2 @setfilename stabs.info
9 * Stabs:: The "stabs" debugging information format.
15 This document describes the stabs debugging symbol tables.
17 Copyright 1992, 1993 Free Software Foundation, Inc.
18 Contributed by Cygnus Support. Written by Julia Menapace.
20 Permission is granted to make and distribute verbatim copies of
21 this manual provided the copyright notice and this permission notice
22 are preserved on all copies.
25 Permission is granted to process this file through Tex and print the
26 results, provided the printed document carries copying permission
27 notice identical to this one except for the removal of this paragraph
28 (this paragraph not being relevant to the printed manual).
31 Permission is granted to copy or distribute modified versions of this
32 manual under the terms of the GPL (for which purpose this text may be
33 regarded as a program in the language TeX).
36 @setchapternewpage odd
39 @title The ``stabs'' debug format
40 @author Julia Menapace, Jim Kingdon, David MacKenzie
41 @author Cygnus Support
44 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
45 \xdef\manvers{\$Revision$} % For use in headers, footers too
47 \hfill Cygnus Support\par
49 \hfill \TeX{}info \texinfoversion\par
53 @vskip 0pt plus 1filll
54 Copyright @copyright{} 1992, 1993 Free Software Foundation, Inc.
55 Contributed by Cygnus Support.
57 Permission is granted to make and distribute verbatim copies of
58 this manual provided the copyright notice and this permission notice
59 are preserved on all copies.
65 @top The "stabs" representation of debugging information
67 This document describes the stabs debugging format.
70 * Overview:: Overview of stabs
71 * Program structure:: Encoding of the structure of the program
72 * Constants:: Constants
74 * Types:: Type definitions
75 * Symbol tables:: Symbol information in symbol tables
76 * Cplusplus:: Appendixes:
77 * Stab types:: Symbol types in a.out files
78 * Symbol descriptors:: Table of symbol descriptors
79 * Type descriptors:: Table of type descriptors
80 * Expanded reference:: Reference information by stab type
81 * Questions:: Questions and anomolies
82 * XCOFF differences:: Differences between GNU stabs in a.out
83 and GNU stabs in XCOFF
84 * Sun differences:: Differences between GNU stabs and Sun
86 * Stabs in ELF:: Stabs in an ELF file.
87 * Symbol Types Index:: Index of symbolic stab symbol type names.
93 @chapter Overview of stabs
95 @dfn{Stabs} refers to a format for information that describes a program
96 to a debugger. This format was apparently invented by
97 @c FIXME! <<name of inventor>> at
98 the University of California at Berkeley, for the @code{pdx} Pascal
99 debugger; the format has spread widely since then.
101 This document is one of the few published sources of documentation on
102 stabs. It is believed to be comprehensive for stabs used by C. The
103 lists of symbol descriptors (@pxref{Symbol descriptors}) and type
104 descriptors (@pxref{Type descriptors}) are believed to be completely
105 comprehensive. Stabs for COBOL-specific features and for variant
106 records (used by Pascal and Modula-2) are poorly documented here.
108 Other sources of information on stabs are @cite{Dbx and Dbxtool
109 Interfaces}, 2nd edition, by Sun, 1988, and @cite{AIX Version 3.2 Files
110 Reference}, Fourth Edition, September 1992, "dbx Stabstring Grammar" in
111 the a.out section, page 2-31. This document is believed to incorporate
112 the information from those two sources except where it explictly directs
113 you to them for more information.
116 * Flow:: Overview of debugging information flow
117 * Stabs format:: Overview of stab format
118 * String field:: The @code{.stabs} @var{string} field
119 * C example:: A simple example in C source
120 * Assembly code:: The simple example at the assembly level
124 @section Overview of debugging information flow
126 The GNU C compiler compiles C source in a @file{.c} file into assembly
127 language in a @file{.s} file, which the assembler translates into
128 a @file{.o} file, which the linker combines with other @file{.o} files and
129 libraries to produce an executable file.
131 With the @samp{-g} option, GCC puts in the @file{.s} file additional
132 debugging information, which is slightly transformed by the assembler
133 and linker, and carried through into the final executable. This
134 debugging information describes features of the source file like line
135 numbers, the types and scopes of variables, and function names,
136 parameters, and scopes.
138 For some object file formats, the debugging information is encapsulated
139 in assembler directives known collectively as @dfn{stab} (symbol table)
140 directives, which are interspersed with the generated code. Stabs are
141 the native format for debugging information in the a.out and XCOFF
142 object file formats. The GNU tools can also emit stabs in the COFF and
143 ECOFF object file formats.
145 The assembler adds the information from stabs to the symbol information
146 it places by default in the symbol table and the string table of the
147 @file{.o} file it is building. The linker consolidates the @file{.o}
148 files into one executable file, with one symbol table and one string
149 table. Debuggers use the symbol and string tables in the executable as
150 a source of debugging information about the program.
153 @section Overview of stab format
155 There are three overall formats for stab assembler directives,
156 differentiated by the first word of the stab. The name of the directive
157 describes which combination of four possible data fields follows. It is
158 either @code{.stabs} (string), @code{.stabn} (number), or @code{.stabd}
159 (dot). IBM's XCOFF assembler uses @code{.stabx} (and some other
160 directives such as @code{.file} and @code{.bi}) instead of
161 @code{.stabs}, @code{.stabn} or @code{.stabd}.
163 The overall format of each class of stab is:
166 .stabs "@var{string}",@var{type},0,@var{desc},@var{value}
167 .stabn @var{type},0,@var{desc},@var{value}
168 .stabd @var{type},0,@var{desc}
169 .stabx "@var{string}",@var{value},@var{type},@var{sdb-type}
172 @c what is the correct term for "current file location"? My AIX
173 @c assembler manual calls it "the value of the current location counter".
174 For @code{.stabn} and @code{.stabd}, there is no @var{string} (the
175 @code{n_strx} field is zero; see @ref{Symbol tables}). For
176 @code{.stabd}, the @var{value} field is implicit and has the value of
177 the current file location. For @code{.stabx}, the @var{sdb-type} field
178 is unused for stabs and can always be set to zero.
180 The number in the @var{type} field gives some basic information about
181 which type of stab this is (or whether it @emph{is} a stab, as opposed
182 to an ordinary symbol). Each valid type number defines a different stab
183 type; further, the stab type defines the exact interpretation of, and
184 possible values for, any remaining @var{string}, @var{desc}, or
185 @var{value} fields present in the stab. @xref{Stab types}, for a list
186 in numeric order of the valid @var{type} field values for stab directives.
189 @section The @code{.stabs} @var{string} field
191 For @code{.stabs} the @var{string} field holds the meat of the
192 debugging information. The generally unstructured nature of this field
193 is what makes stabs extensible. For some stab types the @var{string} field
194 contains only a name. For other stab types the contents can be a great
197 The overall format is of the @var{string} field is:
200 "@var{name}:@var{symbol-descriptor} @var{type-information}"
203 @var{name} is the name of the symbol represented by the stab.
204 @var{name} can be omitted, which means the stab represents an unnamed
205 object. For example, @samp{:t10=*2} defines type 10 as a pointer to
206 type 2, but does not give the type a name. Omitting the @var{name}
207 field is supported by AIX dbx and GDB after about version 4.8, but not
208 other debuggers. GCC sometimes uses a single space as the name instead
209 of omitting the name altogether; apparently that is supported by most
212 The @var{symbol-descriptor} following the @samp{:} is an alphabetic
213 character that tells more specifically what kind of symbol the stab
214 represents. If the @var{symbol-descriptor} is omitted, but type
215 information follows, then the stab represents a local variable. For a
216 list of symbol descriptors, see @ref{Symbol descriptors}. The @samp{c}
217 symbol descriptor is an exception in that it is not followed by type
218 information. @xref{Constants}.
220 @var{type-information} is either a @var{type-number}, or
221 @samp{@var{type-number}=}. A @var{type-number} alone is a type
222 reference, referring directly to a type that has already been defined.
224 The @samp{@var{type-number}=} form is a type definition, where the
225 number represents a new type which is about to be defined. The type
226 definition may refer to other types by number, and those type numbers
227 may be followed by @samp{=} and nested definitions.
229 In a type definition, if the character that follows the equals sign is
230 non-numeric then it is a @var{type-descriptor}, and tells what kind of
231 type is about to be defined. Any other values following the
232 @var{type-descriptor} vary, depending on the @var{type-descriptor}.
233 @xref{Type descriptors}, for a list of @var{type-descriptor} values. If
234 a number follows the @samp{=} then the number is a @var{type-reference}.
235 For a full description of types, @ref{Types}.
237 There is an AIX extension for type attributes. Following the @samp{=}
238 are any number of type attributes. Each one starts with @samp{@@} and
239 ends with @samp{;}. Debuggers, including AIX's dbx and GDB 4.10, skip
240 any type attributes they do not recognize. GDB 4.9 and other versions
241 of dbx may not do this. Because of a conflict with C++
242 (@pxref{Cplusplus}), new attributes should not be defined which begin
243 with a digit, @samp{(}, or @samp{-}; GDB may be unable to distinguish
244 those from the C++ type descriptor @samp{@@}. The attributes are:
247 @item a@var{boundary}
248 @var{boundary} is an integer specifying the alignment. I assume it
249 applies to all variables of this type.
252 Size in bits of a variable of this type.
255 Pointer class (for checking). Not sure what this means, or how
256 @var{integer} is interpreted.
259 Indicate this is a packed type, meaning that structure fields or array
260 elements are placed more closely in memory, to save memory at the
264 All of this can make the @var{string} field quite long. All
265 versions of GDB, and some versions of dbx, can handle arbitrarily long
266 strings. But many versions of dbx cretinously limit the strings to
267 about 80 characters, so compilers which must work with such dbx's need
268 to split the @code{.stabs} directive into several @code{.stabs}
269 directives. Each stab duplicates exactly all but the
270 @var{string} field. The @var{string} field of
271 every stab except the last is marked as continued with a
272 double-backslash at the end. Removing the backslashes and concatenating
273 the @var{string} fields of each stab produces the original,
277 @section A simple example in C source
279 To get the flavor of how stabs describe source information for a C
280 program, let's look at the simple program:
285 printf("Hello world");
289 When compiled with @samp{-g}, the program above yields the following
290 @file{.s} file. Line numbers have been added to make it easier to refer
291 to parts of the @file{.s} file in the description of the stabs that
295 @section The simple example at the assembly level
297 This simple ``hello world'' example demonstrates several of the stab
298 types used to describe C language source files.
302 2 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0
303 3 .stabs "hello.c",100,0,0,Ltext0
306 6 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
307 7 .stabs "char:t2=r2;0;127;",128,0,0,0
308 8 .stabs "long int:t3=r1;-2147483648;2147483647;",128,0,0,0
309 9 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
310 10 .stabs "long unsigned int:t5=r1;0;-1;",128,0,0,0
311 11 .stabs "short int:t6=r1;-32768;32767;",128,0,0,0
312 12 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
313 13 .stabs "short unsigned int:t8=r1;0;65535;",128,0,0,0
314 14 .stabs "long long unsigned int:t9=r1;0;-1;",128,0,0,0
315 15 .stabs "signed char:t10=r1;-128;127;",128,0,0,0
316 16 .stabs "unsigned char:t11=r1;0;255;",128,0,0,0
317 17 .stabs "float:t12=r1;4;0;",128,0,0,0
318 18 .stabs "double:t13=r1;8;0;",128,0,0,0
319 19 .stabs "long double:t14=r1;8;0;",128,0,0,0
320 20 .stabs "void:t15=15",128,0,0,0
323 23 .ascii "Hello, world!\12\0"
338 38 sethi %hi(LC0),%o1
339 39 or %o1,%lo(LC0),%o0
350 50 .stabs "main:F1",36,0,0,_main
351 51 .stabn 192,0,0,LBB2
352 52 .stabn 224,0,0,LBE2
355 @node Program structure
356 @chapter Encoding the structure of the program
358 The elements of the program structure that stabs encode include the name
359 of the main function, the names of the source and include files, the
360 line numbers, procedure names and types, and the beginnings and ends of
364 * Main program:: Indicate what the main program is
365 * Source files:: The path and name of the source file
366 * Include files:: Names of include files
369 * Nested procedures::
374 @section Main program
377 Most languages allow the main program to have any name. The
378 @code{N_MAIN} stab type tells the debugger the name that is used in this
379 program. Only the @var{string} field is significant; it is the name of
380 a function which is the main program. Most C compilers do not use this
381 stab (they expect the debugger to assume that the name is @code{main}),
382 but some C compilers emit an @code{N_MAIN} stab for the @code{main}
386 @section Paths and names of the source files
389 Before any other stabs occur, there must be a stab specifying the source
390 file. This information is contained in a symbol of stab type
391 @code{N_SO}; the @var{string} field contains the name of the file. The
392 @var{value} of the symbol is the start address of the portion of the
393 text section corresponding to that file.
395 With the Sun Solaris2 compiler, the @var{desc} field contains a
396 source-language code.
397 @c Do the debuggers use it? What are the codes? -djm
399 Some compilers (for example, GCC2 and SunOS4 @file{/bin/cc}) also
400 include the directory in which the source was compiled, in a second
401 @code{N_SO} symbol preceding the one containing the file name. This
402 symbol can be distinguished by the fact that it ends in a slash. Code
403 from the @code{cfront} C++ compiler can have additional @code{N_SO} symbols for
404 nonexistent source files after the @code{N_SO} for the real source file;
405 these are believed to contain no useful information.
410 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0 # @r{100 is N_SO}
411 .stabs "hello.c",100,0,0,Ltext0
416 Instead of @code{N_SO} symbols, XCOFF uses a @code{.file} assembler
417 directive which assembles to a standard COFF @code{.file} symbol;
418 explaining this in detail is outside the scope of this document.
421 @section Names of include files
423 There are several schemes for dealing with include files: the
424 traditional @code{N_SOL} approach, Sun's @code{N_BINCL} approach, and the
425 XCOFF @code{C_BINCL} approach (which despite the similar name has little in
426 common with @code{N_BINCL}).
429 An @code{N_SOL} symbol specifies which include file subsequent symbols
430 refer to. The @var{string} field is the name of the file and the
431 @var{value} is the text address corresponding to the start of the
432 previous include file and the start of this one. To specify the main
433 source file again, use an @code{N_SOL} symbol with the name of the main
439 The @code{N_BINCL} approach works as follows. An @code{N_BINCL} symbol
440 specifies the start of an include file. In an object file, only the
441 @var{string} is significant; the Sun linker puts data into some of the
442 other fields. The end of the include file is marked by an
443 @code{N_EINCL} symbol (which has no @var{string} field). In an object
444 file, there is no significant data in the @code{N_EINCL} symbol; the Sun
445 linker puts data into some of the fields. @code{N_BINCL} and
446 @code{N_EINCL} can be nested.
448 If the linker detects that two source files have identical stabs between
449 an @code{N_BINCL} and @code{N_EINCL} pair (as will generally be the case
450 for a header file), then it only puts out the stabs once. Each
451 additional occurance is replaced by an @code{N_EXCL} symbol. I believe
452 the Sun (SunOS4, not sure about Solaris) linker is the only one which
453 supports this feature.
454 @c What do the fields of N_EXCL contain? -djm
458 For the start of an include file in XCOFF, use the @file{.bi} assembler
459 directive, which generates a @code{C_BINCL} symbol. A @file{.ei}
460 directive, which generates a @code{C_EINCL} symbol, denotes the end of
461 the include file. Both directives are followed by the name of the
462 source file in quotes, which becomes the @var{string} for the symbol.
463 The @var{value} of each symbol, produced automatically by the assembler
464 and linker, is the offset into the executable of the beginning
465 (inclusive, as you'd expect) or end (inclusive, as you would not expect)
466 of the portion of the COFF line table that corresponds to this include
467 file. @code{C_BINCL} and @code{C_EINCL} do not nest.
470 @section Line numbers
473 An @code{N_SLINE} symbol represents the start of a source line. The
474 @var{desc} field contains the line number and the @var{value} field
475 contains the code address for the start of that source line. On most
476 machines the address is absolute; for Sun's stabs-in-ELF, it is relative
477 to the function in which the @code{N_SLINE} symbol occurs.
481 GNU documents @code{N_DSLINE} and @code{N_BSLINE} symbols for line
482 numbers in the data or bss segments, respectively. They are identical
483 to @code{N_SLINE} but are relocated differently by the linker. They
484 were intended to be used to describe the source location of a variable
485 declaration, but I believe that GCC2 actually puts the line number in
486 the @var{desc} field of the stab for the variable itself. GDB has been
487 ignoring these symbols (unless they contain a @var{string} field) since
490 For single source lines that generate discontiguous code, such as flow
491 of control statements, there may be more than one line number entry for
492 the same source line. In this case there is a line number entry at the
493 start of each code range, each with the same line number.
495 XCOFF uses COFF line numbers, which are outside the scope of this
496 document, ammeliorated by adequate marking of include files
497 (@pxref{Include files}).
504 @findex N_STSYM, for functions (Sun acc)
505 @findex N_GSYM, for functions (Sun acc)
506 All of the following stabs normally use the @code{N_FUN} symbol type.
507 However, Sun's @code{acc} compiler on SunOS4 uses @code{N_GSYM} and
508 @code{N_STSYM}, which means that the value of the stab for the function
509 is useless and the debugger must get the address of the function from
510 the non-stab symbols instead. BSD Fortran is said to use @code{N_FNAME}
511 with the same restriction; the value of the symbol is not useful (I'm
512 not sure it really does use this, because GDB doesn't handle this and no
515 A function is represented by an @samp{F} symbol descriptor for a global
516 (extern) function, and @samp{f} for a static (local) function. The
517 @var{value} field is the address of the start of the function (absolute
518 for @code{a.out}; relative to the start of the file for Sun's
519 stabs-in-ELF). The type information of the stab represents the return
520 type of the function; thus @samp{foo:f5} means that foo is a function
521 returning type 5. There is no need to try to get the line number of the
522 start of the function from the stab for the function; it is in the next
523 @code{N_SLINE} symbol.
525 @c FIXME: verify whether the "I suspect" below is true or not.
526 Some compilers (such as Sun's Solaris compiler) support an extension for
527 specifying the types of the arguments. I suspect this extension is not
528 used for old (non-prototyped) function definitions in C. If the
529 extension is in use, the type information of the stab for the function
530 is followed by type information for each argument, with each argument
531 preceded by @samp{;}. An argument type of 0 means that additional
532 arguments are being passed, whose types and number may vary (@samp{...}
533 in ANSI C). GDB has tolerated this extension (parsed the syntax, if not
534 necessarily used the information) since at least version 4.8; I don't
535 know whether all versions of dbx tolerate it. The argument types given
536 here are not redundant with the symbols for the formal parameters
537 (@pxref{Parameters}); they are the types of the arguments as they are
538 passed, before any conversions might take place. For example, if a C
539 function which is declared without a prototype takes a @code{float}
540 argument, the value is passed as a @code{double} but then converted to a
541 @code{float}. Debuggers need to use the types given in the arguments
542 when printing values, but when calling the function they need to use the
543 types given in the symbol defining the function.
545 If the return type and types of arguments of a function which is defined
546 in another source file are specified (i.e., a function prototype in ANSI
547 C), traditionally compilers emit no stab; the only way for the debugger
548 to find the information is if the source file where the function is
549 defined was also compiled with debugging symbols. As an extension the
550 Solaris compiler uses symbol descriptor @samp{P} followed by the return
551 type of the function, followed by the arguments, each preceded by
552 @samp{;}, as in a stab with symbol descriptor @samp{f} or @samp{F}.
553 This use of symbol descriptor @samp{P} can be distinguished from its use
554 for register parameters (@pxref{Register parameters}) by the fact that it has
555 symbol type @code{N_FUN}.
557 The AIX documentation also defines symbol descriptor @samp{J} as an
558 internal function. I assume this means a function nested within another
559 function. It also says symbol descriptor @samp{m} is a module in
560 Modula-2 or extended Pascal.
562 Procedures (functions which do not return values) are represented as
563 functions returning the @code{void} type in C. I don't see why this couldn't
564 be used for all languages (inventing a @code{void} type for this purpose if
565 necessary), but the AIX documentation defines @samp{I}, @samp{P}, and
566 @samp{Q} for internal, global, and static procedures, respectively.
567 These symbol descriptors are unusual in that they are not followed by
570 The following example shows a stab for a function @code{main} which
571 returns type number @code{1}. The @code{_main} specified for the value
572 is a reference to an assembler label which is used to fill in the start
573 address of the function.
576 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
579 The stab representing a procedure is located immediately following the
580 code of the procedure. This stab is in turn directly followed by a
581 group of other stabs describing elements of the procedure. These other
582 stabs describe the procedure's parameters, its block local variables, and
585 @node Nested procedures
586 @section Nested procedures
588 For any of the symbol descriptors representing procedures, after the
589 symbol descriptor and the type information is optionally a scope
590 specifier. This consists of a comma, the name of the procedure, another
591 comma, and the name of the enclosing procedure. The first name is local
592 to the scope specified, and seems to be redundant with the name of the
593 symbol (before the @samp{:}). This feature is used by GCC, and
594 presumably Pascal, Modula-2, etc., compilers, for nested functions.
596 If procedures are nested more than one level deep, only the immediately
597 containing scope is specified. For example, this code:
609 return baz (x + 2 * y);
611 return x + bar (3 * x);
619 .stabs "baz:f1,baz,bar",36,0,0,_baz.15 # @r{36 is N_FUN}
620 .stabs "bar:f1,bar,foo",36,0,0,_bar.12
621 .stabs "foo:F1",36,0,0,_foo
624 @node Block structure
625 @section Block structure
629 The program's block structure is represented by the @code{N_LBRAC} (left
630 brace) and the @code{N_RBRAC} (right brace) stab types. The variables
631 defined inside a block precede the @code{N_LBRAC} symbol for most
632 compilers, including GCC. Other compilers, such as the Convex, Acorn
633 RISC machine, and Sun @code{acc} compilers, put the variables after the
634 @code{N_LBRAC} symbol. The @var{value} fields of the @code{N_LBRAC} and
635 @code{N_RBRAC} symbols are the start and end addresses of the code of
636 the block, respectively. For most machines, they are relative to the
637 starting address of this source file. For the Gould NP1, they are
638 absolute. For Sun's stabs-in-ELF, they are relative to the function in
641 The @code{N_LBRAC} and @code{N_RBRAC} stabs that describe the block
642 scope of a procedure are located after the @code{N_FUN} stab that
643 represents the procedure itself.
645 Sun documents the @var{desc} field of @code{N_LBRAC} and
646 @code{N_RBRAC} symbols as containing the nesting level of the block.
647 However, dbx seems to not care, and GCC always sets @var{desc} to
653 The @samp{c} symbol descriptor indicates that this stab represents a
654 constant. This symbol descriptor is an exception to the general rule
655 that symbol descriptors are followed by type information. Instead, it
656 is followed by @samp{=} and one of the following:
660 Boolean constant. @var{value} is a numeric value; I assume it is 0 for
664 Character constant. @var{value} is the numeric value of the constant.
666 @item e @var{type-information} , @var{value}
667 Constant whose value can be represented as integral.
668 @var{type-information} is the type of the constant, as it would appear
669 after a symbol descriptor (@pxref{String field}). @var{value} is the
670 numeric value of the constant. GDB 4.9 does not actually get the right
671 value if @var{value} does not fit in a host @code{int}, but it does not
672 do anything violent, and future debuggers could be extended to accept
673 integers of any size (whether unsigned or not). This constant type is
674 usually documented as being only for enumeration constants, but GDB has
675 never imposed that restriction; I don't know about other debuggers.
678 Integer constant. @var{value} is the numeric value. The type is some
679 sort of generic integer type (for GDB, a host @code{int}); to specify
680 the type explicitly, use @samp{e} instead.
683 Real constant. @var{value} is the real value, which can be @samp{INF}
684 (optionally preceded by a sign) for infinity, @samp{QNAN} for a quiet
685 NaN (not-a-number), or @samp{SNAN} for a signalling NaN. If it is a
686 normal number the format is that accepted by the C library function
690 String constant. @var{string} is a string enclosed in either @samp{'}
691 (in which case @samp{'} characters within the string are represented as
692 @samp{\'} or @samp{"} (in which case @samp{"} characters within the
693 string are represented as @samp{\"}).
695 @item S @var{type-information} , @var{elements} , @var{bits} , @var{pattern}
696 Set constant. @var{type-information} is the type of the constant, as it
697 would appear after a symbol descriptor (@pxref{String field}).
698 @var{elements} is the number of elements in the set (does this means
699 how many bits of @var{pattern} are actually used, which would be
700 redundant with the type, or perhaps the number of bits set in
701 @var{pattern}? I don't get it), @var{bits} is the number of bits in the
702 constant (meaning it specifies the length of @var{pattern}, I think),
703 and @var{pattern} is a hexadecimal representation of the set. AIX
704 documentation refers to a limit of 32 bytes, but I see no reason why
705 this limit should exist. This form could probably be used for arbitrary
706 constants, not just sets; the only catch is that @var{pattern} should be
707 understood to be target, not host, byte order and format.
710 The boolean, character, string, and set constants are not supported by
711 GDB 4.9, but it ignores them. GDB 4.8 and earlier gave an error
712 message and refused to read symbols from the file containing the
715 The above information is followed by @samp{;}.
720 Different types of stabs describe the various ways that variables can be
721 allocated: on the stack, globally, in registers, in common blocks,
722 statically, or as arguments to a function.
725 * Stack variables:: Variables allocated on the stack.
726 * Global variables:: Variables used by more than one source file.
727 * Register variables:: Variables in registers.
728 * Common blocks:: Variables statically allocated together.
729 * Statics:: Variables local to one source file.
730 * Parameters:: Variables for arguments to functions.
733 @node Stack variables
734 @section Automatic variables allocated on the stack
736 If a variable's scope is local to a function and its lifetime is only as
737 long as that function executes (C calls such variables
738 @dfn{automatic}), it can be allocated in a register (@pxref{Register
739 variables}) or on the stack.
742 Each variable allocated on the stack has a stab with the symbol
743 descriptor omitted. Since type information should begin with a digit,
744 @samp{-}, or @samp{(}, only those characters precluded from being used
745 for symbol descriptors. However, the Acorn RISC machine (ARM) is said
746 to get this wrong: it puts out a mere type definition here, without the
747 preceding @samp{@var{type-number}=}. This is a bad idea; there is no
748 guarantee that type descriptors are distinct from symbol descriptors.
749 Stabs for stack variables use the @code{N_LSYM} stab type.
751 The @var{value} of the stab is the offset of the variable within the
752 local variables. On most machines this is an offset from the frame
753 pointer and is negative. The location of the stab specifies which block
754 it is defined in; see @ref{Block structure}.
756 For example, the following C code:
766 produces the following stabs:
769 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
770 .stabs "x:1",128,0,0,-12 # @r{128 is N_LSYM}
771 .stabn 192,0,0,LBB2 # @r{192 is N_LBRAC}
772 .stabn 224,0,0,LBE2 # @r{224 is N_RBRAC}
775 @xref{Procedures} for more information on the @code{N_FUN} stab, and
776 @ref{Block structure} for more information on the @code{N_LBRAC} and
777 @code{N_RBRAC} stabs.
779 @node Global variables
780 @section Global variables
783 A variable whose scope is not specific to just one source file is
784 represented by the @samp{G} symbol descriptor. These stabs use the
785 @code{N_GSYM} stab type. The type information for the stab
786 (@pxref{String field}) gives the type of the variable.
788 For example, the following source code:
795 yields the following assembly code:
798 .stabs "g_foo:G2",32,0,0,0 # @r{32 is N_GSYM}
805 The address of the variable represented by the @code{N_GSYM} is not
806 contained in the @code{N_GSYM} stab. The debugger gets this information
807 from the external symbol for the global variable. In the example above,
808 the @code{.global _g_foo} and @code{_g_foo:} lines tell the assembler to
809 produce an external symbol.
811 @node Register variables
812 @section Register variables
815 @c According to an old version of this manual, AIX uses C_RPSYM instead
816 @c of C_RSYM. I am skeptical; this should be verified.
817 Register variables have their own stab type, @code{N_RSYM}, and their
818 own symbol descriptor, @samp{r}. The stab's @var{value} field contains the
819 number of the register where the variable data will be stored.
820 @c .stabs "name:type",N_RSYM,0,RegSize,RegNumber (Sun doc)
822 AIX defines a separate symbol descriptor @samp{d} for floating point
823 registers. This seems unnecessary; why not just just give floating
824 point registers different register numbers? I have not verified whether
825 the compiler actually uses @samp{d}.
827 If the register is explicitly allocated to a global variable, but not
831 register int g_bar asm ("%g5");
835 then the stab may be emitted at the end of the object file, with
836 the other bss symbols.
839 @section Common blocks
841 A common block is a statically allocated section of memory which can be
842 referred to by several source files. It may contain several variables.
843 I believe Fortran is the only language with this feature.
847 A @code{N_BCOMM} stab begins a common block and an @code{N_ECOMM} stab
848 ends it. The only field that is significant in these two stabs is the
849 @var{string}, which names a normal (non-debugging) symbol that gives the
850 address of the common block.
853 Each stab between the @code{N_BCOMM} and the @code{N_ECOMM} specifies a
854 member of that common block; its @var{value} is the offset within the
855 common block of that variable. The @code{N_ECOML} stab type is
856 documented for this purpose, but Sun's Fortran compiler uses
857 @code{N_GSYM} instead. The test case I looked at had a common block
858 local to a function and it used the @samp{V} symbol descriptor; I assume
859 one would use @samp{S} if not local to a function (that is, if a common
860 block @emph{can} be anything other than local to a function).
863 @section Static variables
865 Initialized static variables are represented by the @samp{S} and
866 @samp{V} symbol descriptors. @samp{S} means file scope static, and
867 @samp{V} means procedure scope static.
869 @c This is probably not worth mentioning; it is only true on the sparc
870 @c for `double' variables which although declared const are actually in
871 @c the data segment (the text segment can't guarantee 8 byte alignment).
873 @c 2.4.5 has a bug in that it uses @code{N_FUN}, so neither dbx nor GDB can
874 @c find the variables)
877 In a.out files, @code{N_STSYM} means the data segment, @code{N_FUN}
878 means the text segment, and @code{N_LCSYM} means the bss segment.
880 For example, the source lines:
883 static const int var_const = 5;
884 static int var_init = 2;
885 static int var_noinit;
889 yield the following stabs:
892 .stabs "var_const:S1",36,0,0,_var_const # @r{36 is N_FUN}
894 .stabs "var_init:S1",38,0,0,_var_init # @r{38 is N_STSYM}
896 .stabs "var_noinit:S1",40,0,0,_var_noinit # @r{40 is N_LCSYM}
899 In XCOFF files, each symbol has a section number, so the stab type
900 need not indicate the segment.
902 In ECOFF files, the storage class is used to specify the section, so the
903 stab type need not indicate the segment.
905 @c In ELF files, it apparently is a big mess. See kludge in dbxread.c
906 @c in GDB. FIXME: Investigate where this kludge comes from.
908 @c This is the place to mention N_ROSYM; I'd rather do so once I can
909 @c coherently explain how this stuff works for stabs-in-ELF.
914 Formal parameters to a function are represented by a stab (or sometimes
915 two; see below) for each parameter. The stabs are in the order in which
916 the debugger should print the parameters (i.e., the order in which the
917 parameters are declared in the source file). The exact form of the stab
918 depends on how the parameter is being passed.
921 Parameters passed on the stack use the symbol descriptor @samp{p} and
922 the @code{N_PSYM} symbol type. The @var{value} of the symbol is an offset
923 used to locate the parameter on the stack; its exact meaning is
924 machine-dependent, but on most machines it is an offset from the frame
927 As a simple example, the code:
938 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
939 .stabs "argc:p1",160,0,0,68 # @r{160 is N_PSYM}
940 .stabs "argv:p20=*21=*2",160,0,0,72
943 The type definition of @code{argv} is interesting because it contains
944 several type definitions. Type 21 is pointer to type 2 (char) and
945 @code{argv} (type 20) is pointer to type 21.
947 @c FIXME: figure out what these mean and describe them coherently.
948 The following are also said to go with @code{N_PSYM}:
951 "name" -> "param_name:#type"
953 -> pF Fortran function parameter
954 -> X (function result variable)
955 -> b (based variable)
957 value -> offset from the argument pointer (positive).
961 * Register parameters::
962 * Local variable parameters::
963 * Reference parameters::
964 * Conformant arrays::
967 @node Register parameters
968 @subsection Passing parameters in registers
970 If the parameter is passed in a register, then traditionally there are
971 two symbols for each argument:
974 .stabs "arg:p1" . . . ; N_PSYM
975 .stabs "arg:r1" . . . ; N_RSYM
978 Debuggers use the second one to find the value, and the first one to
979 know that it is an argument.
982 @findex N_RSYM, for parameters
983 Because that approach is kind of ugly, some compilers use symbol
984 descriptor @samp{P} or @samp{R} to indicate an argument which is in a
985 register. Symbol type @code{C_RPSYM} is used with @samp{R} and
986 @code{N_RSYM} is used with @samp{P}. The symbol @var{value} field is
987 the register number. @samp{P} and @samp{R} mean the same thing; the
988 difference is that @samp{P} is a GNU invention and @samp{R} is an IBM
989 (XCOFF) invention. As of version 4.9, GDB should handle either one.
991 There is at least one case where GCC uses a @samp{p} and @samp{r} pair
992 rather than @samp{P}; this is where the argument is passed in the
993 argument list and then loaded into a register.
995 According to the AIX documentation, symbol descriptor @samp{D} is for a
996 parameter passed in a floating point register. This seems
997 unnecessary---why not just use @samp{R} with a register number which
998 indicates that it's a floating point register? I haven't verified
999 whether the system actually does what the documentation indicates.
1001 @c FIXME: On the hppa this is for any type > 8 bytes, I think, and not
1002 @c for small structures (investigate).
1003 On the sparc and hppa, for a @samp{P} symbol whose type is a structure
1004 or union, the register contains the address of the structure. On the
1005 sparc, this is also true of a @samp{p} and @samp{r} pair (using Sun
1006 @code{cc}) or a @samp{p} symbol. However, if a (small) structure is
1007 really in a register, @samp{r} is used. And, to top it all off, on the
1008 hppa it might be a structure which was passed on the stack and loaded
1009 into a register and for which there is a @samp{p} and @samp{r} pair! I
1010 believe that symbol descriptor @samp{i} is supposed to deal with this
1011 case (it is said to mean "value parameter by reference, indirect
1012 access"; I don't know the source for this information), but I don't know
1013 details or what compilers or debuggers use it, if any (not GDB or GCC).
1014 It is not clear to me whether this case needs to be dealt with
1015 differently than parameters passed by reference (@pxref{Reference parameters}).
1017 @node Local variable parameters
1018 @subsection Storing parameters as local variables
1020 There is a case similar to an argument in a register, which is an
1021 argument that is actually stored as a local variable. Sometimes this
1022 happens when the argument was passed in a register and then the compiler
1023 stores it as a local variable. If possible, the compiler should claim
1024 that it's in a register, but this isn't always done.
1026 @findex N_LSYM, for parameter
1027 Some compilers use the pair of symbols approach described above
1028 (@samp{@var{arg}:p} followed by @samp{@var{arg}:}); this includes GCC1
1029 (not GCC2) on the sparc when passing a small structure and GCC2
1030 (sometimes) when the argument type is @code{float} and it is passed as a
1031 @code{double} and converted to @code{float} by the prologue (in the
1032 latter case the type of the @samp{@var{arg}:p} symbol is @code{double}
1033 and the type of the @samp{@var{arg}:} symbol is @code{float}). GCC, at
1034 least on the 960, uses a single @samp{p} symbol descriptor for an
1035 argument which is stored as a local variable but uses @code{N_LSYM}
1036 instead of @code{N_PSYM}. In this case, the @var{value} of the symbol
1037 is an offset relative to the local variables for that function, not
1038 relative to the arguments; on some machines those are the same thing,
1041 @node Reference parameters
1042 @subsection Passing parameters by reference
1044 If the parameter is passed by reference (e.g., Pascal @code{VAR}
1045 parameters), then the symbol descriptor is @samp{v} if it is in the
1046 argument list, or @samp{a} if it in a register. Other than the fact
1047 that these contain the address of the parameter rather than the
1048 parameter itself, they are identical to @samp{p} and @samp{R},
1049 respectively. I believe @samp{a} is an AIX invention; @samp{v} is
1050 supported by all stabs-using systems as far as I know.
1052 @node Conformant arrays
1053 @subsection Passing conformant array parameters
1055 @c Is this paragraph correct? It is based on piecing together patchy
1056 @c information and some guesswork
1057 Conformant arrays are a feature of Modula-2, and perhaps other
1058 languages, in which the size of an array parameter is not known to the
1059 called function until run-time. Such parameters have two stabs: a
1060 @samp{x} for the array itself, and a @samp{C}, which represents the size
1061 of the array. The @var{value} of the @samp{x} stab is the offset in the
1062 argument list where the address of the array is stored (it this right?
1063 it is a guess); the @var{value} of the @samp{C} stab is the offset in the
1064 argument list where the size of the array (in elements? in bytes?) is
1068 @chapter Defining types
1070 The examples so far have described types as references to previously
1071 defined types, or defined in terms of subranges of or pointers to
1072 previously defined types. This chapter describes the other type
1073 descriptors that may follow the @samp{=} in a type definition.
1076 * Builtin types:: Integers, floating point, void, etc.
1077 * Miscellaneous types:: Pointers, sets, files, etc.
1078 * Cross-references:: Referring to a type not yet defined.
1079 * Subranges:: A type with a specific range.
1080 * Arrays:: An aggregate type of same-typed elements.
1081 * Strings:: Like an array but also has a length.
1082 * Enumerations:: Like an integer but the values have names.
1083 * Structures:: An aggregate type of different-typed elements.
1084 * Typedefs:: Giving a type a name.
1085 * Unions:: Different types sharing storage.
1090 @section Builtin types
1092 Certain types are built in (@code{int}, @code{short}, @code{void},
1093 @code{float}, etc.); the debugger recognizes these types and knows how
1094 to handle them. Thus, don't be surprised if some of the following ways
1095 of specifying builtin types do not specify everything that a debugger
1096 would need to know about the type---in some cases they merely specify
1097 enough information to distinguish the type from other types.
1099 The traditional way to define builtin types is convolunted, so new ways
1100 have been invented to describe them. Sun's @code{acc} uses special
1101 builtin type descriptors (@samp{b} and @samp{R}), and IBM uses negative
1102 type numbers. GDB accepts all three ways, as of version 4.8; dbx just
1103 accepts the traditional builtin types and perhaps one of the other two
1104 formats. The following sections describe each of these formats.
1107 * Traditional builtin types:: Put on your seatbelts and prepare for kludgery
1108 * Builtin type descriptors:: Builtin types with special type descriptors
1109 * Negative type numbers:: Builtin types using negative type numbers
1112 @node Traditional builtin types
1113 @subsection Traditional builtin types
1115 This is the traditional, convoluted method for defining builtin types.
1116 There are several classes of such type definitions: integer, floating
1117 point, and @code{void}.
1120 * Traditional integer types::
1121 * Traditional other types::
1124 @node Traditional integer types
1125 @subsubsection Traditional integer types
1127 Often types are defined as subranges of themselves. If the bounding values
1128 fit within an @code{int}, then they are given normally. For example:
1131 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0 # @r{128 is N_LSYM}
1132 .stabs "char:t2=r2;0;127;",128,0,0,0
1135 Builtin types can also be described as subranges of @code{int}:
1138 .stabs "unsigned short:t6=r1;0;65535;",128,0,0,0
1141 If the lower bound of a subrange is 0 and the upper bound is -1,
1142 the type is an unsigned integral type whose bounds are too
1143 big to describe in an @code{int}. Traditionally this is only used for
1144 @code{unsigned int} and @code{unsigned long}:
1147 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
1150 For larger types, GCC 2.4.5 puts out bounds in octal, with a leading 0.
1151 In this case a negative bound consists of a number which is a 1 bit
1152 followed by a bunch of 0 bits, and a positive bound is one in which a
1153 bunch of bits are 1. All known versions of dbx and GDB version 4 accept
1154 this, but GDB 3.5 refuses to read the whole file containing such
1155 symbols. So GCC 2.3.3 did not output the proper size for these types.
1156 @c FIXME: How about an example?
1158 If the lower bound of a subrange is 0 and the upper bound is negative,
1159 the type is an unsigned integral type whose size in bytes is the
1160 absolute value of the upper bound. I believe this is a Convex
1161 convention for @code{unsigned long long}.
1163 If the lower bound of a subrange is negative and the upper bound is 0,
1164 the type is a signed integral type whose size in bytes is
1165 the absolute value of the lower bound. I believe this is a Convex
1166 convention for @code{long long}. To distinguish this from a legitimate
1167 subrange, the type should be a subrange of itself. I'm not sure whether
1168 this is the case for Convex.
1170 @node Traditional other types
1171 @subsubsection Traditional other types
1173 If the upper bound of a subrange is 0 and the lower bound is positive,
1174 the type is a floating point type, and the lower bound of the subrange
1175 indicates the number of bytes in the type:
1178 .stabs "float:t12=r1;4;0;",128,0,0,0
1179 .stabs "double:t13=r1;8;0;",128,0,0,0
1182 However, GCC writes @code{long double} the same way it writes
1183 @code{double}, so there is no way to distinguish.
1186 .stabs "long double:t14=r1;8;0;",128,0,0,0
1189 Complex types are defined the same way as floating-point types; there is
1190 no way to distinguish a single-precision complex from a double-precision
1191 floating-point type.
1193 The C @code{void} type is defined as itself:
1196 .stabs "void:t15=15",128,0,0,0
1199 I'm not sure how a boolean type is represented.
1201 @node Builtin type descriptors
1202 @subsection Defining builtin types using builtin type descriptors
1204 This is the method used by Sun's @code{acc} for defining builtin types.
1205 These are the type descriptors to define builtin types:
1208 @c FIXME: clean up description of width and offset, once we figure out
1210 @item b @var{signed} @var{char-flag} @var{width} ; @var{offset} ; @var{nbits} ;
1211 Define an integral type. @var{signed} is @samp{u} for unsigned or
1212 @samp{s} for signed. @var{char-flag} is @samp{c} which indicates this
1213 is a character type, or is omitted. I assume this is to distinguish an
1214 integral type from a character type of the same size, for example it
1215 might make sense to set it for the C type @code{wchar_t} so the debugger
1216 can print such variables differently (Solaris does not do this). Sun
1217 sets it on the C types @code{signed char} and @code{unsigned char} which
1218 arguably is wrong. @var{width} and @var{offset} appear to be for small
1219 objects stored in larger ones, for example a @code{short} in an
1220 @code{int} register. @var{width} is normally the number of bytes in the
1221 type. @var{offset} seems to always be zero. @var{nbits} is the number
1222 of bits in the type.
1224 Note that type descriptor @samp{b} used for builtin types conflicts with
1225 its use for Pascal space types (@pxref{Miscellaneous types}); they can
1226 be distinguished because the character following the type descriptor
1227 will be a digit, @samp{(}, or @samp{-} for a Pascal space type, or
1228 @samp{u} or @samp{s} for a builtin type.
1231 Documented by AIX to define a wide character type, but their compiler
1232 actually uses negative type numbers (@pxref{Negative type numbers}).
1234 @item R @var{fp-type} ; @var{bytes} ;
1235 Define a floating point type. @var{fp-type} has one of the following values:
1239 IEEE 32-bit (single precision) floating point format.
1242 IEEE 64-bit (double precision) floating point format.
1244 @item 3 (NF_COMPLEX)
1245 @item 4 (NF_COMPLEX16)
1246 @item 5 (NF_COMPLEX32)
1247 @c "GDB source" really means @file{include/aout/stab_gnu.h}, but trying
1248 @c to put that here got an overfull hbox.
1249 These are for complex numbers. A comment in the GDB source describes
1250 them as Fortran @code{complex}, @code{double complex}, and
1251 @code{complex*16}, respectively, but what does that mean? (i.e., Single
1252 precision? Double precison?).
1254 @item 6 (NF_LDOUBLE)
1255 Long double. This should probably only be used for Sun format
1256 @code{long double}, and new codes should be used for other floating
1257 point formats (@code{NF_DOUBLE} can be used if a @code{long double} is
1258 really just an IEEE double, of course).
1261 @var{bytes} is the number of bytes occupied by the type. This allows a
1262 debugger to perform some operations with the type even if it doesn't
1263 understand @var{fp-type}.
1265 @item g @var{type-information} ; @var{nbits}
1266 Documented by AIX to define a floating type, but their compiler actually
1267 uses negative type numbers (@pxref{Negative type numbers}).
1269 @item c @var{type-information} ; @var{nbits}
1270 Documented by AIX to define a complex type, but their compiler actually
1271 uses negative type numbers (@pxref{Negative type numbers}).
1274 The C @code{void} type is defined as a signed integral type 0 bits long:
1276 .stabs "void:t19=bs0;0;0",128,0,0,0
1278 The Solaris compiler seems to omit the trailing semicolon in this case.
1279 Getting sloppy in this way is not a swift move because if a type is
1280 embedded in a more complex expression it is necessary to be able to tell
1283 I'm not sure how a boolean type is represented.
1285 @node Negative type numbers
1286 @subsection Negative type numbers
1288 This is the method used in XCOFF for defining builtin types.
1289 Since the debugger knows about the builtin types anyway, the idea of
1290 negative type numbers is simply to give a special type number which
1291 indicates the builtin type. There is no stab defining these types.
1293 I'm not sure whether anyone has tried to define what this means if
1294 @code{int} can be other than 32 bits (or if other types can be other than
1295 their customary size). If @code{int} has exactly one size for each
1296 architecture, then it can be handled easily enough, but if the size of
1297 @code{int} can vary according the compiler options, then it gets hairy.
1298 The best way to do this would be to define separate negative type
1299 numbers for 16-bit @code{int} and 32-bit @code{int}; therefore I have
1300 indicated below the customary size (and other format information) for
1301 each type. The information below is currently correct because AIX on
1302 the RS6000 is the only system which uses these type numbers. If these
1303 type numbers start to get used on other systems, I suspect the correct
1304 thing to do is to define a new number in cases where a type does not
1305 have the size and format indicated below (or avoid negative type numbers
1308 Part of the definition of the negative type number is
1309 the name of the type. Types with identical size and format but
1310 different names have different negative type numbers.
1314 @code{int}, 32 bit signed integral type.
1317 @code{char}, 8 bit type holding a character. Both GDB and dbx on AIX
1318 treat this as signed. GCC uses this type whether @code{char} is signed
1319 or not, which seems like a bad idea. The AIX compiler (@code{xlc}) seems to
1320 avoid this type; it uses -5 instead for @code{char}.
1323 @code{short}, 16 bit signed integral type.
1326 @code{long}, 32 bit signed integral type.
1329 @code{unsigned char}, 8 bit unsigned integral type.
1332 @code{signed char}, 8 bit signed integral type.
1335 @code{unsigned short}, 16 bit unsigned integral type.
1338 @code{unsigned int}, 32 bit unsigned integral type.
1341 @code{unsigned}, 32 bit unsigned integral type.
1344 @code{unsigned long}, 32 bit unsigned integral type.
1347 @code{void}, type indicating the lack of a value.
1350 @code{float}, IEEE single precision.
1353 @code{double}, IEEE double precision.
1356 @code{long double}, IEEE double precision. The compiler claims the size
1357 will increase in a future release, and for binary compatibility you have
1358 to avoid using @code{long double}. I hope when they increase it they
1359 use a new negative type number.
1362 @code{integer}. 32 bit signed integral type.
1365 @code{boolean}. 32 bit type. How is the truth value encoded? Is it
1366 the least significant bit or is it a question of whether the whole value
1367 is zero or non-zero?
1370 @code{short real}. IEEE single precision.
1373 @code{real}. IEEE double precision.
1376 @code{stringptr}. @xref{Strings}.
1379 @code{character}, 8 bit unsigned character type.
1382 @code{logical*1}, 8 bit type. This Fortran type has a split
1383 personality in that it is used for boolean variables, but can also be
1384 used for unsigned integers. 0 is false, 1 is true, and other values are
1388 @code{logical*2}, 16 bit type. This Fortran type has a split
1389 personality in that it is used for boolean variables, but can also be
1390 used for unsigned integers. 0 is false, 1 is true, and other values are
1394 @code{logical*4}, 32 bit type. This Fortran type has a split
1395 personality in that it is used for boolean variables, but can also be
1396 used for unsigned integers. 0 is false, 1 is true, and other values are
1400 @code{logical}, 32 bit type. This Fortran type has a split
1401 personality in that it is used for boolean variables, but can also be
1402 used for unsigned integers. 0 is false, 1 is true, and other values are
1406 @code{complex}. A complex type consisting of two IEEE single-precision
1407 floating point values.
1410 @code{complex}. A complex type consisting of two IEEE double-precision
1411 floating point values.
1414 @code{integer*1}, 8 bit signed integral type.
1417 @code{integer*2}, 16 bit signed integral type.
1420 @code{integer*4}, 32 bit signed integral type.
1423 @code{wchar}. Wide character, 16 bits wide, unsigned (what format?
1427 @node Miscellaneous types
1428 @section Miscellaneous types
1431 @item b @var{type-information} ; @var{bytes}
1432 Pascal space type. This is documented by IBM; what does it mean?
1434 This use of the @samp{b} type descriptor can be distinguished
1435 from its use for builtin integral types (@pxref{Builtin type
1436 descriptors}) because the character following the type descriptor is
1437 always a digit, @samp{(}, or @samp{-}.
1439 @item B @var{type-information}
1440 A volatile-qualified version of @var{type-information}. This is
1441 a Sun extension. References and stores to a variable with a
1442 volatile-qualified type must not be optimized or cached; they
1443 must occur as the user specifies them.
1445 @item d @var{type-information}
1446 File of type @var{type-information}. As far as I know this is only used
1449 @item k @var{type-information}
1450 A const-qualified version of @var{type-information}. This is a Sun
1451 extension. A variable with a const-qualified type cannot be modified.
1453 @item M @var{type-information} ; @var{length}
1454 Multiple instance type. The type seems to composed of @var{length}
1455 repetitions of @var{type-information}, for example @code{character*3} is
1456 represented by @samp{M-2;3}, where @samp{-2} is a reference to a
1457 character type (@pxref{Negative type numbers}). I'm not sure how this
1458 differs from an array. This appears to be a Fortran feature.
1459 @var{length} is a bound, like those in range types; see @ref{Subranges}.
1461 @item S @var{type-information}
1462 Pascal set type. @var{type-information} must be a small type such as an
1463 enumeration or a subrange, and the type is a bitmask whose length is
1464 specified by the number of elements in @var{type-information}.
1466 @item * @var{type-information}
1467 Pointer to @var{type-information}.
1470 @node Cross-references
1471 @section Cross-references to other types
1473 A type can be used before it is defined; one common way to deal with
1474 that situation is just to use a type reference to a type which has not
1477 Another way is with the @samp{x} type descriptor, which is followed by
1478 @samp{s} for a structure tag, @samp{u} for a union tag, or @samp{e} for
1479 a enumerator tag, followed by the name of the tag, followed by @samp{:}.
1480 For example, the following C declarations:
1491 .stabs "bar:G16=*17=xsfoo:",32,0,0,0
1494 Not all debuggers support the @samp{x} type descriptor, so on some
1495 machines GCC does not use it. I believe that for the above example it
1496 would just emit a reference to type 17 and never define it, but I
1497 haven't verified that.
1499 Modula-2 imported types, at least on AIX, use the @samp{i} type
1500 descriptor, which is followed by the name of the module from which the
1501 type is imported, followed by @samp{:}, followed by the name of the
1502 type. There is then optionally a comma followed by type information for
1503 the type. This differs from merely naming the type (@pxref{Typedefs}) in
1504 that it identifies the module; I don't understand whether the name of
1505 the type given here is always just the same as the name we are giving
1506 it, or whether this type descriptor is used with a nameless stab
1507 (@pxref{String field}), or what. The symbol ends with @samp{;}.
1510 @section Subrange types
1512 The @samp{r} type descriptor defines a type as a subrange of another
1513 type. It is followed by type information for the type of which it is a
1514 subrange, a semicolon, an integral lower bound, a semicolon, an
1515 integral upper bound, and a semicolon. The AIX documentation does not
1516 specify the trailing semicolon, in an effort to specify array indexes
1517 more cleanly, but a subrange which is not an array index has always
1518 included a trailing semicolon (@pxref{Arrays}).
1520 Instead of an integer, either bound can be one of the following:
1523 @item A @var{offset}
1524 The bound is passed by reference on the stack at offset @var{offset}
1525 from the argument list. @xref{Parameters}, for more information on such
1528 @item T @var{offset}
1529 The bound is passed by value on the stack at offset @var{offset} from
1532 @item a @var{register-number}
1533 The bound is pased by reference in register number
1534 @var{register-number}.
1536 @item t @var{register-number}
1537 The bound is passed by value in register number @var{register-number}.
1543 Subranges are also used for builtin types; see @ref{Traditional builtin types}.
1546 @section Array types
1548 Arrays use the @samp{a} type descriptor. Following the type descriptor
1549 is the type of the index and the type of the array elements. If the
1550 index type is a range type, it ends in a semicolon; otherwise
1551 (for example, if it is a type reference), there does not
1552 appear to be any way to tell where the types are separated. In an
1553 effort to clean up this mess, IBM documents the two types as being
1554 separated by a semicolon, and a range type as not ending in a semicolon
1555 (but this is not right for range types which are not array indexes,
1556 @pxref{Subranges}). I think probably the best solution is to specify
1557 that a semicolon ends a range type, and that the index type and element
1558 type of an array are separated by a semicolon, but that if the index
1559 type is a range type, the extra semicolon can be omitted. GDB (at least
1560 through version 4.9) doesn't support any kind of index type other than a
1561 range anyway; I'm not sure about dbx.
1563 It is well established, and widely used, that the type of the index,
1564 unlike most types found in the stabs, is merely a type definition, not
1565 type information (@pxref{String field}) (that is, it need not start with
1566 @samp{@var{type-number}=} if it is defining a new type). According to a
1567 comment in GDB, this is also true of the type of the array elements; it
1568 gives @samp{ar1;1;10;ar1;1;10;4} as a legitimate way to express a two
1569 dimensional array. According to AIX documentation, the element type
1570 must be type information. GDB accepts either.
1572 The type of the index is often a range type, expressed as the type
1573 descriptor @samp{r} and some parameters. It defines the size of the
1574 array. In the example below, the range @samp{r1;0;2;} defines an index
1575 type which is a subrange of type 1 (integer), with a lower bound of 0
1576 and an upper bound of 2. This defines the valid range of subscripts of
1577 a three-element C array.
1579 For example, the definition:
1582 char char_vec[3] = @{'a','b','c'@};
1586 produces the output:
1589 .stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
1598 If an array is @dfn{packed}, the elements are spaced more
1599 closely than normal, saving memory at the expense of speed. For
1600 example, an array of 3-byte objects might, if unpacked, have each
1601 element aligned on a 4-byte boundary, but if packed, have no padding.
1602 One way to specify that something is packed is with type attributes
1603 (@pxref{String field}). In the case of arrays, another is to use the
1604 @samp{P} type descriptor instead of @samp{a}. Other than specifying a
1605 packed array, @samp{P} is identical to @samp{a}.
1607 @c FIXME-what is it? A pointer?
1608 An open array is represented by the @samp{A} type descriptor followed by
1609 type information specifying the type of the array elements.
1611 @c FIXME: what is the format of this type? A pointer to a vector of pointers?
1612 An N-dimensional dynamic array is represented by
1615 D @var{dimensions} ; @var{type-information}
1618 @c Does dimensions really have this meaning? The AIX documentation
1620 @var{dimensions} is the number of dimensions; @var{type-information}
1621 specifies the type of the array elements.
1623 @c FIXME: what is the format of this type? A pointer to some offsets in
1625 A subarray of an N-dimensional array is represented by
1628 E @var{dimensions} ; @var{type-information}
1631 @c Does dimensions really have this meaning? The AIX documentation
1633 @var{dimensions} is the number of dimensions; @var{type-information}
1634 specifies the type of the array elements.
1639 Some languages, like C or the original Pascal, do not have string types,
1640 they just have related things like arrays of characters. But most
1641 Pascals and various other languages have string types, which are
1642 indicated as follows:
1645 @item n @var{type-information} ; @var{bytes}
1646 @var{bytes} is the maximum length. I'm not sure what
1647 @var{type-information} is; I suspect that it means that this is a string
1648 of @var{type-information} (thus allowing a string of integers, a string
1649 of wide characters, etc., as well as a string of characters). Not sure
1650 what the format of this type is. This is an AIX feature.
1652 @item z @var{type-information} ; @var{bytes}
1653 Just like @samp{n} except that this is a gstring, not an ordinary
1654 string. I don't know the difference.
1657 Pascal Stringptr. What is this? This is an AIX feature.
1661 @section Enumerations
1663 Enumerations are defined with the @samp{e} type descriptor.
1665 @c FIXME: Where does this information properly go? Perhaps it is
1666 @c redundant with something we already explain.
1667 The source line below declares an enumeration type at file scope.
1668 The type definition is located after the @code{N_RBRAC} that marks the end of
1669 the previous procedure's block scope, and before the @code{N_FUN} that marks
1670 the beginning of the next procedure's block scope. Therefore it does not
1671 describe a block local symbol, but a file local one.
1676 enum e_places @{first,second=3,last@};
1680 generates the following stab:
1683 .stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0
1686 The symbol descriptor (@samp{T}) says that the stab describes a
1687 structure, enumeration, or union tag. The type descriptor @samp{e},
1688 following the @samp{22=} of the type definition narrows it down to an
1689 enumeration type. Following the @samp{e} is a list of the elements of
1690 the enumeration. The format is @samp{@var{name}:@var{value},}. The
1691 list of elements ends with @samp{;}.
1693 There is no standard way to specify the size of an enumeration type; it
1694 is determined by the architecture (normally all enumerations types are
1695 32 bits). There should be a way to specify an enumeration type of
1696 another size; type attributes would be one way to do this. @xref{Stabs
1702 The encoding of structures in stabs can be shown with an example.
1704 The following source code declares a structure tag and defines an
1705 instance of the structure in global scope. Then a @code{typedef} equates the
1706 structure tag with a new type. Seperate stabs are generated for the
1707 structure tag, the structure @code{typedef}, and the structure instance. The
1708 stabs for the tag and the @code{typedef} are emited when the definitions are
1709 encountered. Since the structure elements are not initialized, the
1710 stab and code for the structure variable itself is located at the end
1711 of the program in the bss section.
1718 struct s_tag* s_next;
1721 typedef struct s_tag s_typedef;
1724 The structure tag has an @code{N_LSYM} stab type because, like the
1725 enumeration, the symbol has file scope. Like the enumeration, the
1726 symbol descriptor is @samp{T}, for enumeration, structure, or tag type.
1727 The type descriptor @samp{s} following the @samp{16=} of the type
1728 definition narrows the symbol type to structure.
1730 Following the @samp{s} type descriptor is the number of bytes the
1731 structure occupies, followed by a description of each structure element.
1732 The structure element descriptions are of the form @var{name:type, bit
1733 offset from the start of the struct, number of bits in the element}.
1735 @c FIXME: phony line break. Can probably be fixed by using an example
1736 @c with fewer fields.
1739 .stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;
1740 s_char_vec:17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0
1743 In this example, the first two structure elements are previously defined
1744 types. For these, the type following the @samp{@var{name}:} part of the
1745 element description is a simple type reference. The other two structure
1746 elements are new types. In this case there is a type definition
1747 embedded after the @samp{@var{name}:}. The type definition for the
1748 array element looks just like a type definition for a standalone array.
1749 The @code{s_next} field is a pointer to the same kind of structure that
1750 the field is an element of. So the definition of structure type 16
1751 contains a type definition for an element which is a pointer to type 16.
1754 @section Giving a type a name
1756 To give a type a name, use the @samp{t} symbol descriptor. The type
1757 is specified by the type information (@pxref{String field}) for the stab.
1761 .stabs "s_typedef:t16",128,0,0,0 # @r{128 is N_LSYM}
1764 specifies that @code{s_typedef} refers to type number 16. Such stabs
1765 have symbol type @code{N_LSYM} (or @code{C_DECL} for XCOFF).
1767 If you are specifying the tag name for a structure, union, or
1768 enumeration, use the @samp{T} symbol descriptor instead. I believe C is
1769 the only language with this feature.
1771 If the type is an opaque type (I believe this is a Modula-2 feature),
1772 AIX provides a type descriptor to specify it. The type descriptor is
1773 @samp{o} and is followed by a name. I don't know what the name
1774 means---is it always the same as the name of the type, or is this type
1775 descriptor used with a nameless stab (@pxref{String field})? There
1776 optionally follows a comma followed by type information which defines
1777 the type of this type. If omitted, a semicolon is used in place of the
1778 comma and the type information, and the type is much like a generic
1779 pointer type---it has a known size but little else about it is
1793 This code generates a stab for a union tag and a stab for a union
1794 variable. Both use the @code{N_LSYM} stab type. If a union variable is
1795 scoped locally to the procedure in which it is defined, its stab is
1796 located immediately preceding the @code{N_LBRAC} for the procedure's block
1799 The stab for the union tag, however, is located preceding the code for
1800 the procedure in which it is defined. The stab type is @code{N_LSYM}. This
1801 would seem to imply that the union type is file scope, like the struct
1802 type @code{s_tag}. This is not true. The contents and position of the stab
1803 for @code{u_type} do not convey any infomation about its procedure local
1806 @c FIXME: phony line break. Can probably be fixed by using an example
1807 @c with fewer fields.
1810 .stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
1814 The symbol descriptor @samp{T}, following the @samp{name:} means that
1815 the stab describes an enumeration, structure, or union tag. The type
1816 descriptor @samp{u}, following the @samp{23=} of the type definition,
1817 narrows it down to a union type definition. Following the @samp{u} is
1818 the number of bytes in the union. After that is a list of union element
1819 descriptions. Their format is @var{name:type, bit offset into the
1820 union, number of bytes for the element;}.
1822 The stab for the union variable is:
1825 .stabs "an_u:23",128,0,0,-20 # @r{128 is N_LSYM}
1828 @samp{-20} specifies where the variable is stored (@pxref{Stack
1831 @node Function types
1832 @section Function types
1834 Various types can be defined for function variables. These types are
1835 not used in defining functions (@pxref{Procedures}); they are used for
1836 things like pointers to functions.
1838 The simple, traditional, type is type descriptor @samp{f} is followed by
1839 type information for the return type of the function, followed by a
1842 This does not deal with functions for which the number and types of the
1843 parameters are part of the type, as in Modula-2 or ANSI C. AIX provides
1844 extensions to specify these, using the @samp{f}, @samp{F}, @samp{p}, and
1845 @samp{R} type descriptors.
1847 First comes the type descriptor. If it is @samp{f} or @samp{F}, this
1848 type involves a function rather than a procedure, and the type
1849 information for the return type of the function follows, followed by a
1850 comma. Then comes the number of parameters to the function and a
1851 semicolon. Then, for each parameter, there is the name of the parameter
1852 followed by a colon (this is only present for type descriptors @samp{R}
1853 and @samp{F} which represent Pascal function or procedure parameters),
1854 type information for the parameter, a comma, 0 if passed by reference or
1855 1 if passed by value, and a semicolon. The type definition ends with a
1858 For example, this variable definition:
1865 generates the following code:
1868 .stabs "g_pf:G24=*25=f1",32,0,0,0
1869 .common _g_pf,4,"bss"
1872 The variable defines a new type, 24, which is a pointer to another new
1873 type, 25, which is a function returning @code{int}.
1876 @chapter Symbol information in symbol tables
1878 This chapter describes the format of symbol table entries
1879 and how stab assembler directives map to them. It also describes the
1880 transformations that the assembler and linker make on data from stabs.
1883 * Symbol table format::
1884 * Transformations on symbol tables::
1887 @node Symbol table format
1888 @section Symbol table format
1890 Each time the assembler encounters a stab directive, it puts
1891 each field of the stab into a corresponding field in a symbol table
1892 entry of its output file. If the stab contains a @var{string} field, the
1893 symbol table entry for that stab points to a string table entry
1894 containing the string data from the stab. Assembler labels become
1895 relocatable addresses. Symbol table entries in a.out have the format:
1897 @c FIXME: should refer to external, not internal.
1899 struct internal_nlist @{
1900 unsigned long n_strx; /* index into string table of name */
1901 unsigned char n_type; /* type of symbol */
1902 unsigned char n_other; /* misc info (usually empty) */
1903 unsigned short n_desc; /* description field */
1904 bfd_vma n_value; /* value of symbol */
1908 For @code{.stabs} directives, the @code{n_strx} field holds the offset
1909 in bytes from the start of the string table to the string table entry
1910 containing the @var{string} field. For other classes of stabs
1911 (@code{.stabn} and @code{.stabd}) this field is zero.
1913 Symbol table entries with @code{n_type} field values greater than 0x1f
1914 originated as stabs generated by the compiler (with one random
1915 exception). The other entries were placed in the symbol table of the
1916 executable by the assembler or the linker.
1918 @node Transformations on symbol tables
1919 @section Transformations on symbol tables
1921 The linker concatenates object files and does fixups of externally
1924 You can see the transformations made on stab data by the assembler and
1925 linker by examining the symbol table after each pass of the build. To
1926 do this, use @samp{nm -ap}, which dumps the symbol table, including
1927 debugging information, unsorted. For stab entries the columns are:
1928 @var{value}, @var{other}, @var{desc}, @var{type}, @var{string}. For
1929 assembler and linker symbols, the columns are: @var{value}, @var{type},
1932 The low 5 bits of the stab type tell the linker how to relocate the
1933 value of the stab. Thus for stab types like @code{N_RSYM} and
1934 @code{N_LSYM}, where the value is an offset or a register number, the
1935 low 5 bits are @code{N_ABS}, which tells the linker not to relocate the
1938 Where the @var{value} field of a stab contains an assembly language label,
1939 it is transformed by each build step. The assembler turns it into a
1940 relocatable address and the linker turns it into an absolute address.
1943 * Transformations on static variables::
1944 * Transformations on global variables::
1947 @node Transformations on static variables
1948 @subsection Transformations on static variables
1950 This source line defines a static variable at file scope:
1953 static int s_g_repeat
1957 The following stab describes the symbol:
1960 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
1964 The assembler transforms the stab into this symbol table entry in the
1965 @file{.o} file. The location is expressed as a data segment offset.
1968 00000084 - 00 0000 STSYM s_g_repeat:S1
1972 In the symbol table entry from the executable, the linker has made the
1973 relocatable address absolute.
1976 0000e00c - 00 0000 STSYM s_g_repeat:S1
1979 @node Transformations on global variables
1980 @subsection Transformations on global variables
1982 Stabs for global variables do not contain location information. In
1983 this case, the debugger finds location information in the assembler or
1984 linker symbol table entry describing the variable. The source line:
1994 .stabs "g_foo:G2",32,0,0,0
1997 The variable is represented by two symbol table entries in the object
1998 file (see below). The first one originated as a stab. The second one
1999 is an external symbol. The upper case @samp{D} signifies that the
2000 @code{n_type} field of the symbol table contains 7, @code{N_DATA} with
2001 local linkage. The @var{value} field is empty for the stab entry. For
2002 the linker symbol, it contains the relocatable address corresponding to
2006 00000000 - 00 0000 GSYM g_foo:G2
2011 These entries as transformed by the linker. The linker symbol table
2012 entry now holds an absolute address:
2015 00000000 - 00 0000 GSYM g_foo:G2
2021 @chapter GNU C++ stabs
2024 * Basic cplusplus types::
2027 * Methods:: Method definition
2029 * Method modifiers::
2032 * Virtual base classes::
2036 Type descriptors added for C++ descriptions:
2040 method type (@code{##} if minimal debug)
2043 Member (class and variable) type. It is followed by type information
2044 for the offset basetype, a comma, and type information for the type of
2045 the field being pointed to. (FIXME: this is acknowledged to be
2046 gibberish. Can anyone say what really goes here?).
2048 Note that there is a conflict between this and type attributes
2049 (@pxref{String field}); both use type descriptor @samp{@@}.
2050 Fortunately, the @samp{@@} type descriptor used in this C++ sense always
2051 will be followed by a digit, @samp{(}, or @samp{-}, and type attributes
2052 never start with those things.
2055 @node Basic cplusplus types
2056 @section Basic types for C++
2058 << the examples that follow are based on a01.C >>
2061 C++ adds two more builtin types to the set defined for C. These are
2062 the unknown type and the vtable record type. The unknown type, type
2063 16, is defined in terms of itself like the void type.
2065 The vtable record type, type 17, is defined as a structure type and
2066 then as a structure tag. The structure has four fields: delta, index,
2067 pfn, and delta2. pfn is the function pointer.
2069 << In boilerplate $vtbl_ptr_type, what are the fields delta,
2070 index, and delta2 used for? >>
2072 This basic type is present in all C++ programs even if there are no
2073 virtual methods defined.
2076 .stabs "struct_name:sym_desc(type)type_def(17)=type_desc(struct)struct_bytes(8)
2077 elem_name(delta):type_ref(short int),bit_offset(0),field_bits(16);
2078 elem_name(index):type_ref(short int),bit_offset(16),field_bits(16);
2079 elem_name(pfn):type_def(18)=type_desc(ptr to)type_ref(void),
2080 bit_offset(32),field_bits(32);
2081 elem_name(delta2):type_def(short int);bit_offset(32),field_bits(16);;"
2086 .stabs "$vtbl_ptr_type:t17=s8
2087 delta:6,0,16;index:6,16,16;pfn:18=*15,32,32;delta2:6,32,16;;"
2092 .stabs "name:sym_dec(struct tag)type_ref($vtbl_ptr_type)",N_LSYM,NIL,NIL,NIL
2096 .stabs "$vtbl_ptr_type:T17",128,0,0,0
2099 @node Simple classes
2100 @section Simple class definition
2102 The stabs describing C++ language features are an extension of the
2103 stabs describing C. Stabs representing C++ class types elaborate
2104 extensively on the stab format used to describe structure types in C.
2105 Stabs representing class type variables look just like stabs
2106 representing C language variables.
2108 Consider the following very simple class definition.
2114 int Ameth(int in, char other);
2118 The class @code{baseA} is represented by two stabs. The first stab describes
2119 the class as a structure type. The second stab describes a structure
2120 tag of the class type. Both stabs are of stab type @code{N_LSYM}. Since the
2121 stab is not located between an @code{N_FUN} and an @code{N_LBRAC} stab this indicates
2122 that the class is defined at file scope. If it were, then the @code{N_LSYM}
2123 would signify a local variable.
2125 A stab describing a C++ class type is similar in format to a stab
2126 describing a C struct, with each class member shown as a field in the
2127 structure. The part of the struct format describing fields is
2128 expanded to include extra information relevent to C++ class members.
2129 In addition, if the class has multiple base classes or virtual
2130 functions the struct format outside of the field parts is also
2133 In this simple example the field part of the C++ class stab
2134 representing member data looks just like the field part of a C struct
2135 stab. The section on protections describes how its format is
2136 sometimes extended for member data.
2138 The field part of a C++ class stab representing a member function
2139 differs substantially from the field part of a C struct stab. It
2140 still begins with @samp{name:} but then goes on to define a new type number
2141 for the member function, describe its return type, its argument types,
2142 its protection level, any qualifiers applied to the method definition,
2143 and whether the method is virtual or not. If the method is virtual
2144 then the method description goes on to give the vtable index of the
2145 method, and the type number of the first base class defining the
2148 When the field name is a method name it is followed by two colons rather
2149 than one. This is followed by a new type definition for the method.
2150 This is a number followed by an equal sign and the type descriptor
2151 @samp{#}, indicating a method type, and a second @samp{#}, indicating
2152 that this is the @dfn{minimal} type of method definition used by GCC2,
2153 not larger method definitions used by earlier versions of GCC. This is
2154 followed by a type reference showing the return type of the method and a
2157 The format of an overloaded operator method name differs from that of
2158 other methods. It is @samp{op$::@var{operator-name}.} where
2159 @var{operator-name} is the operator name such as @samp{+} or @samp{+=}.
2160 The name ends with a period, and any characters except the period can
2161 occur in the @var{operator-name} string.
2163 The next part of the method description represents the arguments to the
2164 method, preceeded by a colon and ending with a semi-colon. The types of
2165 the arguments are expressed in the same way argument types are expressed
2166 in C++ name mangling. In this example an @code{int} and a @code{char}
2169 This is followed by a number, a letter, and an asterisk or period,
2170 followed by another semicolon. The number indicates the protections
2171 that apply to the member function. Here the 2 means public. The
2172 letter encodes any qualifier applied to the method definition. In
2173 this case, @samp{A} means that it is a normal function definition. The dot
2174 shows that the method is not virtual. The sections that follow
2175 elaborate further on these fields and describe the additional
2176 information present for virtual methods.
2180 .stabs "class_name:sym_desc(type)type_def(20)=type_desc(struct)struct_bytes(4)
2181 field_name(Adat):type(int),bit_offset(0),field_bits(32);
2183 method_name(Ameth)::type_def(21)=type_desc(method)return_type(int);
2184 :arg_types(int char);
2185 protection(public)qualifier(normal)virtual(no);;"
2190 .stabs "baseA:t20=s4Adat:1,0,32;Ameth::21=##1;:ic;2A.;;",128,0,0,0
2192 .stabs "class_name:sym_desc(struct tag)",N_LSYM,NIL,NIL,NIL
2194 .stabs "baseA:T20",128,0,0,0
2197 @node Class instance
2198 @section Class instance
2200 As shown above, describing even a simple C++ class definition is
2201 accomplished by massively extending the stab format used in C to
2202 describe structure types. However, once the class is defined, C stabs
2203 with no modifications can be used to describe class instances. The
2213 yields the following stab describing the class instance. It looks no
2214 different from a standard C stab describing a local variable.
2217 .stabs "name:type_ref(baseA)", N_LSYM, NIL, NIL, frame_ptr_offset
2221 .stabs "AbaseA:20",128,0,0,-20
2225 @section Method defintion
2227 The class definition shown above declares Ameth. The C++ source below
2232 baseA::Ameth(int in, char other)
2239 This method definition yields three stabs following the code of the
2240 method. One stab describes the method itself and following two describe
2241 its parameters. Although there is only one formal argument all methods
2242 have an implicit argument which is the @code{this} pointer. The @code{this}
2243 pointer is a pointer to the object on which the method was called. Note
2244 that the method name is mangled to encode the class name and argument
2245 types. Name mangling is described in the @sc{arm} (@cite{The Annotated
2246 C++ Reference Manual}, by Ellis and Stroustrup, @sc{isbn}
2247 0-201-51459-1); @file{gpcompare.texi} in Cygnus GCC distributions
2248 describes the differences between GNU mangling and @sc{arm}
2250 @c FIXME: Use @xref, especially if this is generally installed in the
2252 @c FIXME: This information should be in a net release, either of GCC or
2253 @c GDB. But gpcompare.texi doesn't seem to be in the FSF GCC.
2256 .stabs "name:symbol_desriptor(global function)return_type(int)",
2257 N_FUN, NIL, NIL, code_addr_of_method_start
2259 .stabs "Ameth__5baseAic:F1",36,0,0,_Ameth__5baseAic
2262 Here is the stab for the @code{this} pointer implicit argument. The
2263 name of the @code{this} pointer is always @code{this}. Type 19, the
2264 @code{this} pointer is defined as a pointer to type 20, @code{baseA},
2265 but a stab defining @code{baseA} has not yet been emited. Since the
2266 compiler knows it will be emited shortly, here it just outputs a cross
2267 reference to the undefined symbol, by prefixing the symbol name with
2271 .stabs "name:sym_desc(register param)type_def(19)=
2272 type_desc(ptr to)type_ref(baseA)=
2273 type_desc(cross-reference to)baseA:",N_RSYM,NIL,NIL,register_number
2275 .stabs "this:P19=*20=xsbaseA:",64,0,0,8
2278 The stab for the explicit integer argument looks just like a parameter
2279 to a C function. The last field of the stab is the offset from the
2280 argument pointer, which in most systems is the same as the frame
2284 .stabs "name:sym_desc(value parameter)type_ref(int)",
2285 N_PSYM,NIL,NIL,offset_from_arg_ptr
2287 .stabs "in:p1",160,0,0,72
2290 << The examples that follow are based on A1.C >>
2293 @section Protections
2296 In the simple class definition shown above all member data and
2297 functions were publicly accessable. The example that follows
2298 contrasts public, protected and privately accessable fields and shows
2299 how these protections are encoded in C++ stabs.
2301 @c FIXME: What does "part of the string" mean?
2302 Protections for class member data are signified by two characters
2303 embedded in the stab defining the class type. These characters are
2304 located after the name: part of the string. @samp{/0} means private,
2305 @samp{/1} means protected, and @samp{/2} means public. If these
2306 characters are omited this means that the member is public. The
2307 following C++ source:
2321 generates the following stab to describe the class type all_data.
2324 .stabs "class_name:sym_desc(type)type_def(19)=type_desc(struct)struct_bytes
2325 data_name:/protection(private)type_ref(int),bit_offset,num_bits;
2326 data_name:/protection(protected)type_ref(char),bit_offset,num_bits;
2327 data_name:(/num omited, private)type_ref(float),bit_offset,num_bits;;"
2332 .stabs "all_data:t19=s12
2333 priv_dat:/01,0,32;prot_dat:/12,32,8;pub_dat:12,64,32;;",128,0,0,0
2336 Protections for member functions are signified by one digit embeded in
2337 the field part of the stab describing the method. The digit is 0 if
2338 private, 1 if protected and 2 if public. Consider the C++ class
2342 class all_methods @{
2344 int priv_meth(int in)@{return in;@};
2346 char protMeth(char in)@{return in;@};
2348 float pubMeth(float in)@{return in;@};
2352 It generates the following stab. The digit in question is to the left
2353 of an @samp{A} in each case. Notice also that in this case two symbol
2354 descriptors apply to the class name struct tag and struct type.
2357 .stabs "class_name:sym_desc(struct tag&type)type_def(21)=
2358 sym_desc(struct)struct_bytes(1)
2359 meth_name::type_def(22)=sym_desc(method)returning(int);
2360 :args(int);protection(private)modifier(normal)virtual(no);
2361 meth_name::type_def(23)=sym_desc(method)returning(char);
2362 :args(char);protection(protected)modifier(normal)virual(no);
2363 meth_name::type_def(24)=sym_desc(method)returning(float);
2364 :args(float);protection(public)modifier(normal)virtual(no);;",
2369 .stabs "all_methods:Tt21=s1priv_meth::22=##1;:i;0A.;protMeth::23=##2;:c;1A.;
2370 pubMeth::24=##12;:f;2A.;;",128,0,0,0
2373 @node Method modifiers
2374 @section Method modifiers (@code{const}, @code{volatile}, @code{const volatile})
2378 In the class example described above all the methods have the normal
2379 modifier. This method modifier information is located just after the
2380 protection information for the method. This field has four possible
2381 character values. Normal methods use @samp{A}, const methods use
2382 @samp{B}, volatile methods use @samp{C}, and const volatile methods use
2383 @samp{D}. Consider the class definition below:
2388 int ConstMeth (int arg) const @{ return arg; @};
2389 char VolatileMeth (char arg) volatile @{ return arg; @};
2390 float ConstVolMeth (float arg) const volatile @{return arg; @};
2394 This class is described by the following stab:
2397 .stabs "class(A):sym_desc(struct)type_def(20)=type_desc(struct)struct_bytes(1)
2398 meth_name(ConstMeth)::type_def(21)sym_desc(method)
2399 returning(int);:arg(int);protection(public)modifier(const)virtual(no);
2400 meth_name(VolatileMeth)::type_def(22)=sym_desc(method)
2401 returning(char);:arg(char);protection(public)modifier(volatile)virt(no)
2402 meth_name(ConstVolMeth)::type_def(23)=sym_desc(method)
2403 returning(float);:arg(float);protection(public)modifer(const volatile)
2404 virtual(no);;", @dots{}
2408 .stabs "A:T20=s1ConstMeth::21=##1;:i;2B.;VolatileMeth::22=##2;:c;2C.;
2409 ConstVolMeth::23=##12;:f;2D.;;",128,0,0,0
2412 @node Virtual methods
2413 @section Virtual methods
2415 << The following examples are based on a4.C >>
2417 The presence of virtual methods in a class definition adds additional
2418 data to the class description. The extra data is appended to the
2419 description of the virtual method and to the end of the class
2420 description. Consider the class definition below:
2426 virtual int A_virt (int arg) @{ return arg; @};
2430 This results in the stab below describing class A. It defines a new
2431 type (20) which is an 8 byte structure. The first field of the class
2432 struct is @samp{Adat}, an integer, starting at structure offset 0 and
2435 The second field in the class struct is not explicitly defined by the
2436 C++ class definition but is implied by the fact that the class
2437 contains a virtual method. This field is the vtable pointer. The
2438 name of the vtable pointer field starts with @samp{$vf} and continues with a
2439 type reference to the class it is part of. In this example the type
2440 reference for class A is 20 so the name of its vtable pointer field is
2441 @samp{$vf20}, followed by the usual colon.
2443 Next there is a type definition for the vtable pointer type (21).
2444 This is in turn defined as a pointer to another new type (22).
2446 Type 22 is the vtable itself, which is defined as an array, indexed by
2447 a range of integers between 0 and 1, and whose elements are of type
2448 17. Type 17 was the vtable record type defined by the boilerplate C++
2449 type definitions, as shown earlier.
2451 The bit offset of the vtable pointer field is 32. The number of bits
2452 in the field are not specified when the field is a vtable pointer.
2454 Next is the method definition for the virtual member function @code{A_virt}.
2455 Its description starts out using the same format as the non-virtual
2456 member functions described above, except instead of a dot after the
2457 @samp{A} there is an asterisk, indicating that the function is virtual.
2458 Since is is virtual some addition information is appended to the end
2459 of the method description.
2461 The first number represents the vtable index of the method. This is a
2462 32 bit unsigned number with the high bit set, followed by a
2465 The second number is a type reference to the first base class in the
2466 inheritence hierarchy defining the virtual member function. In this
2467 case the class stab describes a base class so the virtual function is
2468 not overriding any other definition of the method. Therefore the
2469 reference is to the type number of the class that the stab is
2472 This is followed by three semi-colons. One marks the end of the
2473 current sub-section, one marks the end of the method field, and the
2474 third marks the end of the struct definition.
2476 For classes containing virtual functions the very last section of the
2477 string part of the stab holds a type reference to the first base
2478 class. This is preceeded by @samp{~%} and followed by a final semi-colon.
2481 .stabs "class_name(A):type_def(20)=sym_desc(struct)struct_bytes(8)
2482 field_name(Adat):type_ref(int),bit_offset(0),field_bits(32);
2483 field_name(A virt func ptr):type_def(21)=type_desc(ptr to)type_def(22)=
2484 sym_desc(array)index_type_ref(range of int from 0 to 1);
2485 elem_type_ref(vtbl elem type),
2487 meth_name(A_virt)::typedef(23)=sym_desc(method)returning(int);
2488 :arg_type(int),protection(public)normal(yes)virtual(yes)
2489 vtable_index(1);class_first_defining(A);;;~%first_base(A);",
2493 @c FIXME: bogus line break.
2495 .stabs "A:t20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2496 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2500 @section Inheritence
2502 Stabs describing C++ derived classes include additional sections that
2503 describe the inheritence hierarchy of the class. A derived class stab
2504 also encodes the number of base classes. For each base class it tells
2505 if the base class is virtual or not, and if the inheritence is private
2506 or public. It also gives the offset into the object of the portion of
2507 the object corresponding to each base class.
2509 This additional information is embeded in the class stab following the
2510 number of bytes in the struct. First the number of base classes
2511 appears bracketed by an exclamation point and a comma.
2513 Then for each base type there repeats a series: two digits, a number,
2514 a comma, another number, and a semi-colon.
2516 The first of the two digits is 1 if the base class is virtual and 0 if
2517 not. The second digit is 2 if the derivation is public and 0 if not.
2519 The number following the first two digits is the offset from the start
2520 of the object to the part of the object pertaining to the base class.
2522 After the comma, the second number is a type_descriptor for the base
2523 type. Finally a semi-colon ends the series, which repeats for each
2526 The source below defines three base classes @code{A}, @code{B}, and
2527 @code{C} and the derived class @code{D}.
2534 virtual int A_virt (int arg) @{ return arg; @};
2540 virtual int B_virt (int arg) @{return arg; @};
2546 virtual int C_virt (int arg) @{return arg; @};
2549 class D : A, virtual B, public C @{
2552 virtual int A_virt (int arg ) @{ return arg+1; @};
2553 virtual int B_virt (int arg) @{ return arg+2; @};
2554 virtual int C_virt (int arg) @{ return arg+3; @};
2555 virtual int D_virt (int arg) @{ return arg; @};
2559 Class stabs similar to the ones described earlier are generated for
2562 @c FIXME!!! the linebreaks in the following example probably make the
2563 @c examples literally unusable, but I don't know any other way to get
2564 @c them on the page.
2565 @c One solution would be to put some of the type definitions into
2566 @c separate stabs, even if that's not exactly what the compiler actually
2569 .stabs "A:T20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2570 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2572 .stabs "B:Tt25=s8Bdat:1,0,32;$vf25:21,32;B_virt::26=##1;
2573 :i;2A*-2147483647;25;;;~%25;",128,0,0,0
2575 .stabs "C:Tt28=s8Cdat:1,0,32;$vf28:21,32;C_virt::29=##1;
2576 :i;2A*-2147483647;28;;;~%28;",128,0,0,0
2579 In the stab describing derived class @code{D} below, the information about
2580 the derivation of this class is encoded as follows.
2583 .stabs "derived_class_name:symbol_descriptors(struct tag&type)=
2584 type_descriptor(struct)struct_bytes(32)!num_bases(3),
2585 base_virtual(no)inheritence_public(no)base_offset(0),
2586 base_class_type_ref(A);
2587 base_virtual(yes)inheritence_public(no)base_offset(NIL),
2588 base_class_type_ref(B);
2589 base_virtual(no)inheritence_public(yes)base_offset(64),
2590 base_class_type_ref(C); @dots{}
2593 @c FIXME! fake linebreaks.
2595 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:
2596 1,160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt:
2597 :32:i;2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;
2598 28;;D_virt::32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2601 @node Virtual base classes
2602 @section Virtual base classes
2604 A derived class object consists of a concatination in memory of the data
2605 areas defined by each base class, starting with the leftmost and ending
2606 with the rightmost in the list of base classes. The exception to this
2607 rule is for virtual inheritence. In the example above, class @code{D}
2608 inherits virtually from base class @code{B}. This means that an
2609 instance of a @code{D} object will not contain its own @code{B} part but
2610 merely a pointer to a @code{B} part, known as a virtual base pointer.
2612 In a derived class stab, the base offset part of the derivation
2613 information, described above, shows how the base class parts are
2614 ordered. The base offset for a virtual base class is always given as 0.
2615 Notice that the base offset for @code{B} is given as 0 even though
2616 @code{B} is not the first base class. The first base class @code{A}
2619 The field information part of the stab for class @code{D} describes the field
2620 which is the pointer to the virtual base class @code{B}. The vbase pointer
2621 name is @samp{$vb} followed by a type reference to the virtual base class.
2622 Since the type id for @code{B} in this example is 25, the vbase pointer name
2625 @c FIXME!! fake linebreaks below
2627 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:1,
2628 160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt::32:i;
2629 2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;28;;D_virt:
2630 :32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2633 Following the name and a semicolon is a type reference describing the
2634 type of the virtual base class pointer, in this case 24. Type 24 was
2635 defined earlier as the type of the @code{B} class @code{this} pointer. The
2636 @code{this} pointer for a class is a pointer to the class type.
2639 .stabs "this:P24=*25=xsB:",64,0,0,8
2642 Finally the field offset part of the vbase pointer field description
2643 shows that the vbase pointer is the first field in the @code{D} object,
2644 before any data fields defined by the class. The layout of a @code{D}
2645 class object is a follows, @code{Adat} at 0, the vtable pointer for
2646 @code{A} at 32, @code{Cdat} at 64, the vtable pointer for C at 96, the
2647 virtual base pointer for @code{B} at 128, and @code{Ddat} at 160.
2650 @node Static members
2651 @section Static members
2653 The data area for a class is a concatenation of the space used by the
2654 data members of the class. If the class has virtual methods, a vtable
2655 pointer follows the class data. The field offset part of each field
2656 description in the class stab shows this ordering.
2658 << How is this reflected in stabs? See Cygnus bug #677 for some info. >>
2661 @appendix Table of stab types
2663 The following are all the possible values for the stab @var{type} field, for
2664 @code{a.out} files, in numeric order. This does not apply to XCOFF.
2666 The symbolic names are defined in the file @file{include/aout/stabs.def}.
2669 * Non-stab symbol types:: Types from 0 to 0x1f
2670 * Stab symbol types:: Types from 0x20 to 0xff
2673 @node Non-stab symbol types
2674 @appendixsec Non-stab symbol types
2676 The following types are used by the linker and assembler, not by stab
2677 directives. Since this document does not attempt to describe aspects of
2678 object file format other than the debugging format, no details are
2681 @c Try to get most of these to fit on a single line.
2691 File scope absolute symbol
2693 @item 0x3 N_ABS | N_EXT
2694 External absolute symbol
2697 File scope text symbol
2699 @item 0x5 N_TEXT | N_EXT
2700 External text symbol
2703 File scope data symbol
2705 @item 0x7 N_DATA | N_EXT
2706 External data symbol
2709 File scope BSS symbol
2711 @item 0x9 N_BSS | N_EXT
2715 Same as @code{N_FN}, for Sequent compilers
2718 Symbol is indirected to another symbol
2721 Common---visible after shared library dynamic link
2724 Absolute set element
2727 Text segment set element
2730 Data segment set element
2733 BSS segment set element
2736 Pointer to set vector
2738 @item 0x1e N_WARNING
2739 Print a warning message during linking
2742 File name of a @file{.o} file
2745 @node Stab symbol types
2746 @appendixsec Stab symbol types
2748 The following symbol types indicate that this is a stab. This is the
2749 full list of stab numbers, including stab types that are used in
2750 languages other than C.
2754 Global symbol; see @ref{Global variables}.
2757 Function name (for BSD Fortran); see @ref{Procedures}.
2760 Function name (@pxref{Procedures}) or text segment variable
2764 Data segment file-scope variable; see @ref{Statics}.
2767 BSS segment file-scope variable; see @ref{Statics}.
2770 Name of main routine; see @ref{Main program}.
2772 @c FIXME: discuss this in the Statics node where we talk about
2773 @c the fact that the n_type indicates the section.
2775 Variable in @code{.rodata} section; see @ref{Statics}.
2778 Global symbol (for Pascal); see @ref{N_PC}.
2781 Number of symbols (according to Ultrix V4.0); see @ref{N_NSYMS}.
2784 No DST map; see @ref{N_NOMAP}.
2786 @c FIXME: describe this solaris feature in the body of the text (see
2787 @c comments in include/aout/stab.def).
2789 Object file (Solaris2).
2791 @c See include/aout/stab.def for (a little) more info.
2793 Debugger options (Solaris2).
2796 Register variable; see @ref{Register variables}.
2799 Modula-2 compilation unit; see @ref{N_M2C}.
2802 Line number in text segment; see @ref{Line numbers}.
2805 Line number in data segment; see @ref{Line numbers}.
2808 Line number in bss segment; see @ref{Line numbers}.
2811 Sun source code browser, path to @file{.cb} file; see @ref{N_BROWS}.
2814 GNU Modula2 definition module dependency; see @ref{N_DEFD}.
2817 Function start/body/end line numbers (Solaris2).
2820 GNU C++ exception variable; see @ref{N_EHDECL}.
2823 Modula2 info "for imc" (according to Ultrix V4.0); see @ref{N_MOD2}.
2826 GNU C++ @code{catch} clause; see @ref{N_CATCH}.
2829 Structure of union element; see @ref{N_SSYM}.
2832 Last stab for module (Solaris2).
2835 Path and name of source file; see @ref{Source files}.
2838 Stack variable (@pxref{Stack variables}) or type (@pxref{Typedefs}).
2841 Beginning of an include file (Sun only); see @ref{Include files}.
2844 Name of include file; see @ref{Include files}.
2847 Parameter variable; see @ref{Parameters}.
2850 End of an include file; see @ref{Include files}.
2853 Alternate entry point; see @ref{N_ENTRY}.
2856 Beginning of a lexical block; see @ref{Block structure}.
2859 Place holder for a deleted include file; see @ref{Include files}.
2862 Modula2 scope information (Sun linker); see @ref{N_SCOPE}.
2865 End of a lexical block; see @ref{Block structure}.
2868 Begin named common block; see @ref{Common blocks}.
2871 End named common block; see @ref{Common blocks}.
2874 Member of a common block; see @ref{Common blocks}.
2876 @c FIXME: How does this really work? Move it to main body of document.
2878 Pascal @code{with} statement: type,,0,0,offset (Solaris2).
2881 Gould non-base registers; see @ref{Gould}.
2884 Gould non-base registers; see @ref{Gould}.
2887 Gould non-base registers; see @ref{Gould}.
2890 Gould non-base registers; see @ref{Gould}.
2893 Gould non-base registers; see @ref{Gould}.
2896 @c Restore the default table indent
2901 @node Symbol descriptors
2902 @appendix Table of symbol descriptors
2904 These tell in the @code{.stabs} @var{string} field what kind of symbol
2905 the stab represents. They follow the colon which follows the symbol
2906 name. @xref{String field}, for more information about their use.
2908 @c Please keep this alphabetical
2910 @c In TeX, this looks great, digit is in italics. But makeinfo insists
2911 @c on putting it in `', not realizing that @var should override @code.
2912 @c I don't know of any way to make makeinfo do the right thing. Seems
2913 @c like a makeinfo bug to me.
2917 Variable on the stack; see @ref{Stack variables}.
2920 Parameter passed by reference in register; see @ref{Reference parameters}.
2923 Constant; see @ref{Constants}.
2926 Conformant array bound (Pascal, maybe other languages); @ref{Conformant
2927 arrays}. Name of a caught exception (GNU C++). These can be
2928 distinguished because the latter uses @code{N_CATCH} and the former uses
2929 another symbol type.
2932 Floating point register variable; see @ref{Register variables}.
2935 Parameter in floating point register; see @ref{Register parameters}.
2938 File scope function; see @ref{Procedures}.
2941 Global function; see @ref{Procedures}.
2944 Global variable; see @ref{Global variables}.
2947 @xref{Register parameters}.
2950 Internal (nested) procedure; see @ref{Nested procedures}.
2953 Internal (nested) function; see @ref{Nested procedures}.
2956 Label name (documented by AIX, no further information known).
2959 Module; see @ref{Procedures}.
2962 Argument list parameter; see @ref{Parameters}.
2968 Fortran Function parameter; see @ref{Parameters}.
2971 Unfortunately, three separate meanings have been independently invented
2972 for this symbol descriptor. At least the GNU and Sun uses can be
2973 distinguished by the symbol type. Global Procedure (AIX) (symbol type
2974 used unknown); see @ref{Procedures}. Register parameter (GNU) (symbol
2975 type @code{N_PSYM}); see @ref{Parameters}. Prototype of function
2976 referenced by this file (Sun @code{acc}) (symbol type @code{N_FUN}).
2979 Static Procedure; see @ref{Procedures}.
2982 Register parameter; see @ref{Register parameters}.
2985 Register variable; see @ref{Register variables}.
2988 File scope variable; see @ref{Statics}.
2991 Type name; see @ref{Typedefs}.
2994 Enumeration, structure, or union tag; see @ref{Typedefs}.
2997 Parameter passed by reference; see @ref{Reference parameters}.
3000 Procedure scope static variable; see @ref{Statics}.
3003 Conformant array; see @ref{Conformant arrays}.
3006 Function return variable; see @ref{Parameters}.
3009 @node Type descriptors
3010 @appendix Table of type descriptors
3012 These tell in the @code{.stabs} @var{string} field what kind of type is being
3013 defined. They follow the type number and an equals sign.
3014 @xref{String field}, for more information about their use.
3019 Type reference; see @ref{String field}.
3022 Reference to builtin type; see @ref{Negative type numbers}.
3025 Method (C++); see @ref{Cplusplus}.
3028 Pointer; see @ref{Miscellaneous types}.
3034 Type Attributes (AIX); see @ref{String field}. Member (class and variable)
3035 type (GNU C++); see @ref{Cplusplus}.
3038 Array; see @ref{Arrays}.
3041 Open array; see @ref{Arrays}.
3044 Pascal space type (AIX); see @ref{Miscellaneous types}. Builtin integer
3045 type (Sun); see @ref{Builtin type descriptors}.
3048 Volatile-qualified type; see @ref{Miscellaneous types}.
3051 Complex builtin type; see @ref{Builtin type descriptors}.
3054 COBOL Picture type. See AIX documentation for details.
3057 File type; see @ref{Miscellaneous types}.
3060 N-dimensional dynamic array; see @ref{Arrays}.
3063 Enumeration type; see @ref{Enumerations}.
3066 N-dimensional subarray; see @ref{Arrays}.
3069 Function type; see @ref{Function types}.
3072 Pascal function parameter; see @ref{Function types}
3075 Builtin floating point type; see @ref{Builtin type descriptors}.
3078 COBOL Group. See AIX documentation for details.
3081 Imported type; see @ref{Cross-references}.
3084 Const-qualified type; see @ref{Miscellaneous types}.
3087 COBOL File Descriptor. See AIX documentation for details.
3090 Multiple instance type; see @ref{Miscellaneous types}.
3093 String type; see @ref{Strings}.
3096 Stringptr; see @ref{Strings}.
3099 Opaque type; see @ref{Typedefs}.
3102 Procedure; see @ref{Function types}.
3105 Packed array; see @ref{Arrays}.
3108 Range type; see @ref{Subranges}.
3111 Builtin floating type; see @ref{Builtin type descriptors} (Sun). Pascal
3112 subroutine parameter; see @ref{Function types} (AIX). Detecting this
3113 conflict is possible with careful parsing (hint: a Pascal subroutine
3114 parameter type will always contain a comma, and a builtin type
3115 descriptor never will).
3118 Structure type; see @ref{Structures}.
3121 Set type; see @ref{Miscellaneous types}.
3124 Union; see @ref{Unions}.
3127 Variant record. This is a Pascal and Modula-2 feature which is like a
3128 union within a struct in C. See AIX documentation for details.
3131 Wide character; see @ref{Builtin type descriptors}.
3134 Cross-reference; see @ref{Cross-references}.
3137 gstring; see @ref{Strings}.
3140 @node Expanded reference
3141 @appendix Expanded reference by stab type
3143 @c FIXME: This appendix should go away; see N_PSYM or N_SO for an example.
3145 For a full list of stab types, and cross-references to where they are
3146 described, see @ref{Stab types}. This appendix just duplicates certain
3147 information from the main body of this document; eventually the
3148 information will all be in one place.
3152 The first line is the symbol type (see @file{include/aout/stab.def}).
3154 The second line describes the language constructs the symbol type
3157 The third line is the stab format with the significant stab fields
3158 named and the rest NIL.
3160 Subsequent lines expand upon the meaning and possible values for each
3161 significant stab field. @samp{#} stands in for the type descriptor.
3163 Finally, any further information.
3166 * N_PC:: Pascal global symbol
3167 * N_NSYMS:: Number of symbols
3168 * N_NOMAP:: No DST map
3169 * N_M2C:: Modula-2 compilation unit
3170 * N_BROWS:: Path to .cb file for Sun source code browser
3171 * N_DEFD:: GNU Modula2 definition module dependency
3172 * N_EHDECL:: GNU C++ exception variable
3173 * N_MOD2:: Modula2 information "for imc"
3174 * N_CATCH:: GNU C++ "catch" clause
3175 * N_SSYM:: Structure or union element
3176 * N_ENTRY:: Alternate entry point
3177 * N_SCOPE:: Modula2 scope information (Sun only)
3178 * Gould:: non-base register symbols used on Gould systems
3179 * N_LENG:: Length of preceding entry
3185 @deffn @code{.stabs} N_PC
3187 Global symbol (for Pascal).
3190 "name" -> "symbol_name" <<?>>
3191 value -> supposedly the line number (stab.def is skeptical)
3195 @file{stabdump.c} says:
3197 global pascal symbol: name,,0,subtype,line
3205 @deffn @code{.stabn} N_NSYMS
3207 Number of symbols (according to Ultrix V4.0).
3210 0, files,,funcs,lines (stab.def)
3217 @deffn @code{.stabs} N_NOMAP
3219 No DST map for symbol (according to Ultrix V4.0). I think this means a
3220 variable has been optimized out.
3223 name, ,0,type,ignored (stab.def)
3230 @deffn @code{.stabs} N_M2C
3232 Modula-2 compilation unit.
3235 "string" -> "unit_name,unit_time_stamp[,code_time_stamp]"
3237 value -> 0 (main unit)
3245 @deffn @code{.stabs} N_BROWS
3247 Sun source code browser, path to @file{.cb} file
3250 "path to associated @file{.cb} file"
3252 Note: @var{type} field value overlaps with N_BSLINE.
3258 @deffn @code{.stabn} N_DEFD
3260 GNU Modula2 definition module dependency.
3262 GNU Modula-2 definition module dependency. @var{value} is the modification
3263 time of the definition file. @var{other} is non-zero if it is imported with
3264 the GNU M2 keyword @code{%INITIALIZE}. Perhaps @code{N_M2C} can be used
3265 if there are enough empty fields?
3271 @deffn @code{.stabs} N_EHDECL
3273 GNU C++ exception variable <<?>>.
3275 "@var{string} is variable name"
3277 Note: conflicts with @code{N_MOD2}.
3283 @deffn @code{.stab?} N_MOD2
3285 Modula2 info "for imc" (according to Ultrix V4.0)
3287 Note: conflicts with @code{N_EHDECL} <<?>>
3293 @deffn @code{.stabn} N_CATCH
3295 GNU C++ @code{catch} clause
3297 GNU C++ @code{catch} clause. @code{value} is its address. @code{desc}
3298 is nonzero if this entry is immediately followed by a @code{CAUGHT} stab
3299 saying what exception was caught. Multiple @code{CAUGHT} stabs means
3300 that multiple exceptions can be caught here. If @code{desc} is 0, it
3301 means all exceptions are caught here.
3307 @deffn @code{.stabn} N_SSYM
3309 Structure or union element.
3311 @code{value} is offset in the structure.
3313 <<?looking at structs and unions in C I didn't see these>>
3319 @deffn @code{.stabn} N_ENTRY
3321 Alternate entry point.
3322 @code{value} is its address.
3329 @deffn @code{.stab?} N_SCOPE
3331 Modula2 scope information (Sun linker)
3336 @section Non-base registers on Gould systems
3338 @deffn @code{.stab?} N_NBTEXT
3339 @deffnx @code{.stab?} N_NBDATA
3340 @deffnx @code{.stab?} N_NBBSS
3341 @deffnx @code{.stab?} N_NBSTS
3342 @deffnx @code{.stab?} N_NBLCS
3348 These are used on Gould systems for non-base registers syms.
3350 However, the following values are not the values used by Gould; they are
3351 the values which GNU has been documenting for these values for a long
3352 time, without actually checking what Gould uses. I include these values
3353 only because perhaps some someone actually did something with the GNU
3354 information (I hope not, why GNU knowingly assigned wrong values to
3355 these in the header file is a complete mystery to me).
3358 240 0xf0 N_NBTEXT ??
3359 242 0xf2 N_NBDATA ??
3369 @deffn @code{.stabn} N_LENG
3371 Second symbol entry containing a length-value for the preceding entry.
3372 The @var{value} is the length.
3376 @appendix Questions and anomalies
3380 @c I think this is changed in GCC 2.4.5 to put the line number there.
3381 For GNU C stabs defining local and global variables (@code{N_LSYM} and
3382 @code{N_GSYM}), the @var{desc} field is supposed to contain the source
3383 line number on which the variable is defined. In reality the @var{desc}
3384 field is always 0. (This behavior is defined in @file{dbxout.c} and
3385 putting a line number in @var{desc} is controlled by @samp{#ifdef
3386 WINNING_GDB}, which defaults to false). GDB supposedly uses this
3387 information if you say @samp{list @var{var}}. In reality, @var{var} can
3388 be a variable defined in the program and GDB says @samp{function
3389 @var{var} not defined}.
3392 In GNU C stabs, there seems to be no way to differentiate tag types:
3393 structures, unions, and enums (symbol descriptor @samp{T}) and typedefs
3394 (symbol descriptor @samp{t}) defined at file scope from types defined locally
3395 to a procedure or other more local scope. They all use the @code{N_LSYM}
3396 stab type. Types defined at procedure scope are emited after the
3397 @code{N_RBRAC} of the preceding function and before the code of the
3398 procedure in which they are defined. This is exactly the same as
3399 types defined in the source file between the two procedure bodies.
3400 GDB overcompensates by placing all types in block #1, the block for
3401 symbols of file scope. This is true for default, @samp{-ansi} and
3402 @samp{-traditional} compiler options. (Bugs gcc/1063, gdb/1066.)
3405 What ends the procedure scope? Is it the proc block's @code{N_RBRAC} or the
3406 next @code{N_FUN}? (I believe its the first.)
3409 @c FIXME: This should go with the other stuff about global variables.
3410 Global variable stabs don't have location information. This comes
3411 from the external symbol for the same variable. The external symbol
3412 has a leading underbar on the _name of the variable and the stab does
3413 not. How do we know these two symbol table entries are talking about
3414 the same symbol when their names are different? (Answer: the debugger
3415 knows that external symbols have leading underbars).
3417 @c FIXME: This is absurdly vague; there all kinds of differences, some
3418 @c of which are the same between gnu & sun, and some of which aren't.
3419 @c In particular, I'm pretty sure GCC works with Sun dbx by default.
3421 @c Can GCC be configured to output stabs the way the Sun compiler
3422 @c does, so that their native debugging tools work? <NO?> It doesn't by
3423 @c default. GDB reads either format of stab. (GCC or SunC). How about
3427 @node XCOFF differences
3428 @appendix Differences between GNU stabs in a.out and GNU stabs in XCOFF
3430 @c FIXME: Merge *all* these into the main body of the document.
3431 The AIX/RS6000 native object file format is XCOFF with stabs. This
3432 appendix only covers those differences which are not covered in the main
3433 body of this document.
3437 BSD a.out stab types correspond to AIX XCOFF storage classes. In general
3438 the mapping is @code{N_@var{stabtype}} becomes @code{C_@var{stabtype}}.
3439 Some stab types in a.out are not supported in XCOFF; most of these use
3442 @c FIXME: Get C_* types for the block, figure out whether it is always
3443 @c used (I suspect not), explain clearly, and move to node Statics.
3444 Exception: initialised static @code{N_STSYM} and un-initialized static
3445 @code{N_LCSYM} both map to the @code{C_STSYM} storage class. But the
3446 destinction is preserved because in XCOFF @code{N_STSYM} and
3447 @code{N_LCSYM} must be emited in a named static block. Begin the block
3448 with @samp{.bs s[RW] data_section_name} for @code{N_STSYM} or @samp{.bs
3449 s bss_section_name} for @code{N_LCSYM}. End the block with @samp{.es}.
3451 @c FIXME: I think they are trying to say something about whether the
3452 @c assembler defaults the value to the location counter.
3454 If the XCOFF stab is an @code{N_FUN} (@code{C_FUN}) then follow the
3455 string field with @samp{,.} instead of just @samp{,}.
3458 I think that's it for @file{.s} file differences. They could stand to be
3459 better presented. This is just a list of what I have noticed so far.
3460 There are a @emph{lot} of differences in the information in the symbol
3461 tables of the executable and object files.
3463 Mapping of a.out stab types to XCOFF storage classes:
3466 stab type storage class
3467 -------------------------------
3503 @node Sun differences
3504 @appendix Differences between GNU stabs and Sun native stabs
3506 @c FIXME: Merge all this stuff into the main body of the document.
3510 GNU C stabs define @emph{all} types, file or procedure scope, as
3511 @code{N_LSYM}. Sun doc talks about using @code{N_GSYM} too.
3514 Sun C stabs use type number pairs in the format (@var{a},@var{b}) where
3515 @var{a} is a number starting with 1 and incremented for each sub-source
3516 file in the compilation. @var{b} is a number starting with 1 and
3517 incremented for each new type defined in the compilation. GNU C stabs
3518 use the type number alone, with no source file number.
3522 @appendix Using stabs with the ELF object file format
3524 The ELF object file format allows tools to create object files with
3525 custom sections containing any arbitrary data. To use stabs in ELF
3526 object files, the tools create two custom sections, a section named
3527 @code{.stab} which contains an array of fixed length structures, one
3528 struct per stab, and a section named @code{.stabstr} containing all the
3529 variable length strings that are referenced by stabs in the @code{.stab}
3530 section. The byte order of the stabs binary data matches the byte order
3531 of the ELF file itself, as determined from the @code{EI_DATA} field in
3532 the @code{e_ident} member of the ELF header.
3534 @c Is "source file" the right term for this concept? We don't mean that
3535 @c there is a separate one for include files (but "object file" or
3536 @c "object module" isn't quite right either; the output from ld -r is a
3537 @c single object file but contains many source files).
3538 The first stab in the @code{.stab} section for each source file is
3539 synthetic, generated entirely by the assembler, with no corresponding
3540 @code{.stab} directive as input to the assembler. This stab contains
3541 the following fields:
3545 Offset in the @code{.stabstr} section to the source filename.
3551 Unused field, always zero.
3554 Count of upcoming symbols, i.e., the number of remaining stabs for this
3558 Size of the string table fragment associated with this source file, in
3562 The @code{.stabstr} section always starts with a null byte (so that string
3563 offsets of zero reference a null string), followed by random length strings,
3564 each of which is null byte terminated.
3566 The ELF section header for the @code{.stab} section has its
3567 @code{sh_link} member set to the section number of the @code{.stabstr}
3568 section, and the @code{.stabstr} section has its ELF section
3569 header @code{sh_type} member set to @code{SHT_STRTAB} to mark it as a
3572 Because the linker does not process the @code{.stab} section in any
3573 special way, none of the addresses in the @code{n_value} field of the
3574 stabs are relocated by the linker. Instead they are relative to the
3575 source file (or some entity smaller than a source file, like a
3576 function). To find the address of each section corresponding to a given
3577 source file, the (compiler? assembler?) puts out symbols giving the
3578 address of each section for a given source file. Since these are normal
3579 ELF symbols, the linker can relocate them correctly. They are
3580 named @code{Bbss.bss} for the bss section, @code{Ddata.data} for
3581 the data section, and @code{Drodata.rodata} for the rodata section. I
3582 haven't yet figured out how the debugger gets the address for the text
3585 @node Symbol Types Index
3586 @unnumbered Symbol Types Index