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.
12 The latest version of this document can be found at:
14 git tree: git://git.efficios.com/ctf.git
15 gitweb: http://git.efficios.com/?p=ctf.git
17 A reference implementation of a library to read and write this trace format is
18 being implemented within the BabelTrace project, a converter between trace
19 formats. The development tree is available at:
21 git tree: git://git.efficios.com/babeltrace.git
22 gitweb: http://git.efficios.com/?p=babeltrace.git
25 1. Preliminary definitions
27 - Event Trace: An ordered sequence of events.
28 - Event Stream: An ordered sequence of events, containing a subset of the
30 - Event Packet: A sequence of physically contiguous events within an event
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
35 e.g. event: irq_entry.
36 - An event (or event record) relates to a specific instance of an event
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.
43 2. High-level representation of a trace
45 A trace is divided into multiple event streams. Each event stream contains a
46 subset of the trace event types.
48 The final output of the trace, after its generation and optional transport over
49 the network, is expected to be either on permanent or temporary storage in a
50 virtual file system. Because each event stream is appended to while a trace is
51 being recorded, each is associated with a separate file for output. Therefore,
52 a stored trace can be represented as a directory containing one file per stream.
54 A metadata event stream contains information on trace event types. It describes:
58 - Per-stream event header description.
59 - Per-stream event header selection.
60 - Per-stream event context fields.
62 - Event type to stream mapping.
63 - Event type to name mapping.
64 - Event type to ID mapping.
65 - Event fields description.
70 An event stream is divided in contiguous event packets of variable size. These
71 subdivisions have a variable size. An event packet can contain a certain amount
72 of padding at the end. The rationale for the event stream design choices is
73 explained in Appendix B. Stream Header Rationale.
75 An event stream is divided in contiguous event packets of variable size. These
76 subdivisions have a variable size. An event packet can contain a certain amount
77 of padding at the end. The stream header is repeated at the beginning of each
80 The event stream header will therefore be referred to as the "event packet
81 header" throughout the rest of this document.
86 Types are organized as type classes. Each type class belong to either of two
87 kind of types: basic types or compound types.
91 A basic type is a scalar type, as described in this section. It includes
92 integers, GNU/C bitfields, enumerations, and floating point values.
94 4.1.1 Type inheritance
96 Type specifications can be inherited to allow deriving types from a
97 type class. For example, see the uint32_t named type derived from the "integer"
98 type class below ("Integers" section). Types have a precise binary
99 representation in the trace. A type class has methods to read and write these
100 types, but must be derived into a type to be usable in an event field.
104 We define "byte-packed" types as aligned on the byte size, namely 8-bit.
105 We define "bit-packed" types as following on the next bit, as defined by the
108 All basic types, except bitfields, are either aligned on an architecture-defined
109 specific alignment or byte-packed, depending on the architecture preference.
110 Architectures providing fast unaligned write byte-packed basic types to save
111 space, aligning each type on byte boundaries (8-bit). Architectures with slow
112 unaligned writes align types on specific alignment values. If no specific
113 alignment is declared for a type nor its parents, it is assumed to be bit-packed
114 for bitfields and byte-packed for other types.
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 If a numeric value is encountered between < >, it represents the integer type
278 size used to hold the enumeration, in bits.
280 enum name <integer_type OR size> {
281 somestring = start_value1 ... end_value1,
282 "other string" = start_value2 ... end_value2,
283 yet_another_string, /* will be assigned to end_value2 + 1 */
284 "some other string" = value,
288 If the values are omitted, the enumeration starts at 0 and increment of 1 for
299 Overlapping ranges within a single enumeration are implementation defined.
301 A nameless enumeration can be declared as a field type or as part of a typedef:
303 enum <integer_type> {
310 Compound are aggregation of type declarations. Compound types include
311 structures, variant, arrays, sequences, and strings.
315 Structures are aligned on the largest alignment required by basic types
316 contained within the structure. (This follows the ISO/C standard for structures)
318 Metadata representation of a named structure:
321 field_type field_name;
322 field_type field_name;
329 integer { /* Nameless type */
334 uint64_t second_field_name; /* Named type declared in the metadata */
337 The fields are placed in a sequence next to each other. They each possess a
338 field name, which is a unique identifier within the structure.
340 A nameless structure can be declared as a field type or as part of a typedef:
346 4.2.2 Variants (Discriminated/Tagged Unions)
348 A CTF variant is a selection between different types. A CTF variant must
349 always be defined within the scope of a structure or within fields
350 contained within a structure (defined recursively). A "tag" enumeration
351 field must appear in either the same lexical scope, prior to the variant
352 field (in field declaration order), in an uppermost lexical scope (see
353 Section 7.2.1), or in an uppermost dynamic scope (see Section 7.2.2).
354 The type selection is indicated by the mapping from the enumeration
355 value to the string used as variant type selector. The field to use as
356 tag is specified by the "tag_field", specified between "< >" after the
357 "variant" keyword for unnamed variants, and after "variant name" for
360 The alignment of the variant is the alignment of the type as selected by the tag
361 value for the specific instance of the variant. The alignment of the type
362 containing the variant is independent of the variant alignment. The size of the
363 variant is the size as selected by the tag value for the specific instance of
366 A named variant declaration followed by its definition within a structure
377 enum <integer_type or size> { sel1, sel2, sel3, ... } tag_field;
379 variant name <tag_field> v;
382 An unnamed variant definition within a structure is expressed by the following
386 enum <integer_type or size> { sel1, sel2, sel3, ... } tag_field;
388 variant <tag_field> {
396 Example of a named variant within a sequence that refers to a single tag field:
405 enum <uint2_t> { a, b, c } choice;
406 variant example <choice> v[unsigned int];
409 Example of an unnamed variant:
412 enum <uint2_t> { a, b, c, d } choice;
413 /* Unrelated fields can be added between the variant and its tag */
426 Example of an unnamed variant within an array:
429 enum <uint2_t> { a, b, c } choice;
437 Example of a variant type definition within a structure, where the defined type
438 is then declared within an array of structures. This variant refers to a tag
439 located in an upper lexical scope. This example clearly shows that a variant
440 type definition referring to the tag "x" uses the closest preceding field from
441 the lexical scope of the type definition.
444 enum <uint2_t> { a, b, c, d } x;
446 typedef variant <x> { /*
447 * "x" refers to the preceding "x" enumeration in the
448 * lexical scope of the type definition.
456 enum <int> { x, y, z } x; /* This enumeration is not used by "v". */
457 example_variant v; /*
458 * "v" uses the "enum <uint2_t> { a, b, c, d }"
466 Arrays are fixed-length. Their length is declared in the type declaration within
467 the metadata. They contain an array of "inner type" elements, which can refer to
468 any type not containing the type of the array being declared (no circular
469 dependency). The length is the number of elements in an array.
471 Metadata representation of a named array:
473 typedef elem_type name[length];
475 A nameless array can be declared as a field type within a structure, e.g.:
477 uint8_t field_name[10];
482 Sequences are dynamically-sized arrays. They start with an integer that specify
483 the length of the sequence, followed by an array of "inner type" elements.
484 The length is the number of elements in the sequence.
486 Metadata representation for a named sequence:
488 typedef elem_type name[length_type];
490 A nameless sequence can be declared as a field type, e.g.:
492 long field_name[int];
494 The length type follows the integer types specifications, and the sequence
495 elements follow the "array" specifications.
499 Strings are an array of bytes of variable size and are terminated by a '\0'
500 "NULL" character. Their encoding is described in the metadata. In absence of
501 encoding attribute information, the default encoding is UTF-8.
503 Metadata representation of a named string type:
506 encoding = UTF8 OR ASCII;
509 A nameless string type can be declared as a field type:
511 string field_name; /* Use default UTF8 encoding */
513 5. Event Packet Header
515 The event packet header consists of two part: one is mandatory and have a fixed
516 layout. The second part, the "event packet context", has its layout described in
519 - Aligned on page size. Fixed size. Fields either aligned or packed (depending
520 on the architecture preference).
521 No padding at the end of the event packet header. Native architecture byte
524 Fixed layout (event packet header):
526 - Magic number (CTF magic numbers: 0xC1FC1FC1 and its reverse endianness
527 representation: 0xC11FFCC1) It needs to have a non-symmetric bytewise
528 representation. Used to distinguish between big and little endian traces (this
529 information is determined by knowing the endianness of the architecture
530 reading the trace and comparing the magic number against its value and the
531 reverse, 0xC11FFCC1). This magic number specifies that we use the CTF metadata
532 description language described in this document. Different magic numbers
533 should be used for other metadata description languages.
534 - Trace UUID, used to ensure the event packet match the metadata used.
535 (note: we cannot use a metadata checksum because metadata can be appended to
536 while tracing is active)
537 - Stream ID, used as reference to stream description in metadata.
539 Metadata-defined layout (event packet context):
541 - Event packet content size (in bytes).
542 - Event packet size (in bytes, includes padding).
543 - Event packet content checksum (optional). Checksum excludes the event packet
545 - Per-stream event packet sequence count (to deal with UDP packet loss). The
546 number of significant sequence counter bits should also be present, so
547 wrap-arounds are deal with correctly.
548 - Timestamp at the beginning and timestamp at the end of the event packet.
549 Both timestamps are written in the packet header, but sampled respectively
550 while (or before) writing the first event and while (or after) writing the
551 last event in the packet. The inclusive range between these timestamps should
552 include all event timestamps assigned to events contained within the packet.
553 - Events discarded count
554 - Snapshot of a per-stream free-running counter, counting the number of
555 events discarded that were supposed to be written in the stream prior to
556 the first event in the event packet.
557 * Note: producer-consumer buffer full condition should fill the current
558 event packet with padding so we know exactly where events have been
560 - Lossless compression scheme used for the event packet content. Applied
561 directly to raw data. New types of compression can be added in following
562 versions of the format.
563 0: no compression scheme
567 - Cypher used for the event packet content. Applied after compression.
570 - Checksum scheme used for the event packet content. Applied after encryption.
576 5.1 Event Packet Header Fixed Layout Description
578 struct event_packet_header {
580 uint8_t trace_uuid[16];
584 5.2 Event Packet Context Description
586 Event packet context example. These are declared within the stream declaration
587 in the metadata. All these fields are optional except for "content_size" and
588 "packet_size", which must be present in the context.
590 An example event packet context type:
592 struct event_packet_context {
593 uint64_t timestamp_begin;
594 uint64_t timestamp_end;
596 uint32_t stream_packet_count;
597 uint32_t events_discarded;
599 uint32_t/uint16_t content_size;
600 uint32_t/uint16_t packet_size;
601 uint8_t stream_packet_count_bits; /* Significant counter bits */
602 uint8_t compression_scheme;
603 uint8_t encryption_scheme;
604 uint8_t checksum_scheme;
610 The overall structure of an event is:
612 1 - Stream Packet Context (as specified by the stream metadata)
613 2 - Event Header (as specified by the stream metadata)
614 3 - Stream Event Context (as specified by the stream metadata)
615 4 - Event Context (as specified by the event metadata)
616 5 - Event Payload (as specified by the event metadata)
618 This structure defines an implicit dynamic scoping, where variants
619 located in inner structures (those with a higher number in the listing
620 above) can refer to the fields of outer structures (with lower number in
621 the listing above). See Section 7.2 Metadata Scopes for more detail.
625 Event headers can be described within the metadata. We hereby propose, as an
626 example, two types of events headers. Type 1 accommodates streams with less than
627 31 event IDs. Type 2 accommodates streams with 31 or more event IDs.
629 One major factor can vary between streams: the number of event IDs assigned to
630 a stream. Luckily, this information tends to stay relatively constant (modulo
631 event registration while trace is being recorded), so we can specify different
632 representations for streams containing few event IDs and streams containing
633 many event IDs, so we end up representing the event ID and timestamp as densely
634 as possible in each case.
636 The header is extended in the rare occasions where the information cannot be
637 represented in the ranges available in the standard event header. They are also
638 used in the rare occasions where the data required for a field could not be
639 collected: the flag corresponding to the missing field within the missing_fields
640 array is then set to 1.
642 Types uintX_t represent an X-bit unsigned integer.
645 6.1.1 Type 1 - Few event IDs
647 - Aligned on 32-bit (or 8-bit if byte-packed, depending on the architecture
649 - Native architecture byte ordering.
650 - For "compact" selection
651 - Fixed size: 32 bits.
652 - For "extended" selection
653 - Size depends on the architecture and variant alignment.
655 struct event_header_1 {
658 * id 31 is reserved to indicate an extended header.
660 enum <uint5_t> { compact = 0 ... 30, extended = 31 } id;
666 uint32_t id; /* 32-bit event IDs */
667 uint64_t timestamp; /* 64-bit timestamps */
673 6.1.2 Type 2 - Many event IDs
675 - Aligned on 16-bit (or 8-bit if byte-packed, depending on the architecture
677 - Native architecture byte ordering.
678 - For "compact" selection
679 - Size depends on the architecture and variant alignment.
680 - For "extended" selection
681 - Size depends on the architecture and variant alignment.
683 struct event_header_2 {
685 * id: range: 0 - 65534.
686 * id 65535 is reserved to indicate an extended header.
688 enum <uint16_t> { compact = 0 ... 65534, extended = 65535 } id;
694 uint32_t id; /* 32-bit event IDs */
695 uint64_t timestamp; /* 64-bit timestamps */
703 The event context contains information relative to the current event. The choice
704 and meaning of this information is specified by the metadata "stream" and
705 "event" information. The "stream" context is applied to all events within the
706 stream. The "stream" context structure follows the event header. The "event"
707 context is applied to specific events. Its structure follows the "stream"
710 An example of stream-level event context is to save the event payload size with
711 each event, or to save the current PID with each event. These are declared
712 within the stream declaration within the metadata:
720 uint16_t payload_size;
725 An example of event-specific event context is to declare a bitmap of missing
726 fields, only appended after the stream event context if the extended event
727 header is selected. NR_FIELDS is the number of fields within the event (a
735 uint1_t missing_fields[NR_FIELDS]; /* missing event fields bitmap */
744 An event payload contains fields specific to a given event type. The fields
745 belonging to an event type are described in the event-specific metadata
746 within a structure type.
750 No padding at the end of the event payload. This differs from the ISO/C standard
751 for structures, but follows the CTF standard for structures. In a trace, even
752 though it makes sense to align the beginning of a structure, it really makes no
753 sense to add padding at the end of the structure, because structures are usually
754 not followed by a structure of the same type.
756 This trick can be done by adding a zero-length "end" field at the end of the C
757 structures, and by using the offset of this field rather than using sizeof()
758 when calculating the size of a structure (see Appendix "A. Helper macros").
762 The event payload is aligned on the largest alignment required by types
763 contained within the payload. (This follows the ISO/C standard for structures)
768 The meta-data is located in a stream named "metadata". It is made of "event
769 packets", which each start with an event packet header. The event type within
770 the metadata stream have no event header nor event context. Each event only
771 contains a null-terminated "string" payload, which is a metadata description
772 entry. The events are packed one next to another. Each event packet start with
773 an event packet header, which contains, amongst other fields, the magic number
774 and trace UUID. The trace UUID is represented as a string of hexadecimal digits
777 The metadata can be parsed by reading through the metadata strings, skipping
778 newlines and null-characters. Type names are made of a single identifier, and
779 can be surrounded by prefix/postfix. Text contained within "/*" and "*/", as
780 well as within "//" and end of line, are treated as comments. Boolean values can
781 be represented as true, TRUE, or 1 for true, and false, FALSE, or 0 for false.
784 7.1 Declaration vs Definition
786 A declaration associates a layout to a type, without specifying where
787 this type is located in the event structure hierarchy (see Section 6).
788 This therefore includes typedef, typealias, as well as all type
789 specifiers. In certain circumstances (typedef, structure field and
790 variant field), a declaration is followed by a declarator, which specify
791 the newly defined type name (for typedef), or the field name (for
792 declarations located within structure and variants). Array and sequence,
793 declared with square brackets ("[" "]"), are part of the declarator,
794 similarly to C99. The enumeration type specifier and variant tag name
795 (both specified with "<" ">") are part of the type specifier.
797 A definition associates a type to a location in the event structure
798 hierarchy (see Section 6).
803 CTF metadata uses two different types of scoping: a lexical scope is
804 used for declarations and type definitions, and a dynamic scope is used
805 for variants references to tag fields.
809 Each of "trace", "stream", "event", "struct" and "variant" have their own
810 nestable declaration scope, within which types can be declared using "typedef"
811 and "typealias". A root declaration scope also contains all declarations
812 located outside of any of the aforementioned declarations. An inner
813 declaration scope can refer to type declared within its container
814 lexical scope prior to the inner declaration scope. Redefinition of a
815 typedef or typealias is not valid, although hiding an upper scope
816 typedef or typealias is allowed within a sub-scope.
820 A dynamic scope consists in the lexical scope augmented with the
821 implicit event structure definition hierarchy presented at Section 6.
822 The dynamic scope is only used for variant tag definitions. It is used
823 at definition time to look up the location of the tag field associated
826 Therefore, variants in lower levels in the dynamic scope (e.g. event
827 context) can refer to a tag field located in upper levels (e.g. in the
828 event header) by specifying, in this case, the associated tag with
829 <header.field_name>. This allows, for instance, the event context to
830 define a variant referring to the "id" field of the event header as
833 The target dynamic scope must be specified explicitly when referring to
834 a field outside of the local static scope. The dynamic scope prefixes
837 - Stream Packet Context: <stream.packet.context. >,
838 - Event Header: <stream.event.header. >,
839 - Stream Event Context: <stream.event.context. >,
840 - Event Context: <event.context. >,
841 - Event Payload: <event.fields. >.
843 Multiple declarations of the same field name within a single scope is
844 not valid. It is however valid to re-use the same field name in
845 different scopes. There is no possible conflict, because the dynamic
846 scope must be specified when a variant refers to a tag field located in
847 a different dynamic scope.
849 The information available in the dynamic scopes can be thought of as the
850 current tracing context. At trace production, information about the
851 current context is saved into the specified scope field levels. At trace
852 consumption, for each event, the current trace context is therefore
853 readable by accessing the upper dynamic scopes.
856 7.2 Metadata Examples
858 The grammar representing the CTF metadata is presented in
859 Appendix C. CTF Metadata Grammar. This section presents a rather ligher
860 reading that consists in examples of CTF metadata, with template values:
863 major = value; /* Trace format version */
865 uuid = "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa"; /* Trace UUID */
871 /* Type 1 - Few event IDs; Type 2 - Many event IDs. See section 6.1. */
872 event.header := event_header_1 OR event_header_2;
873 event.context := struct {
876 packet.context := struct {
883 id = value; /* Numeric identifier within the stream */
893 /* More detail on types in section 4. Types */
898 * Type declarations behave similarly to the C standard.
901 typedef aliased_type_prefix aliased_type new_type aliased_type_postfix;
903 /* e.g.: typedef struct example new_type_name[10]; */
908 * The "typealias" declaration can be used to give a name (including
909 * prefix/postfix) to a type. It should also be used to map basic C types
910 * (float, int, unsigned long, ...) to a CTF type. Typealias is a superset of
911 * "typedef": it also allows assignment of a simple variable identifier to a
915 typealias type_class {
917 } : new_type_prefix new_type new_type_postfix;
921 * typealias integer {
927 * typealias integer {
942 enum name <integer_type or size> {
948 * Unnamed types, contained within compound type fields, typedef or typealias.
959 enum <integer_type or size> {
963 typedef type new_type[length];
966 type field_name[length];
969 typedef type new_type[length_type];
972 type field_name[length_type];
984 integer_type field_name:size; /* GNU/C bitfield */
994 The two following macros keep track of the size of a GNU/C structure without
995 padding at the end by placing HEADER_END as the last field. A one byte end field
996 is used for C90 compatibility (C99 flexible arrays could be used here). Note
997 that this does not affect the effective structure size, which should always be
998 calculated with the header_sizeof() helper.
1000 #define HEADER_END char end_field
1001 #define header_sizeof(type) offsetof(typeof(type), end_field)
1004 B. Stream Header Rationale
1006 An event stream is divided in contiguous event packets of variable size. These
1007 subdivisions allow the trace analyzer to perform a fast binary search by time
1008 within the stream (typically requiring to index only the event packet headers)
1009 without reading the whole stream. These subdivisions have a variable size to
1010 eliminate the need to transfer the event packet padding when partially filled
1011 event packets must be sent when streaming a trace for live viewing/analysis.
1012 An event packet can contain a certain amount of padding at the end. Dividing
1013 streams into event packets is also useful for network streaming over UDP and
1014 flight recorder mode tracing (a whole event packet can be swapped out of the
1015 buffer atomically for reading).
1017 The stream header is repeated at the beginning of each event packet to allow
1018 flexibility in terms of:
1020 - streaming support,
1021 - allowing arbitrary buffers to be discarded without making the trace
1023 - allow UDP packet loss handling by either dealing with missing event packet
1024 or asking for re-transmission.
1025 - transparently support flight recorder mode,
1026 - transparently support crash dump.
1028 The event stream header will therefore be referred to as the "event packet
1029 header" throughout the rest of this document.
1031 C. CTF Metadata Grammar
1034 * Common Trace Format (CTF) Metadata Grammar.
1036 * Inspired from the C99 grammar:
1037 * http://www.open-std.org/jtc1/sc22/wg14/www/docs/n1124.pdf (Annex A)
1039 * Specialized for CTF needs by including only constant and declarations from
1040 * C99 (excluding function declarations), and by adding support for variants,
1041 * sequences and CTF-specific specifiers.
1046 1.1) Lexical elements
1089 identifier identifier-nondigit
1092 identifier-nondigit:
1094 universal-character-name
1095 any other implementation-defined characters
1099 [a-zA-Z] /* regular expression */
1102 [0-9] /* regular expression */
1104 1.4) Universal character names
1106 universal-character-name:
1108 \U hex-quad hex-quad
1111 hexadecimal-digit hexadecimal-digit hexadecimal-digit hexadecimal-digit
1117 enumeration-constant
1121 decimal-constant integer-suffix-opt
1122 octal-constant integer-suffix-opt
1123 hexadecimal-constant integer-suffix-opt
1127 decimal-constant digit
1131 octal-constant octal-digit
1133 hexadecimal-constant:
1134 hexadecimal-prefix hexadecimal-digit
1135 hexadecimal-constant hexadecimal-digit
1145 unsigned-suffix long-suffix-opt
1146 unsigned-suffix long-long-suffix
1147 long-suffix unsigned-suffix-opt
1148 long-long-suffix unsigned-suffix-opt
1164 digit-sequence digit
1166 hexadecimal-digit-sequence:
1168 hexadecimal-digit-sequence hexadecimal-digit
1170 enumeration-constant:
1176 L' c-char-sequence '
1180 c-char-sequence c-char
1183 any member of source charset except single-quote ('), backslash
1184 (\), or new-line character.
1188 simple-escape-sequence
1189 octal-escape-sequence
1190 hexadecimal-escape-sequence
1191 universal-character-name
1193 simple-escape-sequence: one of
1194 \' \" \? \\ \a \b \f \n \r \t \v
1196 octal-escape-sequence:
1198 \ octal-digit octal-digit
1199 \ octal-digit octal-digit octal-digit
1201 hexadecimal-escape-sequence:
1202 \x hexadecimal-digit
1203 hexadecimal-escape-sequence hexadecimal-digit
1205 1.6) String literals
1208 " s-char-sequence-opt "
1209 L" s-char-sequence-opt "
1213 s-char-sequence s-char
1216 any member of source charset except double-quote ("), backslash
1217 (\), or new-line character.
1223 [ ] ( ) { } . -> * + - < > : ; ... = ,
1226 2) Phrase structure grammar
1232 ( unary-expression )
1236 postfix-expression [ unary-expression ]
1237 postfix-expression . identifier
1238 postfix-expressoin -> identifier
1242 unary-operator postfix-expression
1244 unary-operator: one of
1247 assignment-operator:
1250 constant-expression:
1253 constant-expression-range:
1254 constant-expression ... constant-expression
1259 declaration-specifiers ;
1260 declaration-specifiers storage-class-specifier declaration-specifiers declarator-list ;
1263 declaration-specifiers:
1264 type-specifier declaration-specifiers-opt
1265 type-qualifier declaration-specifiers-opt
1269 declarator-list , declarator
1271 abstract-declarator-list:
1273 abstract-declarator-list , abstract-declarator
1275 storage-class-specifier:
1297 struct identifier-opt { struct-or-variant-declaration-list-opt }
1300 struct-or-variant-declaration-list:
1301 struct-or-variant-declaration
1302 struct-or-variant-declaration-list struct-or-variant-declaration
1304 struct-or-variant-declaration:
1305 specifier-qualifier-list struct-or-variant-declarator-list ;
1306 declaration-specifiers storage-class-specifier declaration-specifiers declarator-list ;
1307 typealias declaration-specifiers abstract-declarator-list : declaration-specifiers abstract-declarator-list ;
1308 typealias declaration-specifiers abstract-declarator-list : declarator-list ;
1310 specifier-qualifier-list:
1311 type-specifier specifier-qualifier-list-opt
1312 type-qualifier specifier-qualifier-list-opt
1314 struct-or-variant-declarator-list:
1315 struct-or-variant-declarator
1316 struct-or-variant-declarator-list , struct-or-variant-declarator
1318 struct-or-variant-declarator:
1320 declarator-opt : constant-expression
1323 variant identifier-opt variant-tag-opt { struct-or-variant-declaration-list }
1324 variant identifier variant-tag
1330 enum identifier-opt { enumerator-list }
1331 enum identifier-opt { enumerator-list , }
1333 enum identifier-opt < declaration-specifiers > { enumerator-list }
1334 enum identifier-opt < declaration-specifiers > { enumerator-list , }
1335 enum identifier < declaration-specifiers >
1336 enum identifier-opt < integer-constant > { enumerator-list }
1337 enum identifier-opt < integer-constant > { enumerator-list , }
1338 enum identifier < integer-constant >
1342 enumerator-list , enumerator
1345 enumeration-constant
1346 enumeration-constant = constant-expression
1347 enumeration-constant = constant-expression-range
1353 pointer-opt direct-declarator
1358 direct-declarator [ type-specifier ]
1359 direct-declarator [ constant-expression ]
1361 abstract-declarator:
1362 pointer-opt direct-abstract-declarator
1364 direct-abstract-declarator:
1366 ( abstract-declarator )
1367 direct-abstract-declarator [ type-specifier ]
1368 direct-abstract-declarator [ constant-expression ]
1369 direct-abstract-declarator [ ]
1372 * type-qualifier-list-opt
1373 * type-qualifier-list-opt pointer
1375 type-qualifier-list:
1377 type-qualifier-list type-qualifier
1382 2.3) CTF-specific declarations
1385 event { ctf-assignment-expression-list-opt }
1386 stream { ctf-assignment-expression-list-opt }
1387 trace { ctf-assignment-expression-list-opt }
1388 typealias declaration-specifiers abstract-declarator-list : declaration-specifiers abstract-declarator-list ;
1389 typealias declaration-specifiers abstract-declarator-list : declarator-list ;
1392 floating_point { ctf-assignment-expression-list-opt }
1393 integer { ctf-assignment-expression-list-opt }
1394 string { ctf-assignment-expression-list-opt }
1396 ctf-assignment-expression-list:
1397 ctf-assignment-expression
1398 ctf-assignment-expression-list ; ctf-assignment-expression
1400 ctf-assignment-expression:
1401 unary-expression assignment-operator unary-expression
1402 unary-expression type-assignment-operator type-specifier
1403 declaration-specifiers storage-class-specifier declaration-specifiers declarator-list
1404 typealias declaration-specifiers abstract-declarator-list : declaration-specifiers abstract-declarator-list
1405 typealias declaration-specifiers abstract-declarator-list : declarator-list