Add "tagged" union
[ctf.git] / common-trace-format-proposal.txt
CommitLineData
5ba9f198 1
4767a9e7 2RFC: Common Trace Format (CTF) Proposal (pre-v1.7)
5ba9f198
MD
3
4Mathieu Desnoyers, EfficiOS Inc.
5
6The goal of the present document is to propose a trace format that suits the
cc089c3a 7needs of the embedded, telecom, high-performance and kernel communities. It is
5ba9f198 8based on the Common Trace Format Requirements (v1.4) document. It is designed to
cc089c3a
MD
9allow traces to be natively generated by the Linux kernel, Linux user-space
10applications written in C/C++, and hardware components.
11
12The latest version of this document can be found at:
13
14 git tree: git://git.efficios.com/ctf.git
15 gitweb: http://git.efficios.com/?p=ctf.git
5ba9f198
MD
16
17A reference implementation of a library to read and write this trace format is
18being implemented within the BabelTrace project, a converter between trace
19formats. The development tree is available at:
20
21 git tree: git://git.efficios.com/babeltrace.git
22 gitweb: http://git.efficios.com/?p=babeltrace.git
23
24
251. Preliminary definitions
26
3bf79539
MD
27 - Event Trace: An ordered sequence of events.
28 - Event Stream: An ordered sequence of events, containing a subset of the
29 trace event types.
30 - Event Packet: A sequence of physically contiguous events within an event
31 stream.
5ba9f198
MD
32 - Event: This is the basic entry in a trace. (aka: a trace record).
33 - An event identifier (ID) relates to the class (a type) of event within
3bf79539
MD
34 an event stream.
35 e.g. event: irq_entry.
5ba9f198
MD
36 - An event (or event record) relates to a specific instance of an event
37 class.
3bf79539
MD
38 e.g. event: irq_entry, at time X, on CPU Y
39 - Source Architecture: Architecture writing the trace.
40 - Reader Architecture: Architecture reading the trace.
5ba9f198
MD
41
42
432. High-level representation of a trace
44
3bf79539
MD
45A trace is divided into multiple event streams. Each event stream contains a
46subset of the trace event types.
5ba9f198 47
3bf79539
MD
48The final output of the trace, after its generation and optional transport over
49the network, is expected to be either on permanent or temporary storage in a
50virtual file system. Because each event stream is appended to while a trace is
51being recorded, each is associated with a separate file for output. Therefore,
52a stored trace can be represented as a directory containing one file per stream.
5ba9f198 53
3bf79539 54A metadata event stream contains information on trace event types. It describes:
5ba9f198
MD
55
56- Trace version.
57- Types available.
3bf79539
MD
58- Per-stream event header description.
59- Per-stream event header selection.
60- Per-stream event context fields.
5ba9f198 61- Per-event
3bf79539 62 - Event type to stream mapping.
5ba9f198
MD
63 - Event type to name mapping.
64 - Event type to ID mapping.
65 - Event fields description.
66
67
3bf79539 683. Event stream
5ba9f198 69
3bf79539
MD
70An event stream is divided in contiguous event packets of variable size. These
71subdivisions have a variable size. An event packet can contain a certain amount
72of padding at the end. The rationale for the event stream design choices is
73explained in Appendix B. Stream Header Rationale.
5ba9f198 74
3bf79539
MD
75An event stream is divided in contiguous event packets of variable size. These
76subdivisions have a variable size. An event packet can contain a certain amount
77of padding at the end. The stream header is repeated at the beginning of each
78event packet.
5ba9f198 79
3bf79539
MD
80The event stream header will therefore be referred to as the "event packet
81header" throughout the rest of this document.
5ba9f198
MD
82
83
844. Types
85
864.1 Basic types
87
88A basic type is a scalar type, as described in this section.
89
904.1.1 Type inheritance
91
80fd2569
MD
92Type specifications can be inherited to allow deriving types from a
93type class. For example, see the uint32_t named type derived from the "integer"
94type class below ("Integers" section). Types have a precise binary
95representation in the trace. A type class has methods to read and write these
96types, but must be derived into a type to be usable in an event field.
5ba9f198
MD
97
984.1.2 Alignment
99
100We define "byte-packed" types as aligned on the byte size, namely 8-bit.
101We define "bit-packed" types as following on the next bit, as defined by the
102"bitfields" section.
5ba9f198 103
3bf79539
MD
104All basic types, except bitfields, are either aligned on an architecture-defined
105specific alignment or byte-packed, depending on the architecture preference.
106Architectures providing fast unaligned write byte-packed basic types to save
5ba9f198 107space, aligning each type on byte boundaries (8-bit). Architectures with slow
3bf79539
MD
108unaligned writes align types on specific alignment values. If no specific
109alignment is declared for a type nor its parents, it is assumed to be bit-packed
110for bitfields and byte-packed for other types.
5ba9f198 111
3bf79539 112Metadata attribute representation of a specific alignment:
5ba9f198
MD
113
114 align = value; /* value in bits */
115
1164.1.3 Byte order
117
3bf79539
MD
118By default, the native endianness of the source architecture the trace is used.
119Byte order can be overridden for a basic type by specifying a "byte_order"
120attribute. Typical use-case is to specify the network byte order (big endian:
121"be") to save data captured from the network into the trace without conversion.
122If not specified, the byte order is native.
5ba9f198
MD
123
124Metadata representation:
125
126 byte_order = native OR network OR be OR le; /* network and be are aliases */
127
1284.1.4 Size
129
130Type size, in bits, for integers and floats is that returned by "sizeof()" in C
131multiplied by CHAR_BIT.
132We require the size of "char" and "unsigned char" types (CHAR_BIT) to be fixed
133to 8 bits for cross-endianness compatibility.
134
135Metadata representation:
136
137 size = value; (value is in bits)
138
1394.1.5 Integers
140
141Signed integers are represented in two-complement. Integer alignment, size,
142signedness and byte ordering are defined in the metadata. Integers aligned on
143byte size (8-bit) and with length multiple of byte size (8-bit) correspond to
144the C99 standard integers. In addition, integers with alignment and/or size that
145are _not_ a multiple of the byte size are permitted; these correspond to the C99
146standard bitfields, with the added specification that the CTF integer bitfields
147have a fixed binary representation. A MIT-licensed reference implementation of
148the CTF portable bitfields is available at:
149
150 http://git.efficios.com/?p=babeltrace.git;a=blob;f=include/babeltrace/bitfield.h
151
152Binary representation of integers:
153
154- On little and big endian:
155 - Within a byte, high bits correspond to an integer high bits, and low bits
156 correspond to low bits.
157- On little endian:
158 - Integer across multiple bytes are placed from the less significant to the
159 most significant.
160 - Consecutive integers are placed from lower bits to higher bits (even within
161 a byte).
162- On big endian:
163 - Integer across multiple bytes are placed from the most significant to the
164 less significant.
165 - Consecutive integers are placed from higher bits to lower bits (even within
166 a byte).
167
168This binary representation is derived from the bitfield implementation in GCC
169for little and big endian. However, contrary to what GCC does, integers can
170cross units boundaries (no padding is required). Padding can be explicitely
171added (see 4.1.6 GNU/C bitfields) to follow the GCC layout if needed.
172
173Metadata representation:
174
80fd2569 175 integer {
5ba9f198
MD
176 signed = true OR false; /* default false */
177 byte_order = native OR network OR be OR le; /* default native */
178 size = value; /* value in bits, no default */
179 align = value; /* value in bits */
2152348f 180 }
5ba9f198 181
80fd2569 182Example of type inheritance (creation of a uint32_t named type):
5ba9f198 183
80fd2569 184typedef integer {
9e4e34e9 185 size = 32;
5ba9f198
MD
186 signed = false;
187 align = 32;
80fd2569 188} uint32_t;
5ba9f198 189
80fd2569 190Definition of a named 5-bit signed bitfield:
5ba9f198 191
80fd2569 192typedef integer {
5ba9f198
MD
193 size = 5;
194 signed = true;
195 align = 1;
80fd2569 196} int5_t;
5ba9f198
MD
197
1984.1.6 GNU/C bitfields
199
200The GNU/C bitfields follow closely the integer representation, with a
201particularity on alignment: if a bitfield cannot fit in the current unit, the
80fd2569
MD
202unit is padded and the bitfield starts at the following unit. The unit size is
203defined by the size of the type "unit_type".
5ba9f198 204
2152348f 205Metadata representation:
80fd2569
MD
206
207 unit_type name:size:
208
5ba9f198
MD
209As an example, the following structure declared in C compiled by GCC:
210
211struct example {
212 short a:12;
213 short b:5;
214};
215
2152348f
MD
216The example structure is aligned on the largest element (short). The second
217bitfield would be aligned on the next unit boundary, because it would not fit in
218the current unit.
5ba9f198
MD
219
2204.1.7 Floating point
221
222The floating point values byte ordering is defined in the metadata.
223
224Floating point values follow the IEEE 754-2008 standard interchange formats.
225Description of the floating point values include the exponent and mantissa size
226in bits. Some requirements are imposed on the floating point values:
227
228- FLT_RADIX must be 2.
229- mant_dig is the number of digits represented in the mantissa. It is specified
230 by the ISO C99 standard, section 5.2.4, as FLT_MANT_DIG, DBL_MANT_DIG and
231 LDBL_MANT_DIG as defined by <float.h>.
232- exp_dig is the number of digits represented in the exponent. Given that
233 mant_dig is one bit more than its actual size in bits (leading 1 is not
234 needed) and also given that the sign bit always takes one bit, exp_dig can be
235 specified as:
236
237 - sizeof(float) * CHAR_BIT - FLT_MANT_DIG
238 - sizeof(double) * CHAR_BIT - DBL_MANT_DIG
239 - sizeof(long double) * CHAR_BIT - LDBL_MANT_DIG
240
241Metadata representation:
242
80fd2569 243floating_point {
5ba9f198
MD
244 exp_dig = value;
245 mant_dig = value;
246 byte_order = native OR network OR be OR le;
2152348f 247}
5ba9f198
MD
248
249Example of type inheritance:
250
80fd2569 251typedef floating_point {
5ba9f198
MD
252 exp_dig = 8; /* sizeof(float) * CHAR_BIT - FLT_MANT_DIG */
253 mant_dig = 24; /* FLT_MANT_DIG */
254 byte_order = native;
80fd2569 255} float;
5ba9f198
MD
256
257TODO: define NaN, +inf, -inf behavior.
258
2594.1.8 Enumerations
260
261Enumerations are a mapping between an integer type and a table of strings. The
262numerical representation of the enumeration follows the integer type specified
263by the metadata. The enumeration mapping table is detailed in the enumeration
3bf79539
MD
264description within the metadata. The mapping table maps inclusive value ranges
265(or single values) to strings. Instead of being limited to simple
266"value -> string" mappings, these enumerations map
80fd2569 267"[ start_value ... end_value ] -> string", which map inclusive ranges of
3bf79539
MD
268values to strings. An enumeration from the C language can be represented in
269this format by having the same start_value and end_value for each element, which
270is in fact a range of size 1. This single-value range is supported without
4767a9e7 271repeating the start and end values with the value = string declaration.
80fd2569 272
4767a9e7
MD
273If a numeric value is encountered between < >, it represents the integer type
274size used to hold the enumeration, in bits.
275
276enum <integer_type OR size> name {
80fd2569
MD
277 string = start_value1 ... end_value1,
278 "other string" = start_value2 ... end_value2,
279 yet_another_string, /* will be assigned to end_value2 + 1 */
280 "some other string" = value,
281 ...
282};
283
284If the values are omitted, the enumeration starts at 0 and increment of 1 for
285each entry:
286
4767a9e7 287enum <32> name {
80fd2569
MD
288 ZERO,
289 ONE,
290 TWO,
291 TEN = 10,
292 ELEVEN,
3bf79539 293};
5ba9f198 294
80fd2569 295Overlapping ranges within a single enumeration are implementation defined.
5ba9f198 296
2152348f
MD
297A nameless enumeration can be declared as a field type or as part of a typedef:
298
299enum <integer_type> {
300 ...
301}
302
5ba9f198
MD
3034.2 Compound types
304
3054.2.1 Structures
306
307Structures are aligned on the largest alignment required by basic types
308contained within the structure. (This follows the ISO/C standard for structures)
309
80fd2569 310Metadata representation of a named structure:
5ba9f198 311
80fd2569
MD
312struct name {
313 field_type field_name;
314 field_type field_name;
315 ...
316};
5ba9f198
MD
317
318Example:
319
80fd2569
MD
320struct example {
321 integer { /* Nameless type */
322 size = 16;
323 signed = true;
324 align = 16;
325 } first_field_name;
326 uint64_t second_field_name; /* Named type declared in the metadata */
3bf79539 327};
5ba9f198
MD
328
329The fields are placed in a sequence next to each other. They each possess a
330field name, which is a unique identifier within the structure.
331
2152348f 332A nameless structure can be declared as a field type or as part of a typedef:
80fd2569
MD
333
334struct {
335 ...
2152348f 336}
80fd2569 337
77a98c82 3384.2.2 Variants (Discriminated/Tagged Unions)
fcba70d4
MD
339
340A CTF variant is a selection between different types. A CTF variant must always
341be defined within the scope of a structure or within fields contained within a
342structure (defined recursively). A "tag" enumeration field must appear in either
343the same lexical scope or an uppermost scope, prior to the variant field (in
344field declaration order). The type selection is indicated by the mapping from
345the enumeration value to the string used as variant type selector. The field to
346use as tag is specified by the "tag_field", specified between "< >" after the
347"variant" keyword for unnamed variants, and after "variant name" for named
348variants.
349
350The alignment of the variant is the alignment of the type as selected by the tag
351value for the specific instance of the variant. The alignment of the type
352containing the variant is independent of the variant alignment. The size of the
353variant is the size as selected by the tag value for the specific instance of
354the variant.
355
356A named variant declaration followed by its definition within a structure
357declaration:
358
359variant name {
360 field_type sel1;
361 field_type sel2;
362 field_type sel3;
363 ...
364};
365
366struct {
367 enum <integer_type or size> { sel1, sel2, sel3, ... } tag_field;
368 ...
369 variant name <tag_field> v;
370}
371
372An unnamed variant definition within a structure is expressed by the following
373metadata:
374
375struct {
376 enum <integer_type or size> { sel1, sel2, sel3, ... } tag_field;
377 ...
378 variant <tag_field> {
379 field_type sel1;
380 field_type sel2;
381 field_type sel3;
382 ...
383 } v;
384}
385
386Example of a named variant within a sequence that refers to a single tag field:
387
388variant example {
389 uint32_t a;
390 uint64_t b;
391 short c;
392};
393
394struct {
395 enum <uint2_t> { a, b, c } choice;
15850440 396 variant example <choice> v[unsigned int];
fcba70d4
MD
397}
398
399Example of an unnamed variant:
400
401struct {
402 enum <uint2_t> { a, b, c, d } choice;
403 /* Unrelated fields can be added between the variant and its tag */
404 int32_t somevalue;
405 variant <choice> {
406 uint32_t a;
407 uint64_t b;
408 short c;
409 struct {
410 unsigned int field1;
411 uint64_t field2;
412 } d;
413 } s;
414}
415
416Example of an unnamed variant within an array:
417
418struct {
419 enum <uint2_t> { a, b, c } choice;
420 variant <choice> {
421 uint32_t a;
422 uint64_t b;
423 short c;
15850440 424 } v[10];
fcba70d4
MD
425}
426
427Example of a variant type definition within a structure, where the defined type
428is then declared within an array of structures. This variant refers to a tag
429located in an upper lexical scope. This example clearly shows that a variant
430type definition referring to the tag "x" uses the closest preceding field from
431the lexical scope of the type definition.
432
433struct {
434 enum <uint2_t> { a, b, c, d } x;
435
436 typedef variant <x> { /*
437 * "x" refers to the preceding "x" enumeration in the
438 * lexical scope of the type definition.
439 */
440 uint32_t a;
441 uint64_t b;
442 short c;
443 } example_variant;
444
445 struct {
446 enum <int> { x, y, z } x; /* This enumeration is not used by "v". */
447 example_variant v; /*
448 * "v" uses the "enum <uint2_t> { a, b, c, d }"
449 * tag.
450 */
451 } a[10];
452}
453
4544.2.3 Arrays
5ba9f198
MD
455
456Arrays are fixed-length. Their length is declared in the type declaration within
457the metadata. They contain an array of "inner type" elements, which can refer to
458any type not containing the type of the array being declared (no circular
3bf79539 459dependency). The length is the number of elements in an array.
5ba9f198 460
2152348f 461Metadata representation of a named array:
80fd2569
MD
462
463typedef elem_type name[length];
5ba9f198 464
2152348f 465A nameless array can be declared as a field type within a structure, e.g.:
5ba9f198 466
2152348f 467 uint8_t field_name[10];
80fd2569 468
5ba9f198 469
fcba70d4 4704.2.4 Sequences
5ba9f198
MD
471
472Sequences are dynamically-sized arrays. They start with an integer that specify
473the length of the sequence, followed by an array of "inner type" elements.
3bf79539 474The length is the number of elements in the sequence.
5ba9f198 475
2152348f 476Metadata representation for a named sequence:
80fd2569
MD
477
478typedef elem_type name[length_type];
479
480A nameless sequence can be declared as a field type, e.g.:
481
80fd2569
MD
482long field_name[int];
483
484The length type follows the integer types specifications, and the sequence
5ba9f198
MD
485elements follow the "array" specifications.
486
fcba70d4 4874.2.5 Strings
5ba9f198
MD
488
489Strings are an array of bytes of variable size and are terminated by a '\0'
490"NULL" character. Their encoding is described in the metadata. In absence of
491encoding attribute information, the default encoding is UTF-8.
492
80fd2569
MD
493Metadata representation of a named string type:
494
495typedef string {
5ba9f198 496 encoding = UTF8 OR ASCII;
80fd2569 497} name;
5ba9f198 498
80fd2569
MD
499A nameless string type can be declared as a field type:
500
501string field_name; /* Use default UTF8 encoding */
5ba9f198 502
3bf79539
MD
5035. Event Packet Header
504
505The event packet header consists of two part: one is mandatory and have a fixed
506layout. The second part, the "event packet context", has its layout described in
507the metadata.
5ba9f198 508
3bf79539
MD
509- Aligned on page size. Fixed size. Fields either aligned or packed (depending
510 on the architecture preference).
511 No padding at the end of the event packet header. Native architecture byte
5ba9f198 512 ordering.
3bf79539
MD
513
514Fixed layout (event packet header):
515
5ba9f198
MD
516- Magic number (CTF magic numbers: 0xC1FC1FC1 and its reverse endianness
517 representation: 0xC11FFCC1) It needs to have a non-symmetric bytewise
518 representation. Used to distinguish between big and little endian traces (this
519 information is determined by knowing the endianness of the architecture
520 reading the trace and comparing the magic number against its value and the
521 reverse, 0xC11FFCC1). This magic number specifies that we use the CTF metadata
522 description language described in this document. Different magic numbers
523 should be used for other metadata description languages.
3bf79539 524- Trace UUID, used to ensure the event packet match the metadata used.
5ba9f198
MD
525 (note: we cannot use a metadata checksum because metadata can be appended to
526 while tracing is active)
3bf79539
MD
527- Stream ID, used as reference to stream description in metadata.
528
529Metadata-defined layout (event packet context):
530
531- Event packet content size (in bytes).
532- Event packet size (in bytes, includes padding).
533- Event packet content checksum (optional). Checksum excludes the event packet
534 header.
535- Per-stream event packet sequence count (to deal with UDP packet loss). The
536 number of significant sequence counter bits should also be present, so
537 wrap-arounds are deal with correctly.
538- Timestamp at the beginning and timestamp at the end of the event packet.
539 Both timestamps are written in the packet header, but sampled respectively
540 while (or before) writing the first event and while (or after) writing the
541 last event in the packet. The inclusive range between these timestamps should
542 include all event timestamps assigned to events contained within the packet.
5ba9f198 543- Events discarded count
3bf79539
MD
544 - Snapshot of a per-stream free-running counter, counting the number of
545 events discarded that were supposed to be written in the stream prior to
546 the first event in the event packet.
5ba9f198 547 * Note: producer-consumer buffer full condition should fill the current
3bf79539 548 event packet with padding so we know exactly where events have been
5ba9f198 549 discarded.
3bf79539
MD
550- Lossless compression scheme used for the event packet content. Applied
551 directly to raw data. New types of compression can be added in following
552 versions of the format.
5ba9f198
MD
553 0: no compression scheme
554 1: bzip2
555 2: gzip
3bf79539
MD
556 3: xz
557- Cypher used for the event packet content. Applied after compression.
5ba9f198
MD
558 0: no encryption
559 1: AES
3bf79539 560- Checksum scheme used for the event packet content. Applied after encryption.
5ba9f198
MD
561 0: no checksum
562 1: md5
563 2: sha1
564 3: crc32
565
3bf79539
MD
5665.1 Event Packet Header Fixed Layout Description
567
80fd2569
MD
568struct event_packet_header {
569 uint32_t magic;
570 uint8_t trace_uuid[16];
3bf79539 571 uint32_t stream_id;
80fd2569 572};
5ba9f198 573
3bf79539
MD
5745.2 Event Packet Context Description
575
576Event packet context example. These are declared within the stream declaration
577in the metadata. All these fields are optional except for "content_size" and
578"packet_size", which must be present in the context.
579
580An example event packet context type:
581
80fd2569 582struct event_packet_context {
3bf79539
MD
583 uint64_t timestamp_begin;
584 uint64_t timestamp_end;
585 uint32_t checksum;
586 uint32_t stream_packet_count;
587 uint32_t events_discarded;
588 uint32_t cpu_id;
589 uint32_t/uint16_t content_size;
590 uint32_t/uint16_t packet_size;
591 uint8_t stream_packet_count_bits; /* Significant counter bits */
592 uint8_t compression_scheme;
593 uint8_t encryption_scheme;
594 uint8_t checksum;
595};
5ba9f198 596
fcba70d4 597
5ba9f198
MD
5986. Event Structure
599
600The overall structure of an event is:
601
fcba70d4
MD
6021 - Stream Packet Context (as specified by the stream metadata)
6032 - Event Header (as specifed by the stream metadata)
6043 - Stream Event Context (as specified by the stream metadata)
6054 - Event Context (as specified by the event metadata)
6065 - Event Payload (as specified by the event metadata)
5ba9f198 607
fcba70d4 6086.1 Lexical Scope
5ba9f198 609
fcba70d4
MD
610The lexical scope of each structure (stream packet context, header, stream event
611context, event context and payload) is extended in the following way: lower
612levels (e.g. 3) can refer to fields defined in prior levels (e.g. 2 and 1). The
613field in the closest level has priority in case of field name conflict.
614
615This allows, for instance, the event context to define a variant refering to the
616"id" field of the event header as selector.
617
6186.2 Event Header
619
620Event headers can be described within the metadata. We hereby propose, as an
621example, two types of events headers. Type 1 accommodates streams with less than
62231 event IDs. Type 2 accommodates streams with 31 or more event IDs.
5ba9f198 623
3bf79539
MD
624One major factor can vary between streams: the number of event IDs assigned to
625a stream. Luckily, this information tends to stay relatively constant (modulo
5ba9f198 626event registration while trace is being recorded), so we can specify different
3bf79539 627representations for streams containing few event IDs and streams containing
5ba9f198
MD
628many event IDs, so we end up representing the event ID and timestamp as densely
629as possible in each case.
630
fcba70d4
MD
631The header is extended in the rare occasions where the information cannot be
632represented in the ranges available in the standard event header. They are also
3bf79539
MD
633used in the rare occasions where the data required for a field could not be
634collected: the flag corresponding to the missing field within the missing_fields
635array is then set to 1.
5ba9f198
MD
636
637Types uintX_t represent an X-bit unsigned integer.
638
639
fcba70d4 6406.2.1 Type 1 - Few event IDs
5ba9f198
MD
641
642 - Aligned on 32-bit (or 8-bit if byte-packed, depending on the architecture
643 preference).
5ba9f198 644 - Native architecture byte ordering.
fcba70d4
MD
645 - For "compact" selection
646 - Fixed size: 32 bits.
647 - For "extended" selection
648 - Size depends on the architecture and variant alignment.
5ba9f198 649
80fd2569 650struct event_header_1 {
fcba70d4
MD
651 /*
652 * id: range: 0 - 30.
653 * id 31 is reserved to indicate an extended header.
654 */
655 enum <uint5_t> { compact = 0 ... 30, extended = 31 } id;
656 variant <id> {
657 struct {
658 uint27_t timestamp;
659 } compact;
660 struct {
661 uint32_t id; /* 32-bit event IDs */
662 uint64_t timestamp; /* 64-bit timestamps */
663 } extended;
664 } v;
5ba9f198
MD
665};
666
5ba9f198 667
fcba70d4 6686.2.2 Type 2 - Many event IDs
5ba9f198 669
fcba70d4 670 - Aligned on 16-bit (or 8-bit if byte-packed, depending on the architecture
5ba9f198 671 preference).
5ba9f198 672 - Native architecture byte ordering.
fcba70d4
MD
673 - For "compact" selection
674 - Size depends on the architecture and variant alignment.
675 - For "extended" selection
676 - Size depends on the architecture and variant alignment.
5ba9f198 677
80fd2569 678struct event_header_2 {
fcba70d4
MD
679 /*
680 * id: range: 0 - 65534.
681 * id 65535 is reserved to indicate an extended header.
682 */
683 enum <uint16_t> { compact = 0 ... 65534, extended = 65535 } id;
684 variant <id> {
685 struct {
686 uint32_t timestamp;
687 } compact;
688 struct {
689 uint32_t id; /* 32-bit event IDs */
690 uint64_t timestamp; /* 64-bit timestamps */
691 } extended;
692 } v;
5ba9f198
MD
693};
694
5ba9f198
MD
695
6966.2 Event Context
697
698The event context contains information relative to the current event. The choice
fcba70d4
MD
699and meaning of this information is specified by the metadata "stream" and
700"event" information. The "stream" context is applied to all events within the
701stream. The "stream" context structure follows the event header. The "event"
702context is applied to specific events. Its structure follows the "stream"
703context stucture.
5ba9f198 704
fcba70d4
MD
705An example of stream-level event context is to save the event payload size with
706each event, or to save the current PID with each event. These are declared
707within the stream declaration within the metadata:
5ba9f198 708
fcba70d4
MD
709 stream {
710 ...
711 event {
712 ...
713 context = struct {
80fd2569
MD
714 uint pid;
715 uint16_t payload_size;
3bf79539 716 };
fcba70d4
MD
717 }
718 };
719
720An example of event-specific event context is to declare a bitmap of missing
721fields, only appended after the stream event context if the extended event
722header is selected. NR_FIELDS is the number of fields within the event (a
723numeric value).
5ba9f198 724
fcba70d4
MD
725 event {
726 context = struct {
727 variant <id> {
728 struct { } compact;
729 struct {
730 uint1_t missing_fields[NR_FIELDS]; /* missing event fields bitmap */
731 } extended;
732 } v;
733 };
734 ...
735 }
5ba9f198
MD
736
7376.3 Event Payload
738
739An event payload contains fields specific to a given event type. The fields
740belonging to an event type are described in the event-specific metadata
741within a structure type.
742
7436.3.1 Padding
744
745No padding at the end of the event payload. This differs from the ISO/C standard
746for structures, but follows the CTF standard for structures. In a trace, even
747though it makes sense to align the beginning of a structure, it really makes no
748sense to add padding at the end of the structure, because structures are usually
749not followed by a structure of the same type.
750
751This trick can be done by adding a zero-length "end" field at the end of the C
752structures, and by using the offset of this field rather than using sizeof()
3bf79539 753when calculating the size of a structure (see Appendix "A. Helper macros").
5ba9f198
MD
754
7556.3.2 Alignment
756
757The event payload is aligned on the largest alignment required by types
758contained within the payload. (This follows the ISO/C standard for structures)
759
760
5ba9f198
MD
7617. Metadata
762
3bf79539
MD
763The meta-data is located in a stream named "metadata". It is made of "event
764packets", which each start with an event packet header. The event type within
765the metadata stream have no event header nor event context. Each event only
5ba9f198 766contains a null-terminated "string" payload, which is a metadata description
3bf79539
MD
767entry. The events are packed one next to another. Each event packet start with
768an event packet header, which contains, amongst other fields, the magic number
769and trace UUID.
5ba9f198
MD
770
771The metadata can be parsed by reading through the metadata strings, skipping
fcba70d4
MD
772newlines and null-characters. Type names are made of a single identifier, and
773can be surrounded by prefix/postfix. Text contained within "/*" and "*/", as
c6d7abc5
MD
774well as within "//" and end of line, are treated as comments. Boolean values can
775be represented as true, TRUE, or 1 for true, and false, FALSE, or 0 for false.
fcba70d4
MD
776
777The grammar representing the CTF metadata is presented in
778Appendix C. CTF Metadata Grammar.
5ba9f198
MD
779
780trace {
781 major = value; /* Trace format version */
782 minor = value;
3bf79539
MD
783 uuid = value; /* Trace UUID */
784 word_size = value;
785};
5ba9f198 786
3bf79539
MD
787stream {
788 id = stream_id;
77a98c82
MD
789 /* Type 1 - Few event IDs; Type 2 - Many event IDs. See section 6.2. */
790 event.header = event_header_1 OR event_header_2;
791 event.context = struct {
792 ...
3bf79539 793 };
77a98c82
MD
794 packet.context = struct {
795 ...
3bf79539
MD
796 };
797};
5ba9f198
MD
798
799event {
3d13ef1a 800 name = event_name;
3bf79539
MD
801 id = value; /* Numeric identifier within the stream */
802 stream = stream_id;
fcba70d4
MD
803 context = struct {
804 ...
805 };
80fd2569
MD
806 fields = struct {
807 ...
808 };
3bf79539 809};
5ba9f198
MD
810
811/* More detail on types in section 4. Types */
812
3d13ef1a
MD
813/*
814 * Named types:
815 *
fcba70d4
MD
816 * Type declarations behave similarly to the C standard, with the following
817 * added feature: new_type can be preceded by a colon to allow creation of a
818 * type name with prefix/postfix.
3d13ef1a
MD
819 */
820
821typedef aliased_type_prefix aliased_type new_type aliased_type_postfix;
2152348f 822
3d13ef1a 823/* e.g.: typedef struct example new_type_name[10]; */
80fd2569
MD
824
825typedef type_class {
826 ...
fcba70d4 827} : new_type_prefix new_type new_type_postfix;
2152348f 828
3d13ef1a
MD
829/*
830 * e.g.:
831 * typedef integer {
832 * size = 32;
833 * align = 32;
834 * signed = false;
fcba70d4 835 * } : struct page *;
3d13ef1a 836 */
80fd2569
MD
837
838struct name {
3bf79539
MD
839 ...
840};
5ba9f198 841
fcba70d4
MD
842variant name {
843 ...
844};
845
4767a9e7 846enum <integer_type or size> name {
3bf79539
MD
847 ...
848};
849
2152348f
MD
850
851/* Unnamed types, contained within compound type fields or typedef. */
852
80fd2569
MD
853struct {
854 ...
2152348f 855}
5ba9f198 856
fcba70d4
MD
857variant {
858 ...
859}
860
4767a9e7 861enum <integer_type or size> {
80fd2569 862 ...
2152348f
MD
863}
864
865typedef type new_type[length];
3bf79539 866
2152348f
MD
867struct {
868 type field_name[length];
869}
870
871typedef type new_type[length_type];
872
873struct {
874 type field_name[length_type];
875}
876
877integer {
80fd2569 878 ...
2152348f 879}
3bf79539 880
2152348f 881floating_point {
80fd2569 882 ...
2152348f
MD
883}
884
885struct {
886 integer_type field_name:size; /* GNU/C bitfield */
887}
888
889struct {
890 string field_name;
891}
3bf79539 892
fcba70d4 893
3bf79539 894A. Helper macros
5ba9f198
MD
895
896The two following macros keep track of the size of a GNU/C structure without
897padding at the end by placing HEADER_END as the last field. A one byte end field
898is used for C90 compatibility (C99 flexible arrays could be used here). Note
899that this does not affect the effective structure size, which should always be
900calculated with the header_sizeof() helper.
901
902#define HEADER_END char end_field
903#define header_sizeof(type) offsetof(typeof(type), end_field)
3bf79539
MD
904
905
906B. Stream Header Rationale
907
908An event stream is divided in contiguous event packets of variable size. These
909subdivisions allow the trace analyzer to perform a fast binary search by time
910within the stream (typically requiring to index only the event packet headers)
911without reading the whole stream. These subdivisions have a variable size to
912eliminate the need to transfer the event packet padding when partially filled
913event packets must be sent when streaming a trace for live viewing/analysis.
914An event packet can contain a certain amount of padding at the end. Dividing
915streams into event packets is also useful for network streaming over UDP and
916flight recorder mode tracing (a whole event packet can be swapped out of the
917buffer atomically for reading).
918
919The stream header is repeated at the beginning of each event packet to allow
920flexibility in terms of:
921
922 - streaming support,
923 - allowing arbitrary buffers to be discarded without making the trace
924 unreadable,
925 - allow UDP packet loss handling by either dealing with missing event packet
926 or asking for re-transmission.
927 - transparently support flight recorder mode,
928 - transparently support crash dump.
929
930The event stream header will therefore be referred to as the "event packet
931header" throughout the rest of this document.
fcba70d4
MD
932
933C. CTF Metadata Grammar
934
935TODO
This page took 0.103226 seconds and 4 git commands to generate.