces both the host and the guest s 1279 Intel PT traces both the host and the guest so --guest and --host need to be specified.
1304 Without timestamps, --per-thread must be spec 1280 Without timestamps, --per-thread must be specified to distinguish threads.
1305 1281
1306 $ sudo perf kvm --guest --host --guestkallsy 1282 $ sudo perf kvm --guest --host --guestkallsyms $KALLSYMS record --kcore -e intel_pt/tsc=0,mtc=0,cyc=1/k -p 1430 --per-thread
1307 ^C 1283 ^C
1308 [ perf record: Woken up 1 times to write dat 1284 [ perf record: Woken up 1 times to write data ]
1309 [ perf record: Captured and wrote 5.829 MB ] 1285 [ perf record: Captured and wrote 5.829 MB ]
1310 1286
1311 perf script can be used to provide an instruc 1287 perf script can be used to provide an instruction trace
1312 1288
1313 $ perf script --guestkallsyms $KALLSYMS --in !! 1289 $ perf script --guestkallsyms $KALLSYMS --insn-trace --xed -F+ipc | grep -C10 vmresume | head -21
1314 CPU 0/KVM 1440 ffffffff82133cdd __vm 1290 CPU 0/KVM 1440 ffffffff82133cdd __vmx_vcpu_run+0x3d ([kernel.kallsyms]) movq 0x48(%rax), %r9
1315 CPU 0/KVM 1440 ffffffff82133ce1 __vm 1291 CPU 0/KVM 1440 ffffffff82133ce1 __vmx_vcpu_run+0x41 ([kernel.kallsyms]) movq 0x50(%rax), %r10
1316 CPU 0/KVM 1440 ffffffff82133ce5 __vm 1292 CPU 0/KVM 1440 ffffffff82133ce5 __vmx_vcpu_run+0x45 ([kernel.kallsyms]) movq 0x58(%rax), %r11
1317 CPU 0/KVM 1440 ffffffff82133ce9 __vm 1293 CPU 0/KVM 1440 ffffffff82133ce9 __vmx_vcpu_run+0x49 ([kernel.kallsyms]) movq 0x60(%rax), %r12
1318 CPU 0/KVM 1440 ffffffff82133ced __vm 1294 CPU 0/KVM 1440 ffffffff82133ced __vmx_vcpu_run+0x4d ([kernel.kallsyms]) movq 0x68(%rax), %r13
1319 CPU 0/KVM 1440 ffffffff82133cf1 __vm 1295 CPU 0/KVM 1440 ffffffff82133cf1 __vmx_vcpu_run+0x51 ([kernel.kallsyms]) movq 0x70(%rax), %r14
1320 CPU 0/KVM 1440 ffffffff82133cf5 __vm 1296 CPU 0/KVM 1440 ffffffff82133cf5 __vmx_vcpu_run+0x55 ([kernel.kallsyms]) movq 0x78(%rax), %r15
1321 CPU 0/KVM 1440 ffffffff82133cf9 __vm 1297 CPU 0/KVM 1440 ffffffff82133cf9 __vmx_vcpu_run+0x59 ([kernel.kallsyms]) movq (%rax), %rax
1322 CPU 0/KVM 1440 ffffffff82133cfc __vm 1298 CPU 0/KVM 1440 ffffffff82133cfc __vmx_vcpu_run+0x5c ([kernel.kallsyms]) callq 0xffffffff82133c40
1323 CPU 0/KVM 1440 ffffffff82133c40 vmx_ 1299 CPU 0/KVM 1440 ffffffff82133c40 vmx_vmenter+0x0 ([kernel.kallsyms]) jz 0xffffffff82133c46
1324 CPU 0/KVM 1440 ffffffff82133c42 vmx_ 1300 CPU 0/KVM 1440 ffffffff82133c42 vmx_vmenter+0x2 ([kernel.kallsyms]) vmresume IPC: 0.11 (50/445)
1325 :1440 1440 ffffffffbb678b06 nati 1301 :1440 1440 ffffffffbb678b06 native_write_msr+0x6 ([guest.kernel.kallsyms]) nopl %eax, (%rax,%rax,1)
1326 :1440 1440 ffffffffbb678b0b nati 1302 :1440 1440 ffffffffbb678b0b native_write_msr+0xb ([guest.kernel.kallsyms]) retq IPC: 0.04 (2/41)
1327 :1440 1440 ffffffffbb666646 lapi 1303 :1440 1440 ffffffffbb666646 lapic_next_deadline+0x26 ([guest.kernel.kallsyms]) data16 nop
1328 :1440 1440 ffffffffbb666648 lapi 1304 :1440 1440 ffffffffbb666648 lapic_next_deadline+0x28 ([guest.kernel.kallsyms]) xor %eax, %eax
1329 :1440 1440 ffffffffbb66664a lapi 1305 :1440 1440 ffffffffbb66664a lapic_next_deadline+0x2a ([guest.kernel.kallsyms]) popq %rbp
1330 :1440 1440 ffffffffbb66664b lapi 1306 :1440 1440 ffffffffbb66664b lapic_next_deadline+0x2b ([guest.kernel.kallsyms]) retq IPC: 0.16 (4/25)
1331 :1440 1440 ffffffffbb74607f cloc 1307 :1440 1440 ffffffffbb74607f clockevents_program_event+0x8f ([guest.kernel.kallsyms]) test %eax, %eax
1332 :1440 1440 ffffffffbb746081 cloc 1308 :1440 1440 ffffffffbb746081 clockevents_program_event+0x91 ([guest.kernel.kallsyms]) jz 0xffffffffbb74603c IPC: 0.06 (2/30)
1333 :1440 1440 ffffffffbb74603c cloc 1309 :1440 1440 ffffffffbb74603c clockevents_program_event+0x4c ([guest.kernel.kallsyms]) popq %rbx
1334 :1440 1440 ffffffffbb74603d cloc 1310 :1440 1440 ffffffffbb74603d clockevents_program_event+0x4d ([guest.kernel.kallsyms]) popq %r12
1335 1311
1336 Example using VM Time Correlation 1312 Example using VM Time Correlation
1337 1313
1338 Start VM 1314 Start VM
1339 1315
1340 $ sudo virsh start kubuntu20.04 1316 $ sudo virsh start kubuntu20.04
1341 Domain kubuntu20.04 started 1317 Domain kubuntu20.04 started
1342 1318
1343 Mount the guest file system. Note sshfs need 1319 Mount the guest file system. Note sshfs needs -o direct_io to enable reading of proc files. root access is needed to read /proc/kcore.
1344 1320
1345 $ mkdir -p vm0 1321 $ mkdir -p vm0
1346 $ sshfs -o direct_io root@vm0:/ vm0 1322 $ sshfs -o direct_io root@vm0:/ vm0
1347 1323
1348 Copy the guest /proc/kallsyms, /proc/modules 1324 Copy the guest /proc/kallsyms, /proc/modules and /proc/kcore
1349 1325
1350 $ perf buildid-cache -v --kcore vm0/proc/kco 1326 $ perf buildid-cache -v --kcore vm0/proc/kcore
1351 same kcore found in /home/user/.debug/[kerne 1327 same kcore found in /home/user/.debug/[kernel.kcore]/cc9c55a98c5e4ec0aeda69302554aabed5cd6491/2021021312450777
1352 $ KALLSYMS=/home/user/.debug/\[kernel.kcore\ 1328 $ KALLSYMS=/home/user/.debug/\[kernel.kcore\]/cc9c55a98c5e4ec0aeda69302554aabed5cd6491/2021021312450777/kallsyms
1353 1329
1354 Find the VM process 1330 Find the VM process
1355 1331
1356 $ ps -eLl | grep 'KVM\|PID' 1332 $ ps -eLl | grep 'KVM\|PID'
1357 F S UID PID PPID LWP C PRI NI 1333 F S UID PID PPID LWP C PRI NI ADDR SZ WCHAN TTY TIME CMD
1358 3 S 64055 16998 1 17005 13 80 0 1334 3 S 64055 16998 1 17005 13 80 0 - 1818189 - ? 00:00:16 CPU 0/KVM
1359 3 S 64055 16998 1 17006 4 80 0 1335 3 S 64055 16998 1 17006 4 80 0 - 1818189 - ? 00:00:05 CPU 1/KVM
1360 3 S 64055 16998 1 17007 3 80 0 1336 3 S 64055 16998 1 17007 3 80 0 - 1818189 - ? 00:00:04 CPU 2/KVM
1361 3 S 64055 16998 1 17008 4 80 0 1337 3 S 64055 16998 1 17008 4 80 0 - 1818189 - ? 00:00:05 CPU 3/KVM
1362 1338
1363 Start an open-ended perf record, tracing the 1339 Start an open-ended perf record, tracing the VM process, do something on the VM, and then ctrl-C to stop.
1364 IPC can be determined, hence cyc=1 can be add 1340 IPC can be determined, hence cyc=1 can be added.
1365 Only kernel decoding is supported, so 'k' mus 1341 Only kernel decoding is supported, so 'k' must be specified.
1366 Intel PT traces both the host and the guest s 1342 Intel PT traces both the host and the guest so --guest and --host need to be specified.
1367 1343
1368 $ sudo perf kvm --guest --host --guestkallsy 1344 $ sudo perf kvm --guest --host --guestkallsyms $KALLSYMS record --kcore -e intel_pt/cyc=1/k -p 16998
1369 ^C[ perf record: Woken up 1 times to write d 1345 ^C[ perf record: Woken up 1 times to write data ]
1370 [ perf record: Captured and wrote 9.041 MB p 1346 [ perf record: Captured and wrote 9.041 MB perf.data.kvm ]
1371 1347
1372 Now 'perf inject' can be used to determine th 1348 Now 'perf inject' can be used to determine the VMX TCS Offset. Note, Intel PT TSC packets are
1373 only 7-bytes, so the TSC Offset might differ 1349 only 7-bytes, so the TSC Offset might differ from the actual value in the 8th byte. That will
1374 have no effect i.e. the resulting timestamps 1350 have no effect i.e. the resulting timestamps will be correct anyway.
1375 1351
1376 $ perf inject -i perf.data.kvm --vm-time-cor 1352 $ perf inject -i perf.data.kvm --vm-time-correlation=dry-run
1377 ERROR: Unknown TSC Offset for VMCS 0x1bff6a 1353 ERROR: Unknown TSC Offset for VMCS 0x1bff6a
1378 VMCS: 0x1bff6a TSC Offset 0xffffe42722c64c4 1354 VMCS: 0x1bff6a TSC Offset 0xffffe42722c64c41
1379 ERROR: Unknown TSC Offset for VMCS 0x1cbc08 1355 ERROR: Unknown TSC Offset for VMCS 0x1cbc08
1380 VMCS: 0x1cbc08 TSC Offset 0xffffe42722c64c4 1356 VMCS: 0x1cbc08 TSC Offset 0xffffe42722c64c41
1381 ERROR: Unknown TSC Offset for VMCS 0x1c3ce8 1357 ERROR: Unknown TSC Offset for VMCS 0x1c3ce8
1382 VMCS: 0x1c3ce8 TSC Offset 0xffffe42722c64c4 1358 VMCS: 0x1c3ce8 TSC Offset 0xffffe42722c64c41
1383 ERROR: Unknown TSC Offset for VMCS 0x1cbce9 1359 ERROR: Unknown TSC Offset for VMCS 0x1cbce9
1384 VMCS: 0x1cbce9 TSC Offset 0xffffe42722c64c4 1360 VMCS: 0x1cbce9 TSC Offset 0xffffe42722c64c41
1385 1361
1386 Each virtual CPU has a different Virtual Mach 1362 Each virtual CPU has a different Virtual Machine Control Structure (VMCS)
1387 shown above with the calculated TSC Offset. F 1363 shown above with the calculated TSC Offset. For an unchanging TSC Offset
1388 they should all be the same for the same virt 1364 they should all be the same for the same virtual machine.
1389 1365
1390 Now that the TSC Offset is known, it can be p 1366 Now that the TSC Offset is known, it can be provided to 'perf inject'
1391 1367
1392 $ perf inject -i perf.data.kvm --vm-time-cor 1368 $ perf inject -i perf.data.kvm --vm-time-correlation="dry-run 0xffffe42722c64c41"
1393 1369
1394 Note the options for 'perf inject' --vm-time- 1370 Note the options for 'perf inject' --vm-time-correlation are:
1395 1371
1396 [ dry-run ] [ <TSC Offset> [ : <VMCS> [ , <V 1372 [ dry-run ] [ <TSC Offset> [ : <VMCS> [ , <VMCS> ]... ] ]...
1397 1373
1398 So it is possible to specify different TSC Of 1374 So it is possible to specify different TSC Offsets for different VMCS.
1399 The option "dry-run" will cause the file to b 1375 The option "dry-run" will cause the file to be processed but without updating it.
1400 Note it is also possible to get a intel_pt.lo 1376 Note it is also possible to get a intel_pt.log file by adding option --itrace=d
1401 1377
1402 There were no errors so, do it for real 1378 There were no errors so, do it for real
1403 1379
1404 $ perf inject -i perf.data.kvm --vm-time-cor 1380 $ perf inject -i perf.data.kvm --vm-time-correlation=0xffffe42722c64c41 --force
1405 1381
1406 'perf script' can be used to see if there are 1382 'perf script' can be used to see if there are any decoder errors
1407 1383
1408 $ perf script -i perf.data.kvm --guestkallsy 1384 $ perf script -i perf.data.kvm --guestkallsyms $KALLSYMS --itrace=e-o
1409 1385
1410 There were none. 1386 There were none.
1411 1387
1412 'perf script' can be used to provide an instr 1388 'perf script' can be used to provide an instruction trace showing timestamps
1413 1389
1414 $ perf script -i perf.data.kvm --guestkallsy !! 1390 $ perf script -i perf.data.kvm --guestkallsyms $KALLSYMS --insn-trace --xed -F+ipc | grep -C10 vmresume | head -21
1415 CPU 1/KVM 17006 [001] 11500.262865593: 1391 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133cdd __vmx_vcpu_run+0x3d ([kernel.kallsyms]) movq 0x48(%rax), %r9
1416 CPU 1/KVM 17006 [001] 11500.262865593: 1392 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133ce1 __vmx_vcpu_run+0x41 ([kernel.kallsyms]) movq 0x50(%rax), %r10
1417 CPU 1/KVM 17006 [001] 11500.262865593: 1393 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133ce5 __vmx_vcpu_run+0x45 ([kernel.kallsyms]) movq 0x58(%rax), %r11
1418 CPU 1/KVM 17006 [001] 11500.262865593: 1394 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133ce9 __vmx_vcpu_run+0x49 ([kernel.kallsyms]) movq 0x60(%rax), %r12
1419 CPU 1/KVM 17006 [001] 11500.262865593: 1395 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133ced __vmx_vcpu_run+0x4d ([kernel.kallsyms]) movq 0x68(%rax), %r13
1420 CPU 1/KVM 17006 [001] 11500.262865593: 1396 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133cf1 __vmx_vcpu_run+0x51 ([kernel.kallsyms]) movq 0x70(%rax), %r14
1421 CPU 1/KVM 17006 [001] 11500.262865593: 1397 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133cf5 __vmx_vcpu_run+0x55 ([kernel.kallsyms]) movq 0x78(%rax), %r15
1422 CPU 1/KVM 17006 [001] 11500.262865593: 1398 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133cf9 __vmx_vcpu_run+0x59 ([kernel.kallsyms]) movq (%rax), %rax
1423 CPU 1/KVM 17006 [001] 11500.262865593: 1399 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133cfc __vmx_vcpu_run+0x5c ([kernel.kallsyms]) callq 0xffffffff82133c40
1424 CPU 1/KVM 17006 [001] 11500.262865593: 1400 CPU 1/KVM 17006 [001] 11500.262865593: ffffffff82133c40 vmx_vmenter+0x0 ([kernel.kallsyms]) jz 0xffffffff82133c46
1425 CPU 1/KVM 17006 [001] 11500.262866075: 1401 CPU 1/KVM 17006 [001] 11500.262866075: ffffffff82133c42 vmx_vmenter+0x2 ([kernel.kallsyms]) vmresume IPC: 0.05 (40/769)
1426 :17006 17006 [001] 11500.262869216: 1402 :17006 17006 [001] 11500.262869216: ffffffff82200cb0 asm_sysvec_apic_timer_interrupt+0x0 ([guest.kernel.kallsyms]) clac
1427 :17006 17006 [001] 11500.262869216: 1403 :17006 17006 [001] 11500.262869216: ffffffff82200cb3 asm_sysvec_apic_timer_interrupt+0x3 ([guest.kernel.kallsyms]) pushq $0xffffffffffffffff
1428 :17006 17006 [001] 11500.262869216: 1404 :17006 17006 [001] 11500.262869216: ffffffff82200cb5 asm_sysvec_apic_timer_interrupt+0x5 ([guest.kernel.kallsyms]) callq 0xffffffff82201160
1429 :17006 17006 [001] 11500.262869216: 1405 :17006 17006 [001] 11500.262869216: ffffffff82201160 error_entry+0x0 ([guest.kernel.kallsyms]) cld
1430 :17006 17006 [001] 11500.262869216: 1406 :17006 17006 [001] 11500.262869216: ffffffff82201161 error_entry+0x1 ([guest.kernel.kallsyms]) pushq %rsi
1431 :17006 17006 [001] 11500.262869216: 1407 :17006 17006 [001] 11500.262869216: ffffffff82201162 error_entry+0x2 ([guest.kernel.kallsyms]) movq 0x8(%rsp), %rsi
1432 :17006 17006 [001] 11500.262869216: 1408 :17006 17006 [001] 11500.262869216: ffffffff82201167 error_entry+0x7 ([guest.kernel.kallsyms]) movq %rdi, 0x8(%rsp)
1433 :17006 17006 [001] 11500.262869216: 1409 :17006 17006 [001] 11500.262869216: ffffffff8220116c error_entry+0xc ([guest.kernel.kallsyms]) pushq %rdx
1434 :17006 17006 [001] 11500.262869216: 1410 :17006 17006 [001] 11500.262869216: ffffffff8220116d error_entry+0xd ([guest.kernel.kallsyms]) pushq %rcx
1435 :17006 17006 [001] 11500.262869216: 1411 :17006 17006 [001] 11500.262869216: ffffffff8220116e error_entry+0xe ([guest.kernel.kallsyms]) pushq %rax
1436 1412
1437 1413
1438 Tracing Virtual Machines (including user spac 1414 Tracing Virtual Machines (including user space)
1439 --------------------------------------------- 1415 -----------------------------------------------
1440 1416
1441 It is possible to use perf record to record s 1417 It is possible to use perf record to record sideband events within a virtual machine, so that an Intel PT trace on the host can be decoded.
1442 Sideband events from the guest perf.data file 1418 Sideband events from the guest perf.data file can be injected into the host perf.data file using perf inject.
1443 1419
1444 Here is an example of the steps needed: 1420 Here is an example of the steps needed:
1445 1421
1446 On the guest machine: 1422 On the guest machine:
1447 1423
1448 Check that no-kvmclock kernel command line op 1424 Check that no-kvmclock kernel command line option was used to boot:
1449 1425
1450 Note, this is essential to enable time correl 1426 Note, this is essential to enable time correlation between host and guest machines.
1451 1427
1452 $ cat /proc/cmdline 1428 $ cat /proc/cmdline
1453 BOOT_IMAGE=/boot/vmlinuz-5.10.0-16-amd64 roo 1429 BOOT_IMAGE=/boot/vmlinuz-5.10.0-16-amd64 root=UUID=cb49c910-e573-47e0-bce7-79e293df8e1d ro no-kvmclock
1454 1430
1455 There is no BPF support at present so, if pos 1431 There is no BPF support at present so, if possible, disable JIT compiling:
1456 1432
1457 $ echo 0 | sudo tee /proc/sys/net/core/bpf_j 1433 $ echo 0 | sudo tee /proc/sys/net/core/bpf_jit_enable
1458 0 1434 0
1459 1435
1460 Start perf record to collect sideband events: 1436 Start perf record to collect sideband events:
1461 1437
1462 $ sudo perf record -o guest-sideband-testing 1438 $ sudo perf record -o guest-sideband-testing-guest-perf.data --sample-identifier --buildid-all --switch-events --kcore -a -e dummy
1463 1439
1464 On the host machine: 1440 On the host machine:
1465 1441
1466 Start perf record to collect Intel PT trace: 1442 Start perf record to collect Intel PT trace:
1467 1443
1468 Note, the host trace will get very big, very 1444 Note, the host trace will get very big, very fast, so the steps from starting to stopping the host trace really need to be done so that they happen in the shortest time possible.
1469 1445
1470 $ sudo perf record -o guest-sideband-testing 1446 $ sudo perf record -o guest-sideband-testing-host-perf.data -m,64M --kcore -a -e intel_pt/cyc/
1471 1447
1472 On the guest machine: 1448 On the guest machine:
1473 1449
1474 Run a small test case, just 'uname' in this e 1450 Run a small test case, just 'uname' in this example:
1475 1451
1476 $ uname 1452 $ uname
1477 Linux 1453 Linux
1478 1454
1479 On the host machine: 1455 On the host machine:
1480 1456
1481 Stop the Intel PT trace: 1457 Stop the Intel PT trace:
1482 1458
1483 ^C 1459 ^C
1484 [ perf record: Woken up 1 times to write dat 1460 [ perf record: Woken up 1 times to write data ]
1485 [ perf record: Captured and wrote 76.122 MB 1461 [ perf record: Captured and wrote 76.122 MB guest-sideband-testing-host-perf.data ]
1486 1462
1487 On the guest machine: 1463 On the guest machine:
1488 1464
1489 Stop the Intel PT trace: 1465 Stop the Intel PT trace:
1490 1466
1491 ^C 1467 ^C
1492 [ perf record: Woken up 1 times to write dat 1468 [ perf record: Woken up 1 times to write data ]
1493 [ perf record: Captured and wrote 1.247 MB g 1469 [ perf record: Captured and wrote 1.247 MB guest-sideband-testing-guest-perf.data ]
1494 1470
1495 And then copy guest-sideband-testing-guest-pe 1471 And then copy guest-sideband-testing-guest-perf.data to the host (not shown here).
1496 1472
1497 On the host machine: 1473 On the host machine:
1498 1474
1499 With the 2 perf.data recordings, and with the 1475 With the 2 perf.data recordings, and with their ownership changed to the user.
1500 1476
1501 Identify the TSC Offset: 1477 Identify the TSC Offset:
1502 1478
1503 $ perf inject -i guest-sideband-testing-host 1479 $ perf inject -i guest-sideband-testing-host-perf.data --vm-time-correlation=dry-run
1504 VMCS: 0x103fc6 TSC Offset 0xfffffa6ae070cb2 1480 VMCS: 0x103fc6 TSC Offset 0xfffffa6ae070cb20
1505 VMCS: 0x103ff2 TSC Offset 0xfffffa6ae070cb2 1481 VMCS: 0x103ff2 TSC Offset 0xfffffa6ae070cb20
1506 VMCS: 0x10fdaa TSC Offset 0xfffffa6ae070cb2 1482 VMCS: 0x10fdaa TSC Offset 0xfffffa6ae070cb20
1507 VMCS: 0x24d57c TSC Offset 0xfffffa6ae070cb2 1483 VMCS: 0x24d57c TSC Offset 0xfffffa6ae070cb20
1508 1484
1509 Correct Intel PT TSC timestamps for the guest 1485 Correct Intel PT TSC timestamps for the guest machine:
1510 1486
1511 $ perf inject -i guest-sideband-testing-host 1487 $ perf inject -i guest-sideband-testing-host-perf.data --vm-time-correlation=0xfffffa6ae070cb20 --force
1512 1488
1513 Identify the guest machine PID: 1489 Identify the guest machine PID:
1514 1490
1515 $ perf script -i guest-sideband-testing-host 1491 $ perf script -i guest-sideband-testing-host-perf.data --no-itrace --show-task-events | grep KVM
1516 CPU 0/KVM 0 [000] 0.000000: PE 1492 CPU 0/KVM 0 [000] 0.000000: PERF_RECORD_COMM: CPU 0/KVM:13376/13381
1517 CPU 1/KVM 0 [000] 0.000000: PE 1493 CPU 1/KVM 0 [000] 0.000000: PERF_RECORD_COMM: CPU 1/KVM:13376/13382
1518 CPU 2/KVM 0 [000] 0.000000: PE 1494 CPU 2/KVM 0 [000] 0.000000: PERF_RECORD_COMM: CPU 2/KVM:13376/13383
1519 CPU 3/KVM 0 [000] 0.000000: PE 1495 CPU 3/KVM 0 [000] 0.000000: PERF_RECORD_COMM: CPU 3/KVM:13376/13384
1520 1496
1521 Note, the QEMU option -name debug-threads=on 1497 Note, the QEMU option -name debug-threads=on is needed so that thread names
1522 can be used to determine which thread is runn 1498 can be used to determine which thread is running which VCPU as above. libvirt seems to use this by default.
1523 1499
1524 Create a guestmount, assuming the guest machi 1500 Create a guestmount, assuming the guest machine is 'vm_to_test':
1525 1501
1526 $ mkdir -p ~/guestmount/13376 1502 $ mkdir -p ~/guestmount/13376
1527 $ sshfs -o direct_io vm_to_test:/ ~/guestmou 1503 $ sshfs -o direct_io vm_to_test:/ ~/guestmount/13376
1528 1504
1529 Inject the guest perf.data file into the host 1505 Inject the guest perf.data file into the host perf.data file:
1530 1506
1531 Note, due to the guestmount option, guest obj 1507 Note, due to the guestmount option, guest object files and debug files will be copied into the build ID cache from the guest machine, with the notable exception of VDSO.
1532 If needed, VDSO can be copied manually in a f 1508 If needed, VDSO can be copied manually in a fashion similar to that used by the perf-archive script.
1533 1509
1534 $ perf inject -i guest-sideband-testing-host 1510 $ perf inject -i guest-sideband-testing-host-perf.data -o inj --guestmount ~/guestmount --guest-data=guest-sideband-testing-guest-perf.data,13376,0xfffffa6ae070cb20
1535 1511
1536 Show an excerpt from the result. In this cas 1512 Show an excerpt from the result. In this case the CPU and time range have been to chosen to show interaction between guest and host when 'uname' is starting to run on the guest machine:
1537 1513
1538 Notes: 1514 Notes:
1539 1515
1540 - the CPU displayed, [002] in this ca 1516 - the CPU displayed, [002] in this case, is always the host CPU
1541 - events happening in the virtual mac 1517 - events happening in the virtual machine start with VM:13376 VCPU:003, which shows the hypervisor PID 13376 and the VCPU number
1542 - only calls and errors are displayed 1518 - only calls and errors are displayed i.e. --itrace=ce
1543 - branches entering and exiting the v 1519 - branches entering and exiting the virtual machine are split, and show as 2 branches to/from "0 [unknown] ([unknown])"
1544 1520
1545 $ perf script -i inj --itrace=ce -F+machine_ 1521 $ perf script -i inj --itrace=ce -F+machine_pid,+vcpu,+addr,+pid,+tid,-period --ns --time 7919.408803365,7919.408804631 -C 2
1546 CPU 3/KVM 13376/13384 [002] 7919.4088 1522 CPU 3/KVM 13376/13384 [002] 7919.408803365: branches: ffffffffc0f8ebe0 vmx_vcpu_enter_exit+0xc0 ([kernel.kallsyms]) => ffffffffc0f8edc0 __vmx_vcpu_run+0x0 ([kernel.kallsyms])
1547 CPU 3/KVM 13376/13384 [002] 7919.4088 1523 CPU 3/KVM 13376/13384 [002] 7919.408803365: branches: ffffffffc0f8edd5 __vmx_vcpu_run+0x15 ([kernel.kallsyms]) => ffffffffc0f8eca0 vmx_update_host_rsp+0x0 ([kernel.kallsyms])
1548 CPU 3/KVM 13376/13384 [002] 7919.4088 1524 CPU 3/KVM 13376/13384 [002] 7919.408803365: branches: ffffffffc0f8ee1b __vmx_vcpu_run+0x5b ([kernel.kallsyms]) => ffffffffc0f8ed60 vmx_vmenter+0x0 ([kernel.kallsyms])
1549 CPU 3/KVM 13376/13384 [002] 7919.4088 1525 CPU 3/KVM 13376/13384 [002] 7919.408803461: branches: ffffffffc0f8ed62 vmx_vmenter+0x2 ([kernel.kallsyms]) => 0 [unknown] ([unknown])
1550 VM:13376 VCPU:003 uname 3404/340 1526 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408803461: branches: 0 [unknown] ([unknown]) => 7f851c9b5a5c init_cacheinfo+0x3ac (/usr/lib/x86_64-linux-gnu/libc-2.31.so)
1551 VM:13376 VCPU:003 uname 3404/340 1527 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408803567: branches: 7f851c9b5a5a init_cacheinfo+0x3aa (/usr/lib/x86_64-linux-gnu/libc-2.31.so) => 0 [unknown] ([unknown])
1552 CPU 3/KVM 13376/13384 [002] 7919.4088 1528 CPU 3/KVM 13376/13384 [002] 7919.408803567: branches: 0 [unknown] ([unknown]) => ffffffffc0f8ed80 vmx_vmexit+0x0 ([kernel.kallsyms])
1553 CPU 3/KVM 13376/13384 [002] 7919.4088 1529 CPU 3/KVM 13376/13384 [002] 7919.408803596: branches: ffffffffc0f6619a vmx_vcpu_run+0x26a ([kernel.kallsyms]) => ffffffffb2255c60 x86_virt_spec_ctrl+0x0 ([kernel.kallsyms])
1554 CPU 3/KVM 13376/13384 [002] 7919.4088 1530 CPU 3/KVM 13376/13384 [002] 7919.408803801: branches: ffffffffc0f66445 vmx_vcpu_run+0x515 ([kernel.kallsyms]) => ffffffffb2290b30 native_write_msr+0x0 ([kernel.kallsyms])
1555 CPU 3/KVM 13376/13384 [002] 7919.4088 1531 CPU 3/KVM 13376/13384 [002] 7919.408803850: branches: ffffffffc0f661f8 vmx_vcpu_run+0x2c8 ([kernel.kallsyms]) => ffffffffc1092300 kvm_load_host_xsave_state+0x0 ([kernel.kallsyms])
1556 CPU 3/KVM 13376/13384 [002] 7919.4088 1532 CPU 3/KVM 13376/13384 [002] 7919.408803850: branches: ffffffffc1092327 kvm_load_host_xsave_state+0x27 ([kernel.kallsyms]) => ffffffffc1092220 kvm_load_host_xsave_state.part.0+0x0 ([kernel.kallsyms])
1557 CPU 3/KVM 13376/13384 [002] 7919.4088 1533 CPU 3/KVM 13376/13384 [002] 7919.408803862: branches: ffffffffc0f662cf vmx_vcpu_run+0x39f ([kernel.kallsyms]) => ffffffffc0f63f90 vmx_recover_nmi_blocking+0x0 ([kernel.kallsyms])
1558 CPU 3/KVM 13376/13384 [002] 7919.4088 1534 CPU 3/KVM 13376/13384 [002] 7919.408803862: branches: ffffffffc0f662e9 vmx_vcpu_run+0x3b9 ([kernel.kallsyms]) => ffffffffc0f619a0 __vmx_complete_interrupts+0x0 ([kernel.kallsyms])
1559 CPU 3/KVM 13376/13384 [002] 7919.4088 1535 CPU 3/KVM 13376/13384 [002] 7919.408803872: branches: ffffffffc109cfb2 vcpu_enter_guest+0x752 ([kernel.kallsyms]) => ffffffffc0f5f570 vmx_handle_exit_irqoff+0x0 ([kernel.kallsyms])
1560 CPU 3/KVM 13376/13384 [002] 7919.4088 1536 CPU 3/KVM 13376/13384 [002] 7919.408803881: branches: ffffffffc109d028 vcpu_enter_guest+0x7c8 ([kernel.kallsyms]) => ffffffffb234f900 __srcu_read_lock+0x0 ([kernel.kallsyms])
1561 CPU 3/KVM 13376/13384 [002] 7919.4088 1537 CPU 3/KVM 13376/13384 [002] 7919.408803897: branches: ffffffffc109d06f vcpu_enter_guest+0x80f ([kernel.kallsyms]) => ffffffffc0f72e30 vmx_handle_exit+0x0 ([kernel.kallsyms])
1562 CPU 3/KVM 13376/13384 [002] 7919.4088 1538 CPU 3/KVM 13376/13384 [002] 7919.408803897: branches: ffffffffc0f72e3d vmx_handle_exit+0xd ([kernel.kallsyms]) => ffffffffc0f727c0 __vmx_handle_exit+0x0 ([kernel.kallsyms])
1563 CPU 3/KVM 13376/13384 [002] 7919.4088 1539 CPU 3/KVM 13376/13384 [002] 7919.408803897: branches: ffffffffc0f72b15 __vmx_handle_exit+0x355 ([kernel.kallsyms]) => ffffffffc0f60ae0 vmx_flush_pml_buffer+0x0 ([kernel.kallsyms])
1564 CPU 3/KVM 13376/13384 [002] 7919.4088 1540 CPU 3/KVM 13376/13384 [002] 7919.408803903: branches: ffffffffc0f72994 __vmx_handle_exit+0x1d4 ([kernel.kallsyms]) => ffffffffc10b7090 kvm_emulate_cpuid+0x0 ([kernel.kallsyms])
1565 CPU 3/KVM 13376/13384 [002] 7919.4088 1541 CPU 3/KVM 13376/13384 [002] 7919.408803903: branches: ffffffffc10b70f1 kvm_emulate_cpuid+0x61 ([kernel.kallsyms]) => ffffffffc10b6e10 kvm_cpuid+0x0 ([kernel.kallsyms])
1566 CPU 3/KVM 13376/13384 [002] 7919.4088 1542 CPU 3/KVM 13376/13384 [002] 7919.408803941: branches: ffffffffc10b7125 kvm_emulate_cpuid+0x95 ([kernel.kallsyms]) => ffffffffc1093110 kvm_skip_emulated_instruction+0x0 ([kernel.kallsyms])
1567 CPU 3/KVM 13376/13384 [002] 7919.4088 1543 CPU 3/KVM 13376/13384 [002] 7919.408803941: branches: ffffffffc109311f kvm_skip_emulated_instruction+0xf ([kernel.kallsyms]) => ffffffffc0f5e180 vmx_get_rflags+0x0 ([kernel.kallsyms])
1568 CPU 3/KVM 13376/13384 [002] 7919.4088 1544 CPU 3/KVM 13376/13384 [002] 7919.408803951: branches: ffffffffc109312a kvm_skip_emulated_instruction+0x1a ([kernel.kallsyms]) => ffffffffc0f5fd30 vmx_skip_emulated_instruction+0x0 ([kernel.kallsyms])
1569 CPU 3/KVM 13376/13384 [002] 7919.4088 1545 CPU 3/KVM 13376/13384 [002] 7919.408803951: branches: ffffffffc0f5fd79 vmx_skip_emulated_instruction+0x49 ([kernel.kallsyms]) => ffffffffc0f5fb50 skip_emulated_instruction+0x0 ([kernel.kallsyms])
1570 CPU 3/KVM 13376/13384 [002] 7919.4088 1546 CPU 3/KVM 13376/13384 [002] 7919.408803956: branches: ffffffffc0f5fc68 skip_emulated_instruction+0x118 ([kernel.kallsyms]) => ffffffffc0f6a940 vmx_cache_reg+0x0 ([kernel.kallsyms])
1571 CPU 3/KVM 13376/13384 [002] 7919.4088 1547 CPU 3/KVM 13376/13384 [002] 7919.408803964: branches: ffffffffc0f5fc11 skip_emulated_instruction+0xc1 ([kernel.kallsyms]) => ffffffffc0f5f9e0 vmx_set_interrupt_shadow+0x0 ([kernel.kallsyms])
1572 CPU 3/KVM 13376/13384 [002] 7919.4088 1548 CPU 3/KVM 13376/13384 [002] 7919.408803980: branches: ffffffffc109f8b1 vcpu_run+0x71 ([kernel.kallsyms]) => ffffffffc10ad2f0 kvm_cpu_has_pending_timer+0x0 ([kernel.kallsyms])
1573 CPU 3/KVM 13376/13384 [002] 7919.4088 1549 CPU 3/KVM 13376/13384 [002] 7919.408803980: branches: ffffffffc10ad2fb kvm_cpu_has_pending_timer+0xb ([kernel.kallsyms]) => ffffffffc10b0490 apic_has_pending_timer+0x0 ([kernel.kallsyms])
1574 CPU 3/KVM 13376/13384 [002] 7919.4088 1550 CPU 3/KVM 13376/13384 [002] 7919.408803991: branches: ffffffffc109f899 vcpu_run+0x59 ([kernel.kallsyms]) => ffffffffc109c860 vcpu_enter_guest+0x0 ([kernel.kallsyms])
1575 CPU 3/KVM 13376/13384 [002] 7919.4088 1551 CPU 3/KVM 13376/13384 [002] 7919.408803993: branches: ffffffffc109cd4c vcpu_enter_guest+0x4ec ([kernel.kallsyms]) => ffffffffc0f69140 vmx_prepare_switch_to_guest+0x0 ([kernel.kallsyms])
1576 CPU 3/KVM 13376/13384 [002] 7919.4088 1552 CPU 3/KVM 13376/13384 [002] 7919.408803996: branches: ffffffffc109cd7d vcpu_enter_guest+0x51d ([kernel.kallsyms]) => ffffffffb234f930 __srcu_read_unlock+0x0 ([kernel.kallsyms])
1577 CPU 3/KVM 13376/13384 [002] 7919.4088 1553 CPU 3/KVM 13376/13384 [002] 7919.408803996: branches: ffffffffc109cd9c vcpu_enter_guest+0x53c ([kernel.kallsyms]) => ffffffffc0f609b0 vmx_sync_pir_to_irr+0x0 ([kernel.kallsyms])
1578 CPU 3/KVM 13376/13384 [002] 7919.4088 1554 CPU 3/KVM 13376/13384 [002] 7919.408803996: branches: ffffffffc0f60a6d vmx_sync_pir_to_irr+0xbd ([kernel.kallsyms]) => ffffffffc10adc20 kvm_lapic_find_highest_irr+0x0 ([kernel.kallsyms])
1579 CPU 3/KVM 13376/13384 [002] 7919.4088 1555 CPU 3/KVM 13376/13384 [002] 7919.408804010: branches: ffffffffc0f60abd vmx_sync_pir_to_irr+0x10d ([kernel.kallsyms]) => ffffffffc0f60820 vmx_set_rvi+0x0 ([kernel.kallsyms])
1580 CPU 3/KVM 13376/13384 [002] 7919.4088 1556 CPU 3/KVM 13376/13384 [002] 7919.408804019: branches: ffffffffc109ceca vcpu_enter_guest+0x66a ([kernel.kallsyms]) => ffffffffb2249840 fpregs_assert_state_consistent+0x0 ([kernel.kallsyms])
1581 CPU 3/KVM 13376/13384 [002] 7919.4088 1557 CPU 3/KVM 13376/13384 [002] 7919.408804021: branches: ffffffffc109cf10 vcpu_enter_guest+0x6b0 ([kernel.kallsyms]) => ffffffffc0f65f30 vmx_vcpu_run+0x0 ([kernel.kallsyms])
1582 CPU 3/KVM 13376/13384 [002] 7919.4088 1558 CPU 3/KVM 13376/13384 [002] 7919.408804024: branches: ffffffffc0f6603b vmx_vcpu_run+0x10b ([kernel.kallsyms]) => ffffffffb229bed0 __get_current_cr3_fast+0x0 ([kernel.kallsyms])
1583 CPU 3/KVM 13376/13384 [002] 7919.4088 1559 CPU 3/KVM 13376/13384 [002] 7919.408804024: branches: ffffffffc0f66055 vmx_vcpu_run+0x125 ([kernel.kallsyms]) => ffffffffb2253050 cr4_read_shadow+0x0 ([kernel.kallsyms])
1584 CPU 3/KVM 13376/13384 [002] 7919.4088 1560 CPU 3/KVM 13376/13384 [002] 7919.408804030: branches: ffffffffc0f6608d vmx_vcpu_run+0x15d ([kernel.kallsyms]) => ffffffffc10921e0 kvm_load_guest_xsave_state+0x0 ([kernel.kallsyms])
1585 CPU 3/KVM 13376/13384 [002] 7919.4088 1561 CPU 3/KVM 13376/13384 [002] 7919.408804030: branches: ffffffffc1092207 kvm_load_guest_xsave_state+0x27 ([kernel.kallsyms]) => ffffffffc1092110 kvm_load_guest_xsave_state.part.0+0x0 ([kernel.kallsyms])
1586 CPU 3/KVM 13376/13384 [002] 7919.4088 1562 CPU 3/KVM 13376/13384 [002] 7919.408804032: branches: ffffffffc0f660c6 vmx_vcpu_run+0x196 ([kernel.kallsyms]) => ffffffffb22061a0 perf_guest_get_msrs+0x0 ([kernel.kallsyms])
1587 CPU 3/KVM 13376/13384 [002] 7919.4088 1563 CPU 3/KVM 13376/13384 [002] 7919.408804032: branches: ffffffffb22061a9 perf_guest_get_msrs+0x9 ([kernel.kallsyms]) => ffffffffb220cda0 intel_guest_get_msrs+0x0 ([kernel.kallsyms])
1588 CPU 3/KVM 13376/13384 [002] 7919.4088 1564 CPU 3/KVM 13376/13384 [002] 7919.408804039: branches: ffffffffc0f66109 vmx_vcpu_run+0x1d9 ([kernel.kallsyms]) => ffffffffc0f652c0 clear_atomic_switch_msr+0x0 ([kernel.kallsyms])
1589 CPU 3/KVM 13376/13384 [002] 7919.4088 1565 CPU 3/KVM 13376/13384 [002] 7919.408804040: branches: ffffffffc0f66119 vmx_vcpu_run+0x1e9 ([kernel.kallsyms]) => ffffffffc0f73f60 intel_pmu_lbr_is_enabled+0x0 ([kernel.kallsyms])
1590 CPU 3/KVM 13376/13384 [002] 7919.4088 1566 CPU 3/KVM 13376/13384 [002] 7919.408804042: branches: ffffffffc0f73f81 intel_pmu_lbr_is_enabled+0x21 ([kernel.kallsyms]) => ffffffffc10b68e0 kvm_find_cpuid_entry+0x0 ([kernel.kallsyms])
1591 CPU 3/KVM 13376/13384 [002] 7919.4088 1567 CPU 3/KVM 13376/13384 [002] 7919.408804045: branches: ffffffffc0f66454 vmx_vcpu_run+0x524 ([kernel.kallsyms]) => ffffffffc0f61ff0 vmx_update_hv_timer+0x0 ([kernel.kallsyms])
1592 CPU 3/KVM 13376/13384 [002] 7919.4088 1568 CPU 3/KVM 13376/13384 [002] 7919.408804057: branches: ffffffffc0f66142 vmx_vcpu_run+0x212 ([kernel.kallsyms]) => ffffffffc10af100 kvm_wait_lapic_expire+0x0 ([kernel.kallsyms])
1593 CPU 3/KVM 13376/13384 [002] 7919.4088 1569 CPU 3/KVM 13376/13384 [002] 7919.408804057: branches: ffffffffc0f66156 vmx_vcpu_run+0x226 ([kernel.kallsyms]) => ffffffffb2255c60 x86_virt_spec_ctrl+0x0 ([kernel.kallsyms])
1594 CPU 3/KVM 13376/13384 [002] 7919.4088 1570 CPU 3/KVM 13376/13384 [002] 7919.408804057: branches: ffffffffc0f66161 vmx_vcpu_run+0x231 ([kernel.kallsyms]) => ffffffffc0f8eb20 vmx_vcpu_enter_exit+0x0 ([kernel.kallsyms])
1595 CPU 3/KVM 13376/13384 [002] 7919.4088 1571 CPU 3/KVM 13376/13384 [002] 7919.408804057: branches: ffffffffc0f8eb44 vmx_vcpu_enter_exit+0x24 ([kernel.kallsyms]) => ffffffffb2353e10 rcu_note_context_switch+0x0 ([kernel.kallsyms])
1596 CPU 3/KVM 13376/13384 [002] 7919.4088 1572 CPU 3/KVM 13376/13384 [002] 7919.408804057: branches: ffffffffb2353e1c rcu_note_context_switch+0xc ([kernel.kallsyms]) => ffffffffb2353db0 rcu_qs+0x0 ([kernel.kallsyms])
1597 CPU 3/KVM 13376/13384 [002] 7919.4088 1573 CPU 3/KVM 13376/13384 [002] 7919.408804066: branches: ffffffffc0f8ebe0 vmx_vcpu_enter_exit+0xc0 ([kernel.kallsyms]) => ffffffffc0f8edc0 __vmx_vcpu_run+0x0 ([kernel.kallsyms])
1598 CPU 3/KVM 13376/13384 [002] 7919.4088 1574 CPU 3/KVM 13376/13384 [002] 7919.408804066: branches: ffffffffc0f8edd5 __vmx_vcpu_run+0x15 ([kernel.kallsyms]) => ffffffffc0f8eca0 vmx_update_host_rsp+0x0 ([kernel.kallsyms])
1599 CPU 3/KVM 13376/13384 [002] 7919.4088 1575 CPU 3/KVM 13376/13384 [002] 7919.408804066: branches: ffffffffc0f8ee1b __vmx_vcpu_run+0x5b ([kernel.kallsyms]) => ffffffffc0f8ed60 vmx_vmenter+0x0 ([kernel.kallsyms])
1600 CPU 3/KVM 13376/13384 [002] 7919.4088 1576 CPU 3/KVM 13376/13384 [002] 7919.408804162: branches: ffffffffc0f8ed62 vmx_vmenter+0x2 ([kernel.kallsyms]) => 0 [unknown] ([unknown])
1601 VM:13376 VCPU:003 uname 3404/340 1577 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408804162: branches: 0 [unknown] ([unknown]) => 7f851c9b5a5c init_cacheinfo+0x3ac (/usr/lib/x86_64-linux-gnu/libc-2.31.so)
1602 VM:13376 VCPU:003 uname 3404/340 1578 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408804273: branches: 7f851cb7c0e4 _dl_init+0x74 (/usr/lib/x86_64-linux-gnu/ld-2.31.so) => 7f851cb7bf50 call_init.part.0+0x0 (/usr/lib/x86_64-linux-gnu/ld-2.31.so)
1603 VM:13376 VCPU:003 uname 3404/340 1579 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408804526: branches: 55e0c00136f0 _start+0x0 (/usr/bin/uname) => ffffffff83200ac0 asm_exc_page_fault+0x0 ([kernel.kallsyms])
1604 VM:13376 VCPU:003 uname 3404/340 1580 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408804526: branches: ffffffff83200ac3 asm_exc_page_fault+0x3 ([kernel.kallsyms]) => ffffffff83201290 error_entry+0x0 ([kernel.kallsyms])
1605 VM:13376 VCPU:003 uname 3404/340 1581 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408804534: branches: ffffffff832012fa error_entry+0x6a ([kernel.kallsyms]) => ffffffff830b59a0 sync_regs+0x0 ([kernel.kallsyms])
1606 VM:13376 VCPU:003 uname 3404/340 1582 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408804631: branches: ffffffff83200ad9 asm_exc_page_fault+0x19 ([kernel.kallsyms]) => ffffffff830b8210 exc_page_fault+0x0 ([kernel.kallsyms])
1607 VM:13376 VCPU:003 uname 3404/340 1583 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408804631: branches: ffffffff830b82a4 exc_page_fault+0x94 ([kernel.kallsyms]) => ffffffff830b80e0 __kvm_handle_async_pf+0x0 ([kernel.kallsyms])
1608 VM:13376 VCPU:003 uname 3404/340 1584 VM:13376 VCPU:003 uname 3404/3404 [002] 7919.408804631: branches: ffffffff830b80ed __kvm_handle_async_pf+0xd ([kernel.kallsyms]) => ffffffff830b80c0 kvm_read_and_reset_apf_flags+0x0 ([kernel.kallsyms])
1609 1585
1610 1586
1611 Tracing Virtual Machines - Guest Code 1587 Tracing Virtual Machines - Guest Code
1612 ------------------------------------- 1588 -------------------------------------
1613 1589
1614 A common case for KVM test programs is that t 1590 A common case for KVM test programs is that the test program acts as the
1615 hypervisor, creating, running and destroying 1591 hypervisor, creating, running and destroying the virtual machine, and
1616 providing the guest object code from its own 1592 providing the guest object code from its own object code. In this case,
1617 the VM is not running an OS, but only the fun 1593 the VM is not running an OS, but only the functions loaded into it by the
1618 hypervisor test program, and conveniently, lo 1594 hypervisor test program, and conveniently, loaded at the same virtual
1619 addresses. To support that, option "--guest-c 1595 addresses. To support that, option "--guest-code" has been added to perf script
1620 and perf kvm report. 1596 and perf kvm report.
1621 1597
1622 Here is an example tracing a test program fro 1598 Here is an example tracing a test program from the kernel's KVM selftests:
1623 1599
1624 # perf record --kcore -e intel_pt/cyc/ -- to 1600 # perf record --kcore -e intel_pt/cyc/ -- tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test
1625 [ perf record: Woken up 1 times to write dat 1601 [ perf record: Woken up 1 times to write data ]
1626 [ perf record: Captured and wrote 0.280 MB p 1602 [ perf record: Captured and wrote 0.280 MB perf.data ]
1627 # perf script --guest-code --itrace=bep --ns 1603 # perf script --guest-code --itrace=bep --ns -F-period,+addr,+flags
1628 [SNIP] 1604 [SNIP]
1629 tsc_msrs_test 18436 [007] 10897.962087733: 1605 tsc_msrs_test 18436 [007] 10897.962087733: branches: call ffffffffc13b2ff5 __vmx_vcpu_run+0x15 (vmlinux) => ffffffffc13b2f50 vmx_update_host_rsp+0x0 (vmlinux)
1630 tsc_msrs_test 18436 [007] 10897.962087733: 1606 tsc_msrs_test 18436 [007] 10897.962087733: branches: return ffffffffc13b2f5d vmx_update_host_rsp+0xd (vmlinux) => ffffffffc13b2ffa __vmx_vcpu_run+0x1a (vmlinux)
1631 tsc_msrs_test 18436 [007] 10897.962087733: 1607 tsc_msrs_test 18436 [007] 10897.962087733: branches: call ffffffffc13b303b __vmx_vcpu_run+0x5b (vmlinux) => ffffffffc13b2f80 vmx_vmenter+0x0 (vmlinux)
1632 tsc_msrs_test 18436 [007] 10897.962087836: 1608 tsc_msrs_test 18436 [007] 10897.962087836: branches: vmentry ffffffffc13b2f82 vmx_vmenter+0x2 (vmlinux) => 0 [unknown] ([unknown])
1633 [guest/18436] 18436 [007] 10897.962087836: 1609 [guest/18436] 18436 [007] 10897.962087836: branches: vmentry 0 [unknown] ([unknown]) => 402c81 guest_code+0x131 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test)
1634 [guest/18436] 18436 [007] 10897.962087836: 1610 [guest/18436] 18436 [007] 10897.962087836: branches: call 402c81 guest_code+0x131 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test) => 40dba0 ucall+0x0 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test)
1635 [guest/18436] 18436 [007] 10897.962088248: 1611 [guest/18436] 18436 [007] 10897.962088248: branches: vmexit 40dba0 ucall+0x0 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test) => 0 [unknown] ([unknown])
1636 tsc_msrs_test 18436 [007] 10897.962088248: 1612 tsc_msrs_test 18436 [007] 10897.962088248: branches: vmexit 0 [unknown] ([unknown]) => ffffffffc13b2fa0 vmx_vmexit+0x0 (vmlinux)
1637 tsc_msrs_test 18436 [007] 10897.962088248: 1613 tsc_msrs_test 18436 [007] 10897.962088248: branches: jmp ffffffffc13b2fa0 vmx_vmexit+0x0 (vmlinux) => ffffffffc13b2fd2 vmx_vmexit+0x32 (vmlinux)
1638 tsc_msrs_test 18436 [007] 10897.962088256: 1614 tsc_msrs_test 18436 [007] 10897.962088256: branches: return ffffffffc13b2fd2 vmx_vmexit+0x32 (vmlinux) => ffffffffc13b3040 __vmx_vcpu_run+0x60 (vmlinux)
1639 tsc_msrs_test 18436 [007] 10897.962088270: 1615 tsc_msrs_test 18436 [007] 10897.962088270: branches: return ffffffffc13b30b6 __vmx_vcpu_run+0xd6 (vmlinux) => ffffffffc13b2f2e vmx_vcpu_enter_exit+0x4e (vmlinux)
1640 [SNIP] 1616 [SNIP]
1641 tsc_msrs_test 18436 [007] 10897.962089321: 1617 tsc_msrs_test 18436 [007] 10897.962089321: branches: call ffffffffc13b2ff5 __vmx_vcpu_run+0x15 (vmlinux) => ffffffffc13b2f50 vmx_update_host_rsp+0x0 (vmlinux)
1642 tsc_msrs_test 18436 [007] 10897.962089321: 1618 tsc_msrs_test 18436 [007] 10897.962089321: branches: return ffffffffc13b2f5d vmx_update_host_rsp+0xd (vmlinux) => ffffffffc13b2ffa __vmx_vcpu_run+0x1a (vmlinux)
1643 tsc_msrs_test 18436 [007] 10897.962089321: 1619 tsc_msrs_test 18436 [007] 10897.962089321: branches: call ffffffffc13b303b __vmx_vcpu_run+0x5b (vmlinux) => ffffffffc13b2f80 vmx_vmenter+0x0 (vmlinux)
1644 tsc_msrs_test 18436 [007] 10897.962089424: 1620 tsc_msrs_test 18436 [007] 10897.962089424: branches: vmentry ffffffffc13b2f82 vmx_vmenter+0x2 (vmlinux) => 0 [unknown] ([unknown])
1645 [guest/18436] 18436 [007] 10897.962089424: 1621 [guest/18436] 18436 [007] 10897.962089424: branches: vmentry 0 [unknown] ([unknown]) => 40dba0 ucall+0x0 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test)
1646 [guest/18436] 18436 [007] 10897.962089701: 1622 [guest/18436] 18436 [007] 10897.962089701: branches: jmp 40dc1b ucall+0x7b (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test) => 40dc39 ucall+0x99 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test)
1647 [guest/18436] 18436 [007] 10897.962089701: 1623 [guest/18436] 18436 [007] 10897.962089701: branches: jcc 40dc3c ucall+0x9c (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test) => 40dc20 ucall+0x80 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test)
1648 [guest/18436] 18436 [007] 10897.962089701: 1624 [guest/18436] 18436 [007] 10897.962089701: branches: jcc 40dc3c ucall+0x9c (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test) => 40dc20 ucall+0x80 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test)
1649 [guest/18436] 18436 [007] 10897.962089701: 1625 [guest/18436] 18436 [007] 10897.962089701: branches: jcc 40dc37 ucall+0x97 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test) => 40dc50 ucall+0xb0 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test)
1650 [guest/18436] 18436 [007] 10897.962089878: 1626 [guest/18436] 18436 [007] 10897.962089878: branches: vmexit 40dc55 ucall+0xb5 (/home/user/git/work/tools/testing/selftests/kselftest_install/kvm/tsc_msrs_test) => 0 [unknown] ([unknown])
1651 tsc_msrs_test 18436 [007] 10897.962089878: 1627 tsc_msrs_test 18436 [007] 10897.962089878: branches: vmexit 0 [unknown] ([unknown]) => ffffffffc13b2fa0 vmx_vmexit+0x0 (vmlinux)
1652 tsc_msrs_test 18436 [007] 10897.962089878: 1628 tsc_msrs_test 18436 [007] 10897.962089878: branches: jmp ffffffffc13b2fa0 vmx_vmexit+0x0 (vmlinux) => ffffffffc13b2fd2 vmx_vmexit+0x32 (vmlinux)
1653 tsc_msrs_test 18436 [007] 10897.962089887: 1629 tsc_msrs_test 18436 [007] 10897.962089887: branches: return ffffffffc13b2fd2 vmx_vmexit+0x32 (vmlinux) => ffffffffc13b3040 __vmx_vcpu_run+0x60 (vmlinux)
1654 tsc_msrs_test 18436 [007] 10897.962089901: 1630 tsc_msrs_test 18436 [007] 10897.962089901: branches: return ffffffffc13b30b6 __vmx_vcpu_run+0xd6 (vmlinux) => ffffffffc13b2f2e vmx_vcpu_enter_exit+0x4e (vmlinux)
1655 [SNIP] 1631 [SNIP]
1656 1632
1657 # perf kvm --guest-code --guest --host repor 1633 # perf kvm --guest-code --guest --host report -i perf.data --stdio | head -20
1658 1634
1659 # To display the perf.data header info, plea 1635 # To display the perf.data header info, please use --header/--header-only options.
1660 # 1636 #
1661 # 1637 #
1662 # Total Lost Samples: 0 1638 # Total Lost Samples: 0
1663 # 1639 #
1664 # Samples: 12 of event 'instructions' 1640 # Samples: 12 of event 'instructions'
1665 # Event count (approx.): 2274583 1641 # Event count (approx.): 2274583
1666 # 1642 #
1667 # Children Self Command Shared 1643 # Children Self Command Shared Object Symbol
1668 # ........ ........ ............. ....... 1644 # ........ ........ ............. .................... ...........................................
1669 # 1645 #
1670 54.70% 0.00% tsc_msrs_test [kernel. 1646 54.70% 0.00% tsc_msrs_test [kernel.vmlinux] [k] entry_SYSCALL_64_after_hwframe
1671 | 1647 |
1672 ---entry_SYSCALL_64_after_hwframe 1648 ---entry_SYSCALL_64_after_hwframe
1673 do_syscall_64 1649 do_syscall_64
1674 | 1650 |
1675 |--29.44%--syscall_exit_to_use 1651 |--29.44%--syscall_exit_to_user_mode
1676 | exit_to_user_mode_p 1652 | exit_to_user_mode_prepare
1677 | task_work_run 1653 | task_work_run
1678 | __fput 1654 | __fput
1679 1655
1680 1656
1681 Event Trace 1657 Event Trace
1682 ----------- 1658 -----------
1683 1659
1684 Event Trace records information about asynchr 1660 Event Trace records information about asynchronous events, for example interrupts,
1685 faults, VM exits and entries. The informatio 1661 faults, VM exits and entries. The information is recorded in CFE and EVD packets,
1686 and also the Interrupt Flag is recorded on th 1662 and also the Interrupt Flag is recorded on the MODE.Exec packet. The CFE packet
1687 contains a type field to identify one of the 1663 contains a type field to identify one of the following:
1688 1664
1689 1 INTR interrupt, fa 1665 1 INTR interrupt, fault, exception, NMI
1690 2 IRET interrupt ret 1666 2 IRET interrupt return
1691 3 SMI system manage 1667 3 SMI system management interrupt
1692 4 RSM resume from s 1668 4 RSM resume from system management mode
1693 5 SIPI startup inter 1669 5 SIPI startup interprocessor interrupt
1694 6 INIT INIT signal 1670 6 INIT INIT signal
1695 7 VMENTRY VM-Entry 1671 7 VMENTRY VM-Entry
1696 8 VMEXIT VM-Entry 1672 8 VMEXIT VM-Entry
1697 9 VMEXIT_INTR VM-Exit due t 1673 9 VMEXIT_INTR VM-Exit due to interrupt
1698 10 SHUTDOWN Shutdown 1674 10 SHUTDOWN Shutdown
1699 1675
1700 For more details, refer to the Intel 64 and I 1676 For more details, refer to the Intel 64 and IA-32 Architectures Software
1701 Developer Manuals (version 076 or later). 1677 Developer Manuals (version 076 or later).
1702 1678
1703 The capability to do Event Trace is indicated 1679 The capability to do Event Trace is indicated by the
1704 /sys/bus/event_source/devices/intel_pt/caps/e 1680 /sys/bus/event_source/devices/intel_pt/caps/event_trace file.
1705 1681
1706 Event trace is selected for recording using t 1682 Event trace is selected for recording using the "event" config term. e.g.
1707 1683
1708 perf record -e intel_pt/event/u uname 1684 perf record -e intel_pt/event/u uname
1709 1685
1710 Event trace events are output using the --itr 1686 Event trace events are output using the --itrace I option. e.g.
1711 1687
1712 perf script --itrace=Ie 1688 perf script --itrace=Ie
1713 1689
1714 perf script displays events containing CFE ty 1690 perf script displays events containing CFE type, vector and event data,
1715 in the form: 1691 in the form:
1716 1692
1717 evt: hw int (t) cfe: 1693 evt: hw int (t) cfe: INTR IP: 1 vector: 3 PFA: 0x8877665544332211
1718 1694
1719 The IP flag indicates if the event binds to a 1695 The IP flag indicates if the event binds to an IP, which includes any case where
1720 flow control packet generation is enabled, as 1696 flow control packet generation is enabled, as well as when CFE packet IP bit is
1721 set. 1697 set.
1722 1698
1723 perf script displays events containing change 1699 perf script displays events containing changes to the Interrupt Flag in the form:
1724 1700
1725 iflag: t IFLAG 1701 iflag: t IFLAG: 1->0 via branch
1726 1702
1727 where "via branch" indicates a branch (interr 1703 where "via branch" indicates a branch (interrupt or return from interrupt) and
1728 "non branch" indicates an instruction such as 1704 "non branch" indicates an instruction such as CFI, STI or POPF).
1729 1705
1730 In addition, the current state of the interru 1706 In addition, the current state of the interrupt flag is indicated by the presence
1731 or absence of the "D" (interrupt disabled) pe 1707 or absence of the "D" (interrupt disabled) perf script flag. If the interrupt
1732 flag is changed, then the "t" flag is also in 1708 flag is changed, then the "t" flag is also included i.e.
1733 1709
1734 no flag, interrupts enabled I 1710 no flag, interrupts enabled IF=1
1735 t interrupts become disabled IF 1711 t interrupts become disabled IF=1 -> IF=0
1736 D interrupts are disabled IF=0 1712 D interrupts are disabled IF=0
1737 Dt interrupts become enabled IF 1713 Dt interrupts become enabled IF=0 -> IF=1
1738 1714
1739 The intel-pt-events.py script illustrates how 1715 The intel-pt-events.py script illustrates how to access Event Trace information
1740 using a Python script. 1716 using a Python script.
1741 1717
1742 1718
1743 TNT Disable 1719 TNT Disable
1744 ----------- 1720 -----------
1745 1721
1746 TNT packets are disabled using the "notnt" co 1722 TNT packets are disabled using the "notnt" config term. e.g.
1747 1723
1748 perf record -e intel_pt/notnt/u uname 1724 perf record -e intel_pt/notnt/u uname
1749 1725
1750 In that case the --itrace q option is forced 1726 In that case the --itrace q option is forced because walking executable code
1751 to reconstruct the control flow is not possib 1727 to reconstruct the control flow is not possible.
1752 1728
1753 1729
1754 Emulated PTWRITE 1730 Emulated PTWRITE
1755 ---------------- 1731 ----------------
1756 1732
1757 Later perf tools support a method to emulate 1733 Later perf tools support a method to emulate the ptwrite instruction, which
1758 can be useful if hardware does not support th 1734 can be useful if hardware does not support the ptwrite instruction.
1759 1735
1760 Instead of using the ptwrite instruction, a f 1736 Instead of using the ptwrite instruction, a function is used which produces
1761 a trace that encodes the payload data into TN 1737 a trace that encodes the payload data into TNT packets. Here is an example
1762 of the function: 1738 of the function:
1763 1739
1764 #include <stdint.h> 1740 #include <stdint.h>
1765 1741
1766 void perf_emulate_ptwrite(uint64_t x) 1742 void perf_emulate_ptwrite(uint64_t x)
1767 __attribute__((externally_visible, noipa, no 1743 __attribute__((externally_visible, noipa, no_instrument_function, naked));
1768 1744
1769 #define PERF_EMULATE_PTWRITE_8_BITS \ 1745 #define PERF_EMULATE_PTWRITE_8_BITS \
1770 "1: shl %rax\n" \ 1746 "1: shl %rax\n" \
1771 " jc 1f\n" \ 1747 " jc 1f\n" \
1772 "1: shl %rax\n" \ 1748 "1: shl %rax\n" \
1773 " jc 1f\n" \ 1749 " jc 1f\n" \
1774 "1: shl %rax\n" \ 1750 "1: shl %rax\n" \
1775 " jc 1f\n" \ 1751 " jc 1f\n" \
1776 "1: shl %rax\n" \ 1752 "1: shl %rax\n" \
1777 " jc 1f\n" \ 1753 " jc 1f\n" \
1778 "1: shl %rax\n" \ 1754 "1: shl %rax\n" \
1779 " jc 1f\n" \ 1755 " jc 1f\n" \
1780 "1: shl %rax\n" \ 1756 "1: shl %rax\n" \
1781 " jc 1f\n" \ 1757 " jc 1f\n" \
1782 "1: shl %rax\n" \ 1758 "1: shl %rax\n" \
1783 " jc 1f\n" \ 1759 " jc 1f\n" \
1784 "1: shl %rax\n" \ 1760 "1: shl %rax\n" \
1785 " jc 1f\n" 1761 " jc 1f\n"
1786 1762
1787 /* Undefined instruction */ 1763 /* Undefined instruction */
1788 #define PERF_EMULATE_PTWRITE_UD2 ".by 1764 #define PERF_EMULATE_PTWRITE_UD2 ".byte 0x0f, 0x0b\n"
1789 1765
1790 #define PERF_EMULATE_PTWRITE_MAGIC PE 1766 #define PERF_EMULATE_PTWRITE_MAGIC PERF_EMULATE_PTWRITE_UD2 ".ascii \"perf,ptwrite \"\n"
1791 1767
1792 void perf_emulate_ptwrite(uint64_t x __attri 1768 void perf_emulate_ptwrite(uint64_t x __attribute__ ((__unused__)))
1793 { 1769 {
1794 /* Assumes SysV ABI : x passed in r 1770 /* Assumes SysV ABI : x passed in rdi */
1795 __asm__ volatile ( 1771 __asm__ volatile (
1796 "jmp 1f\n" 1772 "jmp 1f\n"
1797 PERF_EMULATE_PTWRITE_MAGIC 1773 PERF_EMULATE_PTWRITE_MAGIC
1798 "1: mov %rdi, %rax\n" 1774 "1: mov %rdi, %rax\n"
1799 PERF_EMULATE_PTWRITE_8_BITS 1775 PERF_EMULATE_PTWRITE_8_BITS
1800 PERF_EMULATE_PTWRITE_8_BITS 1776 PERF_EMULATE_PTWRITE_8_BITS
1801 PERF_EMULATE_PTWRITE_8_BITS 1777 PERF_EMULATE_PTWRITE_8_BITS
1802 PERF_EMULATE_PTWRITE_8_BITS 1778 PERF_EMULATE_PTWRITE_8_BITS
1803 PERF_EMULATE_PTWRITE_8_BITS 1779 PERF_EMULATE_PTWRITE_8_BITS
1804 PERF_EMULATE_PTWRITE_8_BITS 1780 PERF_EMULATE_PTWRITE_8_BITS
1805 PERF_EMULATE_PTWRITE_8_BITS 1781 PERF_EMULATE_PTWRITE_8_BITS
1806 PERF_EMULATE_PTWRITE_8_BITS 1782 PERF_EMULATE_PTWRITE_8_BITS
1807 "1: ret\n" 1783 "1: ret\n"
1808 ); 1784 );
1809 } 1785 }
1810 1786
1811 For example, a test program with the function 1787 For example, a test program with the function above:
1812 1788
1813 #include <stdio.h> 1789 #include <stdio.h>
1814 #include <stdint.h> 1790 #include <stdint.h>
1815 #include <stdlib.h> 1791 #include <stdlib.h>
1816 1792
1817 #include "perf_emulate_ptwrite.h" 1793 #include "perf_emulate_ptwrite.h"
1818 1794
1819 int main(int argc, char *argv[]) 1795 int main(int argc, char *argv[])
1820 { 1796 {
1821 uint64_t x = 0; 1797 uint64_t x = 0;
1822 1798
1823 if (argc > 1) 1799 if (argc > 1)
1824 x = strtoull(argv[1], NULL, 1800 x = strtoull(argv[1], NULL, 0);
1825 perf_emulate_ptwrite(x); 1801 perf_emulate_ptwrite(x);
1826 return 0; 1802 return 0;
1827 } 1803 }
1828 1804
1829 Can be compiled and traced: 1805 Can be compiled and traced:
1830 1806
1831 $ gcc -Wall -Wextra -O3 -g -o eg_ptw eg_ptw. 1807 $ gcc -Wall -Wextra -O3 -g -o eg_ptw eg_ptw.c
1832 $ perf record -e intel_pt//u ./eg_ptw 0x1234 1808 $ perf record -e intel_pt//u ./eg_ptw 0x1234567890abcdef
1833 [ perf record: Woken up 1 times to write dat 1809 [ perf record: Woken up 1 times to write data ]
1834 [ perf record: Captured and wrote 0.017 MB p 1810 [ perf record: Captured and wrote 0.017 MB perf.data ]
1835 $ perf script --itrace=ew 1811 $ perf script --itrace=ew
1836 eg_ptw 19875 [007] 8061.235912: 1812 eg_ptw 19875 [007] 8061.235912: ptwrite: IP: 0 payload: 0x1234567890abcdef 55701249a196 perf_emulate_ptwrite+0x16 (/home/user/eg_ptw)
1837 $ 1813 $
1838 1814
1839 1815
1840 Pipe mode 1816 Pipe mode
1841 --------- 1817 ---------
1842 Pipe mode is a problem for Intel PT and possi 1818 Pipe mode is a problem for Intel PT and possibly other auxtrace users.
1843 It's not recommended to use a pipe as data ou 1819 It's not recommended to use a pipe as data output with Intel PT because
1844 of the following reason. 1820 of the following reason.
1845 1821
1846 Essentially the auxtrace buffers do not behav 1822 Essentially the auxtrace buffers do not behave like the regular perf
1847 event buffers. That is because the head and 1823 event buffers. That is because the head and tail are updated by
1848 software, but in the auxtrace case the data i 1824 software, but in the auxtrace case the data is written by hardware.
1849 So the head and tail do not get updated as da 1825 So the head and tail do not get updated as data is written.
1850 1826
1851 In the Intel PT case, the head and tail are u 1827 In the Intel PT case, the head and tail are updated only when the trace
1852 is disabled by software, for example: 1828 is disabled by software, for example:
1853 - full-trace, system wide : when buffer p 1829 - full-trace, system wide : when buffer passes watermark
1854 - full-trace, not system-wide : when buff 1830 - full-trace, not system-wide : when buffer passes watermark or
1855 context s 1831 context switches
1856 - snapshot mode : as above but also when 1832 - snapshot mode : as above but also when a snapshot is made
1857 - sample mode : as above but also when a 1833 - sample mode : as above but also when a sample is made
1858 1834
1859 That means finished-round ordering doesn't wo 1835 That means finished-round ordering doesn't work. An auxtrace buffer
1860 can turn up that has data that extends back i 1836 can turn up that has data that extends back in time, possibly to the
1861 very beginning of tracing. 1837 very beginning of tracing.
1862 1838
1863 For a perf.data file, that problem is solved 1839 For a perf.data file, that problem is solved by going through the trace
1864 and queuing up the auxtrace buffers in advanc 1840 and queuing up the auxtrace buffers in advance.
1865 1841
1866 For pipe mode, the order of events and timest 1842 For pipe mode, the order of events and timestamps can presumably
1867 be messed up. 1843 be messed up.
1868 1844
1869 1845
1870 EXAMPLE 1846 EXAMPLE
1871 ------- 1847 -------
1872 1848
1873 Examples can be found on perf wiki page "Perf 1849 Examples can be found on perf wiki page "Perf tools support for Intel® Processor Trace":
1874 1850
1875 https://perf.wiki.kernel.org/index.php/Perf_t 1851 https://perf.wiki.kernel.org/index.php/Perf_tools_support_for_Intel%C2%AE_Processor_Trace
1876 1852
1877 1853
1878 SEE ALSO 1854 SEE ALSO
1879 -------- 1855 --------
1880 1856
1881 linkperf:perf-record[1], linkperf:perf-script 1857 linkperf:perf-record[1], linkperf:perf-script[1], linkperf:perf-report[1],
1882 linkperf:perf-inject[1] 1858 linkperf:perf-inject[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.