path: root/cddl/contrib/opensolaris/lib/libctf/common/ctf.5

.\"
.\" This file and its contents are supplied under the terms of the
.\" Common Development and Distribution License ("CDDL"), version 1.0.
.\" You may only use this file in accordance with the terms of version
.\" 1.0 of the CDDL.
.\"
.\" A full copy of the text of the CDDL should have accompanied this
.\" source.  A copy of the CDDL is also available via the Internet at
.\" http://www.illumos.org/license/CDDL.
.\"
.\"
.\" Copyright (c) 2014 Joyent, Inc.
.\"
.Dd February 28, 2022
.Dt CTF 5
.Os
.Sh NAME
.Nm ctf
.Nd Compact C Type Format
.Sh SYNOPSIS
.In sys/ctf.h
.Sh DESCRIPTION
.Nm
is designed to be a compact representation of the C programming
language's type information focused on serving the needs of dynamic
tracing, debuggers, and other in-situ and post-mortem introspection
tools.
.Nm
data is generally included in
.Sy ELF
objects and is tagged as
.Sy SHT_PROGBITS
to ensure that the data is accessible in a running process and in subsequent
core dumps, if generated.
.Lp
The
.Nm
data contained in each file has information about the layout and
sizes of C types, including intrinsic types, enumerations, structures,
typedefs, and unions, that are used by the corresponding
.Sy ELF
object.
The
.Nm
data may also include information about the types of global objects and
the return type and arguments of functions in the symbol table.
.Lp
Because a
.Nm
file is often embedded inside a file, rather than being a standalone
file itself, it may also be referred to as a
.Nm
.Sy container .
.Lp
On
.Fx
systems,
.Nm
data is consumed by
.Xr dtrace 1 .
Programmatic access to
.Nm
data can be obtained through libctf.
.Lp
The
.Nm
file format is broken down into seven different sections.
The first two sections are the
.Sy preamble
and
.Sy header ,
which describe the version of the
.Nm
file, the links it has to other
.Nm
files, and the sizes of the other sections.
The next section is the
.Sy label
section,
which provides a way of identifying similar groups of
.Nm
data across multiple files.
This is followed by the
.Sy object
information section, which describes the types of global
symbols.
The subsequent section is the
.Sy function
information section, which describes the return
types and arguments of functions.
The next section is the
.Sy type
information section, which describes
the format and layout of the C types themselves, and finally the last
section is the
.Sy string
section, which contains the names of types, enumerations, members, and
labels.
.Lp
While strictly speaking, only the
.Sy preamble
and
.Sy header
are required, to be actually useful, both the type and string
sections are necessary.
.Lp
A
.Nm
file may contain all of the type information that it requires, or it
may optionally refer to another
.Nm
file which holds the remaining types.
When a
.Nm
file refers to another file, it is called the
.Sy child
and the file it refers to is called the
.Sy parent .
A given file may only refer to one parent.
This process is called
.Em uniquification
because it ensures each child only has type information that is
unique to it.
A common example of this is that most kernel modules in illumos are uniquified
against the kernel module
.Sy genunix
and the type information that comes from the
.Sy IP
module.
This means that a module only has types that are unique to itself and the most
common types in the kernel are not duplicated.
Uniquification is not used when building kernel modules on
.Fx .
.Sh FILE FORMAT
This documents version
.Em three
of the
.Nm
file format.
The
.Xr ctfconvert 1
and
.Xr ctfmerge 1
utilities emit
.Nm
version 3, and all other applications and libraries can operate on
versions 2 and 3.
.Lp
The file format can be summarized with the following image, the
following sections will cover this in more detail.
.Bd -literal

         +-------------+  0t0
+--------| Preamble    |
|        +-------------+  0t4
|+-------| Header      |
||       +-------------+  0t36 + cth_lbloff
||+------| Labels      |
|||      +-------------+  0t36 + cth_objtoff
|||+-----| Objects     |
||||     +-------------+  0t36 + cth_funcoff
||||+----| Functions   |
|||||    +-------------+  0t36 + cth_typeoff
|||||+---| Types       |
||||||   +-------------+  0t36 + cth_stroff
||||||+--| Strings     |
|||||||  +-------------+  0t36 + cth_stroff + cth_strlen
|||||||
|||||||
|||||||
|||||||    +-- magic -   vers   flags
|||||||    |          |    |      |
|||||||   +------+------+------+------+
+---------| 0xcf | 0xf1 | 0x03 | 0x00 |
 ||||||   +------+------+------+------+
 ||||||   0      1      2      3      4
 ||||||
 ||||||    + parent label        + objects
 ||||||    |       + parent name |     + functions    + strings
 ||||||    |       |     + label |     |      + types |       + strlen
 ||||||    |       |     |       |     |      |       |       |
 ||||||   +------+------+------+------+------+-------+-------+-------+
 +--------| 0x00 | 0x00 | 0x00 | 0x08 | 0x36 | 0x110 | 0x5f4 | 0x611 |
  |||||   +------+------+------+------+------+-------+-------+-------+
  |||||   0x04   0x08   0x0c   0x10   0x14    0x18    0x1c    0x20   0x24
  |||||
  |||||         + Label name
  |||||         |       + Label type
  |||||         |       |       + Next label
  |||||         |       |       |
  |||||       +-------+------+-----+
  +-----------| 0x01  | 0x42 | ... |
   ||||       +-------+------+-----+
   ||||  cth_lbloff   +0x4   +0x8  cth_objtoff
   ||||
   ||||
   |||| Symidx  0t15   0t43   0t44
   ||||       +------+------+------+-----+
   +----------| 0x00 | 0x42 | 0x36 | ... |
    |||       +------+------+------+-----+
    ||| cth_objtoff  +0x4   +0x8   +0xc   cth_funcoff
    |||
    |||        + CTF_TYPE_INFO         + CTF_TYPE_INFO
    |||        |        + Return type  |
    |||        |        |       + arg0 |
    |||       +--------+------+------+-----+
    +---------| 0x2c10 | 0x08 | 0x0c | ... |
     ||       +--------+------+------+-----+
     || cth_funcff     +0x4   +0x8   +0xc  cth_typeoff
     ||
     ||         + ctf_stype_t for type 1
     ||         |  integer           + integer encoding
     ||         |                    |          + ctf_stype_t for type 2
     ||         |                    |          |
     ||       +--------------------+-----------+-----+
     +--------| 0x19 * 0xc01 * 0x0 | 0x1000000 | ... |
      |       +--------------------+-----------+-----+
      | cth_typeoff               +0x0c      +0x10  cth_stroff
      |
      |     +--- str 0
      |     |    +--- str 1       + str 2
      |     |    |                |
      |     v    v                v
      |   +----+---+---+---+----+---+---+---+---+---+----+
      +---| \\0 | i | n | t | \\0 | f | o | o | _ | t | \\0 |
          +----+---+---+---+----+---+---+---+---+---+----+
          0    1   2   3   4    5   6   7   8   9   10   11
.Ed
.Lp
Every
.Nm
file begins with a
.Sy preamble ,
followed by a
.Sy header .
The
.Sy preamble
is defined as follows:
.Bd -literal
typedef struct ctf_preamble {
	uint16_t ctp_magic;	/* magic number (CTF_MAGIC) */
	uint8_t ctp_version;	/* data format version number (CTF_VERSION) */
	uint8_t ctp_flags;	/* flags (see below) */
} ctf_preamble_t;
.Ed
.Pp
The
.Sy preamble
is four bytes long and must be four byte aligned.
This
.Sy preamble
defines the version of the
.Nm
file which defines the format of the rest of the header.
While the header may change in subsequent versions, the preamble will not change
across versions, though the interpretation of its flags may change from
version to version.
The
.Em ctp_magic
member defines the magic number for the
.Nm
file format.
This must always be
.Li 0xcff1 .
If another value is encountered, then the file should not be treated as
a
.Nm
file.
The
.Em ctp_version
member defines the version of the
.Nm
file.
The current version is
.Li 3 .
It is possible to encounter an unsupported version.
In that case, software should not try to parse the format, as it may have
changed.
Finally, the
.Em ctp_flags
member describes aspects of the file which modify its interpretation.
The following flags are currently defined:
.Bd -literal
#define	CTF_F_COMPRESS		0x01
.Ed
.Pp
The flag
.Sy CTF_F_COMPRESS
indicates that the body of the
.Nm
file, all the data following the
.Sy header ,
has been compressed through the
.Sy zlib
library and its
.Sy deflate
algorithm.
If this flag is not present, then the body has not been compressed and no
special action is needed to interpret it.
All offsets into the data as described by
.Sy header ,
always refer to the
.Sy uncompressed
data.
.Lp
In versions two and three of the
.Nm
file format, the
.Sy header
denotes whether or not this
.Nm
file is the child of another
.Nm
file and also indicates the size of the remaining sections.
The structure for the
.Sy header
logically contains a copy of the
.Sy preamble
and the two have a combined size of 36 bytes.
.Bd -literal
typedef struct ctf_header {
	ctf_preamble_t cth_preamble;
	uint32_t cth_parlabel;	/* ref to name of parent lbl uniq'd against */
	uint32_t cth_parname;	/* ref to basename of parent */
	uint32_t cth_lbloff;	/* offset of label section */
	uint32_t cth_objtoff;	/* offset of object section */
	uint32_t cth_funcoff;	/* offset of function section */
	uint32_t cth_typeoff;	/* offset of type section */
	uint32_t cth_stroff;	/* offset of string section */
	uint32_t cth_strlen;	/* length of string section in bytes */
} ctf_header_t;
.Ed
.Pp
After the
.Sy preamble ,
the next two members
.Em cth_parlabel
and
.Em cth_parname ,
are used to identify the parent.
The value of both members are offsets into the
.Sy string
section which point to the start of a null-terminated string.
For more information on the encoding of strings, see the subsection on
.Sx String Identifiers .
If the value of either is zero, then there is no entry for that
member.
If the member
.Em cth_parlabel
is set, then the
.Em ctf_parname
member must be set, otherwise it will not be possible to find the
parent.
If
.Em ctf_parname
is set, it is not necessary to define
.Em cth_parlabel ,
as the parent may not have a label.
For more information on labels and their interpretation, see
.Sx The Label Section .
.Lp
The remaining members (excepting
.Em cth_strlen )
describe the beginning of the corresponding sections.
These offsets are relative to the end of the
.Sy header .
Therefore, something with an offset of 0 is at an offset of thirty-six
bytes relative to the start of the
.Nm
file.
The difference between members indicates the size of the section itself.
Different offsets have different alignment requirements.
The start of the
.Em cth_objtoff
and
.Em cth_funcoff
must be two byte aligned, while the sections
.Em cth_lbloff
and
.Em cth_typeoff
must be four-byte aligned.
The section
.Em cth_stroff
has no alignment requirements.
To calculate the size of a given section, excepting the
.Sy string
section, one should subtract the offset of the section from the following one.
For example, the size of the
.Sy types
section can be calculated by subtracting
.Em cth_typeoff
from
.Em cth_stroff .
.Lp
Finally, the member
.Em cth_strlen
describes the length of the string section itself.
From it, you can also calculate the size of the entire
.Nm
file by adding together the size of the
.Sy ctf_header_t ,
the offset of the string section in
.Em cth_stroff ,
and the size of the string section in
.Em cth_srlen .
.Ss Type Identifiers
Through the
.Nm ctf
data, types are referred to by identifiers.
A given
.Nm
file supports up to 2147483646 (0x7ffffffe) types.
.Nm
version 2 had a much smaller limit of 32767 types.
The first valid type identifier is 0x1.
When a given
.Nm
file is a child, indicated by a non-zero entry for the
.Sy header Ns 's
.Em cth_parname ,
then the first valid type identifier is 0x80000000 and the last is 0xfffffffe.
In this case, type identifiers 0x1 through 0x7ffffffe are references to the
parent.
0x7fffffff and 0xffffffff are not treated as valid type identifiers so as to
enable the use of -1 as an error value.
.Lp
The type identifier zero is a sentinel value used to indicate that there
is no type information available or it is an unknown type.
.Lp
Throughout the file format, the identifier is stored in different sized
values; however, the minimum size to represent a given identifier is a
.Sy uint16_t .
Other consumers of
.Nm
information may use larger or opaque identifiers.
.Ss String Identifiers
String identifiers are always encoded as four byte unsigned integers
which are an offset into a string table.
The
.Nm
format supports two different string tables which have an identifier of
zero or one.
This identifier is stored in the high-order bit of the unsigned four byte
offset.
Therefore, the maximum supported offset into one of these tables is 0x7ffffffff.
.Lp
Table identifier zero, always refers to the
.Sy string
section in the CTF file itself.
String table identifier one refers to an external string table which is the ELF
string table for the ELF symbol table associated with the
.Nm
container.
.Ss Type Encoding
Every
.Nm
type begins with metadata encoded into a
.Sy uint32_t .
This encoded information tells us three different pieces of information:
.Bl -bullet -offset indent -compact
.It
The kind of the type
.It
Whether this type is a root type or not
.It
The length of the variable data
.El
.Lp
The 32 bits that make up the encoding are broken down into six bits
for the kind (bits 26 to 31), one bit for the root type flag (bit 25),
and 25 bits for the length of the variable data.
.Lp
The current version of the file format defines 14 different kinds.
The interpretation of these different kinds will be discussed in the section
.Sx The Type Section .
If a kind is encountered that is not listed below, then it is not a valid
.Nm
file.
The kinds are defined as follows:
.Bd -literal -offset indent
#define	CTF_K_UNKNOWN	0
#define	CTF_K_INTEGER	1
#define	CTF_K_FLOAT	2
#define	CTF_K_POINTER	3
#define	CTF_K_ARRAY	4
#define	CTF_K_FUNCTION	5
#define	CTF_K_STRUCT	6
#define	CTF_K_UNION	7
#define	CTF_K_ENUM	8
#define	CTF_K_FORWARD	9
#define	CTF_K_TYPEDEF	10
#define	CTF_K_VOLATILE	11
#define	CTF_K_CONST	12
#define	CTF_K_RESTRICT	13
.Ed
.Lp
Programs directly reference many types; however, other types are referenced
indirectly because they are part of some other structure.
These types that are referenced directly and used are called
.Sy root
types.
Other types may be used indirectly, for example, a program may reference
a structure directly, but not one of its members which has a type.
That type is not considered a
.Sy root
type.
If a type is a
.Sy root
type, then it will have bit 25 set.
.Lp
The variable length section is specific to each kind and is discussed in the
section
.Sx The Type Section .
.Lp
The following macros are useful for constructing and deconstructing the encoded
type information:
.Bd -literal -offset indent

#define	CTF_V3_MAX_VLEN			0x00ffffff
#define	CTF_V3_INFO_KIND(info)		(((info) & 0xfc000000) >> 26)
#define	CTF_V3_INFO_ISROOT(info)	(((info) & 0x02000000) >> 25)
#define	CTF_V3_INFO_VLEN(info)		(((info) & CTF_V3_MAX_VLEN))

#define	CTF_V3_TYPE_INFO(kind, isroot, vlen) \\
	(((kind) << 26) | (((isroot) ? 1 : 0) << 25) | ((vlen) & CTF_V3_MAX_VLEN))
.Ed
.Ss The Label Section
When consuming
.Nm
data, it is often useful to know whether two different
.Nm
containers come from the same source base and version.
For example, when building illumos, there are many kernel modules that are built
against a single collection of source code.
A label is encoded into the
.Nm
files that corresponds with the particular build.
This ensures that if files on the system were to become mixed up from multiple
releases, that they are not used together by tools, particularly when a child
needs to refer to a type in the parent.
Because they are linked using the type identifiers, if the wrong parent is used
then the wrong type will be encountered.
Note that this mechanism is not currently used on
.Fx .
In particular, kernel modules built on
.Fx
each contain a complete type graph.
.Lp
Each label is encoded in the file format using the following eight byte
structure:
.Bd -literal
typedef struct ctf_lblent {
	uint32_t ctl_label;	/* ref to name of label */
	uint32_t ctl_typeidx;	/* last type associated with this label */
} ctf_lblent_t;
.Ed
.Lp
Each label has two different components, a name and a type identifier.
The name is encoded in the
.Em ctl_label
member which is in the format defined in the section
.Sx String Identifiers .
Generally, the names of all labels are found in the internal string
section.
.Lp
The type identifier encoded in the member
.Em ctl_typeidx
refers to the last type identifier that a label refers to in the current
file.
Labels only refer to types in the current file, if the
.Nm
file is a child, then it will have the same label as its parent;
however, its label will only refer to its types, not its parent's.
.Lp
It is also possible, though rather uncommon, for a
.Nm
file to have multiple labels.
Labels are placed one after another, every eight bytes.
When multiple labels are present, types may only belong to a single label.
.Ss The Object Section
The object section provides a mapping from ELF symbols of type
.Sy STT_OBJECT
in the symbol table to a type identifier.
Every entry in this section is a
.Sy uint32_t
which contains a type identifier as described in the section
.Sx Type Identifiers .
If there is no information for an object, then the type identifier 0x0
is stored for that entry.
.Lp
To walk the object section, you need to have a corresponding
.Sy symbol table
in the ELF object that contains the
.Nm
data.
Not every object is included in this section.
Specifically, when walking the symbol table, an entry is skipped if it matches
any of the following conditions:
.Lp
.Bl -bullet -offset indent -compact
.It
The symbol type is not
.Sy STT_OBJECT
.It
The symbol's section index is
.Sy SHN_UNDEF
.It
The symbol's name offset is zero
.It
The symbol's section index is
.Sy SHN_ABS
and the value of the symbol is zero.
.It
The symbol's name is
.Li _START_
or
.Li _END_ .
These are skipped because they are used for scoping local symbols in
ELF.
.El
.Lp
The following sample code shows an example of iterating the object
section and skipping the correct symbols:
.Bd -literal
#include <gelf.h>
#include <stdio.h>

/*
 * Given the start of the object section in a CTFv3 file, the number of symbols,
 * and the ELF Data sections for the symbol table and the string table, this
 * prints the type identifiers that correspond to objects. Note, a more robust
 * implementation should ensure that they don't walk beyond the end of the CTF
 * object section.
 *
 * An implementation that handles CTFv2 must take into account the fact that
 * type identifiers are 16 bits wide rather than 32 bits wide.
 */
static int
walk_symbols(uint32_t *objtoff, Elf_Data *symdata, Elf_Data *strdata,
    long nsyms)
{
	long i;
	uintptr_t strbase = strdata->d_buf;

	for (i = 1; i < nsyms; i++, objftoff++) {
		const char *name;
		GElf_Sym sym;

		if (gelf_getsym(symdata, i, &sym) == NULL)
			return (1);

		if (GELF_ST_TYPE(sym.st_info) != STT_OBJECT)
			continue;
		if (sym.st_shndx == SHN_UNDEF || sym.st_name == 0)
			continue;
		if (sym.st_shndx == SHN_ABS && sym.st_value == 0)
			continue;
		name = (const char *)(strbase + sym.st_name);
		if (strcmp(name, "_START_") == 0 || strcmp(name, "_END_") == 0)
			continue;

		(void) printf("Symbol %d has type %d\n", i, *objtoff);
	}

	return (0);
}
.Ed
.Ss The Function Section
The function section of the
.Nm
file encodes the types of both the function's arguments and the function's
return value.
Similar to
.Sx The Object Section ,
the function section encodes information for all symbols of type
.Sy STT_FUNCTION ,
excepting those that fit specific criteria.
Unlike with objects, because functions have a variable number of arguments, they
start with a type encoding as defined in
.Sx Type Encoding ,
which is the size of a
.Sy uint32_t .
For functions which have no type information available, they are encoded as
.Li CTF_V3_TYPE_INFO(CTF_K_UNKNOWN, 0, 0) .
Functions with arguments are encoded differently.
Here, the variable length is turned into the number of arguments in the
function.
If a function is a
.Sy varargs
type function, then the number of arguments is increased by one.
Functions with type information are encoded as:
.Li CTF_V3_TYPE_INFO(CTF_K_FUNCTION, 0, nargs) .
.Lp
For functions that have no type information, nothing else is encoded, and the
next function is encoded.
For functions with type information, the next
.Sy uint32_t
is encoded with the type identifier of the return type of the function.
It is followed by each of the type identifiers of the arguments, if any exist,
in the order that they appear in the function.
Therefore, argument 0 is the first type identifier and so on.
When a function has a final varargs argument, that is encoded with the type
identifier of zero.
.Lp
Like
.Sx The Object Section ,
the function section is encoded in the order of the symbol table.
It has similar, but slightly different considerations from objects.
While iterating the symbol table, if any of the following conditions are true,
then the entry is skipped and no corresponding entry is written:
.Lp
.Bl -bullet -offset indent -compact
.It
The symbol type is not
.Sy STT_FUNCTION
.It
The symbol's section index is
.Sy SHN_UNDEF
.It
The symbol's name offset is zero
.It
The symbol's name is
.Li _START_
or
.Li _END_ .
These are skipped because they are used for scoping local symbols in
ELF.
.El
.Ss The Type Section
The type section is the heart of the
.Nm
data.
It encodes all of the information about the types themselves.
The base of the type information comes in two forms, a short form and a long
form, each of which may be followed by a variable number of arguments.
The following definitions describe the short and long forms:
.Bd -literal
#define	CTF_V3_MAX_SIZE		0xfffffffe	/* max size of a type in bytes */
#define	CTF_V3_LSIZE_SENT	0xffffffff	/* sentinel for ctt_size */
#define	CTF_V3_MAX_LSIZE	UINT64_MAX

struct ctf_stype_v3 {
	uint32_t ctt_name;	/* reference to name in string table */
	uint32_t ctt_info;	/* encoded kind, variant length */
	union {
		uint32_t _size;	/* size of entire type in bytes */
		uint32_t _type;	/* reference to another type */
	} _u;
};

struct ctf_type_v3 {
	uint32_t ctt_name;	/* reference to name in string table */
	uint32_t ctt_info;	/* encoded kind, variant length */
	union {
		uint32_t _size;	/* always CTF_LSIZE_SENT */
		uint32_t _type; /* do not use */
	} _u;
	uint32_t ctt_lsizehi;	/* high 32 bits of type size in bytes */
	uint32_t ctt_lsizelo;	/* low 32 bits of type size in bytes */
};

#define	ctt_size _u._size	/* for fundamental types that have a size */
#define	ctt_type _u._type	/* for types that reference another type */
.Ed
.Pp
Type sizes are stored in
.Sy bytes .
The basic small form uses a
.Sy uint32_t
to store the number of bytes.
If the number of bytes in a structure would exceed 0xfffffffe, then the
alternate form, the
.Sy struct ctf_type_v3 ,
is used instead.
To indicate that the larger form is being used, the member
.Em ctt_size
is set to value of
.Sy CTF_V3_LSIZE_SENT
(0xffffffff).
In general, when going through the type section, consumers use the
.Sy struct ctf_type_v3
structure, but pay attention to the value of the member
.Em ctt_size
to determine whether they should increment their scan by the size of
.Sy struct ctf_stype_v3
or
.Sy struct ctf_type_v3 .
Not all kinds of types use
.Sy ctt_size .
Those which do not, will always use the
.Sy struct ctf_stype_v3
structure.
The individual sections for each kind have more information.
.Lp
Types are written out in order.
Therefore the first entry encountered has a type id of 0x1, or 0x8000 if a
child.
The member
.Em ctt_name
is encoded as described in the section
.Sx String Identifiers .
The string that it points to is the name of the type.
If the identifier points to an empty string (one that consists solely of a null
terminator) then the type does not have a name, this is common with anonymous
structures and unions that only have a typedef to name them, as well as
pointers and qualifiers.
.Lp
The next member, the
.Em ctt_info ,
is encoded as described in the section
.Sx Type Encoding .
The type's kind tells us how to interpret the remaining data in the
.Sy struct ctf_type_v3
and any variable length data that may exist.
The rest of this section will be broken down into the interpretation of the
various kinds.
.Ss Encoding of Integers
Integers, which are of type
.Sy CTF_K_INTEGER ,
have no variable length arguments.
Instead, they are followed by a
.Sy uint32_t
which describes their encoding.
All integers must be encoded with a variable length of zero.
The
.Em ctt_size
member describes the length of the integer in bytes.
In general, integer sizes will be rounded up to the closest power of two.
.Lp
The integer encoding contains three different pieces of information:
.Bl -bullet -offset indent -compact
.It
The encoding of the integer
.It
The offset in
.Sy bits
of the type
.It
The size in
.Sy bits
of the type
.El
.Pp
This encoding can be expressed through the following macros:
.Bd -literal -offset indent
#define	CTF_INT_ENCODING(data)	(((data) & 0xff000000) >> 24)
#define	CTF_INT_OFFSET(data)	(((data) & 0x00ff0000) >> 16)
#define	CTF_INT_BITS(data)	(((data) & 0x0000ffff))

#define	CTF_INT_DATA(encoding, offset, bits) \\
	(((encoding) << 24) | ((offset) << 16) | (bits))
.Ed
.Pp
The following flags are defined for the encoding at this time:
.Bd -literal -offset indent
#define	CTF_INT_SIGNED		0x01
#define	CTF_INT_CHAR		0x02
#define	CTF_INT_BOOL		0x04
#define	CTF_INT_VARARGS		0x08
.Ed
.Lp
By default, an integer is considered to be unsigned, unless it has the
.Sy CTF_INT_SIGNED
flag set.
If the flag
.Sy CTF_INT_CHAR
is set, that indicates that the integer is of a type that stores character
data, for example the intrinsic C type
.Sy char
would have the
.Sy CTF_INT_CHAR
flag set.
If the flag
.Sy CTF_INT_BOOL
is set, that indicates that the integer represents a boolean type.
For example, the intrinsic C type
.Sy _Bool
would have the
.Sy CTF_INT_BOOL
flag set.
Finally, the flag
.Sy CTF_INT_VARARGS
indicates that the integer is used as part of a variable number of arguments.
This encoding is rather uncommon.
.Ss Encoding of Floats
Floats, which are of type
.Sy CTF_K_FLOAT ,
are similar to their integer counterparts.
They have no variable length arguments and are followed by a four byte encoding
which describes the kind of float that exists.
The
.Em ctt_size
member is the size, in bytes, of the float.
The float encoding has three different pieces of information inside of it:
.Lp
.Bl -bullet -offset indent -compact
.It
The specific kind of float that exists
.It
The offset in
.Sy bits
of the float
.It
The size in
.Sy bits
of the float
.El
.Lp
This encoding can be expressed through the following macros:
.Bd -literal -offset indent
#define	CTF_FP_ENCODING(data)	(((data) & 0xff000000) >> 24)
#define	CTF_FP_OFFSET(data)	(((data) & 0x00ff0000) >> 16)
#define	CTF_FP_BITS(data)	(((data) & 0x0000ffff))

#define	CTF_FP_DATA(encoding, offset, bits) \\
	(((encoding) << 24) | ((offset) << 16) | (bits))
.Ed
.Lp
Where as the encoding for integers is a series of flags, the encoding for
floats maps to a specific kind of float.
It is not a flag-based value.
The kinds of floats correspond to both their size, and the encoding.
This covers all of the basic C intrinsic floating point types.
The following are the different kinds of floats represented in the encoding:
.Bd -literal -offset indent
#define	CTF_FP_SINGLE	1	/* IEEE 32-bit float encoding */
#define	CTF_FP_DOUBLE	2	/* IEEE 64-bit float encoding */
#define	CTF_FP_CPLX	3	/* Complex encoding */
#define	CTF_FP_DCPLX	4	/* Double complex encoding */
#define	CTF_FP_LDCPLX	5	/* Long double complex encoding */
#define	CTF_FP_LDOUBLE	6	/* Long double encoding */
#define	CTF_FP_INTRVL	7	/* Interval (2x32-bit) encoding */
#define	CTF_FP_DINTRVL	8	/* Double interval (2x64-bit) encoding */
#define	CTF_FP_LDINTRVL	9	/* Long double interval (2x128-bit) encoding */
#define	CTF_FP_IMAGRY	10	/* Imaginary (32-bit) encoding */
#define	CTF_FP_DIMAGRY	11	/* Long imaginary (64-bit) encoding */
#define	CTF_FP_LDIMAGRY	12	/* Long double imaginary (128-bit) encoding */
.Ed
.Ss Encoding of Arrays
Arrays, which are of type
.Sy CTF_K_ARRAY ,
have no variable length arguments.
They are followed by a structure which describes the number of elements in the
array, the type identifier of the elements in the array, and the type identifier
of the index of the array.
With arrays, the
.Em ctt_size
member is set to zero.
The structure that follows an array is defined as:
.Bd -literal
struct ctf_array_v3 {
	uint32_t cta_contents;	/* reference to type of array contents */
	uint32_t cta_index;	/* reference to type of array index */
	uint32_t cta_nelems;	/* number of elements */
};
.Ed
.Lp
The
.Em cta_contents
and
.Em cta_index
members of the
.Sy struct ctf_array_v3
are type identifiers which are encoded as per the section
.Sx Type Identifiers .
The member
.Em cta_nelems
is a simple four byte unsigned count of the number of elements.
This count may be zero when encountering C99's flexible array members.
.Ss Encoding of Functions
Function types, which are of type
.Sy CTF_K_FUNCTION ,
use the variable length list to be the number of arguments in the function.
When the function has a final member which is a varargs, then the argument count
is incremented by one to account for the variable argument.
Here, the
.Em ctt_type
member is encoded with the type identifier of the return type of the function.
Note that the
.Em ctt_size
member is not used here.
.Lp
The variable argument list contains the type identifiers for the arguments of
the function, if any.
Each one is represented by a
.Sy uint32_t
and encoded according to the
.Sx Type Identifiers
section.
If the function's last argument is of type varargs, then it is also written out,
but the type identifier is zero.
This is included in the count of the function's arguments.
In
.Nm
version 2, an extra type identifier may follow the argument and return type
identifiers in order to maintain four-byte alignment for the following type
definition.
Such a type identifier is not included in the argument count and has a value
of zero.
In
.Nm
version 3, four-byte alignment occurs naturally and no padding is used.
.Ss Encoding of Structures and Unions
Structures and Unions, which are encoded with
.Sy CTF_K_STRUCT
and
.Sy CTF_K_UNION
respectively,  are very similar constructs in C.
The main difference between them is the fact that members of a structure
follow one another, where as in a union, all members share the same memory.
They are also very similar in terms of their encoding in
.Nm .
The variable length argument for structures and unions represents the number of
members that they have.
The value of the member
.Em ctt_size
is the size of the structure and union.
There are two different structures which are used to encode members in the
variable list.
When the size of a structure or union is greater than or equal to the large
member threshold, 536870912, then a different structure is used to encode the
member; all members are encoded using the same structure.
The structure for members is as follows:
.Bd -literal
struct ctf_member_v3 {
	uint32_t ctm_name;	/* reference to name in string table */
	uint32_t ctm_type;	/* reference to type of member */
	uint32_t ctm_offset;	/* offset of this member in bits */
};

struct ctf_lmember_v3 {
	uint32_t ctlm_name;	/* reference to name in string table */
	uint32_t ctlm_type;	/* reference to type of member */
	uint32_t ctlm_offsethi;	/* high 32 bits of member offset in bits */
	uint32_t ctlm_offsetlo;	/* low 32 bits of member offset in bits */
};
.Ed
.Lp
Both the
.Em ctm_name
and
.Em ctlm_name
refer to the name of the member.
The name is encoded as an offset into the string table as described by the
section
.Sx String Identifiers .
The members
.Sy ctm_type
and
.Sy ctlm_type
both refer to the type of the member.
They are encoded as per the section
.Sx Type Identifiers .
.Lp
The last piece of information that is present is the offset which describes the
offset in memory at which the member begins.
For unions, this value will always be zero because each member of a union has
an offset of zero.
For structures, this is the offset in
.Sy bits
at which the member begins.
Note that a compiler may lay out a type with padding.
This means that the difference in offset between two consecutive members may be
larger than the size of the member.
When the size of the overall structure is strictly less than 536870912 bytes,
the normal structure,
.Sy struct ctf_member_v3 ,
is used and the offset in bits is stored in the member
.Em ctm_offset .
However, when the size of the structure is greater than or equal to 536870912
bytes, then the number of bits is split into two 32-bit quantities.
One member,
.Em ctlm_offsethi ,
represents the upper 32 bits of the offset, while the other member,
.Em ctlm_offsetlo ,
represents the lower 32 bits of the offset.
These can be joined together to get a 64-bit sized offset in bits by shifting
the member
.Em ctlm_offsethi
to the left by thirty two and then doing a binary or of
.Em ctlm_offsetlo .
.Ss Encoding of Enumerations
Enumerations, noted by the type
.Sy CTF_K_ENUM ,
are similar to structures.
Enumerations use the variable list to note the number of values that the
enumeration contains, which we'll term enumerators.
In C, an enumeration is always equivalent to the intrinsic type
.Sy int ,
thus the value of the member
.Em ctt_size
is always the size of an integer which is determined based on the current model.
For
.Fx
systems, this will always be 4, as an integer is always defined to
be 4 bytes large in both
.Sy ILP32
and
.Sy LP64 ,
regardless of the architecture.
For further details, see
.Xr arch 7 .
.Lp
The enumerators encoded in an enumeration have the following structure in the
variable list:
.Bd -literal
typedef struct ctf_enum {
	uint32_t cte_name;	/* reference to name in string table */
	int32_t cte_value;	/* value associated with this name */
} ctf_enum_t;
.Ed
.Pp
The member
.Em cte_name
refers to the name of the enumerator's value, it is encoded according to the
rules in the section
.Sx String Identifiers .
The member
.Em cte_value
contains the integer value of this enumerator.
.Ss Encoding of Forward References
Forward references, types of kind
.Sy CTF_K_FORWARD ,
in a
.Nm
file refer to types which may not have a definition at all, only a name.
If the
.Nm
file is a child, then it may be that the forward is resolved to an
actual type in the parent, otherwise the definition may be in another
.Nm
container or may not be known at all.
The only member of the
.Sy struct ctf_type_v3
that matters for a forward declaration is the
.Em ctt_name
which points to the name of the forward reference in the string table as
described earlier.
There is no other information recorded for forward references.
.Ss Encoding of Pointers, Typedefs, Volatile, Const, and Restrict
Pointers, typedefs, volatile, const, and restrict are all similar in
.Nm .
They all refer to another type.
In the case of typedefs, they provide an alternate name, while volatile, const,
and restrict change how the type is interpreted in the C programming language.
This covers the
.Nm
kinds
.Sy CTF_K_POINTER ,
.Sy CTF_K_TYPEDEF ,
.Sy CTF_K_VOLATILE ,
.Sy CTF_K_RESTRICT ,
and
.Sy CTF_K_CONST .
.Lp
These types have no variable list entries and use the member
.Em ctt_type
to refer to the base type that they modify.
.Ss Encoding of Unknown Types
Types with the kind
.Sy CTF_K_UNKNOWN
are used to indicate gaps in the type identifier space.
These entries consume an identifier, but do not define anything.
Nothing should refer to these gap identifiers.
.Ss Dependencies Between Types
C types can be imagined as a directed, cyclic, graph.
Structures and unions may refer to each other in a way that creates a cyclic
dependency.
In cases such as these, the entire type section must be read in and processed.
Consumers must not assume that every type can be laid out in dependency order;
they cannot.
.Ss The String Section
The last section of the
.Nm
file is the
.Sy string
section.
This section encodes all of the strings that appear throughout the other
sections.
It is laid out as a series of characters followed by a null terminator.
Generally, all names are written out in ASCII, as most C compilers do not allow
any characters to appear in identifiers outside of a subset of ASCII.
However, any extended characters sets should be written out as a series of UTF-8
bytes.
.Lp
The first entry in the section, at offset zero, is a single null
terminator to reference the empty string.
Following that, each C string should be written out, including the null
terminator.
Offsets that refer to something in this section should refer to the first byte
which begins a string.
Beyond the first byte in the section being the null terminator, the order of
strings is unimportant.
.Ss Data Encoding and ELF Considerations
.Nm
data is generally included in ELF objects which specify information to
identify the architecture and endianness of the file.
A
.Nm
container inside such an object must match the endianness of the ELF object.
Aside from the question of the endian encoding of data, there should be no other
differences between architectures.
While many of the types in this document refer to non-fixed size C integral
types, they are equivalent in the models
.Sy ILP32
and
.Sy LP64 .
If any other model is being used with
.Nm
data that has different sizes, then it must not use the model's sizes for
those integral types and instead use the fixed size equivalents based on an
.Sy ILP32
environment.
.Lp
When placing a
.Nm
container inside of an ELF object, there are certain conventions that are
expected for the purposes of tooling being able to find the
.Nm
data.
In particular, a given ELF object should only contain a single
.Nm
section.
Multiple containers should be merged together into a single one.
.Lp
The
.Nm
file should be included in its own ELF section.
The section's name must be
.Ql .SUNW_ctf .
The type of the section must be
.Sy SHT_PROGBITS .
The section should have a link set to the symbol table and its address
alignment must be 4.
.Sh SEE ALSO
.Xr ctfconvert 1 ,
.Xr ctfdump 1 ,
.Xr ctfmerge 1 ,
.Xr dtrace 1 ,
.Xr elf 3 ,
.Xr gelf 3 ,
.Xr a.out 5 ,
.Xr elf 5 ,
.Xr arch 7