1TRACE-CMD.DAT.V(5) libtracefs Manual TRACE-CMD.DAT.V(5)
2
3
4
6 trace-cmd.dat.v7 - trace-cmd version 7 file format
7
9 trace-cmd.dat ignore
10
12 The trace-cmd(1) utility produces a "trace.dat" file. The file may also
13 be named anything depending if the user specifies a different output
14 name, but it must have a certain binary format. The file is used by
15 trace-cmd to save kernel traces into it and be able to extract the
16 trace from it at a later point (see trace-cmd-report(1)).
17
19 The first three bytes contain the magic value:
20
21 0x17 0x08 0x44
22
23 The next 7 bytes contain the characters:
24
25 "tracing"
26
27 The next set of characters contain a null '\0' terminated string
28 that contains the version of the file:
29
30 "7\0"
31
32 The next 1 byte contains the flags for the file endianess:
33
34 0 = little endian
35 1 = big endian
36
37 The next byte contains the number of bytes per "long" value:
38
39 4 - 32-bit long values
40 8 - 64-bit long values
41
42 Note: This is the long size of the target's user space. Not the
43 kernel space size.
44
45 [ Now all numbers are written in file defined endianess. ]
46
47 The next 4 bytes are a 32-bit word that defines what the traced
48 host machine page size was.
49
50 The compression algorithm header is written next:
51 "name\0version\0"
52 where "name" and "version" are strings, name and version of the
53 compression algorithm used to compress the trace file. If the name
54 is "none", the data in the file is not compressed.
55
56 The next 8 bytes are 64-bit integer, the offset within the file where
57 the first OPTIONS section is located.
58
59 The rest of the file consists of different sections. The only mandatory
60 is the first OPTIONS section, all others are optional. The location and
61 the order of the sections is not strict. Each section starts with a header:
62
64 <2 bytes> unsigned short integer, ID of the section.
65 <2 bytes> unsigned short integer, section flags:
66 1 = the section is compressed.
67 <4 bytes> ID of a string, description of the section.
68 <4 bytes> unsigned integer, size of the section in the file.
69
70 If the section is compressed, the above is the compressed size.
71 The section must be uncompressed on reading. The described format of
72 the sections refers to the uncompressed data.
73
75 Some of the sections in the file may be compressed with the compression algorithm,
76 specified in the compression algorithm header. Compressed sections have a compression
77 header, written after the section header and right before the compressed data:
78 <4 bytes> unsigned int, size of compressed data in this section.
79 <4 bytes> unsigned int, size of uncompressed data.
80 <data> binary compressed data, with the specified size.
81
83 There are two special sections, BUFFER FLYRECORD and BUFFER LATENCY, containing
84 trace data. These sections may be compressed with the compression algorithm, specified
85 in the compression header. Usually the size of these sections is huge, that's why its
86 compression format is different from the other sections. The trace data is compressed
87 in chunks The size of one chunk is specified in the file creation time. The format
88 of compressed trace data is:
89 <4 bytes> unsigned int, count of chunks.
90 Follows the compressed chunks of given count. For each chunk:
91 <4 bytes> unsigned int, size of compressed data in this chunk.
92 <4 bytes> unsigned int, size of uncompressed data, aligned with the trace page size.
93 <data> binary compressed data, with the specified size.
94 These chunks must be uncompressed on reading. The described format of
95 trace data refers to the uncompressed data.
96
98 Section ID: 0
99
100 This is the the only mandatory section in the file. There can be multiple
101 options sections, the first one is located at the offset specified right
102 after the compression algorithm header. The section consists of multiple
103 trace options, each option has the following format:
104 <2 bytes> unsigned short integer, ID of the option.
105 <4 bytes> unsigned integer, size of the option's data.
106 <binary data> bytes of the size specified above, data of the option.
107
108 Options, supported by the trace file version 7:
109
110 DONE: id 0, size 8
111 This option indicates the end of the options section, it is written
112 always as last option. The DONE option data is:
113 <8 bytes> long long unsigned integer, offset in the trace file where
114 the next options section is located. If this offset is 0, then there
115 are no more options sections.
116
117 DATE: id 1, size vary
118 The DATE option data is a null terminated ASCII string, which represents
119 the time difference between trace events timestamps and the Generic Time
120 of Day of the system.
121
122 CPUSTAT: id 2, size vary
123 The CPUSTAT option data is a null terminated ASCII string, the content of the
124 "per_cpu/cpu<id>/stats" file from the trace directory. There is a CPUSTAT option
125 for each CPU.
126
127 BUFFER: id 3, size vary
128 The BUFFER option describes the flyrecord trace data saved in the file, collected
129 from one trace instance. There is BUFFER option for each trace instance. The format
130 of the BUFFER data is:
131 <8 bytes> long long unsigned integer, offset in the trace file where the
132 BUFFER FLYRECORD section is located, containing flyrecord trace data.
133 <string> a null terminated ASCII string, name of the trace instance. Empty string ""
134 is saved as name of the top instance.
135 <string> a null terminated ASCII string, trace clock used for events timestamps in
136 this trace instance.
137 <4 bytes> unsigned integer, size of the trace buffer page.
138 <4 bytes> unsigned integer, count of the CPUs with trace data.
139 For each CPU of the above count:
140 <4 bytes> unsigned integer, ID of the CPU.
141 <8 bytes> long long unsigned integer, offset in the trace file where the trace data
142 for this CPU is located.
143 <8 bytes> long long unsigned integer, size of the trace data for this CPU.
144
145 TRACECLOCK: id 4, size vary
146 The TRACECLOCK option data is a null terminated ASCII string, the content of the
147 "trace_clock" file from the trace directory.
148
149 UNAME: id 5, size vary
150 The UNAME option data is a null terminated ASCII string, identifying the system where
151 the trace data is collected. The string is retrieved by the uname() system call.
152
153 HOOK: id 6, size vary
154 The HOOK option data is a null terminated ASCII string, describing event hooks: custom
155 event matching to connect any two events together.
156
157 OFFSET: id 7, size vary
158 The OFFSET option data is a null terminated ASCII string, representing a fixed time that
159 is added to each event timestamp on reading.
160
161 CPUCOUNT: id 8, size 4
162 The CPUCOUNT option data is:
163 <4 bytes> unsigned integer, number of CPUs in the system.
164
165 VERSION: id 9, size vary
166 The VERSION option data is a null terminated ASCII string, representing the version of
167 the trace-cmd application, used to collect these trace logs.
168
169 PROCMAPS: id 10, size vary
170 The PROCMAPS option data is a null terminated ASCII string, representing the memory map
171 of each traced filtered process. The format of the string is, for each filtered process:
172 <procss ID> <libraries count> <process command> \n
173 <memory start address> <memory end address> <full path of the mapped library file> \n
174 ...
175 separate line for each library, used by this process
176 ...
177 ...
178
179 TRACEID: id 11, size 8
180 The TRACEID option data is a unique identifier of this tracing session:
181 <8 bytes> long long unsigned integer, trace session identifier.
182
183 TIME_SHIFT: id 12, size vary
184 The TIME_SHIFT option stores time synchronization information, collected during host and guest
185 tracing session. Usually it is saved in the guest trace file. This information is used to
186 synchronize guest with host events timestamps, when displaying all files from this tracing
187 session. The format of the TIME_SHIFT option data is:
188 <8 bytes> long long unsigned integer, trace identifier of the peer (usually the host).
189 <4 bytes> unsigned integer, flags specific to the time synchronization protocol, used in this
190 trace session.
191 <4 bytes> unsigned integer, number of traced CPUs. For each CPU, timestamps corrections
192 are recorded:
193 <4 bytes> unsigned integer, count of the recorded timestamps corrections for this CPU.
194 <array of unsigned long long integers of the above count>, times when the corrections are calculated
195 <array of unsigned long long integers of the above count>, corrections offsets
196 <array of unsigned long long integers of the above count>, corrections scaling ratio
197
198 GUEST: id 13, size vary
199 The GUEST option stores information about traced guests in this tracing session. Usually it is
200 saved in the host trace file. There is a separate GUEST option for each traced guest.
201 The information is used when displaying all files from this tracing session. The format of
202 the GUEST option data is:
203 <string> a null terminated ASCII string, name of the guest.
204 <8 bytes> long long unsigned integer, trace identifier of the guest for this session.
205 <4 bytes> unsigned integer, number of guest's CPUs. For each CPU:
206 <4 bytes> unsigned integer, ID of the CPU.
207 <4 bytes> unsigned integer, PID of the host task, emulating this guest CPU.
208
209 TSC2NSEC: id 14, size 16
210 The TSC2NSEC option stores information, used to convert TSC events timestamps to nanoseconds.
211 The format of the TSC2NSEC option data is:
212 <4 bytes> unsigned integer, time multiplier.
213 <4 bytes> unsigned integer, time shift.
214 <8 bytes> unsigned long long integer, time offset.
215
216 HEADER_INFO: id 16, size 8
217 The HEADER_INFO option data is:
218 <8 bytes> long long unsigned integer, offset into the trace file where the HEADER INFO
219 section is located
220
221 FTRACE_EVENTS: id 17, size 8
222 The FTRACE_EVENTS option data is:
223 <8 bytes> long long unsigned integer, offset into the trace file where the
224 FTRACE EVENT FORMATS section is located.
225
226 EVENT_FORMATS: id 18, size 8
227 The EVENT_FORMATS option data is:
228 <8 bytes> long long unsigned integer, offset into the trace file where the EVENT FORMATS
229 section is located.
230
231 KALLSYMS: id 19, size 8
232 The KALLSYMS option data is:
233 <8 bytes> long long unsigned integer, offset into the trace file where the KALLSYMS
234 section is located.
235
236 PRINTK: id 20, size 8
237 The PRINTK option data is:
238 <8 bytes> long long unsigned integer, offset into the trace file where the TRACE_PRINTK
239 section is located.
240
241 CMDLINES: id 21, size 8
242 The CMDLINES option data is:
243 <8 bytes> long long unsigned integer, offset into the trace file where the
244 SAVED COMMAND LINES section is located.
245
246 BUFFER_TEXT: id 22, size
247 The BUFFER_LAT option describes the latency trace data saved in the file. The format
248 of the BUFFER_LAT data is:
249 <8 bytes> long long unsigned integer, offset in the trace file where the
250 BUFFER LATENCY section is located, containing latency trace data.
251 <string> a null terminated ASCII string, name of the trace instance. Empty string ""
252 is saved as name of the top instance.
253 <string> a null terminated ASCII string, trace clock used for events timestamps in
254 this trace instance.
255
257 Section ID: 16
258
259 The first 12 bytes of the section, after the section header, contain the string:
260
261 "header_page\0"
262
263 The next 8 bytes are a 64-bit word containing the size of the
264 page header information stored next.
265
266 The next set of data is of the size read from the previous 8 bytes,
267 and contains the data retrieved from debugfs/tracing/events/header_page.
268
269 Note: The size of the second field \fBcommit\fR contains the target
270 kernel long size. For example:
271
272 field: local_t commit; offset:8; \fBsize:8;\fR signed:1;
273
274 shows the kernel has a 64-bit long.
275
276 The next 13 bytes contain the string:
277
278 "header_event\0"
279
280 The next 8 bytes are a 64-bit word containing the size of the
281 event header information stored next.
282
283 The next set of data is of the size read from the previous 8 bytes
284 and contains the data retrieved from debugfs/tracing/events/header_event.
285
286 This data allows the trace-cmd tool to know if the ring buffer format
287 of the kernel made any changes.
288
290 Section ID: 17
291
292 Directly after the section header comes the information about
293 the Ftrace specific events. These are the events used by the Ftrace plugins
294 and are not enabled by the event tracing.
295
296 The next 4 bytes contain a 32-bit word of the number of Ftrace event
297 format files that are stored in the file.
298
299 For the number of times defined by the previous 4 bytes is the
300 following:
301
302 8 bytes for the size of the Ftrace event format file.
303
304 The Ftrace event format file copied from the target machine:
305 debugfs/tracing/events/ftrace/<event>/format
306
308 Section ID: 18
309
310 Directly after the section header comes the information about
311 the event layout.
312
313 The next 4 bytes are a 32-bit word containing the number of
314 event systems that are stored in the file. These are the
315 directories in debugfs/tracing/events excluding the \fBftrace\fR
316 directory.
317
318 For the number of times defined by the previous 4 bytes is the
319 following:
320
321 A null-terminated string containing the system name.
322
323 4 bytes containing a 32-bit word containing the number
324 of events within the system.
325
326 For the number of times defined in the previous 4 bytes is the
327 following:
328
329 8 bytes for the size of the event format file.
330
331 The event format file copied from the target machine:
332 debugfs/tracing/events/<system>/<event>/format
333
335 Section ID: 19
336
337 Directly after the section header comes the information of the mapping
338 of function addresses to the function names.
339
340 The next 4 bytes are a 32-bit word containing the size of the
341 data holding the function mappings.
342
343 The next set of data is of the size defined by the previous 4 bytes
344 and contains the information from the target machine's file:
345 /proc/kallsyms
346
348 Section ID: 20
349
350 If a developer used trace_printk() within the kernel, it may
351 store the format string outside the ring buffer.
352 This information can be found in:
353 debugfs/tracing/printk_formats
354
355 The next 4 bytes are a 32-bit word containing the size of the
356 data holding the printk formats.
357
358 The next set of data is of the size defined by the previous 4 bytes
359 and contains the information from debugfs/tracing/printk_formats.
360
362 Section ID: 21
363
364 Directly after the section header comes the information mapping
365 a PID to a process name.
366
367 The next 8 bytes contain a 64-bit word that holds the size of the
368 data mapping the PID to a process name.
369
370 The next set of data is of the size defined by the previous 8 bytes
371 and contains the information from debugfs/tracing/saved_cmdlines.
372
374 This section contains flyrecord tracing data, collected in one trace instance.
375 The data is saved per CPU. Each BUFFER FLYRECORD section has a corresponding BUFFER
376 option, containing information about saved CPU's trace data. Padding is placed between
377 the section header and the CPU data, placing the CPU data at a page aligned (target page)
378 position in the file.
379
380 This data is copied directly from the Ftrace ring buffer and is of the
381 same format as the ring buffer specified by the event header files
382 loaded in the header format file.
383
384 The trace-cmd tool will try to \fBmmap(2)\fR the data page by page with the
385 target's page size if possible. If it fails to mmap, it will just read the
386 data instead.
387
388 BUFFER TEXT SECTION
389
390
391 .ft C
392 This section contains latency tracing data, ASCII text taken from the
393 target's debugfs/tracing/trace file.
394
395 STRINGS SECTION
396 .ft
397
398
399 All strings from trace file metadata are stored in string section in the file. The section
400 contains a list of NULL terminated ASCII strings. An ID of the string is used in the file
401 meta data, which is the offset of the actual string into the string section. Strings can be stored
402 into multiple string sections in the file.
403
405 trace-cmd(1), trace-cmd-record(1), trace-cmd-report(1),
406 trace-cmd-start(1), trace-cmd-stop(1), trace-cmd-extract(1),
407 trace-cmd-reset(1), trace-cmd-split(1), trace-cmd-list(1),
408 trace-cmd-listen(1), trace-cmd.dat(5)
409
411 Written by Steven Rostedt, <rostedt@goodmis.org[1]>
412
414 https://git.kernel.org/pub/scm/utils/trace-cmd/trace-cmd.git/
415
417 Copyright (C) 2010 Red Hat, Inc. Free use of this software is granted
418 under the terms of the GNU Public License (GPL).
419
421 1. rostedt@goodmis.org
422 mailto:rostedt@goodmis.org
423
424
425
426libtracefs 04/15/2022 TRACE-CMD.DAT.V(5)