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 A metadata event stream contains information on trace event types
57 expressed in the Trace Stream Description Language (TSDL). It describes:
61 - Per-stream event header description.
62 - Per-stream event header selection.
63 - Per-stream event context fields.
65 - Event type to stream mapping.
66 - Event type to name mapping.
67 - Event type to ID mapping.
68 - Event fields description.
73 An event stream is divided in contiguous event packets of variable size. These
74 subdivisions have a variable size. An event packet can contain a certain
75 amount of padding at the end. The stream header is repeated at the
76 beginning of each event packet. The rationale for the event stream
77 design choices is explained in Appendix B. Stream Header Rationale.
79 The event stream header will therefore be referred to as the "event packet
80 header" throughout the rest of this document.
85 Types are organized as type classes. Each type class belong to either of two
86 kind of types: basic types or compound types.
90 A basic type is a scalar type, as described in this section. It includes
91 integers, GNU/C bitfields, enumerations, and floating point values.
93 4.1.1 Type inheritance
95 Type specifications can be inherited to allow deriving types from a
96 type class. For example, see the uint32_t named type derived from the "integer"
97 type class below ("Integers" section). Types have a precise binary
98 representation in the trace. A type class has methods to read and write these
99 types, but must be derived into a type to be usable in an event field.
103 We define "byte-packed" types as aligned on the byte size, namely 8-bit.
104 We define "bit-packed" types as following on the next bit, as defined by the
107 All basic types, except bitfields, are either aligned on an architecture-defined
108 specific alignment or byte-packed, depending on the architecture preference.
109 Architectures providing fast unaligned write byte-packed basic types to save
110 space, aligning each type on byte boundaries (8-bit). Architectures with slow
111 unaligned writes align types on specific alignment values. If no specific
112 alignment is declared for a type, it is assumed to be bit-packed for
113 integers with size not multiple of 8 bits and for gcc bitfields. All
114 other types are byte-packed.
116 Metadata attribute representation of a specific alignment:
118 align = value; /* value in bits */
122 By default, the native endianness of the source architecture the trace is used.
123 Byte order can be overridden for a basic type by specifying a "byte_order"
124 attribute. Typical use-case is to specify the network byte order (big endian:
125 "be") to save data captured from the network into the trace without conversion.
126 If not specified, the byte order is native.
128 Metadata representation:
130 byte_order = native OR network OR be OR le; /* network and be are aliases */
134 Type size, in bits, for integers and floats is that returned by "sizeof()" in C
135 multiplied by CHAR_BIT.
136 We require the size of "char" and "unsigned char" types (CHAR_BIT) to be fixed
137 to 8 bits for cross-endianness compatibility.
139 Metadata representation:
141 size = value; (value is in bits)
145 Signed integers are represented in two-complement. Integer alignment, size,
146 signedness and byte ordering are defined in the metadata. Integers aligned on
147 byte size (8-bit) and with length multiple of byte size (8-bit) correspond to
148 the C99 standard integers. In addition, integers with alignment and/or size that
149 are _not_ a multiple of the byte size are permitted; these correspond to the C99
150 standard bitfields, with the added specification that the CTF integer bitfields
151 have a fixed binary representation. A MIT-licensed reference implementation of
152 the CTF portable bitfields is available at:
154 http://git.efficios.com/?p=babeltrace.git;a=blob;f=include/babeltrace/bitfield.h
156 Binary representation of integers:
158 - On little and big endian:
159 - Within a byte, high bits correspond to an integer high bits, and low bits
160 correspond to low bits.
162 - Integer across multiple bytes are placed from the less significant to the
164 - Consecutive integers are placed from lower bits to higher bits (even within
167 - Integer across multiple bytes are placed from the most significant to the
169 - Consecutive integers are placed from higher bits to lower bits (even within
172 This binary representation is derived from the bitfield implementation in GCC
173 for little and big endian. However, contrary to what GCC does, integers can
174 cross units boundaries (no padding is required). Padding can be explicitely
175 added (see 4.1.6 GNU/C bitfields) to follow the GCC layout if needed.
177 Metadata representation:
180 signed = true OR false; /* default false */
181 byte_order = native OR network OR be OR le; /* default native */
182 size = value; /* value in bits, no default */
183 align = value; /* value in bits */
186 Example of type inheritance (creation of a uint32_t named type):
194 Definition of a named 5-bit signed bitfield:
202 4.1.6 GNU/C bitfields
204 The GNU/C bitfields follow closely the integer representation, with a
205 particularity on alignment: if a bitfield cannot fit in the current unit, the
206 unit is padded and the bitfield starts at the following unit. The unit size is
207 defined by the size of the type "unit_type".
209 Metadata representation:
213 As an example, the following structure declared in C compiled by GCC:
220 The example structure is aligned on the largest element (short). The second
221 bitfield would be aligned on the next unit boundary, because it would not fit in
226 The floating point values byte ordering is defined in the metadata.
228 Floating point values follow the IEEE 754-2008 standard interchange formats.
229 Description of the floating point values include the exponent and mantissa size
230 in bits. Some requirements are imposed on the floating point values:
232 - FLT_RADIX must be 2.
233 - mant_dig is the number of digits represented in the mantissa. It is specified
234 by the ISO C99 standard, section 5.2.4, as FLT_MANT_DIG, DBL_MANT_DIG and
235 LDBL_MANT_DIG as defined by <float.h>.
236 - exp_dig is the number of digits represented in the exponent. Given that
237 mant_dig is one bit more than its actual size in bits (leading 1 is not
238 needed) and also given that the sign bit always takes one bit, exp_dig can be
241 - sizeof(float) * CHAR_BIT - FLT_MANT_DIG
242 - sizeof(double) * CHAR_BIT - DBL_MANT_DIG
243 - sizeof(long double) * CHAR_BIT - LDBL_MANT_DIG
245 Metadata representation:
250 byte_order = native OR network OR be OR le;
253 Example of type inheritance:
255 typealias floating_point {
256 exp_dig = 8; /* sizeof(float) * CHAR_BIT - FLT_MANT_DIG */
257 mant_dig = 24; /* FLT_MANT_DIG */
261 TODO: define NaN, +inf, -inf behavior.
265 Enumerations are a mapping between an integer type and a table of strings. The
266 numerical representation of the enumeration follows the integer type specified
267 by the metadata. The enumeration mapping table is detailed in the enumeration
268 description within the metadata. The mapping table maps inclusive value ranges
269 (or single values) to strings. Instead of being limited to simple
270 "value -> string" mappings, these enumerations map
271 "[ start_value ... end_value ] -> string", which map inclusive ranges of
272 values to strings. An enumeration from the C language can be represented in
273 this format by having the same start_value and end_value for each element, which
274 is in fact a range of size 1. This single-value range is supported without
275 repeating the start and end values with the value = string declaration.
277 enum name : integer_type {
278 somestring = start_value1 ... end_value1,
279 "other string" = start_value2 ... end_value2,
280 yet_another_string, /* will be assigned to end_value2 + 1 */
281 "some other string" = value,
285 If the values are omitted, the enumeration starts at 0 and increment of 1 for
288 enum name : unsigned int {
296 Overlapping ranges within a single enumeration are implementation defined.
298 A nameless enumeration can be declared as a field type or as part of a typedef:
300 enum : integer_type {
304 Enumerations omitting the container type ": integer_type" use the "int"
305 type (for compatibility with C99). The "int" type must be previously
308 typealias integer { size = 32; align = 32; signed = true } := int;
317 Compound are aggregation of type declarations. Compound types include
318 structures, variant, arrays, sequences, and strings.
322 Structures are aligned on the largest alignment required by basic types
323 contained within the structure. (This follows the ISO/C standard for structures)
325 Metadata representation of a named structure:
328 field_type field_name;
329 field_type field_name;
336 integer { /* Nameless type */
341 uint64_t second_field_name; /* Named type declared in the metadata */
344 The fields are placed in a sequence next to each other. They each possess a
345 field name, which is a unique identifier within the structure.
347 A nameless structure can be declared as a field type or as part of a typedef:
353 4.2.2 Variants (Discriminated/Tagged Unions)
355 A CTF variant is a selection between different types. A CTF variant must
356 always be defined within the scope of a structure or within fields
357 contained within a structure (defined recursively). A "tag" enumeration
358 field must appear in either the same lexical scope, prior to the variant
359 field (in field declaration order), in an uppermost lexical scope (see
360 Section 7.3.1), or in an uppermost dynamic scope (see Section 7.3.2).
361 The type selection is indicated by the mapping from the enumeration
362 value to the string used as variant type selector. The field to use as
363 tag is specified by the "tag_field", specified between "< >" after the
364 "variant" keyword for unnamed variants, and after "variant name" for
367 The alignment of the variant is the alignment of the type as selected by the tag
368 value for the specific instance of the variant. The alignment of the type
369 containing the variant is independent of the variant alignment. The size of the
370 variant is the size as selected by the tag value for the specific instance of
373 A named variant declaration followed by its definition within a structure
384 enum : integer_type { sel1, sel2, sel3, ... } tag_field;
386 variant name <tag_field> v;
389 An unnamed variant definition within a structure is expressed by the following
393 enum : integer_type { sel1, sel2, sel3, ... } tag_field;
395 variant <tag_field> {
403 Example of a named variant within a sequence that refers to a single tag field:
412 enum : uint2_t { a, b, c } choice;
413 variant example <choice> v[unsigned int];
416 Example of an unnamed variant:
419 enum : uint2_t { a, b, c, d } choice;
420 /* Unrelated fields can be added between the variant and its tag */
433 Example of an unnamed variant within an array:
436 enum : uint2_t { a, b, c } choice;
444 Example of a variant type definition within a structure, where the defined type
445 is then declared within an array of structures. This variant refers to a tag
446 located in an upper lexical scope. This example clearly shows that a variant
447 type definition referring to the tag "x" uses the closest preceding field from
448 the lexical scope of the type definition.
451 enum : uint2_t { a, b, c, d } x;
453 typedef variant <x> { /*
454 * "x" refers to the preceding "x" enumeration in the
455 * lexical scope of the type definition.
463 enum : int { x, y, z } x; /* This enumeration is not used by "v". */
464 example_variant v; /*
465 * "v" uses the "enum : uint2_t { a, b, c, d }"
473 Arrays are fixed-length. Their length is declared in the type declaration within
474 the metadata. They contain an array of "inner type" elements, which can refer to
475 any type not containing the type of the array being declared (no circular
476 dependency). The length is the number of elements in an array.
478 Metadata representation of a named array:
480 typedef elem_type name[length];
482 A nameless array can be declared as a field type within a structure, e.g.:
484 uint8_t field_name[10];
489 Sequences are dynamically-sized arrays. They start with an integer that specify
490 the length of the sequence, followed by an array of "inner type" elements.
491 The length is the number of elements in the sequence.
493 Metadata representation for a named sequence:
495 typedef elem_type name[length_type];
497 A nameless sequence can be declared as a field type, e.g.:
499 long field_name[int];
501 The length type follows the integer types specifications, and the sequence
502 elements follow the "array" specifications.
506 Strings are an array of bytes of variable size and are terminated by a '\0'
507 "NULL" character. Their encoding is described in the metadata. In absence of
508 encoding attribute information, the default encoding is UTF-8.
510 Metadata representation of a named string type:
513 encoding = UTF8 OR ASCII;
516 A nameless string type can be declared as a field type:
518 string field_name; /* Use default UTF8 encoding */
520 5. Event Packet Header
522 The event packet header consists of two part: one is mandatory and have a fixed
523 layout. The second part, the "event packet context", has its layout described in
526 - Aligned on page size. Fixed size. Fields either aligned or packed (depending
527 on the architecture preference).
528 No padding at the end of the event packet header. Native architecture byte
531 Fixed layout (event packet header):
533 - Magic number (CTF magic number: 0xC1FC1FC1 This magic number specifies
534 that we use the CTF metadata description language described in this
535 document. Different magic numbers should be used for other metadata
536 description languages.
537 - Trace UUID, used to ensure the event packet match the metadata used.
538 (note: we cannot use a metadata checksum because metadata can be appended to
539 while tracing is active)
540 - Stream ID, used as reference to stream description in metadata.
542 Metadata-defined layout (event packet context):
544 - Event packet content size (in bytes).
545 - Event packet size (in bytes, includes padding).
546 - Event packet content checksum (optional). Checksum excludes the event packet
548 - Per-stream event packet sequence count (to deal with UDP packet loss). The
549 number of significant sequence counter bits should also be present, so
550 wrap-arounds are dealt with correctly.
551 - Timestamp at the beginning and timestamp at the end of the event packet.
552 Both timestamps are written in the packet header, but sampled respectively
553 while (or before) writing the first event and while (or after) writing the
554 last event in the packet. The inclusive range between these timestamps should
555 include all event timestamps assigned to events contained within the packet.
556 - Events discarded count
557 - Snapshot of a per-stream free-running counter, counting the number of
558 events discarded that were supposed to be written in the stream prior to
559 the first event in the event packet.
560 * Note: producer-consumer buffer full condition should fill the current
561 event packet with padding so we know exactly where events have been
563 - Lossless compression scheme used for the event packet content. Applied
564 directly to raw data. New types of compression can be added in following
565 versions of the format.
566 0: no compression scheme
570 - Cypher used for the event packet content. Applied after compression.
573 - Checksum scheme used for the event packet content. Applied after encryption.
579 5.1 Event Packet Header Fixed Layout Description
581 The event packet header layout is indicated by the trace packet.header
582 field. Here is an example structure type for the packet header with the
583 fields typically expected:
585 struct event_packet_header {
587 uint8_t trace_uuid[16];
593 packet.header := struct event_packet_header;
596 If the trace_uuid is not present, no validation that the metadata
597 actually corresponds to the stream is performed.
599 If the stream_id packet header field is missing, the trace can only
600 contain a single stream. Its "id" field can be left out, and its events
601 don't need to declare a "stream_id" field.
604 5.2 Event Packet Context Description
606 Event packet context example. These are declared within the stream declaration
607 in the metadata. All these fields are optional. If the packet size field is
608 missing, the whole stream only contains a single packet. If the content
609 size field is missing, the packet is filled (no padding).
611 An example event packet context type:
613 struct event_packet_context {
614 uint64_t timestamp_begin;
615 uint64_t timestamp_end;
617 uint32_t stream_packet_count;
618 uint32_t events_discarded;
620 uint32_t/uint16_t content_size;
621 uint32_t/uint16_t packet_size;
622 uint8_t stream_packet_count_bits; /* Significant counter bits */
623 uint8_t compression_scheme;
624 uint8_t encryption_scheme;
625 uint8_t checksum_scheme;
631 The overall structure of an event is:
633 1 - Stream Packet Context (as specified by the stream metadata)
634 2 - Event Header (as specified by the stream metadata)
635 3 - Stream Event Context (as specified by the stream metadata)
636 4 - Event Context (as specified by the event metadata)
637 5 - Event Payload (as specified by the event metadata)
639 This structure defines an implicit dynamic scoping, where variants
640 located in inner structures (those with a higher number in the listing
641 above) can refer to the fields of outer structures (with lower number in
642 the listing above). See Section 7.3 TSDL Scopes for more detail.
646 Event headers can be described within the metadata. We hereby propose, as an
647 example, two types of events headers. Type 1 accommodates streams with less than
648 31 event IDs. Type 2 accommodates streams with 31 or more event IDs.
650 One major factor can vary between streams: the number of event IDs assigned to
651 a stream. Luckily, this information tends to stay relatively constant (modulo
652 event registration while trace is being recorded), so we can specify different
653 representations for streams containing few event IDs and streams containing
654 many event IDs, so we end up representing the event ID and timestamp as densely
655 as possible in each case.
657 The header is extended in the rare occasions where the information cannot be
658 represented in the ranges available in the standard event header. They are also
659 used in the rare occasions where the data required for a field could not be
660 collected: the flag corresponding to the missing field within the missing_fields
661 array is then set to 1.
663 Types uintX_t represent an X-bit unsigned integer.
666 6.1.1 Type 1 - Few event IDs
668 - Aligned on 32-bit (or 8-bit if byte-packed, depending on the architecture
670 - Native architecture byte ordering.
671 - For "compact" selection
672 - Fixed size: 32 bits.
673 - For "extended" selection
674 - Size depends on the architecture and variant alignment.
676 struct event_header_1 {
679 * id 31 is reserved to indicate an extended header.
681 enum : uint5_t { compact = 0 ... 30, extended = 31 } id;
687 uint32_t id; /* 32-bit event IDs */
688 uint64_t timestamp; /* 64-bit timestamps */
694 6.1.2 Type 2 - Many event IDs
696 - Aligned on 16-bit (or 8-bit if byte-packed, depending on the architecture
698 - Native architecture byte ordering.
699 - For "compact" selection
700 - Size depends on the architecture and variant alignment.
701 - For "extended" selection
702 - Size depends on the architecture and variant alignment.
704 struct event_header_2 {
706 * id: range: 0 - 65534.
707 * id 65535 is reserved to indicate an extended header.
709 enum : uint16_t { compact = 0 ... 65534, extended = 65535 } id;
715 uint32_t id; /* 32-bit event IDs */
716 uint64_t timestamp; /* 64-bit timestamps */
724 The event context contains information relative to the current event. The choice
725 and meaning of this information is specified by the metadata "stream" and
726 "event" information. The "stream" context is applied to all events within the
727 stream. The "stream" context structure follows the event header. The "event"
728 context is applied to specific events. Its structure follows the "stream"
731 An example of stream-level event context is to save the event payload size with
732 each event, or to save the current PID with each event. These are declared
733 within the stream declaration within the metadata:
741 uint16_t payload_size;
746 An example of event-specific event context is to declare a bitmap of missing
747 fields, only appended after the stream event context if the extended event
748 header is selected. NR_FIELDS is the number of fields within the event (a
756 uint1_t missing_fields[NR_FIELDS]; /* missing event fields bitmap */
765 An event payload contains fields specific to a given event type. The fields
766 belonging to an event type are described in the event-specific metadata
767 within a structure type.
771 No padding at the end of the event payload. This differs from the ISO/C standard
772 for structures, but follows the CTF standard for structures. In a trace, even
773 though it makes sense to align the beginning of a structure, it really makes no
774 sense to add padding at the end of the structure, because structures are usually
775 not followed by a structure of the same type.
777 This trick can be done by adding a zero-length "end" field at the end of the C
778 structures, and by using the offset of this field rather than using sizeof()
779 when calculating the size of a structure (see Appendix "A. Helper macros").
783 The event payload is aligned on the largest alignment required by types
784 contained within the payload. (This follows the ISO/C standard for structures)
787 7. Trace Stream Description Language (TSDL)
789 The Trace Stream Description Language (TSDL) allows expression of the
790 binary trace streams layout in a C99-like Domain Specific Language
796 The trace stream layout description is located in the trace meta-data.
797 The meta-data is itself located in a stream identified by its name:
800 It is made of "event packets", which each start with an event packet
801 header. The event type within the metadata stream have no event header
802 nor event context. Each event only contains a "string" payload without
803 any null-character. The events are packed one next to another. Each
804 event packet start with an event packet header, which contains, amongst
805 other fields, the magic number, trace UUID and packet length. In the
806 event packet header, the trace UUID is represented as an array of bytes.
807 Within the string-based metadata description, the trace UUID is
808 represented as a string of hexadecimal digits and dashes "-".
810 The metadata can be parsed by reading characters within the metadata
811 stream, for each packet starting after the packet header, for the length
812 of the packet payload specified in the header. Text contained within
813 "/*" and "*/", as well as within "//" and end of line, are treated as
814 comments. Boolean values can be represented as true, TRUE, or 1 for
815 true, and false, FALSE, or 0 for false.
818 7.2 Declaration vs Definition
820 A declaration associates a layout to a type, without specifying where
821 this type is located in the event structure hierarchy (see Section 6).
822 This therefore includes typedef, typealias, as well as all type
823 specifiers. In certain circumstances (typedef, structure field and
824 variant field), a declaration is followed by a declarator, which specify
825 the newly defined type name (for typedef), or the field name (for
826 declarations located within structure and variants). Array and sequence,
827 declared with square brackets ("[" "]"), are part of the declarator,
828 similarly to C99. The enumeration base type is specified by
829 ": enum_base", which is part of the type specifier. The variant tag
830 name, specified between "<" ">", is also part of the type specifier.
832 A definition associates a type to a location in the event structure
833 hierarchy (see Section 6). This association is denoted by ":=", as shown
839 TSDL uses two different types of scoping: a lexical scope is used for
840 declarations and type definitions, and a dynamic scope is used for
841 variants references to tag fields.
845 Each of "trace", "stream", "event", "struct" and "variant" have their own
846 nestable declaration scope, within which types can be declared using "typedef"
847 and "typealias". A root declaration scope also contains all declarations
848 located outside of any of the aforementioned declarations. An inner
849 declaration scope can refer to type declared within its container
850 lexical scope prior to the inner declaration scope. Redefinition of a
851 typedef or typealias is not valid, although hiding an upper scope
852 typedef or typealias is allowed within a sub-scope.
856 A dynamic scope consists in the lexical scope augmented with the
857 implicit event structure definition hierarchy presented at Section 6.
858 The dynamic scope is only used for variant tag definitions. It is used
859 at definition time to look up the location of the tag field associated
862 Therefore, variants in lower levels in the dynamic scope (e.g. event
863 context) can refer to a tag field located in upper levels (e.g. in the
864 event header) by specifying, in this case, the associated tag with
865 <header.field_name>. This allows, for instance, the event context to
866 define a variant referring to the "id" field of the event header as
869 The target dynamic scope must be specified explicitly when referring to
870 a field outside of the local static scope. The dynamic scope prefixes
873 - Trace Packet Header: <trace.packet.header. >,
874 - Stream Packet Context: <stream.packet.context. >,
875 - Event Header: <stream.event.header. >,
876 - Stream Event Context: <stream.event.context. >,
877 - Event Context: <event.context. >,
878 - Event Payload: <event.fields. >.
880 Multiple declarations of the same field name within a single scope is
881 not valid. It is however valid to re-use the same field name in
882 different scopes. There is no possible conflict, because the dynamic
883 scope must be specified when a variant refers to a tag field located in
884 a different dynamic scope.
886 The information available in the dynamic scopes can be thought of as the
887 current tracing context. At trace production, information about the
888 current context is saved into the specified scope field levels. At trace
889 consumption, for each event, the current trace context is therefore
890 readable by accessing the upper dynamic scopes.
895 The grammar representing the TSDL metadata is presented in Appendix C.
896 TSDL Grammar. This section presents a rather ligher reading that
897 consists in examples of TSDL metadata, with template values.
899 The stream "id" can be left out if there is only one stream in the
900 trace. The event "id" field can be left out if there is only one event
904 major = value; /* Trace format version */
906 uuid = "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa"; /* Trace UUID */
907 byte_order = be OR le; /* Endianness (required) */
908 packet.header := struct {
910 uint8_t trace_uuid[16];
917 /* Type 1 - Few event IDs; Type 2 - Many event IDs. See section 6.1. */
918 event.header := event_header_1 OR event_header_2;
919 event.context := struct {
922 packet.context := struct {
929 id = value; /* Numeric identifier within the stream */
939 /* More detail on types in section 4. Types */
944 * Type declarations behave similarly to the C standard.
947 typedef aliased_type_specifiers new_type_declarators;
949 /* e.g.: typedef struct example new_type_name[10]; */
954 * The "typealias" declaration can be used to give a name (including
955 * pointer declarator specifier) to a type. It should also be used to
956 * map basic C types (float, int, unsigned long, ...) to a CTF type.
957 * Typealias is a superset of "typedef": it also allows assignment of a
958 * simple variable identifier to a type.
961 typealias type_class {
963 } := type_specifiers type_declarator;
967 * typealias integer {
971 * } := struct page *;
973 * typealias integer {
988 enum name : integer_type {
994 * Unnamed types, contained within compound type fields, typedef or typealias.
1005 enum : integer_type {
1009 typedef type new_type[length];
1012 type field_name[length];
1015 typedef type new_type[length_type];
1018 type field_name[length_type];
1030 integer_type field_name:size; /* GNU/C bitfield */
1040 The two following macros keep track of the size of a GNU/C structure without
1041 padding at the end by placing HEADER_END as the last field. A one byte end field
1042 is used for C90 compatibility (C99 flexible arrays could be used here). Note
1043 that this does not affect the effective structure size, which should always be
1044 calculated with the header_sizeof() helper.
1046 #define HEADER_END char end_field
1047 #define header_sizeof(type) offsetof(typeof(type), end_field)
1050 B. Stream Header Rationale
1052 An event stream is divided in contiguous event packets of variable size. These
1053 subdivisions allow the trace analyzer to perform a fast binary search by time
1054 within the stream (typically requiring to index only the event packet headers)
1055 without reading the whole stream. These subdivisions have a variable size to
1056 eliminate the need to transfer the event packet padding when partially filled
1057 event packets must be sent when streaming a trace for live viewing/analysis.
1058 An event packet can contain a certain amount of padding at the end. Dividing
1059 streams into event packets is also useful for network streaming over UDP and
1060 flight recorder mode tracing (a whole event packet can be swapped out of the
1061 buffer atomically for reading).
1063 The stream header is repeated at the beginning of each event packet to allow
1064 flexibility in terms of:
1066 - streaming support,
1067 - allowing arbitrary buffers to be discarded without making the trace
1069 - allow UDP packet loss handling by either dealing with missing event packet
1070 or asking for re-transmission.
1071 - transparently support flight recorder mode,
1072 - transparently support crash dump.
1074 The event stream header will therefore be referred to as the "event packet
1075 header" throughout the rest of this document.
1081 * Common Trace Format (CTF) Trace Stream Description Language (TSDL) Grammar.
1083 * Inspired from the C99 grammar:
1084 * http://www.open-std.org/jtc1/sc22/wg14/www/docs/n1124.pdf (Annex A)
1085 * and c++1x grammar (draft)
1086 * http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2011/n3291.pdf (Annex A)
1088 * Specialized for CTF needs by including only constant and declarations from
1089 * C99 (excluding function declarations), and by adding support for variants,
1090 * sequences and CTF-specific specifiers. Enumeration container types
1091 * semantic is inspired from c++1x enum-base.
1096 1.1) Lexical elements
1139 identifier identifier-nondigit
1142 identifier-nondigit:
1144 universal-character-name
1145 any other implementation-defined characters
1149 [a-zA-Z] /* regular expression */
1152 [0-9] /* regular expression */
1154 1.4) Universal character names
1156 universal-character-name:
1158 \U hex-quad hex-quad
1161 hexadecimal-digit hexadecimal-digit hexadecimal-digit hexadecimal-digit
1167 enumeration-constant
1171 decimal-constant integer-suffix-opt
1172 octal-constant integer-suffix-opt
1173 hexadecimal-constant integer-suffix-opt
1177 decimal-constant digit
1181 octal-constant octal-digit
1183 hexadecimal-constant:
1184 hexadecimal-prefix hexadecimal-digit
1185 hexadecimal-constant hexadecimal-digit
1195 unsigned-suffix long-suffix-opt
1196 unsigned-suffix long-long-suffix
1197 long-suffix unsigned-suffix-opt
1198 long-long-suffix unsigned-suffix-opt
1214 digit-sequence digit
1216 hexadecimal-digit-sequence:
1218 hexadecimal-digit-sequence hexadecimal-digit
1220 enumeration-constant:
1226 L' c-char-sequence '
1230 c-char-sequence c-char
1233 any member of source charset except single-quote ('), backslash
1234 (\), or new-line character.
1238 simple-escape-sequence
1239 octal-escape-sequence
1240 hexadecimal-escape-sequence
1241 universal-character-name
1243 simple-escape-sequence: one of
1244 \' \" \? \\ \a \b \f \n \r \t \v
1246 octal-escape-sequence:
1248 \ octal-digit octal-digit
1249 \ octal-digit octal-digit octal-digit
1251 hexadecimal-escape-sequence:
1252 \x hexadecimal-digit
1253 hexadecimal-escape-sequence hexadecimal-digit
1255 1.6) String literals
1258 " s-char-sequence-opt "
1259 L" s-char-sequence-opt "
1263 s-char-sequence s-char
1266 any member of source charset except double-quote ("), backslash
1267 (\), or new-line character.
1273 [ ] ( ) { } . -> * + - < > : ; ... = ,
1276 2) Phrase structure grammar
1282 ( unary-expression )
1286 postfix-expression [ unary-expression ]
1287 postfix-expression . identifier
1288 postfix-expressoin -> identifier
1292 unary-operator postfix-expression
1294 unary-operator: one of
1297 assignment-operator:
1300 type-assignment-operator:
1303 constant-expression:
1306 constant-expression-range:
1307 constant-expression ... constant-expression
1312 declaration-specifiers declarator-list-opt ;
1315 declaration-specifiers:
1316 storage-class-specifier declaration-specifiers-opt
1317 type-specifier declaration-specifiers-opt
1318 type-qualifier declaration-specifiers-opt
1322 declarator-list , declarator
1324 abstract-declarator-list:
1326 abstract-declarator-list , abstract-declarator
1328 storage-class-specifier:
1351 struct identifier-opt { struct-or-variant-declaration-list-opt }
1354 struct-or-variant-declaration-list:
1355 struct-or-variant-declaration
1356 struct-or-variant-declaration-list struct-or-variant-declaration
1358 struct-or-variant-declaration:
1359 specifier-qualifier-list struct-or-variant-declarator-list ;
1360 declaration-specifiers storage-class-specifier declaration-specifiers declarator-list ;
1361 typealias declaration-specifiers abstract-declarator-list := declaration-specifiers abstract-declarator-list ;
1362 typealias declaration-specifiers abstract-declarator-list := declarator-list ;
1364 specifier-qualifier-list:
1365 type-specifier specifier-qualifier-list-opt
1366 type-qualifier specifier-qualifier-list-opt
1368 struct-or-variant-declarator-list:
1369 struct-or-variant-declarator
1370 struct-or-variant-declarator-list , struct-or-variant-declarator
1372 struct-or-variant-declarator:
1374 declarator-opt : constant-expression
1377 variant identifier-opt variant-tag-opt { struct-or-variant-declaration-list }
1378 variant identifier variant-tag
1384 enum identifier-opt { enumerator-list }
1385 enum identifier-opt { enumerator-list , }
1387 enum identifier-opt : declaration-specifiers { enumerator-list }
1388 enum identifier-opt : declaration-specifiers { enumerator-list , }
1392 enumerator-list , enumerator
1395 enumeration-constant
1396 enumeration-constant = constant-expression
1397 enumeration-constant = constant-expression-range
1403 pointer-opt direct-declarator
1408 direct-declarator [ type-specifier ]
1409 direct-declarator [ constant-expression ]
1411 abstract-declarator:
1412 pointer-opt direct-abstract-declarator
1414 direct-abstract-declarator:
1416 ( abstract-declarator )
1417 direct-abstract-declarator [ type-specifier ]
1418 direct-abstract-declarator [ constant-expression ]
1419 direct-abstract-declarator [ ]
1422 * type-qualifier-list-opt
1423 * type-qualifier-list-opt pointer
1425 type-qualifier-list:
1427 type-qualifier-list type-qualifier
1432 2.3) CTF-specific declarations
1435 event { ctf-assignment-expression-list-opt }
1436 stream { ctf-assignment-expression-list-opt }
1437 trace { ctf-assignment-expression-list-opt }
1438 typealias declaration-specifiers abstract-declarator-list := declaration-specifiers abstract-declarator-list ;
1439 typealias declaration-specifiers abstract-declarator-list := declarator-list ;
1442 floating_point { ctf-assignment-expression-list-opt }
1443 integer { ctf-assignment-expression-list-opt }
1444 string { ctf-assignment-expression-list-opt }
1446 ctf-assignment-expression-list:
1447 ctf-assignment-expression
1448 ctf-assignment-expression-list ; ctf-assignment-expression
1450 ctf-assignment-expression:
1451 unary-expression assignment-operator unary-expression
1452 unary-expression type-assignment-operator type-specifier
1453 declaration-specifiers storage-class-specifier declaration-specifiers declarator-list
1454 typealias declaration-specifiers abstract-declarator-list := declaration-specifiers abstract-declarator-list
1455 typealias declaration-specifiers abstract-declarator-list := declarator-list