The "stabs" representation of debugging information ¶

1.1 Overview of Debugging Information Flow ¶

The GNU C compiler compiles C source in a .c file into assembly language in a .s file, which the assembler translates into a .o file, which the linker combines with other .o files and libraries to produce an executable file.

With the ‘-g’ option, GCC puts in the .s file additional debugging information, which is slightly transformed by the assembler and linker, and carried through into the final executable. This debugging information describes features of the source file like line numbers, the types and scopes of variables, and function names, parameters, and scopes.

For some object file formats, the debugging information is encapsulated in assembler directives known collectively as stab (symbol table) directives, which are interspersed with the generated code. Stabs are the native format for debugging information in the a.out and XCOFF object file formats. The GNU tools can also emit stabs in the COFF and ECOFF object file formats.

The assembler adds the information from stabs to the symbol information it places by default in the symbol table and the string table of the .o file it is building. The linker consolidates the .o files into one executable file, with one symbol table and one string table. Debuggers use the symbol and string tables in the executable as a source of debugging information about the program.

1.2 Overview of Stab Format ¶

There are three overall formats for stab assembler directives, differentiated by the first word of the stab. The name of the directive describes which combination of four possible data fields follows. It is either .stabs (string), .stabn (number), or .stabd (dot). IBM’s XCOFF assembler uses .stabx (and some other directives such as .file and .bi) instead of .stabs, .stabn or .stabd.

The overall format of each class of stab is:

.stabs "string",type,other,desc,value
.stabn type,other,desc,value
.stabd type,other,desc
.stabx "string",value,type,sdb-type

For .stabn and .stabd, there is no string (the n_strx field is zero; see Symbol Information in Symbol Tables). For .stabd, the value field is implicit and has the value of the current file location. For .stabx, the sdb-type field is unused for stabs and can always be set to zero. The other field is almost always unused and can be set to zero.

The number in the type field gives some basic information about which type of stab this is (or whether it is a stab, as opposed to an ordinary symbol). Each valid type number defines a different stab type; further, the stab type defines the exact interpretation of, and possible values for, any remaining string, desc, or value fields present in the stab. See Table of Stab Types, for a list in numeric order of the valid type field values for stab directives.

1.3 The String Field ¶

For most stabs the string field holds the meat of the debugging information. The flexible nature of this field is what makes stabs extensible. For some stab types the string field contains only a name. For other stab types the contents can be a great deal more complex.

The overall format of the string field for most stab types is:

"name:symbol-descriptor type-information"

name is the name of the symbol represented by the stab; it can contain a pair of colons (see Defining a Symbol Within Another Type). name can be omitted, which means the stab represents an unnamed object. For example, ‘:t10=*2’ defines type 10 as a pointer to type 2, but does not give the type a name. Omitting the name field is supported by AIX dbx and GDB after about version 4.8, but not other debuggers. GCC sometimes uses a single space as the name instead of omitting the name altogether; apparently that is supported by most debuggers.

The symbol-descriptor following the ‘:’ is an alphabetic character that tells more specifically what kind of symbol the stab represents. If the symbol-descriptor is omitted, but type information follows, then the stab represents a local variable. For a list of symbol descriptors, see Table of Symbol Descriptors. The ‘c’ symbol descriptor is an exception in that it is not followed by type information. See Constants.

type-information is either a type-number, or ‘type-number=’. A type-number alone is a type reference, referring directly to a type that has already been defined.

The ‘type-number=’ form is a type definition, where the number represents a new type which is about to be defined. The type definition may refer to other types by number, and those type numbers may be followed by ‘=’ and nested definitions. Also, the Lucid compiler will repeat ‘type-number=’ more than once if it wants to define several type numbers at once.

In a type definition, if the character that follows the equals sign is non-numeric then it is a type-descriptor, and tells what kind of type is about to be defined. Any other values following the type-descriptor vary, depending on the type-descriptor. See Table of Type Descriptors, for a list of type-descriptor values. If a number follows the ‘=’ then the number is a type-reference. For a full description of types, Defining Types.

A type-number is often a single number. The GNU and Sun tools additionally permit a type-number to be a pair (file-number,filetype-number) (the parentheses appear in the string, and serve to distinguish the two cases). The file-number is 0 for the base source file, 1 for the first included file, 2 for the next, and so on. The filetype-number is a number starting with 1 which is incremented for each new type defined in the file. (Separating the file number and the type number permits the N_BINCL optimization to succeed more often; see Names of Include Files).

There is an AIX extension for type attributes. Following the ‘=’ are any number of type attributes. Each one starts with ‘@’ and ends with ‘;’. Debuggers, including AIX’s dbx and GDB 4.10, skip any type attributes they do not recognize. GDB 4.9 and other versions of dbx may not do this. Because of a conflict with C++ (see GNU C++ Stabs), new attributes should not be defined which begin with a digit, ‘(’, or ‘-’; GDB may be unable to distinguish those from the C++ type descriptor ‘@’. The attributes are:

aboundary

boundary is an integer specifying the alignment. I assume it applies to all variables of this type.

pinteger

Pointer class (for checking). Not sure what this means, or how integer is interpreted.

P

Indicate this is a packed type, meaning that structure fields or array elements are placed more closely in memory, to save memory at the expense of speed.

ssize

Size in bits of a variable of this type. This is fully supported by GDB 4.11 and later.

S

Indicate that this type is a string instead of an array of characters, or a bitstring instead of a set. It doesn’t change the layout of the data being represented, but does enable the debugger to know which type it is.

V

Indicate that this type is a vector instead of an array. The only major difference between vectors and arrays is that vectors are passed by value instead of by reference (vector coprocessor extension).

All of this can make the string field quite long. All versions of GDB, and some versions of dbx, can handle arbitrarily long strings. But many versions of dbx (or assemblers or linkers, I’m not sure which) cretinously limit the strings to about 80 characters, so compilers which must work with such systems need to split the .stabs directive into several .stabs directives. Each stab duplicates every field except the string field. The string field of every stab except the last is marked as continued with a backslash at the end (in the assembly code this may be written as a double backslash, depending on the assembler). Removing the backslashes and concatenating the string fields of each stab produces the original, long string. Just to be incompatible (or so they don’t have to worry about what the assembler does with backslashes), AIX can use ‘?’ instead of backslash.

1.4 A Simple Example in C Source ¶

To get the flavor of how stabs describe source information for a C program, let’s look at the simple program:

main()
{
        printf("Hello world");
}

When compiled with ‘-g’, the program above yields the following .s file. Line numbers have been added to make it easier to refer to parts of the .s file in the description of the stabs that follows.

1.5 The Simple Example at the Assembly Level ¶

This simple “hello world” example demonstrates several of the stab types used to describe C language source files.

1  gcc2_compiled.:
2  .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0
3  .stabs "hello.c",100,0,0,Ltext0
4  .text
5  Ltext0:
6  .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
7  .stabs "char:t2=r2;0;127;",128,0,0,0
8  .stabs "long int:t3=r1;-2147483648;2147483647;",128,0,0,0
9  .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
10 .stabs "long unsigned int:t5=r1;0;-1;",128,0,0,0
11 .stabs "short int:t6=r1;-32768;32767;",128,0,0,0
12 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
13 .stabs "short unsigned int:t8=r1;0;65535;",128,0,0,0
14 .stabs "long long unsigned int:t9=r1;0;-1;",128,0,0,0
15 .stabs "signed char:t10=r1;-128;127;",128,0,0,0
16 .stabs "unsigned char:t11=r1;0;255;",128,0,0,0
17 .stabs "float:t12=r1;4;0;",128,0,0,0
18 .stabs "double:t13=r1;8;0;",128,0,0,0
19 .stabs "long double:t14=r1;8;0;",128,0,0,0
20 .stabs "void:t15=15",128,0,0,0
21      .align 4
22 LC0:
23      .ascii "Hello, world!\12\0"
24      .align 4
25      .global _main
26      .proc 1
27 _main:
28 .stabn 68,0,4,LM1
29 LM1:
30      !#PROLOGUE# 0
31      save %sp,-136,%sp
32      !#PROLOGUE# 1
33      call ___main,0
34      nop
35 .stabn 68,0,5,LM2
36 LM2:
37 LBB2:
38      sethi %hi(LC0),%o1
39      or %o1,%lo(LC0),%o0
40      call _printf,0
41      nop
42 .stabn 68,0,6,LM3
43 LM3:
44 LBE2:
45 .stabn 68,0,6,LM4
46 LM4:
47 L1:
48      ret
49      restore
50 .stabs "main:F1",36,0,0,_main
51 .stabn 192,0,0,LBB2
52 .stabn 224,0,0,LBE2

2.1 Main Program ¶

Most languages allow the main program to have any name. The N_MAIN stab type tells the debugger the name that is used in this program. Only the string field is significant; it is the name of a function which is the main program. Most C compilers do not use this stab (they expect the debugger to assume that the name is main), but some C compilers emit an N_MAIN stab for the main function. I’m not sure how XCOFF handles this.

2.2 Paths and Names of the Source Files ¶

Before any other stabs occur, there must be a stab specifying the source file. This information is contained in a symbol of stab type N_SO; the string field contains the name of the file. The value of the symbol is the start address of the portion of the text section corresponding to that file.

Some compilers use the desc field to indicate the language of the source file. Sun’s compilers started this usage, and the first constants are derived from their documentation. Languages added by gcc/gdb start at 0x32 to avoid conflict with languages Sun may add in the future. A desc field with a value 0 indicates that no language has been specified via this mechanism.

N_SO_AS (0x1)

Assembly language

N_SO_C (0x2)

K&R traditional C

N_SO_ANSI_C (0x3)

ANSI C

N_SO_CC (0x4)

C++

N_SO_FORTRAN (0x5)

Fortran

N_SO_PASCAL (0x6)

Pascal

N_SO_FORTRAN90 (0x7)

Fortran90

N_SO_OBJC (0x32)

Objective-C

N_SO_OBJCPLUS (0x33)

Objective-C++

Some compilers (for example, GCC2 and SunOS4 /bin/cc) also include the directory in which the source was compiled, in a second N_SO symbol preceding the one containing the file name. This symbol can be distinguished by the fact that it ends in a slash. Code from the cfront C++ compiler can have additional N_SO symbols for nonexistent source files after the N_SO for the real source file; these are believed to contain no useful information.

For example:

.stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0     # 100 is N_SO
.stabs "hello.c",100,0,0,Ltext0
        .text
Ltext0:

Instead of N_SO symbols, XCOFF uses a .file assembler directive which assembles to a C_FILE symbol; explaining this in detail is outside the scope of this document.

If it is useful to indicate the end of a source file, this is done with an N_SO symbol with an empty string for the name. The value is the address of the end of the text section for the file. For some systems, there is no indication of the end of a source file, and you just need to figure it ended when you see an N_SO for a different source file, or a symbol ending in .o (which at least some linkers insert to mark the start of a new .o file).

2.3 Names of Include Files ¶

There are several schemes for dealing with include files: the traditional N_SOL approach, Sun’s N_BINCL approach, and the XCOFF C_BINCL approach (which despite the similar name has little in common with N_BINCL).

An N_SOL symbol specifies which include file subsequent symbols refer to. The string field is the name of the file and the value is the text address corresponding to the end of the previous include file and the start of this one. To specify the main source file again, use an N_SOL symbol with the name of the main source file.

The N_BINCL approach works as follows. An N_BINCL symbol specifies the start of an include file. In an object file, only the string is significant; the linker puts data into some of the other fields. The end of the include file is marked by an N_EINCL symbol (which has no string field). In an object file, there is no significant data in the N_EINCL symbol. N_BINCL and N_EINCL can be nested.

If the linker detects that two source files have identical stabs between an N_BINCL and N_EINCL pair (as will generally be the case for a header file), then it only puts out the stabs once. Each additional occurrence is replaced by an N_EXCL symbol. I believe the GNU linker and the Sun (both SunOS4 and Solaris) linker are the only ones which supports this feature.

A linker which supports this feature will set the value of a N_BINCL symbol to the total of all the characters in the stabs strings included in the header file, omitting any file numbers. The value of an N_EXCL symbol is the same as the value of the N_BINCL symbol it replaces. This information can be used to match up N_EXCL and N_BINCL symbols which have the same filename. The N_EINCL value, and the values of the other and description fields for all three, appear to always be zero.

For the start of an include file in XCOFF, use the .bi assembler directive, which generates a C_BINCL symbol. A .ei directive, which generates a C_EINCL symbol, denotes the end of the include file. Both directives are followed by the name of the source file in quotes, which becomes the string for the symbol. The value of each symbol, produced automatically by the assembler and linker, is the offset into the executable of the beginning (inclusive, as you’d expect) or end (inclusive, as you would not expect) of the portion of the COFF line table that corresponds to this include file. C_BINCL and C_EINCL do not nest.

2.4 Line Numbers ¶

An N_SLINE symbol represents the start of a source line. The desc field contains the line number and the value contains the code address for the start of that source line. On most machines the address is absolute; for stabs in sections (see Using Stabs in Their Own Sections), it is relative to the function in which the N_SLINE symbol occurs.

GNU documents N_DSLINE and N_BSLINE symbols for line numbers in the data or bss segments, respectively. They are identical to N_SLINE but are relocated differently by the linker. They were intended to be used to describe the source location of a variable declaration, but I believe that GCC2 actually puts the line number in the desc field of the stab for the variable itself. GDB has been ignoring these symbols (unless they contain a string field) since at least GDB 3.5.

For single source lines that generate discontiguous code, such as flow of control statements, there may be more than one line number entry for the same source line. In this case there is a line number entry at the start of each code range, each with the same line number.

XCOFF does not use stabs for line numbers. Instead, it uses COFF line numbers (which are outside the scope of this document). Standard COFF line numbers cannot deal with include files, but in XCOFF this is fixed with the C_BINCL method of marking include files (see Names of Include Files).

2.5 Procedures ¶

All of the following stabs normally use the N_FUN symbol type. However, Sun’s acc compiler on SunOS4 uses N_GSYM and N_STSYM, which means that the value of the stab for the function is useless and the debugger must get the address of the function from the non-stab symbols instead. On systems where non-stab symbols have leading underscores, the stabs will lack underscores and the debugger needs to know about the leading underscore to match up the stab and the non-stab symbol. BSD Fortran is said to use N_FNAME with the same restriction; the value of the symbol is not useful (I’m not sure it really does use this, because GDB doesn’t handle this and no one has complained).

A function is represented by an ‘F’ symbol descriptor for a global (extern) function, and ‘f’ for a static (local) function. For a.out, the value of the symbol is the address of the start of the function; it is already relocated. For stabs in ELF, the SunPRO compiler version 2.0.1 and GCC put out an address which gets relocated by the linker. In a future release SunPRO is planning to put out zero, in which case the address can be found from the ELF (non-stab) symbol. Because looking things up in the ELF symbols would probably be slow, I’m not sure how to find which symbol of that name is the right one, and this doesn’t provide any way to deal with nested functions, it would probably be better to make the value of the stab an address relative to the start of the file, or just absolute. See Having the Linker Relocate Stabs in ELF for more information on linker relocation of stabs in ELF files. For XCOFF, the stab uses the C_FUN storage class and the value of the stab is meaningless; the address of the function can be found from the csect symbol (XTY_LD/XMC_PR).

The type information of the stab represents the return type of the function; thus ‘foo:f5’ means that foo is a function returning type 5. There is no need to try to get the line number of the start of the function from the stab for the function; it is in the next N_SLINE symbol.

Some compilers (such as Sun’s Solaris compiler) support an extension for specifying the types of the arguments. I suspect this extension is not used for old (non-prototyped) function definitions in C. If the extension is in use, the type information of the stab for the function is followed by type information for each argument, with each argument preceded by ‘;’. An argument type of 0 means that additional arguments are being passed, whose types and number may vary (‘...’ in ANSI C). GDB has tolerated this extension (parsed the syntax, if not necessarily used the information) since at least version 4.8; I don’t know whether all versions of dbx tolerate it. The argument types given here are not redundant with the symbols for the formal parameters (see Parameters); they are the types of the arguments as they are passed, before any conversions might take place. For example, if a C function which is declared without a prototype takes a float argument, the value is passed as a double but then converted to a float. Debuggers need to use the types given in the arguments when printing values, but when calling the function they need to use the types given in the symbol defining the function.

If the return type and types of arguments of a function which is defined in another source file are specified (i.e., a function prototype in ANSI C), traditionally compilers emit no stab; the only way for the debugger to find the information is if the source file where the function is defined was also compiled with debugging symbols. As an extension the Solaris compiler uses symbol descriptor ‘P’ followed by the return type of the function, followed by the arguments, each preceded by ‘;’, as in a stab with symbol descriptor ‘f’ or ‘F’. This use of symbol descriptor ‘P’ can be distinguished from its use for register parameters (see Passing Parameters in Registers) by the fact that it has symbol type N_FUN.

The AIX documentation also defines symbol descriptor ‘J’ as an internal function. I assume this means a function nested within another function. It also says symbol descriptor ‘m’ is a module in Modula-2 or extended Pascal.

Procedures (functions which do not return values) are represented as functions returning the void type in C. I don’t see why this couldn’t be used for all languages (inventing a void type for this purpose if necessary), but the AIX documentation defines ‘I’, ‘P’, and ‘Q’ for internal, global, and static procedures, respectively. These symbol descriptors are unusual in that they are not followed by type information.

The following example shows a stab for a function main which returns type number 1. The _main specified for the value is a reference to an assembler label which is used to fill in the start address of the function.

.stabs "main:F1",36,0,0,_main      # 36 is N_FUN

The stab representing a procedure is located immediately following the code of the procedure. This stab is in turn directly followed by a group of other stabs describing elements of the procedure. These other stabs describe the procedure’s parameters, its block local variables, and its block structure.

If functions can appear in different sections, then the debugger may not be able to find the end of a function. Recent versions of GCC will mark the end of a function with an N_FUN symbol with an empty string for the name. The value is the address of the end of the current function. Without such a symbol, there is no indication of the address of the end of a function, and you must assume that it ended at the starting address of the next function or at the end of the text section for the program.

2.6 Nested Procedures ¶

For any of the symbol descriptors representing procedures, after the symbol descriptor and the type information is optionally a scope specifier. This consists of a comma, the name of the procedure, another comma, and the name of the enclosing procedure. The first name is local to the scope specified, and seems to be redundant with the name of the symbol (before the ‘:’). This feature is used by GCC, and presumably Pascal, Modula-2, etc., compilers, for nested functions.

If procedures are nested more than one level deep, only the immediately containing scope is specified. For example, this code:

int
foo (int x)
{
  int bar (int y)
    {
      int baz (int z)
        {
          return x + y + z;
        }
      return baz (x + 2 * y);
    }
  return x + bar (3 * x);
}

produces the stabs:

.stabs "baz:f1,baz,bar",36,0,0,_baz.15         # 36 is N_FUN
.stabs "bar:f1,bar,foo",36,0,0,_bar.12
.stabs "foo:F1",36,0,0,_foo

2.7 Block Structure ¶

The program’s block structure is represented by the N_LBRAC (left brace) and the N_RBRAC (right brace) stab types. The variables defined inside a block precede the N_LBRAC symbol for most compilers, including GCC. Other compilers, such as the Convex, Acorn RISC machine, and Sun acc compilers, put the variables after the N_LBRAC symbol. The values of the N_LBRAC and N_RBRAC symbols are the start and end addresses of the code of the block, respectively. For most machines, they are relative to the starting address of this source file. For the Gould NP1, they are absolute. For stabs in sections (see Using Stabs in Their Own Sections), they are relative to the function in which they occur.

The N_LBRAC and N_RBRAC stabs that describe the block scope of a procedure are located after the N_FUN stab that represents the procedure itself.

Sun documents the desc field of N_LBRAC and N_RBRAC symbols as containing the nesting level of the block. However, dbx seems to not care, and GCC always sets desc to zero.

For XCOFF, block scope is indicated with C_BLOCK symbols. If the name of the symbol is ‘.bb’, then it is the beginning of the block; if the name of the symbol is ‘.be’; it is the end of the block.

2.8 Alternate Entry Points ¶

Some languages, like Fortran, have the ability to enter procedures at some place other than the beginning. One can declare an alternate entry point. The N_ENTRY stab is for this; however, the Sun FORTRAN compiler doesn’t use it. According to AIX documentation, only the name of a C_ENTRY stab is significant; the address of the alternate entry point comes from the corresponding external symbol. A previous revision of this document said that the value of an N_ENTRY stab was the address of the alternate entry point, but I don’t know the source for that information.

4.1 Automatic Variables Allocated on the Stack ¶

If a variable’s scope is local to a function and its lifetime is only as long as that function executes (C calls such variables automatic), it can be allocated in a register (see Register Variables) or on the stack.

Each variable allocated on the stack has a stab with the symbol descriptor omitted. Since type information should begin with a digit, ‘-’, or ‘(’, only those characters precluded from being used for symbol descriptors. However, the Acorn RISC machine (ARM) is said to get this wrong: it puts out a mere type definition here, without the preceding ‘type-number=’. This is a bad idea; there is no guarantee that type descriptors are distinct from symbol descriptors. Stabs for stack variables use the N_LSYM stab type, or C_LSYM for XCOFF.

The value of the stab is the offset of the variable within the local variables. On most machines this is an offset from the frame pointer and is negative. The location of the stab specifies which block it is defined in; see Block Structure.

For example, the following C code:

int
main ()
{
  int x;
}

produces the following stabs:

.stabs "main:F1",36,0,0,_main   # 36 is N_FUN
.stabs "x:1",128,0,0,-12        # 128 is N_LSYM
.stabn 192,0,0,LBB2             # 192 is N_LBRAC
.stabn 224,0,0,LBE2             # 224 is N_RBRAC

See Procedures for more information on the N_FUN stab, and Block Structure for more information on the N_LBRAC and N_RBRAC stabs.

4.2 Global Variables ¶

A variable whose scope is not specific to just one source file is represented by the ‘G’ symbol descriptor. These stabs use the N_GSYM stab type (C_GSYM for XCOFF). The type information for the stab (see The String Field) gives the type of the variable.

For example, the following source code:

char g_foo = 'c';

yields the following assembly code:

.stabs "g_foo:G2",32,0,0,0     # 32 is N_GSYM
     .global _g_foo
     .data
_g_foo:
     .byte 99

The address of the variable represented by the N_GSYM is not contained in the N_GSYM stab. The debugger gets this information from the external symbol for the global variable. In the example above, the .global _g_foo and _g_foo: lines tell the assembler to produce an external symbol.

Some compilers, like GCC, output N_GSYM stabs only once, where the variable is defined. Other compilers, like SunOS4 /bin/cc, output a N_GSYM stab for each compilation unit which references the variable.

4.3 Register Variables ¶

Register variables have their own stab type, N_RSYM (C_RSYM for XCOFF), and their own symbol descriptor, ‘r’. The stab’s value is the number of the register where the variable data will be stored.

AIX defines a separate symbol descriptor ‘d’ for floating point registers. This seems unnecessary; why not just just give floating point registers different register numbers? I have not verified whether the compiler actually uses ‘d’.

If the register is explicitly allocated to a global variable, but not initialized, as in:

register int g_bar asm ("%g5");

then the stab may be emitted at the end of the object file, with the other bss symbols.

4.4 Common Blocks ¶

A common block is a statically allocated section of memory which can be referred to by several source files. It may contain several variables. I believe Fortran is the only language with this feature.

A N_BCOMM stab begins a common block and an N_ECOMM stab ends it. The only field that is significant in these two stabs is the string, which names a normal (non-debugging) symbol that gives the address of the common block. According to IBM documentation, only the N_BCOMM has the name of the common block (even though their compiler actually puts it both places).

The stabs for the members of the common block are between the N_BCOMM and the N_ECOMM; the value of each stab is the offset within the common block of that variable. IBM uses the C_ECOML stab type, and there is a corresponding N_ECOML stab type, but Sun’s Fortran compiler uses N_GSYM instead. The variables within a common block use the ‘V’ symbol descriptor (I believe this is true of all Fortran variables). Other stabs (at least type declarations using C_DECL) can also be between the N_BCOMM and the N_ECOMM.

4.5 Static Variables ¶

Initialized static variables are represented by the ‘S’ and ‘V’ symbol descriptors. ‘S’ means file scope static, and ‘V’ means procedure scope static. One exception: in XCOFF, IBM’s xlc compiler always uses ‘V’, and whether it is file scope or not is distinguished by whether the stab is located within a function.

In a.out files, N_STSYM means the data section, N_FUN means the text section, and N_LCSYM means the bss section. For those systems with a read-only data section separate from the text section (Solaris), N_ROSYM means the read-only data section.

For example, the source lines:

static const int var_const = 5;
static int var_init = 2;
static int var_noinit;

yield the following stabs:

.stabs "var_const:S1",36,0,0,_var_const      # 36 is N_FUN
...
.stabs "var_init:S1",38,0,0,_var_init        # 38 is N_STSYM
...
.stabs "var_noinit:S1",40,0,0,_var_noinit    # 40 is N_LCSYM

In XCOFF files, the stab type need not indicate the section; C_STSYM can be used for all statics. Also, each static variable is enclosed in a static block. A C_BSTAT (emitted with a ‘.bs’ assembler directive) symbol begins the static block; its value is the symbol number of the csect symbol whose value is the address of the static block, its section is the section of the variables in that static block, and its name is ‘.bs’. A C_ESTAT (emitted with a ‘.es’ assembler directive) symbol ends the static block; its name is ‘.es’ and its value and section are ignored.

In ECOFF files, the storage class is used to specify the section, so the stab type need not indicate the section.

In ELF files, for the SunPRO compiler version 2.0.1, symbol descriptor ‘S’ means that the address is absolute (the linker relocates it) and symbol descriptor ‘V’ means that the address is relative to the start of the relevant section for that compilation unit. SunPRO has plans to have the linker stop relocating stabs; I suspect that their the debugger gets the address from the corresponding ELF (not stab) symbol. I’m not sure how to find which symbol of that name is the right one. The clean way to do all this would be to have the value of a symbol descriptor ‘S’ symbol be an offset relative to the start of the file, just like everything else, but that introduces obvious compatibility problems. For more information on linker stab relocation, See Having the Linker Relocate Stabs in ELF.

4.6 Fortran Based Variables ¶

Fortran (at least, the Sun and SGI dialects of FORTRAN-77) has a feature which allows allocating arrays with malloc, but which avoids blurring the line between arrays and pointers the way that C does. In stabs such a variable uses the ‘b’ symbol descriptor.

For example, the Fortran declarations

real foo, foo10(10), foo10_5(10,5)
pointer (foop, foo)
pointer (foo10p, foo10)
pointer (foo105p, foo10_5)

produce the stabs

foo:b6
foo10:bar3;1;10;6
foo10_5:bar3;1;5;ar3;1;10;6

In this example, real is type 6 and type 3 is an integral type which is the type of the subscripts of the array (probably integer).

The ‘b’ symbol descriptor is like ‘V’ in that it denotes a statically allocated symbol whose scope is local to a function; see See Static Variables. The value of the symbol, instead of being the address of the variable itself, is the address of a pointer to that variable. So in the above example, the value of the foo stab is the address of a pointer to a real, the value of the foo10 stab is the address of a pointer to a 10-element array of reals, and the value of the foo10_5 stab is the address of a pointer to a 5-element array of 10-element arrays of reals.

4.7 Parameters ¶

Formal parameters to a function are represented by a stab (or sometimes two; see below) for each parameter. The stabs are in the order in which the debugger should print the parameters (i.e., the order in which the parameters are declared in the source file). The exact form of the stab depends on how the parameter is being passed.

Parameters passed on the stack use the symbol descriptor ‘p’ and the N_PSYM symbol type (or C_PSYM for XCOFF). The value of the symbol is an offset used to locate the parameter on the stack; its exact meaning is machine-dependent, but on most machines it is an offset from the frame pointer.

As a simple example, the code:

main (argc, argv)
     int argc;
     char **argv;

produces the stabs:

.stabs "main:F1",36,0,0,_main                 # 36 is N_FUN
.stabs "argc:p1",160,0,0,68                   # 160 is N_PSYM
.stabs "argv:p20=*21=*2",160,0,0,72

The type definition of argv is interesting because it contains several type definitions. Type 21 is pointer to type 2 (char) and argv (type 20) is pointer to type 21.

The following symbol descriptors are also said to go with N_PSYM. The value of the symbol is said to be an offset from the argument pointer (I’m not sure whether this is true or not).

pP (<<??>>)
pF Fortran function parameter
X  (function result variable)

• Passing Parameters in Registers
• Storing Parameters as Local Variables
• Passing Parameters by Reference
• Passing Conformant Array Parameters

.stabs "arg:p1" . . .       ; N_PSYM
.stabs "arg:r1" . . .       ; N_RSYM

void
subr (f)
     float f;
{

.stabs "f:p13",160,0,3,8   # 160 is N_PSYM, here 13 is double
.stabs "f:r12",64,0,3,0    # 64 is N_RSYM, here 12 is float

extern int fprintf (FILE *stream, char *format, ...);
...
fprintf (stdout, "%d\n", x);

5.1 Builtin Types ¶

Certain types are built in (int, short, void, float, etc.); the debugger recognizes these types and knows how to handle them. Thus, don’t be surprised if some of the following ways of specifying builtin types do not specify everything that a debugger would need to know about the type—in some cases they merely specify enough information to distinguish the type from other types.

The traditional way to define builtin types is convoluted, so new ways have been invented to describe them. Sun’s acc uses special builtin type descriptors (‘b’ and ‘R’), and IBM uses negative type numbers. GDB accepts all three ways, as of version 4.8; dbx just accepts the traditional builtin types and perhaps one of the other two formats. The following sections describe each of these formats.

• Traditional Builtin Types
• Defining Builtin Types Using Builtin Type Descriptors
• Negative Type Numbers

.stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0    # 128 is N_LSYM
.stabs "char:t2=r2;0;127;",128,0,0,0

.stabs "unsigned short:t6=r1;0;65535;",128,0,0,0

.stabs "unsigned int:t4=r1;0;-1;",128,0,0,0

long int:t3=r1;001000000000000000000000;000777777777777777777777;
long unsigned int:t5=r1;000000000000000000000000;001777777777777777777777;

.stabs "float:t12=r1;4;0;",128,0,0,0
.stabs "double:t13=r1;8;0;",128,0,0,0

.stabs "long double:t14=r1;8;0;",128,0,0,0

.stabs "void:t15=15",128,0,0,0

.stabs "void:t19=bs0;0;0",128,0,0,0

.stabs "boolean:t10=@s8;-16",128,0,0,0

.stabs "boolean:t10=@s64;-16",128,0,0,0

.stabs "CARDINAL:t10=-8",128,0,0,0

5.2 Miscellaneous Types ¶

b type-information ; bytes

Pascal space type. This is documented by IBM; what does it mean?

This use of the ‘b’ type descriptor can be distinguished from its use for builtin integral types (see Defining Builtin Types Using Builtin Type Descriptors) because the character following the type descriptor is always a digit, ‘(’, or ‘-’.

B type-information

A volatile-qualified version of type-information. This is a Sun extension. References and stores to a variable with a volatile-qualified type must not be optimized or cached; they must occur as the user specifies them.

d type-information

File of type type-information. As far as I know this is only used by Pascal.

k type-information

A const-qualified version of type-information. This is a Sun extension. A variable with a const-qualified type cannot be modified.

M type-information ; length

Multiple instance type. The type seems to composed of length repetitions of type-information, for example character*3 is represented by ‘M-2;3’, where ‘-2’ is a reference to a character type (see Negative Type Numbers). I’m not sure how this differs from an array. This appears to be a Fortran feature. length is a bound, like those in range types; see Subrange Types.

S type-information

Pascal set type. type-information must be a small type such as an enumeration or a subrange, and the type is a bitmask whose length is specified by the number of elements in type-information.

In CHILL, if it is a bitstring instead of a set, also use the ‘S’ type attribute (see The String Field).

* type-information

Pointer to type-information.

5.3 Cross-References to Other Types ¶

A type can be used before it is defined; one common way to deal with that situation is just to use a type reference to a type which has not yet been defined.

Another way is with the ‘x’ type descriptor, which is followed by ‘s’ for a structure tag, ‘u’ for a union tag, or ‘e’ for a enumerator tag, followed by the name of the tag, followed by ‘:’. If the name contains ‘::’ between a ‘<’ and ‘>’ pair (for C++ templates), such a ‘::’ does not end the name—only a single ‘:’ ends the name; see Defining a Symbol Within Another Type.

For example, the following C declarations:

struct foo;
struct foo *bar;

produce:

.stabs "bar:G16=*17=xsfoo:",32,0,0,0

Not all debuggers support the ‘x’ type descriptor, so on some machines GCC does not use it. I believe that for the above example it would just emit a reference to type 17 and never define it, but I haven’t verified that.

Modula-2 imported types, at least on AIX, use the ‘i’ type descriptor, which is followed by the name of the module from which the type is imported, followed by ‘:’, followed by the name of the type. There is then optionally a comma followed by type information for the type. This differs from merely naming the type (see Giving a Type a Name) in that it identifies the module; I don’t understand whether the name of the type given here is always just the same as the name we are giving it, or whether this type descriptor is used with a nameless stab (see The String Field), or what. The symbol ends with ‘;’.

5.4 Subrange Types ¶

The ‘r’ type descriptor defines a type as a subrange of another type. It is followed by type information for the type of which it is a subrange, a semicolon, an integral lower bound, a semicolon, an integral upper bound, and a semicolon. The AIX documentation does not specify the trailing semicolon, in an effort to specify array indexes more cleanly, but a subrange which is not an array index has always included a trailing semicolon (see Array Types).

Instead of an integer, either bound can be one of the following:

A offset

The bound is passed by reference on the stack at offset offset from the argument list. See Parameters, for more information on such offsets.

T offset

The bound is passed by value on the stack at offset offset from the argument list.

a register-number

The bound is passed by reference in register number register-number.

t register-number

The bound is passed by value in register number register-number.

J

There is no bound.

Subranges are also used for builtin types; see Traditional Builtin Types.

5.5 Array Types ¶

Arrays use the ‘a’ type descriptor. Following the type descriptor is the type of the index and the type of the array elements. If the index type is a range type, it ends in a semicolon; otherwise (for example, if it is a type reference), there does not appear to be any way to tell where the types are separated. In an effort to clean up this mess, IBM documents the two types as being separated by a semicolon, and a range type as not ending in a semicolon (but this is not right for range types which are not array indexes, see Subrange Types). I think probably the best solution is to specify that a semicolon ends a range type, and that the index type and element type of an array are separated by a semicolon, but that if the index type is a range type, the extra semicolon can be omitted. GDB (at least through version 4.9) doesn’t support any kind of index type other than a range anyway; I’m not sure about dbx.

It is well established, and widely used, that the type of the index, unlike most types found in the stabs, is merely a type definition, not type information (see The String Field) (that is, it need not start with ‘type-number=’ if it is defining a new type). According to a comment in GDB, this is also true of the type of the array elements; it gives ‘ar1;1;10;ar1;1;10;4’ as a legitimate way to express a two dimensional array. According to AIX documentation, the element type must be type information. GDB accepts either.

The type of the index is often a range type, expressed as the type descriptor ‘r’ and some parameters. It defines the size of the array. In the example below, the range ‘r1;0;2;’ defines an index type which is a subrange of type 1 (integer), with a lower bound of 0 and an upper bound of 2. This defines the valid range of subscripts of a three-element C array.

For example, the definition:

char char_vec[3] = {'a','b','c'};

produces the output:

.stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
     .global _char_vec
     .align 4
_char_vec:
     .byte 97
     .byte 98
     .byte 99

If an array is packed, the elements are spaced more closely than normal, saving memory at the expense of speed. For example, an array of 3-byte objects might, if unpacked, have each element aligned on a 4-byte boundary, but if packed, have no padding. One way to specify that something is packed is with type attributes (see The String Field). In the case of arrays, another is to use the ‘P’ type descriptor instead of ‘a’. Other than specifying a packed array, ‘P’ is identical to ‘a’.

An open array is represented by the ‘A’ type descriptor followed by type information specifying the type of the array elements.

An N-dimensional dynamic array is represented by

D dimensions ; type-information

dimensions is the number of dimensions; type-information specifies the type of the array elements.

A subarray of an N-dimensional array is represented by

E dimensions ; type-information

dimensions is the number of dimensions; type-information specifies the type of the array elements.

5.6 Strings ¶

Some languages, like C or the original Pascal, do not have string types, they just have related things like arrays of characters. But most Pascals and various other languages have string types, which are indicated as follows:

n type-information ; bytes

bytes is the maximum length. I’m not sure what type-information is; I suspect that it means that this is a string of type-information (thus allowing a string of integers, a string of wide characters, etc., as well as a string of characters). Not sure what the format of this type is. This is an AIX feature.

z type-information ; bytes

Just like ‘n’ except that this is a gstring, not an ordinary string. I don’t know the difference.

N

Pascal Stringptr. What is this? This is an AIX feature.

Languages, such as CHILL which have a string type which is basically just an array of characters use the ‘S’ type attribute (see The String Field).

5.7 Enumerations ¶

Enumerations are defined with the ‘e’ type descriptor.

The source line below declares an enumeration type at file scope. The type definition is located after the N_RBRAC that marks the end of the previous procedure’s block scope, and before the N_FUN that marks the beginning of the next procedure’s block scope. Therefore it does not describe a block local symbol, but a file local one.

The source line:

enum e_places {first,second=3,last};

generates the following stab:

.stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0

The symbol descriptor (‘T’) says that the stab describes a structure, enumeration, or union tag. The type descriptor ‘e’, following the ‘22=’ of the type definition narrows it down to an enumeration type. Following the ‘e’ is a list of the elements of the enumeration. The format is ‘name:value,’. The list of elements ends with ‘;’. The fact that value is specified as an integer can cause problems if the value is large. GCC 2.5.2 tries to output it in octal in that case with a leading zero, which is probably a good thing, although GDB 4.11 supports octal only in cases where decimal is perfectly good. Negative decimal values are supported by both GDB and dbx.

There is no standard way to specify the size of an enumeration type; it is determined by the architecture (normally all enumerations types are 32 bits). Type attributes can be used to specify an enumeration type of another size for debuggers which support them; see The String Field.

Enumeration types are unusual in that they define symbols for the enumeration values (first, second, and third in the above example), and even though these symbols are visible in the file as a whole (rather than being in a more local namespace like structure member names), they are defined in the type definition for the enumeration type rather than each having their own symbol. In order to be fast, GDB will only get symbols from such types (in its initial scan of the stabs) if the type is the first thing defined after a ‘T’ or ‘t’ symbol descriptor (the above example fulfills this requirement). If the type does not have a name, the compiler should emit it in a nameless stab (see The String Field); GCC does this.

5.8 Structures ¶

The encoding of structures in stabs can be shown with an example.

The following source code declares a structure tag and defines an instance of the structure in global scope. Then a typedef equates the structure tag with a new type. Separate stabs are generated for the structure tag, the structure typedef, and the structure instance. The stabs for the tag and the typedef are emitted when the definitions are encountered. Since the structure elements are not initialized, the stab and code for the structure variable itself is located at the end of the program in the bss section.

struct s_tag {
  int   s_int;
  float s_float;
  char  s_char_vec[8];
  struct s_tag* s_next;
} g_an_s;

typedef struct s_tag s_typedef;

The structure tag has an N_LSYM stab type because, like the enumeration, the symbol has file scope. Like the enumeration, the symbol descriptor is ‘T’, for enumeration, structure, or tag type. The type descriptor ‘s’ following the ‘16=’ of the type definition narrows the symbol type to structure.

Following the ‘s’ type descriptor is the number of bytes the structure occupies, followed by a description of each structure element. The structure element descriptions are of the form ‘name:type, bit offset from the start of the struct, number of bits in the element’.

# 128 is N_LSYM
.stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;
        s_char_vec:17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0

In this example, the first two structure elements are previously defined types. For these, the type following the ‘name:’ part of the element description is a simple type reference. The other two structure elements are new types. In this case there is a type definition embedded after the ‘name:’. The type definition for the array element looks just like a type definition for a stand-alone array. The s_next field is a pointer to the same kind of structure that the field is an element of. So the definition of structure type 16 contains a type definition for an element which is a pointer to type 16.

If a field is a static member (this is a C++ feature in which a single variable appears to be a field of every structure of a given type) it still starts out with the field name, a colon, and the type, but then instead of a comma, bit position, comma, and bit size, there is a colon followed by the name of the variable which each such field refers to.

If the structure has methods (a C++ feature), they follow the non-method fields; see GNU C++ Stabs.

5.9 Giving a Type a Name ¶

To give a type a name, use the ‘t’ symbol descriptor. The type is specified by the type information (see The String Field) for the stab. For example,

.stabs "s_typedef:t16",128,0,0,0     # 128 is N_LSYM

specifies that s_typedef refers to type number 16. Such stabs have symbol type N_LSYM (or C_DECL for XCOFF). (The Sun documentation mentions using N_GSYM in some cases).

If you are specifying the tag name for a structure, union, or enumeration, use the ‘T’ symbol descriptor instead. I believe C is the only language with this feature.

If the type is an opaque type (I believe this is a Modula-2 feature), AIX provides a type descriptor to specify it. The type descriptor is ‘o’ and is followed by a name. I don’t know what the name means—is it always the same as the name of the type, or is this type descriptor used with a nameless stab (see The String Field)? There optionally follows a comma followed by type information which defines the type of this type. If omitted, a semicolon is used in place of the comma and the type information, and the type is much like a generic pointer type—it has a known size but little else about it is specified.

5.10 Unions ¶

union u_tag {
  int  u_int;
  float u_float;
  char* u_char;
} an_u;

This code generates a stab for a union tag and a stab for a union variable. Both use the N_LSYM stab type. If a union variable is scoped locally to the procedure in which it is defined, its stab is located immediately preceding the N_LBRAC for the procedure’s block start.

The stab for the union tag, however, is located preceding the code for the procedure in which it is defined. The stab type is N_LSYM. This would seem to imply that the union type is file scope, like the struct type s_tag. This is not true. The contents and position of the stab for u_type do not convey any information about its procedure local scope.

# 128 is N_LSYM
.stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
       128,0,0,0

The symbol descriptor ‘T’, following the ‘name:’ means that the stab describes an enumeration, structure, or union tag. The type descriptor ‘u’, following the ‘23=’ of the type definition, narrows it down to a union type definition. Following the ‘u’ is the number of bytes in the union. After that is a list of union element descriptions. Their format is ‘name:type, bit offset into the union, number of bytes for the element;’.

The stab for the union variable is:

.stabs "an_u:23",128,0,0,-20     # 128 is N_LSYM

‘-20’ specifies where the variable is stored (see Automatic Variables Allocated on the Stack).

5.11 Function Types ¶

Various types can be defined for function variables. These types are not used in defining functions (see Procedures); they are used for things like pointers to functions.

The simple, traditional, type is type descriptor ‘f’ is followed by type information for the return type of the function, followed by a semicolon.

This does not deal with functions for which the number and types of the parameters are part of the type, as in Modula-2 or ANSI C. AIX provides extensions to specify these, using the ‘f’, ‘F’, ‘p’, and ‘R’ type descriptors.

First comes the type descriptor. If it is ‘f’ or ‘F’, this type involves a function rather than a procedure, and the type information for the return type of the function follows, followed by a comma. Then comes the number of parameters to the function and a semicolon. Then, for each parameter, there is the name of the parameter followed by a colon (this is only present for type descriptors ‘R’ and ‘F’ which represent Pascal function or procedure parameters), type information for the parameter, a comma, 0 if passed by reference or 1 if passed by value, and a semicolon. The type definition ends with a semicolon.

For example, this variable definition:

int (*g_pf)();

generates the following code:

.stabs "g_pf:G24=*25=f1",32,0,0,0
    .common _g_pf,4,"bss"

The variable defines a new type, 24, which is a pointer to another new type, 25, which is a function returning int.

7.1 Symbol Table Format ¶

Each time the assembler encounters a stab directive, it puts each field of the stab into a corresponding field in a symbol table entry of its output file. If the stab contains a string field, the symbol table entry for that stab points to a string table entry containing the string data from the stab. Assembler labels become relocatable addresses. Symbol table entries in a.out have the format:

struct internal_nlist {
  unsigned long n_strx;         /* index into string table of name */
  unsigned char n_type;         /* type of symbol */
  unsigned char n_other;        /* misc info (usually empty) */
  unsigned short n_desc;        /* description field */
  bfd_vma n_value;              /* value of symbol */
};

If the stab has a string, the n_strx field holds the offset in bytes of the string within the string table. The string is terminated by a NUL character. If the stab lacks a string (for example, it was produced by a .stabn or .stabd directive), the n_strx field is zero.

Symbol table entries with n_type field values greater than 0x1f originated as stabs generated by the compiler (with one random exception). The other entries were placed in the symbol table of the executable by the assembler or the linker.

7.2 Transformations on Symbol Tables ¶

The linker concatenates object files and does fixups of externally defined symbols.

You can see the transformations made on stab data by the assembler and linker by examining the symbol table after each pass of the build. To do this, use ‘nm -ap’, which dumps the symbol table, including debugging information, unsorted. For stab entries the columns are: value, other, desc, type, string. For assembler and linker symbols, the columns are: value, type, string.

The low 5 bits of the stab type tell the linker how to relocate the value of the stab. Thus for stab types like N_RSYM and N_LSYM, where the value is an offset or a register number, the low 5 bits are N_ABS, which tells the linker not to relocate the value.

Where the value of a stab contains an assembly language label, it is transformed by each build step. The assembler turns it into a relocatable address and the linker turns it into an absolute address.

• Transformations on Static Variables
• Transformations on Global Variables
• Transformations of Stabs in separate sections

static int s_g_repeat

.stabs "s_g_repeat:S1",38,0,0,_s_g_repeat

00000084 - 00 0000 STSYM s_g_repeat:S1

0000e00c - 00 0000 STSYM s_g_repeat:S1

char g_foo = 'c';

.stabs "g_foo:G2",32,0,0,0

00000000 - 00 0000  GSYM g_foo:G2
00000080 D _g_foo

00000000 - 00 0000  GSYM g_foo:G2
...
0000e008 D _g_foo

static int ld = 5;

.Ddata.data:
...
        .stabs "ld:V(0,3)",0x26,0,4,.L18-Ddata.data    # 0x26 is N_STSYM
...
.L18:
        .align 4
        .word 0x5

Symnum n_type n_othr n_desc n_value  n_strx String
31     STSYM  0      4      00000004 680    ld:V(0,3)

Symnum n_type n_othr n_desc n_value  n_strx String
31     STSYM  0      4      00000004 680    ld:V(0,3)

8.1 C++ Class Names ¶

In C++, a class name which is declared with class, struct, or union, is not only a tag, as in C, but also a type name. Thus there should be stabs with both ‘t’ and ‘T’ symbol descriptors (see Giving a Type a Name).

To save space, there is a special abbreviation for this case. If the ‘T’ symbol descriptor is followed by ‘t’, then the stab defines both a type name and a tag.

For example, the C++ code

struct foo {int x;};

can be represented as either

.stabs "foo:T19=s4x:1,0,32;;",128,0,0,0       # 128 is N_LSYM
.stabs "foo:t19",128,0,0,0

or

.stabs "foo:Tt19=s4x:1,0,32;;",128,0,0,0

8.2 Defining a Symbol Within Another Type ¶

In C++, a symbol (such as a type name) can be defined within another type.

In stabs, this is sometimes represented by making the name of a symbol which contains ‘::’. Such a pair of colons does not end the name of the symbol, the way a single colon would (see The String Field). I’m not sure how consistently used or well thought out this mechanism is. So that a pair of colons in this position always has this meaning, ‘:’ cannot be used as a symbol descriptor.

For example, if the string for a stab is ‘foo::bar::baz:t5=*6’, then foo::bar::baz is the name of the symbol, ‘t’ is the symbol descriptor, and ‘5=*6’ is the type information.

8.3 Basic Types For C++ ¶

<< the examples that follow are based on a01.C >>

C++ adds two more builtin types to the set defined for C. These are the unknown type and the vtable record type. The unknown type, type 16, is defined in terms of itself like the void type.

The vtable record type, type 17, is defined as a structure type and then as a structure tag. The structure has four fields: delta, index, pfn, and delta2. pfn is the function pointer.

<< In boilerplate $vtbl_ptr_type, what are the fields delta, index, and delta2 used for? >>

This basic type is present in all C++ programs even if there are no virtual methods defined.

.stabs "struct_name:sym_desc(type)type_def(17)=type_desc(struct)struct_bytes(8)
        elem_name(delta):type_ref(short int),bit_offset(0),field_bits(16);
        elem_name(index):type_ref(short int),bit_offset(16),field_bits(16);
        elem_name(pfn):type_def(18)=type_desc(ptr to)type_ref(void),
                                    bit_offset(32),field_bits(32);
        elem_name(delta2):type_def(short int);bit_offset(32),field_bits(16);;"
        N_LSYM, NIL, NIL

.stabs "$vtbl_ptr_type:t17=s8
        delta:6,0,16;index:6,16,16;pfn:18=*15,32,32;delta2:6,32,16;;"
        ,128,0,0,0

.stabs "name:sym_dec(struct tag)type_ref($vtbl_ptr_type)",N_LSYM,NIL,NIL,NIL

.stabs "$vtbl_ptr_type:T17",128,0,0,0

8.4 Simple Class Definition ¶

The stabs describing C++ language features are an extension of the stabs describing C. Stabs representing C++ class types elaborate extensively on the stab format used to describe structure types in C. Stabs representing class type variables look just like stabs representing C language variables.

Consider the following very simple class definition.

class baseA {
public:
        int Adat;
        int Ameth(int in, char other);
};

The class baseA is represented by two stabs. The first stab describes the class as a structure type. The second stab describes a structure tag of the class type. Both stabs are of stab type N_LSYM. Since the stab is not located between an N_FUN and an N_LBRAC stab this indicates that the class is defined at file scope. If it were, then the N_LSYM would signify a local variable.

A stab describing a C++ class type is similar in format to a stab describing a C struct, with each class member shown as a field in the structure. The part of the struct format describing fields is expanded to include extra information relevant to C++ class members. In addition, if the class has multiple base classes or virtual functions the struct format outside of the field parts is also augmented.

In this simple example the field part of the C++ class stab representing member data looks just like the field part of a C struct stab. The section on protections describes how its format is sometimes extended for member data.

The field part of a C++ class stab representing a member function differs substantially from the field part of a C struct stab. It still begins with ‘name:’ but then goes on to define a new type number for the member function, describe its return type, its argument types, its protection level, any qualifiers applied to the method definition, and whether the method is virtual or not. If the method is virtual then the method description goes on to give the vtable index of the method, and the type number of the first base class defining the method.

When the field name is a method name it is followed by two colons rather than one. This is followed by a new type definition for the method. This is a number followed by an equal sign and the type of the method. Normally this will be a type declared using the ‘#’ type descriptor; see The ‘#’ Type Descriptor; static member functions are declared using the ‘f’ type descriptor instead; see Function Types.

The format of an overloaded operator method name differs from that of other methods. It is ‘op$::operator-name.’ where operator-name is the operator name such as ‘+’ or ‘+=’. The name ends with a period, and any characters except the period can occur in the operator-name string.

The next part of the method description represents the arguments to the method, preceded by a colon and ending with a semi-colon. The types of the arguments are expressed in the same way argument types are expressed in C++ name mangling. In this example an int and a char map to ‘ic’.

This is followed by a number, a letter, and an asterisk or period, followed by another semicolon. The number indicates the protections that apply to the member function. Here the 2 means public. The letter encodes any qualifier applied to the method definition. In this case, ‘A’ means that it is a normal function definition. The dot shows that the method is not virtual. The sections that follow elaborate further on these fields and describe the additional information present for virtual methods.

.stabs "class_name:sym_desc(type)type_def(20)=type_desc(struct)struct_bytes(4)
        field_name(Adat):type(int),bit_offset(0),field_bits(32);

        method_name(Ameth)::type_def(21)=type_desc(method)return_type(int);
        :arg_types(int char);
        protection(public)qualifier(normal)virtual(no);;"
        N_LSYM,NIL,NIL,NIL

.stabs "baseA:t20=s4Adat:1,0,32;Ameth::21=##1;:ic;2A.;;",128,0,0,0

.stabs "class_name:sym_desc(struct tag)",N_LSYM,NIL,NIL,NIL

.stabs "baseA:T20",128,0,0,0

8.5 Class Instance ¶

As shown above, describing even a simple C++ class definition is accomplished by massively extending the stab format used in C to describe structure types. However, once the class is defined, C stabs with no modifications can be used to describe class instances. The following source:

main () {
        baseA AbaseA;
}

yields the following stab describing the class instance. It looks no different from a standard C stab describing a local variable.

.stabs "name:type_ref(baseA)", N_LSYM, NIL, NIL, frame_ptr_offset

.stabs "AbaseA:20",128,0,0,-20

8.6 Method Definition ¶

The class definition shown above declares Ameth. The C++ source below defines Ameth:

int
baseA::Ameth(int in, char other)
{
        return in;
};

This method definition yields three stabs following the code of the method. One stab describes the method itself and following two describe its parameters. Although there is only one formal argument all methods have an implicit argument which is the this pointer. The this pointer is a pointer to the object on which the method was called. Note that the method name is mangled to encode the class name and argument types. Name mangling is described in the ARM (The Annotated C++ Reference Manual, by Ellis and Stroustrup, ISBN 0-201-51459-1); gpcompare.texi in Cygnus GCC distributions describes the differences between GNU mangling and ARM mangling.

.stabs "name:symbol_descriptor(global function)return_type(int)",
        N_FUN, NIL, NIL, code_addr_of_method_start

.stabs "Ameth__5baseAic:F1",36,0,0,_Ameth__5baseAic

Here is the stab for the this pointer implicit argument. The name of the this pointer is always this. Type 19, the this pointer is defined as a pointer to type 20, baseA, but a stab defining baseA has not yet been emitted. Since the compiler knows it will be emitted shortly, here it just outputs a cross reference to the undefined symbol, by prefixing the symbol name with ‘xs’.

.stabs "name:sym_desc(register param)type_def(19)=
        type_desc(ptr to)type_ref(baseA)=
        type_desc(cross-reference to)baseA:",N_RSYM,NIL,NIL,register_number

.stabs "this:P19=*20=xsbaseA:",64,0,0,8

The stab for the explicit integer argument looks just like a parameter to a C function. The last field of the stab is the offset from the argument pointer, which in most systems is the same as the frame pointer.

.stabs "name:sym_desc(value parameter)type_ref(int)",
        N_PSYM,NIL,NIL,offset_from_arg_ptr

.stabs "in:p1",160,0,0,72

<< The examples that follow are based on A1.C >>

8.7 The ‘#’ Type Descriptor ¶

This is used to describe a class method. This is a function which takes an extra argument as its first argument, for the this pointer.

If the ‘#’ is immediately followed by another ‘#’, the second one will be followed by the return type and a semicolon. The class and argument types are not specified, and must be determined by demangling the name of the method if it is available.

Otherwise, the single ‘#’ is followed by the class type, a comma, the return type, a comma, and zero or more parameter types separated by commas. The list of arguments is terminated by a semicolon. In the debugging output generated by gcc, a final argument type of void indicates a method which does not take a variable number of arguments. If the final argument type of void does not appear, the method was declared with an ellipsis.

Note that although such a type will normally be used to describe fields in structures, unions, or classes, for at least some versions of the compiler it can also be used in other contexts.

8.8 The ‘@’ Type Descriptor ¶

The ‘@’ type descriptor is used for a pointer-to-non-static-member-data type. It is followed by type information for the class (or union), a comma, and type information for the member data.

The following C++ source:

typedef int A::*int_in_a;

generates the following stab:

.stabs "int_in_a:t20=21=@19,1",128,0,0,0

Note that there is a conflict between this and type attributes (see The String Field); both use type descriptor ‘@’. Fortunately, the ‘@’ type descriptor used in this C++ sense always will be followed by a digit, ‘(’, or ‘-’, and type attributes never start with those things.

8.9 Protections ¶

In the simple class definition shown above all member data and functions were publicly accessible. The example that follows contrasts public, protected and privately accessible fields and shows how these protections are encoded in C++ stabs.

If the character following the ‘field-name:’ part of the string is ‘/’, then the next character is the visibility. ‘0’ means private, ‘1’ means protected, and ‘2’ means public. Debuggers should ignore visibility characters they do not recognize, and assume a reasonable default (such as public) (GDB 4.11 does not, but this should be fixed in the next GDB release). If no visibility is specified the field is public. The visibility ‘9’ means that the field has been optimized out and is public (there is no way to specify an optimized out field with a private or protected visibility). Visibility ‘9’ is not supported by GDB 4.11; this should be fixed in the next GDB release.

The following C++ source:

class vis {
private:
        int   priv;
protected:
        char  prot;
public:
        float pub;
};

generates the following stab:

# 128 is N_LSYM
.stabs "vis:T19=s12priv:/01,0,32;prot:/12,32,8;pub:12,64,32;;",128,0,0,0

‘vis:T19=s12’ indicates that type number 19 is a 12 byte structure named vis The priv field has public visibility (‘/0’), type int (‘1’), and offset and size ‘,0,32;’. The prot field has protected visibility (‘/1’), type char (‘2’) and offset and size ‘,32,8;’. The pub field has type float (‘12’), and offset and size ‘,64,32;’.

Protections for member functions are signified by one digit embedded in the field part of the stab describing the method. The digit is 0 if private, 1 if protected and 2 if public. Consider the C++ class definition below:

class all_methods {
private:
        int   priv_meth(int in){return in;};
protected:
        char  protMeth(char in){return in;};
public:
        float pubMeth(float in){return in;};
};

It generates the following stab. The digit in question is to the left of an ‘A’ in each case. Notice also that in this case two symbol descriptors apply to the class name struct tag and struct type.

.stabs "class_name:sym_desc(struct tag&type)type_def(21)=
        sym_desc(struct)struct_bytes(1)
        meth_name::type_def(22)=sym_desc(method)returning(int);
        :args(int);protection(private)modifier(normal)virtual(no);
        meth_name::type_def(23)=sym_desc(method)returning(char);
        :args(char);protection(protected)modifier(normal)virtual(no);
        meth_name::type_def(24)=sym_desc(method)returning(float);
        :args(float);protection(public)modifier(normal)virtual(no);;",
        N_LSYM,NIL,NIL,NIL

.stabs "all_methods:Tt21=s1priv_meth::22=##1;:i;0A.;protMeth::23=##2;:c;1A.;
        pubMeth::24=##12;:f;2A.;;",128,0,0,0

8.10 Method Modifiers (const, volatile, const volatile) ¶

<< based on a6.C >>

In the class example described above all the methods have the normal modifier. This method modifier information is located just after the protection information for the method. This field has four possible character values. Normal methods use ‘A’, const methods use ‘B’, volatile methods use ‘C’, and const volatile methods use ‘D’. Consider the class definition below:

class A {
public:
        int ConstMeth (int arg) const { return arg; };
        char VolatileMeth (char arg) volatile { return arg; };
        float ConstVolMeth (float arg) const volatile {return arg; };
};

This class is described by the following stab:

.stabs "class(A):sym_desc(struct)type_def(20)=type_desc(struct)struct_bytes(1)
        meth_name(ConstMeth)::type_def(21)sym_desc(method)
        returning(int);:arg(int);protection(public)modifier(const)virtual(no);
        meth_name(VolatileMeth)::type_def(22)=sym_desc(method)
        returning(char);:arg(char);protection(public)modifier(volatile)virt(no)
        meth_name(ConstVolMeth)::type_def(23)=sym_desc(method)
        returning(float);:arg(float);protection(public)modifier(const volatile)
        virtual(no);;", ...

.stabs "A:T20=s1ConstMeth::21=##1;:i;2B.;VolatileMeth::22=##2;:c;2C.;
             ConstVolMeth::23=##12;:f;2D.;;",128,0,0,0

8.11 Virtual Methods ¶

<< The following examples are based on a4.C >>

The presence of virtual methods in a class definition adds additional data to the class description. The extra data is appended to the description of the virtual method and to the end of the class description. Consider the class definition below:

class A {
public:
        int Adat;
        virtual int A_virt (int arg) { return arg; };
};

This results in the stab below describing class A. It defines a new type (20) which is an 8 byte structure. The first field of the class struct is ‘Adat’, an integer, starting at structure offset 0 and occupying 32 bits.

The second field in the class struct is not explicitly defined by the C++ class definition but is implied by the fact that the class contains a virtual method. This field is the vtable pointer. The name of the vtable pointer field starts with ‘$vf’ and continues with a type reference to the class it is part of. In this example the type reference for class A is 20 so the name of its vtable pointer field is ‘$vf20’, followed by the usual colon.

Next there is a type definition for the vtable pointer type (21). This is in turn defined as a pointer to another new type (22).

Type 22 is the vtable itself, which is defined as an array, indexed by a range of integers between 0 and 1, and whose elements are of type 17. Type 17 was the vtable record type defined by the boilerplate C++ type definitions, as shown earlier.

The bit offset of the vtable pointer field is 32. The number of bits in the field are not specified when the field is a vtable pointer.

Next is the method definition for the virtual member function A_virt. Its description starts out using the same format as the non-virtual member functions described above, except instead of a dot after the ‘A’ there is an asterisk, indicating that the function is virtual. Since is is virtual some addition information is appended to the end of the method description.

The first number represents the vtable index of the method. This is a 32 bit unsigned number with the high bit set, followed by a semi-colon.

The second number is a type reference to the first base class in the inheritance hierarchy defining the virtual member function. In this case the class stab describes a base class so the virtual function is not overriding any other definition of the method. Therefore the reference is to the type number of the class that the stab is describing (20).

This is followed by three semi-colons. One marks the end of the current sub-section, one marks the end of the method field, and the third marks the end of the struct definition.

For classes containing virtual functions the very last section of the string part of the stab holds a type reference to the first base class. This is preceded by ‘~%’ and followed by a final semi-colon.

.stabs "class_name(A):type_def(20)=sym_desc(struct)struct_bytes(8)
        field_name(Adat):type_ref(int),bit_offset(0),field_bits(32);
        field_name(A virt func ptr):type_def(21)=type_desc(ptr to)type_def(22)=
        sym_desc(array)index_type_ref(range of int from 0 to 1);
        elem_type_ref(vtbl elem type),
        bit_offset(32);
        meth_name(A_virt)::typedef(23)=sym_desc(method)returning(int);
        :arg_type(int),protection(public)normal(yes)virtual(yes)
        vtable_index(1);class_first_defining(A);;;~%first_base(A);",
        N_LSYM,NIL,NIL,NIL

.stabs "A:t20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
        A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0

8.12 Inheritance ¶

Stabs describing C++ derived classes include additional sections that describe the inheritance hierarchy of the class. A derived class stab also encodes the number of base classes. For each base class it tells if the base class is virtual or not, and if the inheritance is private or public. It also gives the offset into the object of the portion of the object corresponding to each base class.

This additional information is embedded in the class stab following the number of bytes in the struct. First the number of base classes appears bracketed by an exclamation point and a comma.

Then for each base type there repeats a series: a virtual character, a visibility character, a number, a comma, another number, and a semi-colon.

The virtual character is ‘1’ if the base class is virtual and ‘0’ if not. The visibility character is ‘2’ if the derivation is public, ‘1’ if it is protected, and ‘0’ if it is private. Debuggers should ignore virtual or visibility characters they do not recognize, and assume a reasonable default (such as public and non-virtual) (GDB 4.11 does not, but this should be fixed in the next GDB release).

The number following the virtual and visibility characters is the offset from the start of the object to the part of the object pertaining to the base class.

After the comma, the second number is a type_descriptor for the base type. Finally a semi-colon ends the series, which repeats for each base class.

The source below defines three base classes A, B, and C and the derived class D.

class A {
public:
        int Adat;
        virtual int A_virt (int arg) { return arg; };
};

class B {
public:
        int B_dat;
        virtual int B_virt (int arg) {return arg; };
};

class C {
public:
        int Cdat;
        virtual int C_virt (int arg) {return arg; };
};

class D : A, virtual B, public C {
public:
        int Ddat;
        virtual int A_virt (int arg ) { return arg+1; };
        virtual int B_virt (int arg)  { return arg+2; };
        virtual int C_virt (int arg)  { return arg+3; };
        virtual int D_virt (int arg)  { return arg; };
};

Class stabs similar to the ones described earlier are generated for each base class.

.stabs "A:T20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
        A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0

.stabs "B:Tt25=s8Bdat:1,0,32;$vf25:21,32;B_virt::26=##1;
        :i;2A*-2147483647;25;;;~%25;",128,0,0,0

.stabs "C:Tt28=s8Cdat:1,0,32;$vf28:21,32;C_virt::29=##1;
        :i;2A*-2147483647;28;;;~%28;",128,0,0,0

In the stab describing derived class D below, the information about the derivation of this class is encoded as follows.

.stabs "derived_class_name:symbol_descriptors(struct tag&type)=
        type_descriptor(struct)struct_bytes(32)!num_bases(3),
        base_virtual(no)inheritance_public(no)base_offset(0),
        base_class_type_ref(A);
        base_virtual(yes)inheritance_public(no)base_offset(NIL),
        base_class_type_ref(B);
        base_virtual(no)inheritance_public(yes)base_offset(64),
        base_class_type_ref(C); ...

.stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:
        1,160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt:
        :32:i;2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;
        28;;D_virt::32:i;2A*-2147483646;31;;;~%20;",128,0,0,0

8.13 Virtual Base Classes ¶

A derived class object consists of a concatenation in memory of the data areas defined by each base class, starting with the leftmost and ending with the rightmost in the list of base classes. The exception to this rule is for virtual inheritance. In the example above, class D inherits virtually from base class B. This means that an instance of a D object will not contain its own B part but merely a pointer to a B part, known as a virtual base pointer.

In a derived class stab, the base offset part of the derivation information, described above, shows how the base class parts are ordered. The base offset for a virtual base class is always given as 0. Notice that the base offset for B is given as 0 even though B is not the first base class. The first base class A starts at offset 0.

The field information part of the stab for class D describes the field which is the pointer to the virtual base class B. The vbase pointer name is ‘$vb’ followed by a type reference to the virtual base class. Since the type id for B in this example is 25, the vbase pointer name is ‘$vb25’.

.stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:1,
       160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt::32:i;
       2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;28;;D_virt:
       :32:i;2A*-2147483646;31;;;~%20;",128,0,0,0

Following the name and a semicolon is a type reference describing the type of the virtual base class pointer, in this case 24. Type 24 was defined earlier as the type of the B class this pointer. The this pointer for a class is a pointer to the class type.

.stabs "this:P24=*25=xsB:",64,0,0,8

Finally the field offset part of the vbase pointer field description shows that the vbase pointer is the first field in the D object, before any data fields defined by the class. The layout of a D class object is a follows, Adat at 0, the vtable pointer for A at 32, Cdat at 64, the vtable pointer for C at 96, the virtual base pointer for B at 128, and Ddat at 160.

8.14 Static Members ¶

The data area for a class is a concatenation of the space used by the data members of the class. If the class has virtual methods, a vtable pointer follows the class data. The field offset part of each field description in the class stab shows this ordering.

<< How is this reflected in stabs? See Cygnus bug #677 for some info. >>