2 RFC: Common Trace Format (CTF) Proposal (pre-v1.7)
4 Mathieu Desnoyers, EfficiOS Inc.
6 The goal of the present document is to propose a trace format that suits the
7 needs of the embedded, telecom, high-performance and kernel communities. It is
8 based on the Common Trace Format Requirements (v1.4) document. It is designed to
9 allow traces to be natively generated by the Linux kernel, Linux user-space
10 applications written in C/C++, and hardware components. One major element of
11 CTF is the Trace Stream Description Language (TSDL) which flexibility
12 enables description of various binary trace stream layouts.
14 The latest version of this document can be found at:
16 git tree: git://git.efficios.com/ctf.git
17 gitweb: http://git.efficios.com/?p=ctf.git
19 A reference implementation of a library to read and write this trace format is
20 being implemented within the BabelTrace project, a converter between trace
21 formats. The development tree is available at:
23 git tree: git://git.efficios.com/babeltrace.git
24 gitweb: http://git.efficios.com/?p=babeltrace.git
27 1. Preliminary definitions
29 - Event Trace: An ordered sequence of events.
30 - Event Stream: An ordered sequence of events, containing a subset of the
32 - Event Packet: A sequence of physically contiguous events within an event
34 - Event: This is the basic entry in a trace. (aka: a trace record).
35 - An event identifier (ID) relates to the class (a type) of event within
37 e.g. event: irq_entry.
38 - An event (or event record) relates to a specific instance of an event
40 e.g. event: irq_entry, at time X, on CPU Y
41 - Source Architecture: Architecture writing the trace.
42 - Reader Architecture: Architecture reading the trace.
45 2. High-level representation of a trace
47 A trace is divided into multiple event streams. Each event stream contains a
48 subset of the trace event types.
50 The final output of the trace, after its generation and optional transport over
51 the network, is expected to be either on permanent or temporary storage in a
52 virtual file system. Because each event stream is appended to while a trace is
53 being recorded, each is associated with a separate file for output. Therefore,
54 a stored trace can be represented as a directory containing one file per stream.
56 Meta-data description associated with the trace contains information on
57 trace event types expressed in the Trace Stream Description Language
58 (TSDL). This language describes:
62 - Per-trace event header description.
63 - Per-stream event header description.
64 - Per-stream event context description.
66 - Event type to stream mapping.
67 - Event type to name mapping.
68 - Event type to ID mapping.
69 - Event context description.
70 - Event fields description.
75 An event stream can be divided into contiguous event packets of variable
76 size. These subdivisions have a variable size. An event packet can
77 contain a certain amount of padding at the end. The stream header is
78 repeated at the beginning of each event packet. The rationale for the
79 event stream design choices is explained in Appendix B. Stream Header
82 The event stream header will therefore be referred to as the "event packet
83 header" throughout the rest of this document.
88 Types are organized as type classes. Each type class belong to either of two
89 kind of types: basic types or compound types.
93 A basic type is a scalar type, as described in this section. It includes
94 integers, GNU/C bitfields, enumerations, and floating point values.
96 4.1.1 Type inheritance
98 Type specifications can be inherited to allow deriving types from a
99 type class. For example, see the uint32_t named type derived from the "integer"
100 type class below ("Integers" section). Types have a precise binary
101 representation in the trace. A type class has methods to read and write these
102 types, but must be derived into a type to be usable in an event field.
106 We define "byte-packed" types as aligned on the byte size, namely 8-bit.
107 We define "bit-packed" types as following on the next bit, as defined by the
110 Each basic type must specify its alignment, in bits. Examples of
111 possible alignments are: bit-packed (align = 1), byte-packed (align =
112 8), or word-aligned (e.g. align = 32 or align = 64). The choice depends
113 on the architecture preference and compactness vs performance trade-offs
114 of the implementation. Architectures providing fast unaligned write
115 byte-packed basic types to save space, aligning each type on byte
116 boundaries (8-bit). Architectures with slow unaligned writes align types
117 on specific alignment values. If no specific alignment is declared for a
118 type, it is assumed to be bit-packed for integers with size not multiple
119 of 8 bits and for gcc bitfields. All other basic types are byte-packed
120 by default. It is however recommended to always specify the alignment
121 explicitly. Alignment values must be power of two. Compound types are
122 aligned as specified in their individual specification.
124 TSDL meta-data attribute representation of a specific alignment:
126 align = value; /* value in bits */
130 By default, the native endianness of the source architecture the trace is used.
131 Byte order can be overridden for a basic type by specifying a "byte_order"
132 attribute. Typical use-case is to specify the network byte order (big endian:
133 "be") to save data captured from the network into the trace without conversion.
134 If not specified, the byte order is native.
136 TSDL meta-data representation:
138 byte_order = native OR network OR be OR le; /* network and be are aliases */
142 Type size, in bits, for integers and floats is that returned by "sizeof()" in C
143 multiplied by CHAR_BIT.
144 We require the size of "char" and "unsigned char" types (CHAR_BIT) to be fixed
145 to 8 bits for cross-endianness compatibility.
147 TSDL meta-data representation:
149 size = value; (value is in bits)
153 Signed integers are represented in two-complement. Integer alignment,
154 size, signedness and byte ordering are defined in the TSDL meta-data.
155 Integers aligned on byte size (8-bit) and with length multiple of byte
156 size (8-bit) correspond to the C99 standard integers. In addition,
157 integers with alignment and/or size that are _not_ a multiple of the
158 byte size are permitted; these correspond to the C99 standard bitfields,
159 with the added specification that the CTF integer bitfields have a fixed
160 binary representation. A MIT-licensed reference implementation of the
161 CTF portable bitfields is available at:
163 http://git.efficios.com/?p=babeltrace.git;a=blob;f=include/babeltrace/bitfield.h
165 Binary representation of integers:
167 - On little and big endian:
168 - Within a byte, high bits correspond to an integer high bits, and low bits
169 correspond to low bits.
171 - Integer across multiple bytes are placed from the less significant to the
173 - Consecutive integers are placed from lower bits to higher bits (even within
176 - Integer across multiple bytes are placed from the most significant to the
178 - Consecutive integers are placed from higher bits to lower bits (even within
181 This binary representation is derived from the bitfield implementation in GCC
182 for little and big endian. However, contrary to what GCC does, integers can
183 cross units boundaries (no padding is required). Padding can be explicitly
184 added (see 4.1.6 GNU/C bitfields) to follow the GCC layout if needed.
186 TSDL meta-data representation:
189 signed = true OR false; /* default false */
190 byte_order = native OR network OR be OR le; /* default native */
191 size = value; /* value in bits, no default */
192 align = value; /* value in bits */
193 /* based used for pretty-printing output, default: decimal. */
194 base = decimal OR dec OR OR d OR i OR u OR 10 OR hexadecimal OR hex OR x OR X OR p OR 16
195 OR octal OR oct OR o OR 8 OR binary OR b OR 2;
196 /* character encoding, default: none */
197 encoding = none or UTF8 or ASCII;
200 Example of type inheritance (creation of a uint32_t named type):
208 Definition of a named 5-bit signed bitfield:
216 The character encoding field can be used to specify that the integer
217 must be printed as a text character when read. e.g.:
227 4.1.6 GNU/C bitfields
229 The GNU/C bitfields follow closely the integer representation, with a
230 particularity on alignment: if a bitfield cannot fit in the current unit, the
231 unit is padded and the bitfield starts at the following unit. The unit size is
232 defined by the size of the type "unit_type".
234 TSDL meta-data representation:
238 As an example, the following structure declared in C compiled by GCC:
245 The example structure is aligned on the largest element (short). The second
246 bitfield would be aligned on the next unit boundary, because it would not fit in
251 The floating point values byte ordering is defined in the TSDL meta-data.
253 Floating point values follow the IEEE 754-2008 standard interchange formats.
254 Description of the floating point values include the exponent and mantissa size
255 in bits. Some requirements are imposed on the floating point values:
257 - FLT_RADIX must be 2.
258 - mant_dig is the number of digits represented in the mantissa. It is specified
259 by the ISO C99 standard, section 5.2.4, as FLT_MANT_DIG, DBL_MANT_DIG and
260 LDBL_MANT_DIG as defined by <float.h>.
261 - exp_dig is the number of digits represented in the exponent. Given that
262 mant_dig is one bit more than its actual size in bits (leading 1 is not
263 needed) and also given that the sign bit always takes one bit, exp_dig can be
266 - sizeof(float) * CHAR_BIT - FLT_MANT_DIG
267 - sizeof(double) * CHAR_BIT - DBL_MANT_DIG
268 - sizeof(long double) * CHAR_BIT - LDBL_MANT_DIG
270 TSDL meta-data representation:
275 byte_order = native OR network OR be OR le;
279 Example of type inheritance:
281 typealias floating_point {
282 exp_dig = 8; /* sizeof(float) * CHAR_BIT - FLT_MANT_DIG */
283 mant_dig = 24; /* FLT_MANT_DIG */
288 TODO: define NaN, +inf, -inf behavior.
290 Bit-packed, byte-packed or larger alignments can be used for floating
291 point values, similarly to integers.
295 Enumerations are a mapping between an integer type and a table of strings. The
296 numerical representation of the enumeration follows the integer type specified
297 by the meta-data. The enumeration mapping table is detailed in the enumeration
298 description within the meta-data. The mapping table maps inclusive value
299 ranges (or single values) to strings. Instead of being limited to simple
300 "value -> string" mappings, these enumerations map
301 "[ start_value ... end_value ] -> string", which map inclusive ranges of
302 values to strings. An enumeration from the C language can be represented in
303 this format by having the same start_value and end_value for each element, which
304 is in fact a range of size 1. This single-value range is supported without
305 repeating the start and end values with the value = string declaration.
307 enum name : integer_type {
308 somestring = start_value1 ... end_value1,
309 "other string" = start_value2 ... end_value2,
310 yet_another_string, /* will be assigned to end_value2 + 1 */
311 "some other string" = value,
315 If the values are omitted, the enumeration starts at 0 and increment of 1 for
318 enum name : unsigned int {
326 Overlapping ranges within a single enumeration are implementation defined.
328 A nameless enumeration can be declared as a field type or as part of a typedef:
330 enum : integer_type {
334 Enumerations omitting the container type ": integer_type" use the "int"
335 type (for compatibility with C99). The "int" type must be previously
338 typealias integer { size = 32; align = 32; signed = true } := int;
347 Compound are aggregation of type declarations. Compound types include
348 structures, variant, arrays, sequences, and strings.
352 Structures are aligned on the largest alignment required by basic types
353 contained within the structure. (This follows the ISO/C standard for structures)
355 TSDL meta-data representation of a named structure:
358 field_type field_name;
359 field_type field_name;
366 integer { /* Nameless type */
371 uint64_t second_field_name; /* Named type declared in the meta-data */
374 The fields are placed in a sequence next to each other. They each possess a
375 field name, which is a unique identifier within the structure.
377 A nameless structure can be declared as a field type or as part of a typedef:
383 Alignment for a structure compound type can be forced to a minimum value
384 by adding an "align" specifier after the declaration of a structure
385 body. This attribute is read as: align(value). The value is specified in
386 bits. The structure will be aligned on the maximum value between this
387 attribute and the alignment required by the basic types contained within
394 4.2.2 Variants (Discriminated/Tagged Unions)
396 A CTF variant is a selection between different types. A CTF variant must
397 always be defined within the scope of a structure or within fields
398 contained within a structure (defined recursively). A "tag" enumeration
399 field must appear in either the same lexical scope, prior to the variant
400 field (in field declaration order), in an upper lexical scope (see
401 Section 7.3.1), or in an upper dynamic scope (see Section 7.3.2). The
402 type selection is indicated by the mapping from the enumeration value to
403 the string used as variant type selector. The field to use as tag is
404 specified by the "tag_field", specified between "< >" after the
405 "variant" keyword for unnamed variants, and after "variant name" for
408 The alignment of the variant is the alignment of the type as selected by the tag
409 value for the specific instance of the variant. The alignment of the type
410 containing the variant is independent of the variant alignment. The size of the
411 variant is the size as selected by the tag value for the specific instance of
414 A named variant declaration followed by its definition within a structure
425 enum : integer_type { sel1, sel2, sel3, ... } tag_field;
427 variant name <tag_field> v;
430 An unnamed variant definition within a structure is expressed by the following
434 enum : integer_type { sel1, sel2, sel3, ... } tag_field;
436 variant <tag_field> {
444 Example of a named variant within a sequence that refers to a single tag field:
453 enum : uint2_t { a, b, c } choice;
455 variant example <choice> v[seqlen];
458 Example of an unnamed variant:
461 enum : uint2_t { a, b, c, d } choice;
462 /* Unrelated fields can be added between the variant and its tag */
475 Example of an unnamed variant within an array:
478 enum : uint2_t { a, b, c } choice;
486 Example of a variant type definition within a structure, where the defined type
487 is then declared within an array of structures. This variant refers to a tag
488 located in an upper lexical scope. This example clearly shows that a variant
489 type definition referring to the tag "x" uses the closest preceding field from
490 the lexical scope of the type definition.
493 enum : uint2_t { a, b, c, d } x;
495 typedef variant <x> { /*
496 * "x" refers to the preceding "x" enumeration in the
497 * lexical scope of the type definition.
505 enum : int { x, y, z } x; /* This enumeration is not used by "v". */
506 example_variant v; /*
507 * "v" uses the "enum : uint2_t { a, b, c, d }"
515 Arrays are fixed-length. Their length is declared in the type
516 declaration within the meta-data. They contain an array of "inner type"
517 elements, which can refer to any type not containing the type of the
518 array being declared (no circular dependency). The length is the number
519 of elements in an array.
521 TSDL meta-data representation of a named array:
523 typedef elem_type name[length];
525 A nameless array can be declared as a field type within a structure, e.g.:
527 uint8_t field_name[10];
529 Arrays are always aligned on their element alignment requirement.
533 Sequences are dynamically-sized arrays. They refer to a a "length"
534 unsigned integer field, which must appear in either the same lexical scope,
535 prior to the sequence field (in field declaration order), in an upper
536 lexical scope (see Section 7.3.1), or in an upper dynamic scope (see
537 Section 7.3.2). This length field represents the number of elements in
538 the sequence. The sequence per se is an array of "inner type" elements.
540 TSDL meta-data representation for a sequence type definition:
543 unsigned int length_field;
544 typedef elem_type typename[length_field];
545 typename seq_field_name;
548 A sequence can also be declared as a field type, e.g.:
551 unsigned int length_field;
552 long seq_field_name[length_field];
555 Multiple sequences can refer to the same length field, and these length
556 fields can be in a different upper dynamic scope:
558 e.g., assuming the stream.event.header defines:
563 event.header := struct {
572 long seq_a[stream.event.header.seq_len];
573 char seq_b[stream.event.header.seq_len];
577 The sequence elements follow the "array" specifications.
581 Strings are an array of bytes of variable size and are terminated by a '\0'
582 "NULL" character. Their encoding is described in the TSDL meta-data. In
583 absence of encoding attribute information, the default encoding is
586 TSDL meta-data representation of a named string type:
589 encoding = UTF8 OR ASCII;
592 A nameless string type can be declared as a field type:
594 string field_name; /* Use default UTF8 encoding */
596 Strings are always aligned on byte size.
598 5. Event Packet Header
600 The event packet header consists of two parts: the "event packet header"
601 is the same for all streams of a trace. The second part, the "event
602 packet context", is described on a per-stream basis. Both are described
603 in the TSDL meta-data. The packets are aligned on architecture-page-sized
606 Event packet header (all fields are optional, specified by TSDL meta-data):
608 - Magic number (CTF magic number: 0xC1FC1FC1) specifies that this is a
609 CTF packet. This magic number is optional, but when present, it should
610 come at the very beginning of the packet.
611 - Trace UUID, used to ensure the event packet match the meta-data used.
612 (note: we cannot use a meta-data checksum in every cases instead of a
613 UUID because meta-data can be appended to while tracing is active)
614 This field is optional.
615 - Stream ID, used as reference to stream description in meta-data.
616 This field is optional if there is only one stream description in the
617 meta-data, but becomes required if there are more than one stream in
618 the TSDL meta-data description.
620 Event packet context (all fields are optional, specified by TSDL meta-data):
622 - Event packet content size (in bytes).
623 - Event packet size (in bytes, includes padding).
624 - Event packet content checksum (optional). Checksum excludes the event packet
626 - Per-stream event packet sequence count (to deal with UDP packet loss). The
627 number of significant sequence counter bits should also be present, so
628 wrap-arounds are dealt with correctly.
629 - Time-stamp at the beginning and time-stamp at the end of the event packet.
630 Both timestamps are written in the packet header, but sampled respectively
631 while (or before) writing the first event and while (or after) writing the
632 last event in the packet. The inclusive range between these timestamps should
633 include all event timestamps assigned to events contained within the packet.
634 - Events discarded count
635 - Snapshot of a per-stream free-running counter, counting the number of
636 events discarded that were supposed to be written in the stream prior to
637 the first event in the event packet.
638 * Note: producer-consumer buffer full condition should fill the current
639 event packet with padding so we know exactly where events have been
641 - Lossless compression scheme used for the event packet content. Applied
642 directly to raw data. New types of compression can be added in following
643 versions of the format.
644 0: no compression scheme
648 - Cypher used for the event packet content. Applied after compression.
651 - Checksum scheme used for the event packet content. Applied after encryption.
657 5.1 Event Packet Header Description
659 The event packet header layout is indicated by the trace packet.header
660 field. Here is a recommended structure type for the packet header with
661 the fields typically expected (although these fields are each optional):
663 struct event_packet_header {
671 packet.header := struct event_packet_header;
674 If the magic number is not present, tools such as "file" will have no
675 mean to discover the file type.
677 If the uuid is not present, no validation that the meta-data actually
678 corresponds to the stream is performed.
680 If the stream_id packet header field is missing, the trace can only
681 contain a single stream. Its "id" field can be left out, and its events
682 don't need to declare a "stream_id" field.
685 5.2 Event Packet Context Description
687 Event packet context example. These are declared within the stream declaration
688 in the meta-data. All these fields are optional. If the packet size field is
689 missing, the whole stream only contains a single packet. If the content
690 size field is missing, the packet is filled (no padding). The content
691 and packet sizes include all headers.
693 An example event packet context type:
695 struct event_packet_context {
696 uint64_t timestamp_begin;
697 uint64_t timestamp_end;
699 uint32_t stream_packet_count;
700 uint32_t events_discarded;
702 uint32_t/uint16_t content_size;
703 uint32_t/uint16_t packet_size;
704 uint8_t stream_packet_count_bits; /* Significant counter bits */
705 uint8_t compression_scheme;
706 uint8_t encryption_scheme;
707 uint8_t checksum_scheme;
713 The overall structure of an event is:
715 1 - Stream Packet Context (as specified by the stream meta-data)
716 2 - Event Header (as specified by the stream meta-data)
717 3 - Stream Event Context (as specified by the stream meta-data)
718 4 - Event Context (as specified by the event meta-data)
719 5 - Event Payload (as specified by the event meta-data)
721 This structure defines an implicit dynamic scoping, where variants
722 located in inner structures (those with a higher number in the listing
723 above) can refer to the fields of outer structures (with lower number in
724 the listing above). See Section 7.3 TSDL Scopes for more detail.
728 Event headers can be described within the meta-data. We hereby propose, as an
729 example, two types of events headers. Type 1 accommodates streams with less than
730 31 event IDs. Type 2 accommodates streams with 31 or more event IDs.
732 One major factor can vary between streams: the number of event IDs assigned to
733 a stream. Luckily, this information tends to stay relatively constant (modulo
734 event registration while trace is being recorded), so we can specify different
735 representations for streams containing few event IDs and streams containing
736 many event IDs, so we end up representing the event ID and time-stamp as
737 densely as possible in each case.
739 The header is extended in the rare occasions where the information cannot be
740 represented in the ranges available in the standard event header. They are also
741 used in the rare occasions where the data required for a field could not be
742 collected: the flag corresponding to the missing field within the missing_fields
743 array is then set to 1.
745 Types uintX_t represent an X-bit unsigned integer, as declared with
748 typealias integer { size = X; align = X; signed = false } := uintX_t;
752 typealias integer { size = X; align = 1; signed = false } := uintX_t;
754 6.1.1 Type 1 - Few event IDs
756 - Aligned on 32-bit (or 8-bit if byte-packed, depending on the architecture
758 - Native architecture byte ordering.
759 - For "compact" selection
760 - Fixed size: 32 bits.
761 - For "extended" selection
762 - Size depends on the architecture and variant alignment.
764 struct event_header_1 {
767 * id 31 is reserved to indicate an extended header.
769 enum : uint5_t { compact = 0 ... 30, extended = 31 } id;
775 uint32_t id; /* 32-bit event IDs */
776 uint64_t timestamp; /* 64-bit timestamps */
779 } align(32); /* or align(8) */
782 6.1.2 Type 2 - Many event IDs
784 - Aligned on 16-bit (or 8-bit if byte-packed, depending on the architecture
786 - Native architecture byte ordering.
787 - For "compact" selection
788 - Size depends on the architecture and variant alignment.
789 - For "extended" selection
790 - Size depends on the architecture and variant alignment.
792 struct event_header_2 {
794 * id: range: 0 - 65534.
795 * id 65535 is reserved to indicate an extended header.
797 enum : uint16_t { compact = 0 ... 65534, extended = 65535 } id;
803 uint32_t id; /* 32-bit event IDs */
804 uint64_t timestamp; /* 64-bit timestamps */
807 } align(16); /* or align(8) */
812 The event context contains information relative to the current event.
813 The choice and meaning of this information is specified by the TSDL
814 stream and event meta-data descriptions. The stream context is applied
815 to all events within the stream. The stream context structure follows
816 the event header. The event context is applied to specific events. Its
817 structure follows the stream context structure.
819 An example of stream-level event context is to save the event payload size with
820 each event, or to save the current PID with each event. These are declared
821 within the stream declaration within the meta-data:
825 event.context := struct {
827 uint16_t payload_size;
831 An example of event-specific event context is to declare a bitmap of missing
832 fields, only appended after the stream event context if the extended event
833 header is selected. NR_FIELDS is the number of fields within the event (a
841 uint1_t missing_fields[NR_FIELDS]; /* missing event fields bitmap */
850 An event payload contains fields specific to a given event type. The fields
851 belonging to an event type are described in the event-specific meta-data
852 within a structure type.
856 No padding at the end of the event payload. This differs from the ISO/C standard
857 for structures, but follows the CTF standard for structures. In a trace, even
858 though it makes sense to align the beginning of a structure, it really makes no
859 sense to add padding at the end of the structure, because structures are usually
860 not followed by a structure of the same type.
862 This trick can be done by adding a zero-length "end" field at the end of the C
863 structures, and by using the offset of this field rather than using sizeof()
864 when calculating the size of a structure (see Appendix "A. Helper macros").
868 The event payload is aligned on the largest alignment required by types
869 contained within the payload. (This follows the ISO/C standard for structures)
872 7. Trace Stream Description Language (TSDL)
874 The Trace Stream Description Language (TSDL) allows expression of the
875 binary trace streams layout in a C99-like Domain Specific Language
881 The trace stream layout description is located in the trace meta-data.
882 The meta-data is itself located in a stream identified by its name:
885 The meta-data description can be expressed in two different formats:
886 text-only and packet-based. The text-only description facilitates
887 generation of meta-data and provides a convenient way to enter the
888 meta-data information by hand. The packet-based meta-data provides the
889 CTF stream packet facilities (checksumming, compression, encryption,
890 network-readiness) for meta-data stream generated and transported by a
893 The text-only meta-data file is a plain text TSDL description.
895 The packet-based meta-data is made of "meta-data packets", which each
896 start with a meta-data packet header. The packet-based meta-data
897 description is detected by reading the magic number "0x75D11D57" at the
898 beginning of the file. This magic number is also used to detect the
899 endianness of the architecture by trying to read the CTF magic number
900 and its counterpart in reversed endianness. The events within the
901 meta-data stream have no event header nor event context. Each event only
902 contains a "sequence" payload, which is a sequence of bits using the
903 "trace.packet.header.content_size" field as a placeholder for its length
904 (the packet header size should be substracted). The formatting of this
905 sequence of bits is a plain-text representation of the TSDL description.
906 Each meta-data packet start with a special packet header, specific to
907 the meta-data stream, which contains, exactly:
909 struct metadata_packet_header {
910 uint32_t magic; /* 0x75D11D57 */
911 uint8_t uuid[16]; /* Unique Universal Identifier */
912 uint32_t checksum; /* 0 if unused */
913 uint32_t content_size; /* in bits */
914 uint32_t packet_size; /* in bits */
915 uint8_t compression_scheme; /* 0 if unused */
916 uint8_t encryption_scheme; /* 0 if unused */
917 uint8_t checksum_scheme; /* 0 if unused */
920 The packet-based meta-data can be converted to a text-only meta-data by
921 concatenating all the strings in contains.
923 In the textual representation of the meta-data, the text contained
924 within "/*" and "*/", as well as within "//" and end of line, are
925 treated as comments. Boolean values can be represented as true, TRUE,
926 or 1 for true, and false, FALSE, or 0 for false. Within the string-based
927 meta-data description, the trace UUID is represented as a string of
928 hexadecimal digits and dashes "-". In the event packet header, the trace
929 UUID is represented as an array of bytes.
932 7.2 Declaration vs Definition
934 A declaration associates a layout to a type, without specifying where
935 this type is located in the event structure hierarchy (see Section 6).
936 This therefore includes typedef, typealias, as well as all type
937 specifiers. In certain circumstances (typedef, structure field and
938 variant field), a declaration is followed by a declarator, which specify
939 the newly defined type name (for typedef), or the field name (for
940 declarations located within structure and variants). Array and sequence,
941 declared with square brackets ("[" "]"), are part of the declarator,
942 similarly to C99. The enumeration base type is specified by
943 ": enum_base", which is part of the type specifier. The variant tag
944 name, specified between "<" ">", is also part of the type specifier.
946 A definition associates a type to a location in the event structure
947 hierarchy (see Section 6). This association is denoted by ":=", as shown
953 TSDL uses two different types of scoping: a lexical scope is used for
954 declarations and type definitions, and a dynamic scope is used for
955 variants references to tag fields and for sequence references to length
960 Each of "trace", "stream", "event", "struct" and "variant" have their own
961 nestable declaration scope, within which types can be declared using "typedef"
962 and "typealias". A root declaration scope also contains all declarations
963 located outside of any of the aforementioned declarations. An inner
964 declaration scope can refer to type declared within its container
965 lexical scope prior to the inner declaration scope. Redefinition of a
966 typedef or typealias is not valid, although hiding an upper scope
967 typedef or typealias is allowed within a sub-scope.
971 A dynamic scope consists in the lexical scope augmented with the
972 implicit event structure definition hierarchy presented at Section 6.
973 The dynamic scope is used for variant tag and sequence length
974 definitions. It is used at definition time to look up the location of
975 the tag field associated with a variant, and to lookup up the location
976 of the length field associated with a sequence.
978 Therefore, variants (or sequences) in lower levels in the dynamic scope
979 (e.g. event context) can refer to a tag (or length) field located in
980 upper levels (e.g. in the event header) by specifying, in this case, the
981 associated tag with <header.field_name>. This allows, for instance, the
982 event context to define a variant referring to the "id" field of the
983 event header as selector.
985 The target dynamic scope must be specified explicitly when referring to
986 a field outside of the local static scope. The dynamic scope prefixes
989 - Trace Packet Header: <trace.packet.header. >,
990 - Stream Packet Context: <stream.packet.context. >,
991 - Event Header: <stream.event.header. >,
992 - Stream Event Context: <stream.event.context. >,
993 - Event Context: <event.context. >,
994 - Event Payload: <event.fields. >.
996 Multiple declarations of the same field name within a single scope is
997 not valid. It is however valid to re-use the same field name in
998 different scopes. There is no possible conflict, because the dynamic
999 scope must be specified when a variant refers to a tag field located in
1000 a different dynamic scope.
1002 The information available in the dynamic scopes can be thought of as the
1003 current tracing context. At trace production, information about the
1004 current context is saved into the specified scope field levels. At trace
1005 consumption, for each event, the current trace context is therefore
1006 readable by accessing the upper dynamic scopes.
1011 The grammar representing the TSDL meta-data is presented in Appendix C.
1012 TSDL Grammar. This section presents a rather lighter reading that
1013 consists in examples of TSDL meta-data, with template values.
1015 The stream "id" can be left out if there is only one stream in the
1016 trace. The event "id" field can be left out if there is only one event
1020 major = value; /* Trace format version */
1022 uuid = "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa"; /* Trace UUID */
1023 byte_order = be OR le; /* Endianness (required) */
1024 packet.header := struct {
1033 /* Type 1 - Few event IDs; Type 2 - Many event IDs. See section 6.1. */
1034 event.header := event_header_1 OR event_header_2;
1035 event.context := struct {
1038 packet.context := struct {
1045 id = value; /* Numeric identifier within the stream */
1046 stream_id = stream_id;
1055 /* More detail on types in section 4. Types */
1060 * Type declarations behave similarly to the C standard.
1063 typedef aliased_type_specifiers new_type_declarators;
1065 /* e.g.: typedef struct example new_type_name[10]; */
1070 * The "typealias" declaration can be used to give a name (including
1071 * pointer declarator specifier) to a type. It should also be used to
1072 * map basic C types (float, int, unsigned long, ...) to a CTF type.
1073 * Typealias is a superset of "typedef": it also allows assignment of a
1074 * simple variable identifier to a type.
1077 typealias type_class {
1079 } := type_specifiers type_declarator;
1083 * typealias integer {
1087 * } := struct page *;
1089 * typealias integer {
1104 enum name : integer_type {
1110 * Unnamed types, contained within compound type fields, typedef or typealias.
1125 enum : integer_type {
1129 typedef type new_type[length];
1132 type field_name[length];
1135 typedef type new_type[length_type];
1138 type field_name[length_type];
1150 integer_type field_name:size; /* GNU/C bitfield */
1160 The two following macros keep track of the size of a GNU/C structure without
1161 padding at the end by placing HEADER_END as the last field. A one byte end field
1162 is used for C90 compatibility (C99 flexible arrays could be used here). Note
1163 that this does not affect the effective structure size, which should always be
1164 calculated with the header_sizeof() helper.
1166 #define HEADER_END char end_field
1167 #define header_sizeof(type) offsetof(typeof(type), end_field)
1170 B. Stream Header Rationale
1172 An event stream is divided in contiguous event packets of variable size. These
1173 subdivisions allow the trace analyzer to perform a fast binary search by time
1174 within the stream (typically requiring to index only the event packet headers)
1175 without reading the whole stream. These subdivisions have a variable size to
1176 eliminate the need to transfer the event packet padding when partially filled
1177 event packets must be sent when streaming a trace for live viewing/analysis.
1178 An event packet can contain a certain amount of padding at the end. Dividing
1179 streams into event packets is also useful for network streaming over UDP and
1180 flight recorder mode tracing (a whole event packet can be swapped out of the
1181 buffer atomically for reading).
1183 The stream header is repeated at the beginning of each event packet to allow
1184 flexibility in terms of:
1186 - streaming support,
1187 - allowing arbitrary buffers to be discarded without making the trace
1189 - allow UDP packet loss handling by either dealing with missing event packet
1190 or asking for re-transmission.
1191 - transparently support flight recorder mode,
1192 - transparently support crash dump.
1198 * Common Trace Format (CTF) Trace Stream Description Language (TSDL) Grammar.
1200 * Inspired from the C99 grammar:
1201 * http://www.open-std.org/jtc1/sc22/wg14/www/docs/n1124.pdf (Annex A)
1202 * and c++1x grammar (draft)
1203 * http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2011/n3291.pdf (Annex A)
1205 * Specialized for CTF needs by including only constant and declarations from
1206 * C99 (excluding function declarations), and by adding support for variants,
1207 * sequences and CTF-specific specifiers. Enumeration container types
1208 * semantic is inspired from c++1x enum-base.
1213 1.1) Lexical elements
1257 identifier identifier-nondigit
1260 identifier-nondigit:
1262 universal-character-name
1263 any other implementation-defined characters
1267 [a-zA-Z] /* regular expression */
1270 [0-9] /* regular expression */
1272 1.4) Universal character names
1274 universal-character-name:
1276 \U hex-quad hex-quad
1279 hexadecimal-digit hexadecimal-digit hexadecimal-digit hexadecimal-digit
1285 enumeration-constant
1289 decimal-constant integer-suffix-opt
1290 octal-constant integer-suffix-opt
1291 hexadecimal-constant integer-suffix-opt
1295 decimal-constant digit
1299 octal-constant octal-digit
1301 hexadecimal-constant:
1302 hexadecimal-prefix hexadecimal-digit
1303 hexadecimal-constant hexadecimal-digit
1313 unsigned-suffix long-suffix-opt
1314 unsigned-suffix long-long-suffix
1315 long-suffix unsigned-suffix-opt
1316 long-long-suffix unsigned-suffix-opt
1330 enumeration-constant:
1336 L' c-char-sequence '
1340 c-char-sequence c-char
1343 any member of source charset except single-quote ('), backslash
1344 (\), or new-line character.
1348 simple-escape-sequence
1349 octal-escape-sequence
1350 hexadecimal-escape-sequence
1351 universal-character-name
1353 simple-escape-sequence: one of
1354 \' \" \? \\ \a \b \f \n \r \t \v
1356 octal-escape-sequence:
1358 \ octal-digit octal-digit
1359 \ octal-digit octal-digit octal-digit
1361 hexadecimal-escape-sequence:
1362 \x hexadecimal-digit
1363 hexadecimal-escape-sequence hexadecimal-digit
1365 1.6) String literals
1368 " s-char-sequence-opt "
1369 L" s-char-sequence-opt "
1373 s-char-sequence s-char
1376 any member of source charset except double-quote ("), backslash
1377 (\), or new-line character.
1383 [ ] ( ) { } . -> * + - < > : ; ... = ,
1386 2) Phrase structure grammar
1392 ( unary-expression )
1396 postfix-expression [ unary-expression ]
1397 postfix-expression . identifier
1398 postfix-expressoin -> identifier
1402 unary-operator postfix-expression
1404 unary-operator: one of
1407 assignment-operator:
1410 type-assignment-operator:
1413 constant-expression-range:
1414 unary-expression ... unary-expression
1419 declaration-specifiers declarator-list-opt ;
1422 declaration-specifiers:
1423 storage-class-specifier declaration-specifiers-opt
1424 type-specifier declaration-specifiers-opt
1425 type-qualifier declaration-specifiers-opt
1429 declarator-list , declarator
1431 abstract-declarator-list:
1433 abstract-declarator-list , abstract-declarator
1435 storage-class-specifier:
1458 align ( unary-expression )
1461 struct identifier-opt { struct-or-variant-declaration-list-opt } align-attribute-opt
1462 struct identifier align-attribute-opt
1464 struct-or-variant-declaration-list:
1465 struct-or-variant-declaration
1466 struct-or-variant-declaration-list struct-or-variant-declaration
1468 struct-or-variant-declaration:
1469 specifier-qualifier-list struct-or-variant-declarator-list ;
1470 declaration-specifiers-opt storage-class-specifier declaration-specifiers-opt declarator-list ;
1471 typealias declaration-specifiers abstract-declarator-list type-assignment-operator declaration-specifiers abstract-declarator-list ;
1472 typealias declaration-specifiers abstract-declarator-list type-assignment-operator declarator-list ;
1474 specifier-qualifier-list:
1475 type-specifier specifier-qualifier-list-opt
1476 type-qualifier specifier-qualifier-list-opt
1478 struct-or-variant-declarator-list:
1479 struct-or-variant-declarator
1480 struct-or-variant-declarator-list , struct-or-variant-declarator
1482 struct-or-variant-declarator:
1484 declarator-opt : unary-expression
1487 variant identifier-opt variant-tag-opt { struct-or-variant-declaration-list }
1488 variant identifier variant-tag
1494 enum identifier-opt { enumerator-list }
1495 enum identifier-opt { enumerator-list , }
1497 enum identifier-opt : declaration-specifiers { enumerator-list }
1498 enum identifier-opt : declaration-specifiers { enumerator-list , }
1502 enumerator-list , enumerator
1505 enumeration-constant
1506 enumeration-constant assignment-operator unary-expression
1507 enumeration-constant assignment-operator constant-expression-range
1513 pointer-opt direct-declarator
1518 direct-declarator [ unary-expression ]
1520 abstract-declarator:
1521 pointer-opt direct-abstract-declarator
1523 direct-abstract-declarator:
1525 ( abstract-declarator )
1526 direct-abstract-declarator [ unary-expression ]
1527 direct-abstract-declarator [ ]
1530 * type-qualifier-list-opt
1531 * type-qualifier-list-opt pointer
1533 type-qualifier-list:
1535 type-qualifier-list type-qualifier
1540 2.3) CTF-specific declarations
1543 event { ctf-assignment-expression-list-opt }
1544 stream { ctf-assignment-expression-list-opt }
1545 trace { ctf-assignment-expression-list-opt }
1546 typealias declaration-specifiers abstract-declarator-list type-assignment-operator declaration-specifiers abstract-declarator-list
1547 typealias declaration-specifiers abstract-declarator-list type-assignment-operator declarator-list
1550 floating_point { ctf-assignment-expression-list-opt }
1551 integer { ctf-assignment-expression-list-opt }
1552 string { ctf-assignment-expression-list-opt }
1555 ctf-assignment-expression-list:
1556 ctf-assignment-expression ;
1557 ctf-assignment-expression-list ctf-assignment-expression ;
1559 ctf-assignment-expression:
1560 unary-expression assignment-operator unary-expression
1561 unary-expression type-assignment-operator type-specifier
1562 declaration-specifiers-opt storage-class-specifier declaration-specifiers-opt declarator-list
1563 typealias declaration-specifiers abstract-declarator-list type-assignment-operator declaration-specifiers abstract-declarator-list
1564 typealias declaration-specifiers abstract-declarator-list type-assignment-operator declarator-list