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, Jim Kingdon,
21 Permission is granted to make and distribute verbatim copies of
22 this manual provided the copyright notice and this permission notice
23 are preserved on all copies.
26 Permission is granted to process this file through Tex and print the
27 results, provided the printed document carries copying permission
28 notice identical to this one except for the removal of this paragraph
29 (this paragraph not being relevant to the printed manual).
32 Permission is granted to copy or distribute modified versions of this
33 manual under the terms of the GPL (for which purpose this text may be
34 regarded as a program in the language TeX).
37 @setchapternewpage odd
40 @title The ``stabs'' debug format
41 @author Julia Menapace, Jim Kingdon, David MacKenzie
42 @author Cygnus Support
45 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
46 \xdef\manvers{\$Revision$} % For use in headers, footers too
48 \hfill Cygnus Support\par
50 \hfill \TeX{}info \texinfoversion\par
54 @vskip 0pt plus 1filll
55 Copyright @copyright{} 1992, 1993 Free Software Foundation, Inc.
56 Contributed by Cygnus Support.
58 Permission is granted to make and distribute verbatim copies of
59 this manual provided the copyright notice and this permission notice
60 are preserved on all copies.
66 @top The "stabs" representation of debugging information
68 This document describes the stabs debugging format.
71 * Overview:: Overview of stabs
72 * Program Structure:: Encoding of the structure of the program
73 * Constants:: Constants
75 * Types:: Type definitions
76 * Symbol Tables:: Symbol information in symbol tables
77 * Cplusplus:: Appendixes:
78 * Stab Types:: Symbol types in a.out files
79 * Symbol Descriptors:: Table of symbol descriptors
80 * Type Descriptors:: Table of type descriptors
81 * Expanded Reference:: Reference information by stab type
82 * Questions:: Questions and anomolies
83 * XCOFF Differences:: Differences between GNU stabs in a.out
84 and GNU stabs in XCOFF
85 * Sun Differences:: Differences between GNU stabs and Sun
87 * Stabs In ELF:: Stabs in an ELF file.
88 * Symbol Types Index:: Index of symbolic stab symbol type names.
94 @chapter Overview of Stabs
96 @dfn{Stabs} refers to a format for information that describes a program
97 to a debugger. This format was apparently invented by
98 @c FIXME! <<name of inventor>> at
99 the University of California at Berkeley, for the @code{pdx} Pascal
100 debugger; the format has spread widely since then.
102 This document is one of the few published sources of documentation on
103 stabs. It is believed to be comprehensive for stabs used by C. The
104 lists of symbol descriptors (@pxref{Symbol Descriptors}) and type
105 descriptors (@pxref{Type Descriptors}) are believed to be completely
106 comprehensive. Stabs for COBOL-specific features and for variant
107 records (used by Pascal and Modula-2) are poorly documented here.
109 Other sources of information on stabs are @cite{Dbx and Dbxtool
110 Interfaces}, 2nd edition, by Sun, 1988, and @cite{AIX Version 3.2 Files
111 Reference}, Fourth Edition, September 1992, "dbx Stabstring Grammar" in
112 the a.out section, page 2-31. This document is believed to incorporate
113 the information from those two sources except where it explictly directs
114 you to them for more information.
117 * Flow:: Overview of debugging information flow
118 * Stabs Format:: Overview of stab format
119 * String Field:: The string field
120 * C Example:: A simple example in C source
121 * Assembly Code:: The simple example at the assembly level
125 @section Overview of Debugging Information Flow
127 The GNU C compiler compiles C source in a @file{.c} file into assembly
128 language in a @file{.s} file, which the assembler translates into
129 a @file{.o} file, which the linker combines with other @file{.o} files and
130 libraries to produce an executable file.
132 With the @samp{-g} option, GCC puts in the @file{.s} file additional
133 debugging information, which is slightly transformed by the assembler
134 and linker, and carried through into the final executable. This
135 debugging information describes features of the source file like line
136 numbers, the types and scopes of variables, and function names,
137 parameters, and scopes.
139 For some object file formats, the debugging information is encapsulated
140 in assembler directives known collectively as @dfn{stab} (symbol table)
141 directives, which are interspersed with the generated code. Stabs are
142 the native format for debugging information in the a.out and XCOFF
143 object file formats. The GNU tools can also emit stabs in the COFF and
144 ECOFF object file formats.
146 The assembler adds the information from stabs to the symbol information
147 it places by default in the symbol table and the string table of the
148 @file{.o} file it is building. The linker consolidates the @file{.o}
149 files into one executable file, with one symbol table and one string
150 table. Debuggers use the symbol and string tables in the executable as
151 a source of debugging information about the program.
154 @section Overview of Stab Format
156 There are three overall formats for stab assembler directives,
157 differentiated by the first word of the stab. The name of the directive
158 describes which combination of four possible data fields follows. It is
159 either @code{.stabs} (string), @code{.stabn} (number), or @code{.stabd}
160 (dot). IBM's XCOFF assembler uses @code{.stabx} (and some other
161 directives such as @code{.file} and @code{.bi}) instead of
162 @code{.stabs}, @code{.stabn} or @code{.stabd}.
164 The overall format of each class of stab is:
167 .stabs "@var{string}",@var{type},@var{other},@var{desc},@var{value}
168 .stabn @var{type},@var{other},@var{desc},@var{value}
169 .stabd @var{type},@var{other},@var{desc}
170 .stabx "@var{string}",@var{value},@var{type},@var{sdb-type}
173 @c what is the correct term for "current file location"? My AIX
174 @c assembler manual calls it "the value of the current location counter".
175 For @code{.stabn} and @code{.stabd}, there is no @var{string} (the
176 @code{n_strx} field is zero; see @ref{Symbol Tables}). For
177 @code{.stabd}, the @var{value} field is implicit and has the value of
178 the current file location. For @code{.stabx}, the @var{sdb-type} field
179 is unused for stabs and can always be set to zero. The @var{other}
180 field is almost always unused and can be set to zero.
182 The number in the @var{type} field gives some basic information about
183 which type of stab this is (or whether it @emph{is} a stab, as opposed
184 to an ordinary symbol). Each valid type number defines a different stab
185 type; further, the stab type defines the exact interpretation of, and
186 possible values for, any remaining @var{string}, @var{desc}, or
187 @var{value} fields present in the stab. @xref{Stab Types}, for a list
188 in numeric order of the valid @var{type} field values for stab directives.
191 @section The String Field
193 For most stabs the string field holds the meat of the
194 debugging information. The flexible nature of this field
195 is what makes stabs extensible. For some stab types the string field
196 contains only a name. For other stab types the contents can be a great
199 The overall format of the string field for most stab types is:
202 "@var{name}:@var{symbol-descriptor} @var{type-information}"
205 @var{name} is the name of the symbol represented by the stab.
206 @var{name} can be omitted, which means the stab represents an unnamed
207 object. For example, @samp{:t10=*2} defines type 10 as a pointer to
208 type 2, but does not give the type a name. Omitting the @var{name}
209 field is supported by AIX dbx and GDB after about version 4.8, but not
210 other debuggers. GCC sometimes uses a single space as the name instead
211 of omitting the name altogether; apparently that is supported by most
214 The @var{symbol-descriptor} following the @samp{:} is an alphabetic
215 character that tells more specifically what kind of symbol the stab
216 represents. If the @var{symbol-descriptor} is omitted, but type
217 information follows, then the stab represents a local variable. For a
218 list of symbol descriptors, see @ref{Symbol Descriptors}. The @samp{c}
219 symbol descriptor is an exception in that it is not followed by type
220 information. @xref{Constants}.
222 @var{type-information} is either a @var{type-number}, or
223 @samp{@var{type-number}=}. A @var{type-number} alone is a type
224 reference, referring directly to a type that has already been defined.
226 The @samp{@var{type-number}=} form is a type definition, where the
227 number represents a new type which is about to be defined. The type
228 definition may refer to other types by number, and those type numbers
229 may be followed by @samp{=} and nested definitions.
231 In a type definition, if the character that follows the equals sign is
232 non-numeric then it is a @var{type-descriptor}, and tells what kind of
233 type is about to be defined. Any other values following the
234 @var{type-descriptor} vary, depending on the @var{type-descriptor}.
235 @xref{Type Descriptors}, for a list of @var{type-descriptor} values. If
236 a number follows the @samp{=} then the number is a @var{type-reference}.
237 For a full description of types, @ref{Types}.
239 There is an AIX extension for type attributes. Following the @samp{=}
240 are any number of type attributes. Each one starts with @samp{@@} and
241 ends with @samp{;}. Debuggers, including AIX's dbx and GDB 4.10, skip
242 any type attributes they do not recognize. GDB 4.9 and other versions
243 of dbx may not do this. Because of a conflict with C++
244 (@pxref{Cplusplus}), new attributes should not be defined which begin
245 with a digit, @samp{(}, or @samp{-}; GDB may be unable to distinguish
246 those from the C++ type descriptor @samp{@@}. The attributes are:
249 @item a@var{boundary}
250 @var{boundary} is an integer specifying the alignment. I assume it
251 applies to all variables of this type.
254 Size in bits of a variable of this type.
257 Pointer class (for checking). Not sure what this means, or how
258 @var{integer} is interpreted.
261 Indicate this is a packed type, meaning that structure fields or array
262 elements are placed more closely in memory, to save memory at the
266 All of this can make the string field quite long. All
267 versions of GDB, and some versions of dbx, can handle arbitrarily long
268 strings. But many versions of dbx cretinously limit the strings to
269 about 80 characters, so compilers which must work with such dbx's need
270 to split the @code{.stabs} directive into several @code{.stabs}
271 directives. Each stab duplicates exactly all but the
272 string field. The string field of
273 every stab except the last is marked as continued with a
274 double-backslash at the end. Removing the backslashes and concatenating
275 the string fields of each stab produces the original,
279 @section A Simple Example in C Source
281 To get the flavor of how stabs describe source information for a C
282 program, let's look at the simple program:
287 printf("Hello world");
291 When compiled with @samp{-g}, the program above yields the following
292 @file{.s} file. Line numbers have been added to make it easier to refer
293 to parts of the @file{.s} file in the description of the stabs that
297 @section The Simple Example at the Assembly Level
299 This simple ``hello world'' example demonstrates several of the stab
300 types used to describe C language source files.
304 2 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0
305 3 .stabs "hello.c",100,0,0,Ltext0
308 6 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
309 7 .stabs "char:t2=r2;0;127;",128,0,0,0
310 8 .stabs "long int:t3=r1;-2147483648;2147483647;",128,0,0,0
311 9 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
312 10 .stabs "long unsigned int:t5=r1;0;-1;",128,0,0,0
313 11 .stabs "short int:t6=r1;-32768;32767;",128,0,0,0
314 12 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
315 13 .stabs "short unsigned int:t8=r1;0;65535;",128,0,0,0
316 14 .stabs "long long unsigned int:t9=r1;0;-1;",128,0,0,0
317 15 .stabs "signed char:t10=r1;-128;127;",128,0,0,0
318 16 .stabs "unsigned char:t11=r1;0;255;",128,0,0,0
319 17 .stabs "float:t12=r1;4;0;",128,0,0,0
320 18 .stabs "double:t13=r1;8;0;",128,0,0,0
321 19 .stabs "long double:t14=r1;8;0;",128,0,0,0
322 20 .stabs "void:t15=15",128,0,0,0
325 23 .ascii "Hello, world!\12\0"
340 38 sethi %hi(LC0),%o1
341 39 or %o1,%lo(LC0),%o0
352 50 .stabs "main:F1",36,0,0,_main
353 51 .stabn 192,0,0,LBB2
354 52 .stabn 224,0,0,LBE2
357 @node Program Structure
358 @chapter Encoding the Structure of the Program
360 The elements of the program structure that stabs encode include the name
361 of the main function, the names of the source and include files, the
362 line numbers, procedure names and types, and the beginnings and ends of
366 * Main Program:: Indicate what the main program is
367 * Source Files:: The path and name of the source file
368 * Include Files:: Names of include files
371 * Nested Procedures::
376 @section Main Program
379 Most languages allow the main program to have any name. The
380 @code{N_MAIN} stab type tells the debugger the name that is used in this
381 program. Only the string field is significant; it is the name of
382 a function which is the main program. Most C compilers do not use this
383 stab (they expect the debugger to assume that the name is @code{main}),
384 but some C compilers emit an @code{N_MAIN} stab for the @code{main}
388 @section Paths and Names of the Source Files
391 Before any other stabs occur, there must be a stab specifying the source
392 file. This information is contained in a symbol of stab type
393 @code{N_SO}; the string field contains the name of the file. The
394 value of the symbol is the start address of the portion of the
395 text section corresponding to that file.
397 With the Sun Solaris2 compiler, the desc field contains a
398 source-language code.
399 @c Do the debuggers use it? What are the codes? -djm
401 Some compilers (for example, GCC2 and SunOS4 @file{/bin/cc}) also
402 include the directory in which the source was compiled, in a second
403 @code{N_SO} symbol preceding the one containing the file name. This
404 symbol can be distinguished by the fact that it ends in a slash. Code
405 from the @code{cfront} C++ compiler can have additional @code{N_SO} symbols for
406 nonexistent source files after the @code{N_SO} for the real source file;
407 these are believed to contain no useful information.
412 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0 # @r{100 is N_SO}
413 .stabs "hello.c",100,0,0,Ltext0
418 Instead of @code{N_SO} symbols, XCOFF uses a @code{.file} assembler
419 directive which assembles to a standard COFF @code{.file} symbol;
420 explaining this in detail is outside the scope of this document.
423 @section Names of Include Files
425 There are several schemes for dealing with include files: the
426 traditional @code{N_SOL} approach, Sun's @code{N_BINCL} approach, and the
427 XCOFF @code{C_BINCL} approach (which despite the similar name has little in
428 common with @code{N_BINCL}).
431 An @code{N_SOL} symbol specifies which include file subsequent symbols
432 refer to. The string field is the name of the file and the
433 value is the text address corresponding to the start of the
434 previous include file and the start of this one. To specify the main
435 source file again, use an @code{N_SOL} symbol with the name of the main
441 The @code{N_BINCL} approach works as follows. An @code{N_BINCL} symbol
442 specifies the start of an include file. In an object file, only the
443 string is significant; the Sun linker puts data into some of the
444 other fields. The end of the include file is marked by an
445 @code{N_EINCL} symbol (which has no string field). In an object
446 file, there is no significant data in the @code{N_EINCL} symbol; the Sun
447 linker puts data into some of the fields. @code{N_BINCL} and
448 @code{N_EINCL} can be nested.
450 If the linker detects that two source files have identical stabs between
451 an @code{N_BINCL} and @code{N_EINCL} pair (as will generally be the case
452 for a header file), then it only puts out the stabs once. Each
453 additional occurance is replaced by an @code{N_EXCL} symbol. I believe
454 the Sun (SunOS4, not sure about Solaris) linker is the only one which
455 supports this feature.
456 @c What do the fields of N_EXCL contain? -djm
460 For the start of an include file in XCOFF, use the @file{.bi} assembler
461 directive, which generates a @code{C_BINCL} symbol. A @file{.ei}
462 directive, which generates a @code{C_EINCL} symbol, denotes the end of
463 the include file. Both directives are followed by the name of the
464 source file in quotes, which becomes the string for the symbol.
465 The value of each symbol, produced automatically by the assembler
466 and linker, is the offset into the executable of the beginning
467 (inclusive, as you'd expect) or end (inclusive, as you would not expect)
468 of the portion of the COFF line table that corresponds to this include
469 file. @code{C_BINCL} and @code{C_EINCL} do not nest.
472 @section Line Numbers
475 An @code{N_SLINE} symbol represents the start of a source line. The
476 desc field contains the line number and the value
477 contains the code address for the start of that source line. On most
478 machines the address is absolute; for Sun's stabs-in-ELF, it is relative
479 to the function in which the @code{N_SLINE} symbol occurs.
483 GNU documents @code{N_DSLINE} and @code{N_BSLINE} symbols for line
484 numbers in the data or bss segments, respectively. They are identical
485 to @code{N_SLINE} but are relocated differently by the linker. They
486 were intended to be used to describe the source location of a variable
487 declaration, but I believe that GCC2 actually puts the line number in
488 the desc field of the stab for the variable itself. GDB has been
489 ignoring these symbols (unless they contain a string field) since
492 For single source lines that generate discontiguous code, such as flow
493 of control statements, there may be more than one line number entry for
494 the same source line. In this case there is a line number entry at the
495 start of each code range, each with the same line number.
497 XCOFF does not use stabs for line numbers. Instead, it uses COFF line
498 numbers (which are outside the scope of this document). Standard COFF
499 line numbers cannot deal with include files, but in XCOFF this is fixed
500 with the @code{C_BINCL} method of marking include files (@pxref{Include
506 @findex N_FUN, for functions
508 @findex N_STSYM, for functions (Sun acc)
509 @findex N_GSYM, for functions (Sun acc)
510 All of the following stabs normally use the @code{N_FUN} symbol type.
511 However, Sun's @code{acc} compiler on SunOS4 uses @code{N_GSYM} and
512 @code{N_STSYM}, which means that the value of the stab for the function
513 is useless and the debugger must get the address of the function from
514 the non-stab symbols instead. BSD Fortran is said to use @code{N_FNAME}
515 with the same restriction; the value of the symbol is not useful (I'm
516 not sure it really does use this, because GDB doesn't handle this and no
519 A function is represented by an @samp{F} symbol descriptor for a global
520 (extern) function, and @samp{f} for a static (local) function. The
521 value is the address of the start of the function (absolute
522 for @code{a.out}; relative to the start of the file for Sun's
523 stabs-in-ELF). The type information of the stab represents the return
524 type of the function; thus @samp{foo:f5} means that foo is a function
525 returning type 5. There is no need to try to get the line number of the
526 start of the function from the stab for the function; it is in the next
527 @code{N_SLINE} symbol.
529 @c FIXME: verify whether the "I suspect" below is true or not.
530 Some compilers (such as Sun's Solaris compiler) support an extension for
531 specifying the types of the arguments. I suspect this extension is not
532 used for old (non-prototyped) function definitions in C. If the
533 extension is in use, the type information of the stab for the function
534 is followed by type information for each argument, with each argument
535 preceded by @samp{;}. An argument type of 0 means that additional
536 arguments are being passed, whose types and number may vary (@samp{...}
537 in ANSI C). GDB has tolerated this extension (parsed the syntax, if not
538 necessarily used the information) since at least version 4.8; I don't
539 know whether all versions of dbx tolerate it. The argument types given
540 here are not redundant with the symbols for the formal parameters
541 (@pxref{Parameters}); they are the types of the arguments as they are
542 passed, before any conversions might take place. For example, if a C
543 function which is declared without a prototype takes a @code{float}
544 argument, the value is passed as a @code{double} but then converted to a
545 @code{float}. Debuggers need to use the types given in the arguments
546 when printing values, but when calling the function they need to use the
547 types given in the symbol defining the function.
549 If the return type and types of arguments of a function which is defined
550 in another source file are specified (i.e., a function prototype in ANSI
551 C), traditionally compilers emit no stab; the only way for the debugger
552 to find the information is if the source file where the function is
553 defined was also compiled with debugging symbols. As an extension the
554 Solaris compiler uses symbol descriptor @samp{P} followed by the return
555 type of the function, followed by the arguments, each preceded by
556 @samp{;}, as in a stab with symbol descriptor @samp{f} or @samp{F}.
557 This use of symbol descriptor @samp{P} can be distinguished from its use
558 for register parameters (@pxref{Register Parameters}) by the fact that it has
559 symbol type @code{N_FUN}.
561 The AIX documentation also defines symbol descriptor @samp{J} as an
562 internal function. I assume this means a function nested within another
563 function. It also says symbol descriptor @samp{m} is a module in
564 Modula-2 or extended Pascal.
566 Procedures (functions which do not return values) are represented as
567 functions returning the @code{void} type in C. I don't see why this couldn't
568 be used for all languages (inventing a @code{void} type for this purpose if
569 necessary), but the AIX documentation defines @samp{I}, @samp{P}, and
570 @samp{Q} for internal, global, and static procedures, respectively.
571 These symbol descriptors are unusual in that they are not followed by
574 The following example shows a stab for a function @code{main} which
575 returns type number @code{1}. The @code{_main} specified for the value
576 is a reference to an assembler label which is used to fill in the start
577 address of the function.
580 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
583 The stab representing a procedure is located immediately following the
584 code of the procedure. This stab is in turn directly followed by a
585 group of other stabs describing elements of the procedure. These other
586 stabs describe the procedure's parameters, its block local variables, and
589 @node Nested Procedures
590 @section Nested Procedures
592 For any of the symbol descriptors representing procedures, after the
593 symbol descriptor and the type information is optionally a scope
594 specifier. This consists of a comma, the name of the procedure, another
595 comma, and the name of the enclosing procedure. The first name is local
596 to the scope specified, and seems to be redundant with the name of the
597 symbol (before the @samp{:}). This feature is used by GCC, and
598 presumably Pascal, Modula-2, etc., compilers, for nested functions.
600 If procedures are nested more than one level deep, only the immediately
601 containing scope is specified. For example, this code:
613 return baz (x + 2 * y);
615 return x + bar (3 * x);
623 .stabs "baz:f1,baz,bar",36,0,0,_baz.15 # @r{36 is N_FUN}
624 .stabs "bar:f1,bar,foo",36,0,0,_bar.12
625 .stabs "foo:F1",36,0,0,_foo
628 @node Block Structure
629 @section Block Structure
633 The program's block structure is represented by the @code{N_LBRAC} (left
634 brace) and the @code{N_RBRAC} (right brace) stab types. The variables
635 defined inside a block precede the @code{N_LBRAC} symbol for most
636 compilers, including GCC. Other compilers, such as the Convex, Acorn
637 RISC machine, and Sun @code{acc} compilers, put the variables after the
638 @code{N_LBRAC} symbol. The values of the @code{N_LBRAC} and
639 @code{N_RBRAC} symbols are the start and end addresses of the code of
640 the block, respectively. For most machines, they are relative to the
641 starting address of this source file. For the Gould NP1, they are
642 absolute. For Sun's stabs-in-ELF, they are relative to the function in
645 The @code{N_LBRAC} and @code{N_RBRAC} stabs that describe the block
646 scope of a procedure are located after the @code{N_FUN} stab that
647 represents the procedure itself.
649 Sun documents the desc field of @code{N_LBRAC} and
650 @code{N_RBRAC} symbols as containing the nesting level of the block.
651 However, dbx seems to not care, and GCC always sets desc to
657 The @samp{c} symbol descriptor indicates that this stab represents a
658 constant. This symbol descriptor is an exception to the general rule
659 that symbol descriptors are followed by type information. Instead, it
660 is followed by @samp{=} and one of the following:
664 Boolean constant. @var{value} is a numeric value; I assume it is 0 for
668 Character constant. @var{value} is the numeric value of the constant.
670 @item e @var{type-information} , @var{value}
671 Constant whose value can be represented as integral.
672 @var{type-information} is the type of the constant, as it would appear
673 after a symbol descriptor (@pxref{String Field}). @var{value} is the
674 numeric value of the constant. GDB 4.9 does not actually get the right
675 value if @var{value} does not fit in a host @code{int}, but it does not
676 do anything violent, and future debuggers could be extended to accept
677 integers of any size (whether unsigned or not). This constant type is
678 usually documented as being only for enumeration constants, but GDB has
679 never imposed that restriction; I don't know about other debuggers.
682 Integer constant. @var{value} is the numeric value. The type is some
683 sort of generic integer type (for GDB, a host @code{int}); to specify
684 the type explicitly, use @samp{e} instead.
687 Real constant. @var{value} is the real value, which can be @samp{INF}
688 (optionally preceded by a sign) for infinity, @samp{QNAN} for a quiet
689 NaN (not-a-number), or @samp{SNAN} for a signalling NaN. If it is a
690 normal number the format is that accepted by the C library function
694 String constant. @var{string} is a string enclosed in either @samp{'}
695 (in which case @samp{'} characters within the string are represented as
696 @samp{\'} or @samp{"} (in which case @samp{"} characters within the
697 string are represented as @samp{\"}).
699 @item S @var{type-information} , @var{elements} , @var{bits} , @var{pattern}
700 Set constant. @var{type-information} is the type of the constant, as it
701 would appear after a symbol descriptor (@pxref{String Field}).
702 @var{elements} is the number of elements in the set (does this means
703 how many bits of @var{pattern} are actually used, which would be
704 redundant with the type, or perhaps the number of bits set in
705 @var{pattern}? I don't get it), @var{bits} is the number of bits in the
706 constant (meaning it specifies the length of @var{pattern}, I think),
707 and @var{pattern} is a hexadecimal representation of the set. AIX
708 documentation refers to a limit of 32 bytes, but I see no reason why
709 this limit should exist. This form could probably be used for arbitrary
710 constants, not just sets; the only catch is that @var{pattern} should be
711 understood to be target, not host, byte order and format.
714 The boolean, character, string, and set constants are not supported by
715 GDB 4.9, but it ignores them. GDB 4.8 and earlier gave an error
716 message and refused to read symbols from the file containing the
719 The above information is followed by @samp{;}.
724 Different types of stabs describe the various ways that variables can be
725 allocated: on the stack, globally, in registers, in common blocks,
726 statically, or as arguments to a function.
729 * Stack Variables:: Variables allocated on the stack.
730 * Global Variables:: Variables used by more than one source file.
731 * Register Variables:: Variables in registers.
732 * Common Blocks:: Variables statically allocated together.
733 * Statics:: Variables local to one source file.
734 * Based Variables:: Fortran pointer based variables.
735 * Parameters:: Variables for arguments to functions.
738 @node Stack Variables
739 @section Automatic Variables Allocated on the Stack
741 If a variable's scope is local to a function and its lifetime is only as
742 long as that function executes (C calls such variables
743 @dfn{automatic}), it can be allocated in a register (@pxref{Register
744 Variables}) or on the stack.
747 Each variable allocated on the stack has a stab with the symbol
748 descriptor omitted. Since type information should begin with a digit,
749 @samp{-}, or @samp{(}, only those characters precluded from being used
750 for symbol descriptors. However, the Acorn RISC machine (ARM) is said
751 to get this wrong: it puts out a mere type definition here, without the
752 preceding @samp{@var{type-number}=}. This is a bad idea; there is no
753 guarantee that type descriptors are distinct from symbol descriptors.
754 Stabs for stack variables use the @code{N_LSYM} stab type.
756 The value of the stab is the offset of the variable within the
757 local variables. On most machines this is an offset from the frame
758 pointer and is negative. The location of the stab specifies which block
759 it is defined in; see @ref{Block Structure}.
761 For example, the following C code:
771 produces the following stabs:
774 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
775 .stabs "x:1",128,0,0,-12 # @r{128 is N_LSYM}
776 .stabn 192,0,0,LBB2 # @r{192 is N_LBRAC}
777 .stabn 224,0,0,LBE2 # @r{224 is N_RBRAC}
780 @xref{Procedures} for more information on the @code{N_FUN} stab, and
781 @ref{Block Structure} for more information on the @code{N_LBRAC} and
782 @code{N_RBRAC} stabs.
784 @node Global Variables
785 @section Global Variables
788 A variable whose scope is not specific to just one source file is
789 represented by the @samp{G} symbol descriptor. These stabs use the
790 @code{N_GSYM} stab type. The type information for the stab
791 (@pxref{String Field}) gives the type of the variable.
793 For example, the following source code:
800 yields the following assembly code:
803 .stabs "g_foo:G2",32,0,0,0 # @r{32 is N_GSYM}
810 The address of the variable represented by the @code{N_GSYM} is not
811 contained in the @code{N_GSYM} stab. The debugger gets this information
812 from the external symbol for the global variable. In the example above,
813 the @code{.global _g_foo} and @code{_g_foo:} lines tell the assembler to
814 produce an external symbol.
816 @node Register Variables
817 @section Register Variables
820 @c According to an old version of this manual, AIX uses C_RPSYM instead
821 @c of C_RSYM. I am skeptical; this should be verified.
822 Register variables have their own stab type, @code{N_RSYM}, and their
823 own symbol descriptor, @samp{r}. The stab's value is the
824 number of the register where the variable data will be stored.
825 @c .stabs "name:type",N_RSYM,0,RegSize,RegNumber (Sun doc)
827 AIX defines a separate symbol descriptor @samp{d} for floating point
828 registers. This seems unnecessary; why not just just give floating
829 point registers different register numbers? I have not verified whether
830 the compiler actually uses @samp{d}.
832 If the register is explicitly allocated to a global variable, but not
836 register int g_bar asm ("%g5");
840 then the stab may be emitted at the end of the object file, with
841 the other bss symbols.
844 @section Common Blocks
846 A common block is a statically allocated section of memory which can be
847 referred to by several source files. It may contain several variables.
848 I believe Fortran is the only language with this feature.
852 A @code{N_BCOMM} stab begins a common block and an @code{N_ECOMM} stab
853 ends it. The only field that is significant in these two stabs is the
854 string, which names a normal (non-debugging) symbol that gives the
855 address of the common block.
858 Each stab between the @code{N_BCOMM} and the @code{N_ECOMM} specifies a
859 member of that common block; its value is the offset within the
860 common block of that variable. The @code{N_ECOML} stab type is
861 documented for this purpose, but Sun's Fortran compiler uses
862 @code{N_GSYM} instead. The test case I looked at had a common block
863 local to a function and it used the @samp{V} symbol descriptor; I assume
864 one would use @samp{S} if not local to a function (that is, if a common
865 block @emph{can} be anything other than local to a function).
868 @section Static Variables
870 Initialized static variables are represented by the @samp{S} and
871 @samp{V} symbol descriptors. @samp{S} means file scope static, and
872 @samp{V} means procedure scope static.
874 @c This is probably not worth mentioning; it is only true on the sparc
875 @c for `double' variables which although declared const are actually in
876 @c the data segment (the text segment can't guarantee 8 byte alignment).
878 @c 2.4.5 has a bug in that it uses @code{N_FUN}, so neither dbx nor GDB can
879 @c find the variables)
882 @findex N_FUN, for variables
884 In a.out files, @code{N_STSYM} means the data section, @code{N_FUN}
885 means the text section, and @code{N_LCSYM} means the bss section. For
886 those systems with a read-only data section separate from the text
887 section (Solaris), @code{N_ROSYM} means the read-only data section.
889 For example, the source lines:
892 static const int var_const = 5;
893 static int var_init = 2;
894 static int var_noinit;
898 yield the following stabs:
901 .stabs "var_const:S1",36,0,0,_var_const # @r{36 is N_FUN}
903 .stabs "var_init:S1",38,0,0,_var_init # @r{38 is N_STSYM}
905 .stabs "var_noinit:S1",40,0,0,_var_noinit # @r{40 is N_LCSYM}
908 In XCOFF files, each symbol has a section number, so the stab type
909 need not indicate the section.
911 In ECOFF files, the storage class is used to specify the section, so the
912 stab type need not indicate the section.
914 In ELF files, for Solaris 2.1, symbol descriptor @samp{S} means that the
915 address is absolute (ld relocates it) and symbol descriptor @samp{V}
916 means that the address is relative to the start of the relevant section
917 for that compilation unit. I don't know what it does for @samp{S} stabs
918 on Solaris 2.3 (in which ld no longer relocates stabs). For more
919 information on ld stab relocation, @xref{Stabs In ELF}.
921 @node Based Variables
922 @section Fortran Based Variables
924 Fortran (at least, the Sun and SGI dialects of FORTRAN-77) has a feature
925 which allows allocating arrays with @code{malloc}, but which avoids
926 blurring the line between arrays and pointers the way that C does. In
927 stabs such a variable uses the @samp{b} symbol descriptor.
929 For example, the Fortran declarations
932 real foo, foo10(10), foo10_5(10,5)
934 pointer (foo10p, foo10)
935 pointer (foo105p, foo10_5)
943 foo10_5:bar3;1;5;ar3;1;10;6
946 In this example, @code{real} is type 6 and type 3 is an integral type
947 which is the type of the subscripts of the array (probably
950 The @samp{b} symbol descriptor is like @samp{V} in that it denotes a
951 statically allocated symbol whose scope is local to a function; see
952 @xref{Statics}. The value of the symbol, instead of being the address
953 of the variable itself, is the address of a pointer to that variable.
954 So in the above example, the value of the @code{foo} stab is the address
955 of a pointer to a real, the value of the @code{foo10} stab is the
956 address of a pointer to a 10-element array of reals, and the value of
957 the @code{foo10_5} stab is the address of a pointer to a 5-element array
958 of 10-element arrays of reals.
963 Formal parameters to a function are represented by a stab (or sometimes
964 two; see below) for each parameter. The stabs are in the order in which
965 the debugger should print the parameters (i.e., the order in which the
966 parameters are declared in the source file). The exact form of the stab
967 depends on how the parameter is being passed.
970 Parameters passed on the stack use the symbol descriptor @samp{p} and
971 the @code{N_PSYM} symbol type. The value of the symbol is an offset
972 used to locate the parameter on the stack; its exact meaning is
973 machine-dependent, but on most machines it is an offset from the frame
976 As a simple example, the code:
987 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
988 .stabs "argc:p1",160,0,0,68 # @r{160 is N_PSYM}
989 .stabs "argv:p20=*21=*2",160,0,0,72
992 The type definition of @code{argv} is interesting because it contains
993 several type definitions. Type 21 is pointer to type 2 (char) and
994 @code{argv} (type 20) is pointer to type 21.
996 @c FIXME: figure out what these mean and describe them coherently.
997 The following symbol descriptors are also said to go with @code{N_PSYM}.
998 The value of the symbol is said to be an offset from the argument
999 pointer (I'm not sure whether this is true or not).
1003 pF Fortran function parameter
1004 X (function result variable)
1008 * Register Parameters::
1009 * Local Variable Parameters::
1010 * Reference Parameters::
1011 * Conformant Arrays::
1014 @node Register Parameters
1015 @subsection Passing Parameters in Registers
1017 If the parameter is passed in a register, then traditionally there are
1018 two symbols for each argument:
1021 .stabs "arg:p1" . . . ; N_PSYM
1022 .stabs "arg:r1" . . . ; N_RSYM
1025 Debuggers use the second one to find the value, and the first one to
1026 know that it is an argument.
1029 @findex N_RSYM, for parameters
1030 Because that approach is kind of ugly, some compilers use symbol
1031 descriptor @samp{P} or @samp{R} to indicate an argument which is in a
1032 register. Symbol type @code{C_RPSYM} is used with @samp{R} and
1033 @code{N_RSYM} is used with @samp{P}. The symbol's value is
1034 the register number. @samp{P} and @samp{R} mean the same thing; the
1035 difference is that @samp{P} is a GNU invention and @samp{R} is an IBM
1036 (XCOFF) invention. As of version 4.9, GDB should handle either one.
1038 There is at least one case where GCC uses a @samp{p} and @samp{r} pair
1039 rather than @samp{P}; this is where the argument is passed in the
1040 argument list and then loaded into a register.
1042 According to the AIX documentation, symbol descriptor @samp{D} is for a
1043 parameter passed in a floating point register. This seems
1044 unnecessary---why not just use @samp{R} with a register number which
1045 indicates that it's a floating point register? I haven't verified
1046 whether the system actually does what the documentation indicates.
1048 @c FIXME: On the hppa this is for any type > 8 bytes, I think, and not
1049 @c for small structures (investigate).
1050 On the sparc and hppa, for a @samp{P} symbol whose type is a structure
1051 or union, the register contains the address of the structure. On the
1052 sparc, this is also true of a @samp{p} and @samp{r} pair (using Sun
1053 @code{cc}) or a @samp{p} symbol. However, if a (small) structure is
1054 really in a register, @samp{r} is used. And, to top it all off, on the
1055 hppa it might be a structure which was passed on the stack and loaded
1056 into a register and for which there is a @samp{p} and @samp{r} pair! I
1057 believe that symbol descriptor @samp{i} is supposed to deal with this
1058 case (it is said to mean "value parameter by reference, indirect
1059 access"; I don't know the source for this information), but I don't know
1060 details or what compilers or debuggers use it, if any (not GDB or GCC).
1061 It is not clear to me whether this case needs to be dealt with
1062 differently than parameters passed by reference (@pxref{Reference Parameters}).
1064 @node Local Variable Parameters
1065 @subsection Storing Parameters as Local Variables
1067 There is a case similar to an argument in a register, which is an
1068 argument that is actually stored as a local variable. Sometimes this
1069 happens when the argument was passed in a register and then the compiler
1070 stores it as a local variable. If possible, the compiler should claim
1071 that it's in a register, but this isn't always done.
1073 @findex N_LSYM, for parameter
1074 Some compilers use the pair of symbols approach described above
1075 (@samp{@var{arg}:p} followed by @samp{@var{arg}:}); this includes GCC1
1076 (not GCC2) on the sparc when passing a small structure and GCC2
1077 (sometimes) when the argument type is @code{float} and it is passed as a
1078 @code{double} and converted to @code{float} by the prologue (in the
1079 latter case the type of the @samp{@var{arg}:p} symbol is @code{double}
1080 and the type of the @samp{@var{arg}:} symbol is @code{float}).
1082 GCC, at least on the 960, has another solution to the same problem. It
1083 uses a single @samp{p} symbol descriptor for an argument which is stored
1084 as a local variable but uses @code{N_LSYM} instead of @code{N_PSYM}. In
1085 this case, the value of the symbol is an offset relative to the local
1086 variables for that function, not relative to the arguments; on some
1087 machines those are the same thing, but not on all.
1089 @c This is mostly just background info; the part that logically belongs
1090 @c here is the last sentence.
1091 On the VAX or on other machines in which the calling convention includes
1092 the number of words of arguments actually passed, the debugger (GDB at
1093 least) uses the parameter symbols to keep track of whether it needs to
1094 print nameless arguments in addition to the formal parameters which it
1095 has printed because each one has a stab. For example, in
1098 extern int fprintf (FILE *stream, char *format, @dots{});
1100 fprintf (stdout, "%d\n", x);
1103 there are stabs for @code{stream} and @code{format}. On most machines,
1104 the debugger can only print those two arguments (because it has no way
1105 of knowing that additional arguments were passed), but on the VAX or
1106 other machines with a calling convention which indicates the number of
1107 words of arguments, the debugger can print all three arguments. To do
1108 so, the parameter symbol (symbol descriptor @samp{p}) (not necessarily
1109 @samp{r} or symbol descriptor omitted symbols) needs to contain the
1110 actual type as passed (for example, @code{double} not @code{float} if it
1111 is passed as a double and converted to a float).
1113 @node Reference Parameters
1114 @subsection Passing Parameters by Reference
1116 If the parameter is passed by reference (e.g., Pascal @code{VAR}
1117 parameters), then the symbol descriptor is @samp{v} if it is in the
1118 argument list, or @samp{a} if it in a register. Other than the fact
1119 that these contain the address of the parameter rather than the
1120 parameter itself, they are identical to @samp{p} and @samp{R},
1121 respectively. I believe @samp{a} is an AIX invention; @samp{v} is
1122 supported by all stabs-using systems as far as I know.
1124 @node Conformant Arrays
1125 @subsection Passing Conformant Array Parameters
1127 @c Is this paragraph correct? It is based on piecing together patchy
1128 @c information and some guesswork
1129 Conformant arrays are a feature of Modula-2, and perhaps other
1130 languages, in which the size of an array parameter is not known to the
1131 called function until run-time. Such parameters have two stabs: a
1132 @samp{x} for the array itself, and a @samp{C}, which represents the size
1133 of the array. The value of the @samp{x} stab is the offset in the
1134 argument list where the address of the array is stored (it this right?
1135 it is a guess); the value of the @samp{C} stab is the offset in the
1136 argument list where the size of the array (in elements? in bytes?) is
1140 @chapter Defining Types
1142 The examples so far have described types as references to previously
1143 defined types, or defined in terms of subranges of or pointers to
1144 previously defined types. This chapter describes the other type
1145 descriptors that may follow the @samp{=} in a type definition.
1148 * Builtin Types:: Integers, floating point, void, etc.
1149 * Miscellaneous Types:: Pointers, sets, files, etc.
1150 * Cross-References:: Referring to a type not yet defined.
1151 * Subranges:: A type with a specific range.
1152 * Arrays:: An aggregate type of same-typed elements.
1153 * Strings:: Like an array but also has a length.
1154 * Enumerations:: Like an integer but the values have names.
1155 * Structures:: An aggregate type of different-typed elements.
1156 * Typedefs:: Giving a type a name.
1157 * Unions:: Different types sharing storage.
1162 @section Builtin Types
1164 Certain types are built in (@code{int}, @code{short}, @code{void},
1165 @code{float}, etc.); the debugger recognizes these types and knows how
1166 to handle them. Thus, don't be surprised if some of the following ways
1167 of specifying builtin types do not specify everything that a debugger
1168 would need to know about the type---in some cases they merely specify
1169 enough information to distinguish the type from other types.
1171 The traditional way to define builtin types is convolunted, so new ways
1172 have been invented to describe them. Sun's @code{acc} uses special
1173 builtin type descriptors (@samp{b} and @samp{R}), and IBM uses negative
1174 type numbers. GDB accepts all three ways, as of version 4.8; dbx just
1175 accepts the traditional builtin types and perhaps one of the other two
1176 formats. The following sections describe each of these formats.
1179 * Traditional Builtin Types:: Put on your seatbelts and prepare for kludgery
1180 * Builtin Type Descriptors:: Builtin types with special type descriptors
1181 * Negative Type Numbers:: Builtin types using negative type numbers
1184 @node Traditional Builtin Types
1185 @subsection Traditional Builtin Types
1187 This is the traditional, convoluted method for defining builtin types.
1188 There are several classes of such type definitions: integer, floating
1189 point, and @code{void}.
1192 * Traditional Integer Types::
1193 * Traditional Other Types::
1196 @node Traditional Integer Types
1197 @subsubsection Traditional Integer Types
1199 Often types are defined as subranges of themselves. If the bounding values
1200 fit within an @code{int}, then they are given normally. For example:
1203 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0 # @r{128 is N_LSYM}
1204 .stabs "char:t2=r2;0;127;",128,0,0,0
1207 Builtin types can also be described as subranges of @code{int}:
1210 .stabs "unsigned short:t6=r1;0;65535;",128,0,0,0
1213 If the lower bound of a subrange is 0 and the upper bound is -1,
1214 the type is an unsigned integral type whose bounds are too
1215 big to describe in an @code{int}. Traditionally this is only used for
1216 @code{unsigned int} and @code{unsigned long}:
1219 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
1222 For larger types, GCC 2.4.5 puts out bounds in octal, with a leading 0.
1223 In this case a negative bound consists of a number which is a 1 bit
1224 followed by a bunch of 0 bits, and a positive bound is one in which a
1225 bunch of bits are 1. All known versions of dbx and GDB version 4 accept
1226 this, but GDB 3.5 refuses to read the whole file containing such
1227 symbols. So GCC 2.3.3 did not output the proper size for these types.
1228 @c FIXME: How about an example?
1230 If the lower bound of a subrange is 0 and the upper bound is negative,
1231 the type is an unsigned integral type whose size in bytes is the
1232 absolute value of the upper bound. I believe this is a Convex
1233 convention for @code{unsigned long long}.
1235 If the lower bound of a subrange is negative and the upper bound is 0,
1236 the type is a signed integral type whose size in bytes is
1237 the absolute value of the lower bound. I believe this is a Convex
1238 convention for @code{long long}. To distinguish this from a legitimate
1239 subrange, the type should be a subrange of itself. I'm not sure whether
1240 this is the case for Convex.
1242 @node Traditional Other Types
1243 @subsubsection Traditional Other Types
1245 If the upper bound of a subrange is 0 and the lower bound is positive,
1246 the type is a floating point type, and the lower bound of the subrange
1247 indicates the number of bytes in the type:
1250 .stabs "float:t12=r1;4;0;",128,0,0,0
1251 .stabs "double:t13=r1;8;0;",128,0,0,0
1254 However, GCC writes @code{long double} the same way it writes
1255 @code{double}, so there is no way to distinguish.
1258 .stabs "long double:t14=r1;8;0;",128,0,0,0
1261 Complex types are defined the same way as floating-point types; there is
1262 no way to distinguish a single-precision complex from a double-precision
1263 floating-point type.
1265 The C @code{void} type is defined as itself:
1268 .stabs "void:t15=15",128,0,0,0
1271 I'm not sure how a boolean type is represented.
1273 @node Builtin Type Descriptors
1274 @subsection Defining Builtin Types Using Builtin Type Descriptors
1276 This is the method used by Sun's @code{acc} for defining builtin types.
1277 These are the type descriptors to define builtin types:
1280 @c FIXME: clean up description of width and offset, once we figure out
1282 @item b @var{signed} @var{char-flag} @var{width} ; @var{offset} ; @var{nbits} ;
1283 Define an integral type. @var{signed} is @samp{u} for unsigned or
1284 @samp{s} for signed. @var{char-flag} is @samp{c} which indicates this
1285 is a character type, or is omitted. I assume this is to distinguish an
1286 integral type from a character type of the same size, for example it
1287 might make sense to set it for the C type @code{wchar_t} so the debugger
1288 can print such variables differently (Solaris does not do this). Sun
1289 sets it on the C types @code{signed char} and @code{unsigned char} which
1290 arguably is wrong. @var{width} and @var{offset} appear to be for small
1291 objects stored in larger ones, for example a @code{short} in an
1292 @code{int} register. @var{width} is normally the number of bytes in the
1293 type. @var{offset} seems to always be zero. @var{nbits} is the number
1294 of bits in the type.
1296 Note that type descriptor @samp{b} used for builtin types conflicts with
1297 its use for Pascal space types (@pxref{Miscellaneous Types}); they can
1298 be distinguished because the character following the type descriptor
1299 will be a digit, @samp{(}, or @samp{-} for a Pascal space type, or
1300 @samp{u} or @samp{s} for a builtin type.
1303 Documented by AIX to define a wide character type, but their compiler
1304 actually uses negative type numbers (@pxref{Negative Type Numbers}).
1306 @item R @var{fp-type} ; @var{bytes} ;
1307 Define a floating point type. @var{fp-type} has one of the following values:
1311 IEEE 32-bit (single precision) floating point format.
1314 IEEE 64-bit (double precision) floating point format.
1316 @item 3 (NF_COMPLEX)
1317 @item 4 (NF_COMPLEX16)
1318 @item 5 (NF_COMPLEX32)
1319 @c "GDB source" really means @file{include/aout/stab_gnu.h}, but trying
1320 @c to put that here got an overfull hbox.
1321 These are for complex numbers. A comment in the GDB source describes
1322 them as Fortran @code{complex}, @code{double complex}, and
1323 @code{complex*16}, respectively, but what does that mean? (i.e., Single
1324 precision? Double precison?).
1326 @item 6 (NF_LDOUBLE)
1327 Long double. This should probably only be used for Sun format
1328 @code{long double}, and new codes should be used for other floating
1329 point formats (@code{NF_DOUBLE} can be used if a @code{long double} is
1330 really just an IEEE double, of course).
1333 @var{bytes} is the number of bytes occupied by the type. This allows a
1334 debugger to perform some operations with the type even if it doesn't
1335 understand @var{fp-type}.
1337 @item g @var{type-information} ; @var{nbits}
1338 Documented by AIX to define a floating type, but their compiler actually
1339 uses negative type numbers (@pxref{Negative Type Numbers}).
1341 @item c @var{type-information} ; @var{nbits}
1342 Documented by AIX to define a complex type, but their compiler actually
1343 uses negative type numbers (@pxref{Negative Type Numbers}).
1346 The C @code{void} type is defined as a signed integral type 0 bits long:
1348 .stabs "void:t19=bs0;0;0",128,0,0,0
1350 The Solaris compiler seems to omit the trailing semicolon in this case.
1351 Getting sloppy in this way is not a swift move because if a type is
1352 embedded in a more complex expression it is necessary to be able to tell
1355 I'm not sure how a boolean type is represented.
1357 @node Negative Type Numbers
1358 @subsection Negative Type Numbers
1360 This is the method used in XCOFF for defining builtin types.
1361 Since the debugger knows about the builtin types anyway, the idea of
1362 negative type numbers is simply to give a special type number which
1363 indicates the builtin type. There is no stab defining these types.
1365 There are several subtle issues with negative type numbers.
1367 One is the size of the type. A builtin type (for example the C types
1368 @code{int} or @code{long}) might have different sizes depending on
1369 compiler options, the target architecture, the ABI, etc. This issue
1370 doesn't come up for IBM tools since (so far) they just target the
1371 RS/6000; the sizes indicated below for each size are what the IBM
1372 RS/6000 tools use. To deal with differing sizes, either define separate
1373 negative type numbers for each size (which works but requires changing
1374 the debugger, and, unless you get both AIX dbx and GDB to accept the
1375 change, introduces an incompatibility), or use a type attribute
1376 (@pxref{String Field}) to define a new type with the appropriate size
1377 (which merely requires a debugger which understands type attributes,
1378 like AIX dbx). For example,
1381 .stabs "boolean:t10=@@s8;-16",128,0,0,0
1384 defines an 8-bit boolean type, and
1387 .stabs "boolean:t10=@@s64;-16",128,0,0,0
1390 defines a 64-bit boolean type.
1392 A similar issue is the format of the type. This comes up most often for
1393 floating-point types, which could have various formats (particularly
1394 extended doubles, which vary quite a bit even among IEEE systems).
1395 Again, it is best to define a new negative type number for each
1396 different format; changing the format based on the target system has
1397 various problems. One such problem is that the Alpha has both VAX and
1398 IEEE floating types. One can easily imagine one library using the VAX
1399 types and another library in the same executable using the IEEE types.
1400 Another example is that the interpretation of whether a boolean is true
1401 or false can be based on the least significant bit, most significant
1402 bit, whether it is zero, etc., and different compilers (or different
1403 options to the same compiler) might provide different kinds of boolean.
1405 The last major issue is the names of the types. The name of a given
1406 type depends @emph{only} on the negative type number given; these do not
1407 vary depending on the language, the target system, or anything else.
1408 One can always define separate type numbers---in the following list you
1409 will see for example separate @code{int} and @code{integer*4} types
1410 which are identical except for the name. But compatibility can be
1411 maintained by not inventing new negative type numbers and instead just
1412 defining a new type with a new name. For example:
1415 .stabs "CARDINAL:t10=-8",128,0,0,0
1418 Here is the list of negative type numbers. The phrase @dfn{integral
1419 type} is used to mean twos-complement (I strongly suspect that all
1420 machines which use stabs use twos-complement; most machines use
1421 twos-complement these days).
1425 @code{int}, 32 bit signed integral type.
1428 @code{char}, 8 bit type holding a character. Both GDB and dbx on AIX
1429 treat this as signed. GCC uses this type whether @code{char} is signed
1430 or not, which seems like a bad idea. The AIX compiler (@code{xlc}) seems to
1431 avoid this type; it uses -5 instead for @code{char}.
1434 @code{short}, 16 bit signed integral type.
1437 @code{long}, 32 bit signed integral type.
1440 @code{unsigned char}, 8 bit unsigned integral type.
1443 @code{signed char}, 8 bit signed integral type.
1446 @code{unsigned short}, 16 bit unsigned integral type.
1449 @code{unsigned int}, 32 bit unsigned integral type.
1452 @code{unsigned}, 32 bit unsigned integral type.
1455 @code{unsigned long}, 32 bit unsigned integral type.
1458 @code{void}, type indicating the lack of a value.
1461 @code{float}, IEEE single precision.
1464 @code{double}, IEEE double precision.
1467 @code{long double}, IEEE double precision. The compiler claims the size
1468 will increase in a future release, and for binary compatibility you have
1469 to avoid using @code{long double}. I hope when they increase it they
1470 use a new negative type number.
1473 @code{integer}. 32 bit signed integral type.
1476 @code{boolean}. 32 bit type. How is the truth value encoded? Is it
1477 the least significant bit or is it a question of whether the whole value
1478 is zero or non-zero?
1481 @code{short real}. IEEE single precision.
1484 @code{real}. IEEE double precision.
1487 @code{stringptr}. @xref{Strings}.
1490 @code{character}, 8 bit unsigned character type.
1493 @code{logical*1}, 8 bit type. This Fortran type has a split
1494 personality in that it is used for boolean variables, but can also be
1495 used for unsigned integers. 0 is false, 1 is true, and other values are
1499 @code{logical*2}, 16 bit type. This Fortran type has a split
1500 personality in that it is used for boolean variables, but can also be
1501 used for unsigned integers. 0 is false, 1 is true, and other values are
1505 @code{logical*4}, 32 bit type. This Fortran type has a split
1506 personality in that it is used for boolean variables, but can also be
1507 used for unsigned integers. 0 is false, 1 is true, and other values are
1511 @code{logical}, 32 bit type. This Fortran type has a split
1512 personality in that it is used for boolean variables, but can also be
1513 used for unsigned integers. 0 is false, 1 is true, and other values are
1517 @code{complex}. A complex type consisting of two IEEE single-precision
1518 floating point values.
1521 @code{complex}. A complex type consisting of two IEEE double-precision
1522 floating point values.
1525 @code{integer*1}, 8 bit signed integral type.
1528 @code{integer*2}, 16 bit signed integral type.
1531 @code{integer*4}, 32 bit signed integral type.
1534 @code{wchar}. Wide character, 16 bits wide, unsigned (what format?
1538 @node Miscellaneous Types
1539 @section Miscellaneous Types
1542 @item b @var{type-information} ; @var{bytes}
1543 Pascal space type. This is documented by IBM; what does it mean?
1545 This use of the @samp{b} type descriptor can be distinguished
1546 from its use for builtin integral types (@pxref{Builtin Type
1547 Descriptors}) because the character following the type descriptor is
1548 always a digit, @samp{(}, or @samp{-}.
1550 @item B @var{type-information}
1551 A volatile-qualified version of @var{type-information}. This is
1552 a Sun extension. References and stores to a variable with a
1553 volatile-qualified type must not be optimized or cached; they
1554 must occur as the user specifies them.
1556 @item d @var{type-information}
1557 File of type @var{type-information}. As far as I know this is only used
1560 @item k @var{type-information}
1561 A const-qualified version of @var{type-information}. This is a Sun
1562 extension. A variable with a const-qualified type cannot be modified.
1564 @item M @var{type-information} ; @var{length}
1565 Multiple instance type. The type seems to composed of @var{length}
1566 repetitions of @var{type-information}, for example @code{character*3} is
1567 represented by @samp{M-2;3}, where @samp{-2} is a reference to a
1568 character type (@pxref{Negative Type Numbers}). I'm not sure how this
1569 differs from an array. This appears to be a Fortran feature.
1570 @var{length} is a bound, like those in range types; see @ref{Subranges}.
1572 @item S @var{type-information}
1573 Pascal set type. @var{type-information} must be a small type such as an
1574 enumeration or a subrange, and the type is a bitmask whose length is
1575 specified by the number of elements in @var{type-information}.
1577 @item * @var{type-information}
1578 Pointer to @var{type-information}.
1581 @node Cross-References
1582 @section Cross-References to Other Types
1584 A type can be used before it is defined; one common way to deal with
1585 that situation is just to use a type reference to a type which has not
1588 Another way is with the @samp{x} type descriptor, which is followed by
1589 @samp{s} for a structure tag, @samp{u} for a union tag, or @samp{e} for
1590 a enumerator tag, followed by the name of the tag, followed by @samp{:}.
1591 For example, the following C declarations:
1602 .stabs "bar:G16=*17=xsfoo:",32,0,0,0
1605 Not all debuggers support the @samp{x} type descriptor, so on some
1606 machines GCC does not use it. I believe that for the above example it
1607 would just emit a reference to type 17 and never define it, but I
1608 haven't verified that.
1610 Modula-2 imported types, at least on AIX, use the @samp{i} type
1611 descriptor, which is followed by the name of the module from which the
1612 type is imported, followed by @samp{:}, followed by the name of the
1613 type. There is then optionally a comma followed by type information for
1614 the type. This differs from merely naming the type (@pxref{Typedefs}) in
1615 that it identifies the module; I don't understand whether the name of
1616 the type given here is always just the same as the name we are giving
1617 it, or whether this type descriptor is used with a nameless stab
1618 (@pxref{String Field}), or what. The symbol ends with @samp{;}.
1621 @section Subrange Types
1623 The @samp{r} type descriptor defines a type as a subrange of another
1624 type. It is followed by type information for the type of which it is a
1625 subrange, a semicolon, an integral lower bound, a semicolon, an
1626 integral upper bound, and a semicolon. The AIX documentation does not
1627 specify the trailing semicolon, in an effort to specify array indexes
1628 more cleanly, but a subrange which is not an array index has always
1629 included a trailing semicolon (@pxref{Arrays}).
1631 Instead of an integer, either bound can be one of the following:
1634 @item A @var{offset}
1635 The bound is passed by reference on the stack at offset @var{offset}
1636 from the argument list. @xref{Parameters}, for more information on such
1639 @item T @var{offset}
1640 The bound is passed by value on the stack at offset @var{offset} from
1643 @item a @var{register-number}
1644 The bound is pased by reference in register number
1645 @var{register-number}.
1647 @item t @var{register-number}
1648 The bound is passed by value in register number @var{register-number}.
1654 Subranges are also used for builtin types; see @ref{Traditional Builtin Types}.
1657 @section Array Types
1659 Arrays use the @samp{a} type descriptor. Following the type descriptor
1660 is the type of the index and the type of the array elements. If the
1661 index type is a range type, it ends in a semicolon; otherwise
1662 (for example, if it is a type reference), there does not
1663 appear to be any way to tell where the types are separated. In an
1664 effort to clean up this mess, IBM documents the two types as being
1665 separated by a semicolon, and a range type as not ending in a semicolon
1666 (but this is not right for range types which are not array indexes,
1667 @pxref{Subranges}). I think probably the best solution is to specify
1668 that a semicolon ends a range type, and that the index type and element
1669 type of an array are separated by a semicolon, but that if the index
1670 type is a range type, the extra semicolon can be omitted. GDB (at least
1671 through version 4.9) doesn't support any kind of index type other than a
1672 range anyway; I'm not sure about dbx.
1674 It is well established, and widely used, that the type of the index,
1675 unlike most types found in the stabs, is merely a type definition, not
1676 type information (@pxref{String Field}) (that is, it need not start with
1677 @samp{@var{type-number}=} if it is defining a new type). According to a
1678 comment in GDB, this is also true of the type of the array elements; it
1679 gives @samp{ar1;1;10;ar1;1;10;4} as a legitimate way to express a two
1680 dimensional array. According to AIX documentation, the element type
1681 must be type information. GDB accepts either.
1683 The type of the index is often a range type, expressed as the type
1684 descriptor @samp{r} and some parameters. It defines the size of the
1685 array. In the example below, the range @samp{r1;0;2;} defines an index
1686 type which is a subrange of type 1 (integer), with a lower bound of 0
1687 and an upper bound of 2. This defines the valid range of subscripts of
1688 a three-element C array.
1690 For example, the definition:
1693 char char_vec[3] = @{'a','b','c'@};
1697 produces the output:
1700 .stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
1709 If an array is @dfn{packed}, the elements are spaced more
1710 closely than normal, saving memory at the expense of speed. For
1711 example, an array of 3-byte objects might, if unpacked, have each
1712 element aligned on a 4-byte boundary, but if packed, have no padding.
1713 One way to specify that something is packed is with type attributes
1714 (@pxref{String Field}). In the case of arrays, another is to use the
1715 @samp{P} type descriptor instead of @samp{a}. Other than specifying a
1716 packed array, @samp{P} is identical to @samp{a}.
1718 @c FIXME-what is it? A pointer?
1719 An open array is represented by the @samp{A} type descriptor followed by
1720 type information specifying the type of the array elements.
1722 @c FIXME: what is the format of this type? A pointer to a vector of pointers?
1723 An N-dimensional dynamic array is represented by
1726 D @var{dimensions} ; @var{type-information}
1729 @c Does dimensions really have this meaning? The AIX documentation
1731 @var{dimensions} is the number of dimensions; @var{type-information}
1732 specifies the type of the array elements.
1734 @c FIXME: what is the format of this type? A pointer to some offsets in
1736 A subarray of an N-dimensional array is represented by
1739 E @var{dimensions} ; @var{type-information}
1742 @c Does dimensions really have this meaning? The AIX documentation
1744 @var{dimensions} is the number of dimensions; @var{type-information}
1745 specifies the type of the array elements.
1750 Some languages, like C or the original Pascal, do not have string types,
1751 they just have related things like arrays of characters. But most
1752 Pascals and various other languages have string types, which are
1753 indicated as follows:
1756 @item n @var{type-information} ; @var{bytes}
1757 @var{bytes} is the maximum length. I'm not sure what
1758 @var{type-information} is; I suspect that it means that this is a string
1759 of @var{type-information} (thus allowing a string of integers, a string
1760 of wide characters, etc., as well as a string of characters). Not sure
1761 what the format of this type is. This is an AIX feature.
1763 @item z @var{type-information} ; @var{bytes}
1764 Just like @samp{n} except that this is a gstring, not an ordinary
1765 string. I don't know the difference.
1768 Pascal Stringptr. What is this? This is an AIX feature.
1772 @section Enumerations
1774 Enumerations are defined with the @samp{e} type descriptor.
1776 @c FIXME: Where does this information properly go? Perhaps it is
1777 @c redundant with something we already explain.
1778 The source line below declares an enumeration type at file scope.
1779 The type definition is located after the @code{N_RBRAC} that marks the end of
1780 the previous procedure's block scope, and before the @code{N_FUN} that marks
1781 the beginning of the next procedure's block scope. Therefore it does not
1782 describe a block local symbol, but a file local one.
1787 enum e_places @{first,second=3,last@};
1791 generates the following stab:
1794 .stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0
1797 The symbol descriptor (@samp{T}) says that the stab describes a
1798 structure, enumeration, or union tag. The type descriptor @samp{e},
1799 following the @samp{22=} of the type definition narrows it down to an
1800 enumeration type. Following the @samp{e} is a list of the elements of
1801 the enumeration. The format is @samp{@var{name}:@var{value},}. The
1802 list of elements ends with @samp{;}.
1804 There is no standard way to specify the size of an enumeration type; it
1805 is determined by the architecture (normally all enumerations types are
1806 32 bits). There should be a way to specify an enumeration type of
1807 another size; type attributes would be one way to do this. @xref{Stabs
1813 The encoding of structures in stabs can be shown with an example.
1815 The following source code declares a structure tag and defines an
1816 instance of the structure in global scope. Then a @code{typedef} equates the
1817 structure tag with a new type. Seperate stabs are generated for the
1818 structure tag, the structure @code{typedef}, and the structure instance. The
1819 stabs for the tag and the @code{typedef} are emited when the definitions are
1820 encountered. Since the structure elements are not initialized, the
1821 stab and code for the structure variable itself is located at the end
1822 of the program in the bss section.
1829 struct s_tag* s_next;
1832 typedef struct s_tag s_typedef;
1835 The structure tag has an @code{N_LSYM} stab type because, like the
1836 enumeration, the symbol has file scope. Like the enumeration, the
1837 symbol descriptor is @samp{T}, for enumeration, structure, or tag type.
1838 The type descriptor @samp{s} following the @samp{16=} of the type
1839 definition narrows the symbol type to structure.
1841 Following the @samp{s} type descriptor is the number of bytes the
1842 structure occupies, followed by a description of each structure element.
1843 The structure element descriptions are of the form @var{name:type, bit
1844 offset from the start of the struct, number of bits in the element}.
1846 @c FIXME: phony line break. Can probably be fixed by using an example
1847 @c with fewer fields.
1850 .stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;
1851 s_char_vec:17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0
1854 In this example, the first two structure elements are previously defined
1855 types. For these, the type following the @samp{@var{name}:} part of the
1856 element description is a simple type reference. The other two structure
1857 elements are new types. In this case there is a type definition
1858 embedded after the @samp{@var{name}:}. The type definition for the
1859 array element looks just like a type definition for a standalone array.
1860 The @code{s_next} field is a pointer to the same kind of structure that
1861 the field is an element of. So the definition of structure type 16
1862 contains a type definition for an element which is a pointer to type 16.
1865 @section Giving a Type a Name
1867 To give a type a name, use the @samp{t} symbol descriptor. The type
1868 is specified by the type information (@pxref{String Field}) for the stab.
1872 .stabs "s_typedef:t16",128,0,0,0 # @r{128 is N_LSYM}
1875 specifies that @code{s_typedef} refers to type number 16. Such stabs
1876 have symbol type @code{N_LSYM} (or @code{C_DECL} for XCOFF).
1878 If you are specifying the tag name for a structure, union, or
1879 enumeration, use the @samp{T} symbol descriptor instead. I believe C is
1880 the only language with this feature.
1882 If the type is an opaque type (I believe this is a Modula-2 feature),
1883 AIX provides a type descriptor to specify it. The type descriptor is
1884 @samp{o} and is followed by a name. I don't know what the name
1885 means---is it always the same as the name of the type, or is this type
1886 descriptor used with a nameless stab (@pxref{String Field})? There
1887 optionally follows a comma followed by type information which defines
1888 the type of this type. If omitted, a semicolon is used in place of the
1889 comma and the type information, and the type is much like a generic
1890 pointer type---it has a known size but little else about it is
1904 This code generates a stab for a union tag and a stab for a union
1905 variable. Both use the @code{N_LSYM} stab type. If a union variable is
1906 scoped locally to the procedure in which it is defined, its stab is
1907 located immediately preceding the @code{N_LBRAC} for the procedure's block
1910 The stab for the union tag, however, is located preceding the code for
1911 the procedure in which it is defined. The stab type is @code{N_LSYM}. This
1912 would seem to imply that the union type is file scope, like the struct
1913 type @code{s_tag}. This is not true. The contents and position of the stab
1914 for @code{u_type} do not convey any infomation about its procedure local
1917 @c FIXME: phony line break. Can probably be fixed by using an example
1918 @c with fewer fields.
1921 .stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
1925 The symbol descriptor @samp{T}, following the @samp{name:} means that
1926 the stab describes an enumeration, structure, or union tag. The type
1927 descriptor @samp{u}, following the @samp{23=} of the type definition,
1928 narrows it down to a union type definition. Following the @samp{u} is
1929 the number of bytes in the union. After that is a list of union element
1930 descriptions. Their format is @var{name:type, bit offset into the
1931 union, number of bytes for the element;}.
1933 The stab for the union variable is:
1936 .stabs "an_u:23",128,0,0,-20 # @r{128 is N_LSYM}
1939 @samp{-20} specifies where the variable is stored (@pxref{Stack
1942 @node Function Types
1943 @section Function Types
1945 Various types can be defined for function variables. These types are
1946 not used in defining functions (@pxref{Procedures}); they are used for
1947 things like pointers to functions.
1949 The simple, traditional, type is type descriptor @samp{f} is followed by
1950 type information for the return type of the function, followed by a
1953 This does not deal with functions for which the number and types of the
1954 parameters are part of the type, as in Modula-2 or ANSI C. AIX provides
1955 extensions to specify these, using the @samp{f}, @samp{F}, @samp{p}, and
1956 @samp{R} type descriptors.
1958 First comes the type descriptor. If it is @samp{f} or @samp{F}, this
1959 type involves a function rather than a procedure, and the type
1960 information for the return type of the function follows, followed by a
1961 comma. Then comes the number of parameters to the function and a
1962 semicolon. Then, for each parameter, there is the name of the parameter
1963 followed by a colon (this is only present for type descriptors @samp{R}
1964 and @samp{F} which represent Pascal function or procedure parameters),
1965 type information for the parameter, a comma, 0 if passed by reference or
1966 1 if passed by value, and a semicolon. The type definition ends with a
1969 For example, this variable definition:
1976 generates the following code:
1979 .stabs "g_pf:G24=*25=f1",32,0,0,0
1980 .common _g_pf,4,"bss"
1983 The variable defines a new type, 24, which is a pointer to another new
1984 type, 25, which is a function returning @code{int}.
1987 @chapter Symbol Information in Symbol Tables
1989 This chapter describes the format of symbol table entries
1990 and how stab assembler directives map to them. It also describes the
1991 transformations that the assembler and linker make on data from stabs.
1994 * Symbol Table Format::
1995 * Transformations On Symbol Tables::
1998 @node Symbol Table Format
1999 @section Symbol Table Format
2001 Each time the assembler encounters a stab directive, it puts
2002 each field of the stab into a corresponding field in a symbol table
2003 entry of its output file. If the stab contains a string field, the
2004 symbol table entry for that stab points to a string table entry
2005 containing the string data from the stab. Assembler labels become
2006 relocatable addresses. Symbol table entries in a.out have the format:
2008 @c FIXME: should refer to external, not internal.
2010 struct internal_nlist @{
2011 unsigned long n_strx; /* index into string table of name */
2012 unsigned char n_type; /* type of symbol */
2013 unsigned char n_other; /* misc info (usually empty) */
2014 unsigned short n_desc; /* description field */
2015 bfd_vma n_value; /* value of symbol */
2019 If the stab has a string, the @code{n_strx} field holds the offset in
2020 bytes of the string within the string table. The string is terminated
2021 by a NUL character. If the stab lacks a string (for example, it was
2022 produced by a @code{.stabn} or @code{.stabd} directive), the
2023 @code{n_strx} field is zero.
2025 Symbol table entries with @code{n_type} field values greater than 0x1f
2026 originated as stabs generated by the compiler (with one random
2027 exception). The other entries were placed in the symbol table of the
2028 executable by the assembler or the linker.
2030 @node Transformations On Symbol Tables
2031 @section Transformations on Symbol Tables
2033 The linker concatenates object files and does fixups of externally
2036 You can see the transformations made on stab data by the assembler and
2037 linker by examining the symbol table after each pass of the build. To
2038 do this, use @samp{nm -ap}, which dumps the symbol table, including
2039 debugging information, unsorted. For stab entries the columns are:
2040 @var{value}, @var{other}, @var{desc}, @var{type}, @var{string}. For
2041 assembler and linker symbols, the columns are: @var{value}, @var{type},
2044 The low 5 bits of the stab type tell the linker how to relocate the
2045 value of the stab. Thus for stab types like @code{N_RSYM} and
2046 @code{N_LSYM}, where the value is an offset or a register number, the
2047 low 5 bits are @code{N_ABS}, which tells the linker not to relocate the
2050 Where the value of a stab contains an assembly language label,
2051 it is transformed by each build step. The assembler turns it into a
2052 relocatable address and the linker turns it into an absolute address.
2055 * Transformations On Static Variables::
2056 * Transformations On Global Variables::
2057 * ELF Transformations:: In ELF, things are a bit different.
2060 @node Transformations On Static Variables
2061 @subsection Transformations on Static Variables
2063 This source line defines a static variable at file scope:
2066 static int s_g_repeat
2070 The following stab describes the symbol:
2073 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
2077 The assembler transforms the stab into this symbol table entry in the
2078 @file{.o} file. The location is expressed as a data segment offset.
2081 00000084 - 00 0000 STSYM s_g_repeat:S1
2085 In the symbol table entry from the executable, the linker has made the
2086 relocatable address absolute.
2089 0000e00c - 00 0000 STSYM s_g_repeat:S1
2092 @node Transformations On Global Variables
2093 @subsection Transformations on Global Variables
2095 Stabs for global variables do not contain location information. In
2096 this case, the debugger finds location information in the assembler or
2097 linker symbol table entry describing the variable. The source line:
2107 .stabs "g_foo:G2",32,0,0,0
2110 The variable is represented by two symbol table entries in the object
2111 file (see below). The first one originated as a stab. The second one
2112 is an external symbol. The upper case @samp{D} signifies that the
2113 @code{n_type} field of the symbol table contains 7, @code{N_DATA} with
2114 local linkage. The stab's value is zero since the value is not used for
2115 @code{N_GSYM} stabs. The value of the linker symbol is the relocatable
2116 address corresponding to the variable.
2119 00000000 - 00 0000 GSYM g_foo:G2
2124 These entries as transformed by the linker. The linker symbol table
2125 entry now holds an absolute address:
2128 00000000 - 00 0000 GSYM g_foo:G2
2133 @node ELF Transformations
2134 @subsection Transformations of Stabs in ELF Files
2136 For ELF files, use @code{objdump --stabs} instead of @code{nm} to show
2137 the stabs in an object or executable file. @code{objdump} is a GNU
2138 utility; Sun does not provide any equivalent.
2140 The following example is for a stab whose value is an address is
2141 relative to the compilation unit (@pxref{Stabs In ELF}). For example,
2148 appears within a function, then the assembly language output from the
2154 .stabs "ld:V(0,3)",0x26,0,4,.L18-Ddata.data # @r{0x26 is N_STSYM}
2161 Because the value is formed by subtracting one symbol from another, the
2162 value is absolute, not relocatable, and so the object file contains
2165 Symnum n_type n_othr n_desc n_value n_strx String
2166 31 STSYM 0 4 00000004 680 ld:V(0,3)
2169 without any relocations, and the executable file also contains
2172 Symnum n_type n_othr n_desc n_value n_strx String
2173 31 STSYM 0 4 00000004 680 ld:V(0,3)
2177 @chapter GNU C++ Stabs
2180 * Basic Cplusplus Types::
2183 * Methods:: Method definition
2185 * Method Modifiers::
2188 * Virtual Base Classes::
2192 Type descriptors added for C++ descriptions:
2196 method type (@code{##} if minimal debug)
2199 Member (class and variable) type. It is followed by type information
2200 for the offset basetype, a comma, and type information for the type of
2201 the field being pointed to. (FIXME: this is acknowledged to be
2202 gibberish. Can anyone say what really goes here?).
2204 Note that there is a conflict between this and type attributes
2205 (@pxref{String Field}); both use type descriptor @samp{@@}.
2206 Fortunately, the @samp{@@} type descriptor used in this C++ sense always
2207 will be followed by a digit, @samp{(}, or @samp{-}, and type attributes
2208 never start with those things.
2211 @node Basic Cplusplus Types
2212 @section Basic Types For C++
2214 << the examples that follow are based on a01.C >>
2217 C++ adds two more builtin types to the set defined for C. These are
2218 the unknown type and the vtable record type. The unknown type, type
2219 16, is defined in terms of itself like the void type.
2221 The vtable record type, type 17, is defined as a structure type and
2222 then as a structure tag. The structure has four fields: delta, index,
2223 pfn, and delta2. pfn is the function pointer.
2225 << In boilerplate $vtbl_ptr_type, what are the fields delta,
2226 index, and delta2 used for? >>
2228 This basic type is present in all C++ programs even if there are no
2229 virtual methods defined.
2232 .stabs "struct_name:sym_desc(type)type_def(17)=type_desc(struct)struct_bytes(8)
2233 elem_name(delta):type_ref(short int),bit_offset(0),field_bits(16);
2234 elem_name(index):type_ref(short int),bit_offset(16),field_bits(16);
2235 elem_name(pfn):type_def(18)=type_desc(ptr to)type_ref(void),
2236 bit_offset(32),field_bits(32);
2237 elem_name(delta2):type_def(short int);bit_offset(32),field_bits(16);;"
2242 .stabs "$vtbl_ptr_type:t17=s8
2243 delta:6,0,16;index:6,16,16;pfn:18=*15,32,32;delta2:6,32,16;;"
2248 .stabs "name:sym_dec(struct tag)type_ref($vtbl_ptr_type)",N_LSYM,NIL,NIL,NIL
2252 .stabs "$vtbl_ptr_type:T17",128,0,0,0
2255 @node Simple Classes
2256 @section Simple Class Definition
2258 The stabs describing C++ language features are an extension of the
2259 stabs describing C. Stabs representing C++ class types elaborate
2260 extensively on the stab format used to describe structure types in C.
2261 Stabs representing class type variables look just like stabs
2262 representing C language variables.
2264 Consider the following very simple class definition.
2270 int Ameth(int in, char other);
2274 The class @code{baseA} is represented by two stabs. The first stab describes
2275 the class as a structure type. The second stab describes a structure
2276 tag of the class type. Both stabs are of stab type @code{N_LSYM}. Since the
2277 stab is not located between an @code{N_FUN} and an @code{N_LBRAC} stab this indicates
2278 that the class is defined at file scope. If it were, then the @code{N_LSYM}
2279 would signify a local variable.
2281 A stab describing a C++ class type is similar in format to a stab
2282 describing a C struct, with each class member shown as a field in the
2283 structure. The part of the struct format describing fields is
2284 expanded to include extra information relevent to C++ class members.
2285 In addition, if the class has multiple base classes or virtual
2286 functions the struct format outside of the field parts is also
2289 In this simple example the field part of the C++ class stab
2290 representing member data looks just like the field part of a C struct
2291 stab. The section on protections describes how its format is
2292 sometimes extended for member data.
2294 The field part of a C++ class stab representing a member function
2295 differs substantially from the field part of a C struct stab. It
2296 still begins with @samp{name:} but then goes on to define a new type number
2297 for the member function, describe its return type, its argument types,
2298 its protection level, any qualifiers applied to the method definition,
2299 and whether the method is virtual or not. If the method is virtual
2300 then the method description goes on to give the vtable index of the
2301 method, and the type number of the first base class defining the
2304 When the field name is a method name it is followed by two colons rather
2305 than one. This is followed by a new type definition for the method.
2306 This is a number followed by an equal sign and the type descriptor
2307 @samp{#}, indicating a method type, and a second @samp{#}, indicating
2308 that this is the @dfn{minimal} type of method definition used by GCC2,
2309 not larger method definitions used by earlier versions of GCC. This is
2310 followed by a type reference showing the return type of the method and a
2313 The format of an overloaded operator method name differs from that of
2314 other methods. It is @samp{op$::@var{operator-name}.} where
2315 @var{operator-name} is the operator name such as @samp{+} or @samp{+=}.
2316 The name ends with a period, and any characters except the period can
2317 occur in the @var{operator-name} string.
2319 The next part of the method description represents the arguments to the
2320 method, preceeded by a colon and ending with a semi-colon. The types of
2321 the arguments are expressed in the same way argument types are expressed
2322 in C++ name mangling. In this example an @code{int} and a @code{char}
2325 This is followed by a number, a letter, and an asterisk or period,
2326 followed by another semicolon. The number indicates the protections
2327 that apply to the member function. Here the 2 means public. The
2328 letter encodes any qualifier applied to the method definition. In
2329 this case, @samp{A} means that it is a normal function definition. The dot
2330 shows that the method is not virtual. The sections that follow
2331 elaborate further on these fields and describe the additional
2332 information present for virtual methods.
2336 .stabs "class_name:sym_desc(type)type_def(20)=type_desc(struct)struct_bytes(4)
2337 field_name(Adat):type(int),bit_offset(0),field_bits(32);
2339 method_name(Ameth)::type_def(21)=type_desc(method)return_type(int);
2340 :arg_types(int char);
2341 protection(public)qualifier(normal)virtual(no);;"
2346 .stabs "baseA:t20=s4Adat:1,0,32;Ameth::21=##1;:ic;2A.;;",128,0,0,0
2348 .stabs "class_name:sym_desc(struct tag)",N_LSYM,NIL,NIL,NIL
2350 .stabs "baseA:T20",128,0,0,0
2353 @node Class Instance
2354 @section Class Instance
2356 As shown above, describing even a simple C++ class definition is
2357 accomplished by massively extending the stab format used in C to
2358 describe structure types. However, once the class is defined, C stabs
2359 with no modifications can be used to describe class instances. The
2369 yields the following stab describing the class instance. It looks no
2370 different from a standard C stab describing a local variable.
2373 .stabs "name:type_ref(baseA)", N_LSYM, NIL, NIL, frame_ptr_offset
2377 .stabs "AbaseA:20",128,0,0,-20
2381 @section Method Definition
2383 The class definition shown above declares Ameth. The C++ source below
2388 baseA::Ameth(int in, char other)
2395 This method definition yields three stabs following the code of the
2396 method. One stab describes the method itself and following two describe
2397 its parameters. Although there is only one formal argument all methods
2398 have an implicit argument which is the @code{this} pointer. The @code{this}
2399 pointer is a pointer to the object on which the method was called. Note
2400 that the method name is mangled to encode the class name and argument
2401 types. Name mangling is described in the @sc{arm} (@cite{The Annotated
2402 C++ Reference Manual}, by Ellis and Stroustrup, @sc{isbn}
2403 0-201-51459-1); @file{gpcompare.texi} in Cygnus GCC distributions
2404 describes the differences between GNU mangling and @sc{arm}
2406 @c FIXME: Use @xref, especially if this is generally installed in the
2408 @c FIXME: This information should be in a net release, either of GCC or
2409 @c GDB. But gpcompare.texi doesn't seem to be in the FSF GCC.
2412 .stabs "name:symbol_desriptor(global function)return_type(int)",
2413 N_FUN, NIL, NIL, code_addr_of_method_start
2415 .stabs "Ameth__5baseAic:F1",36,0,0,_Ameth__5baseAic
2418 Here is the stab for the @code{this} pointer implicit argument. The
2419 name of the @code{this} pointer is always @code{this}. Type 19, the
2420 @code{this} pointer is defined as a pointer to type 20, @code{baseA},
2421 but a stab defining @code{baseA} has not yet been emited. Since the
2422 compiler knows it will be emited shortly, here it just outputs a cross
2423 reference to the undefined symbol, by prefixing the symbol name with
2427 .stabs "name:sym_desc(register param)type_def(19)=
2428 type_desc(ptr to)type_ref(baseA)=
2429 type_desc(cross-reference to)baseA:",N_RSYM,NIL,NIL,register_number
2431 .stabs "this:P19=*20=xsbaseA:",64,0,0,8
2434 The stab for the explicit integer argument looks just like a parameter
2435 to a C function. The last field of the stab is the offset from the
2436 argument pointer, which in most systems is the same as the frame
2440 .stabs "name:sym_desc(value parameter)type_ref(int)",
2441 N_PSYM,NIL,NIL,offset_from_arg_ptr
2443 .stabs "in:p1",160,0,0,72
2446 << The examples that follow are based on A1.C >>
2449 @section Protections
2452 In the simple class definition shown above all member data and
2453 functions were publicly accessable. The example that follows
2454 contrasts public, protected and privately accessable fields and shows
2455 how these protections are encoded in C++ stabs.
2457 @c FIXME: What does "part of the string" mean?
2458 Protections for class member data are signified by two characters
2459 embedded in the stab defining the class type. These characters are
2460 located after the name: part of the string. @samp{/0} means private,
2461 @samp{/1} means protected, and @samp{/2} means public. If these
2462 characters are omited this means that the member is public. The
2463 following C++ source:
2477 generates the following stab to describe the class type all_data.
2480 .stabs "class_name:sym_desc(type)type_def(19)=type_desc(struct)struct_bytes
2481 data_name:/protection(private)type_ref(int),bit_offset,num_bits;
2482 data_name:/protection(protected)type_ref(char),bit_offset,num_bits;
2483 data_name:(/num omited, private)type_ref(float),bit_offset,num_bits;;"
2488 .stabs "all_data:t19=s12
2489 priv_dat:/01,0,32;prot_dat:/12,32,8;pub_dat:12,64,32;;",128,0,0,0
2492 Protections for member functions are signified by one digit embeded in
2493 the field part of the stab describing the method. The digit is 0 if
2494 private, 1 if protected and 2 if public. Consider the C++ class
2498 class all_methods @{
2500 int priv_meth(int in)@{return in;@};
2502 char protMeth(char in)@{return in;@};
2504 float pubMeth(float in)@{return in;@};
2508 It generates the following stab. The digit in question is to the left
2509 of an @samp{A} in each case. Notice also that in this case two symbol
2510 descriptors apply to the class name struct tag and struct type.
2513 .stabs "class_name:sym_desc(struct tag&type)type_def(21)=
2514 sym_desc(struct)struct_bytes(1)
2515 meth_name::type_def(22)=sym_desc(method)returning(int);
2516 :args(int);protection(private)modifier(normal)virtual(no);
2517 meth_name::type_def(23)=sym_desc(method)returning(char);
2518 :args(char);protection(protected)modifier(normal)virual(no);
2519 meth_name::type_def(24)=sym_desc(method)returning(float);
2520 :args(float);protection(public)modifier(normal)virtual(no);;",
2525 .stabs "all_methods:Tt21=s1priv_meth::22=##1;:i;0A.;protMeth::23=##2;:c;1A.;
2526 pubMeth::24=##12;:f;2A.;;",128,0,0,0
2529 @node Method Modifiers
2530 @section Method Modifiers (@code{const}, @code{volatile}, @code{const volatile})
2534 In the class example described above all the methods have the normal
2535 modifier. This method modifier information is located just after the
2536 protection information for the method. This field has four possible
2537 character values. Normal methods use @samp{A}, const methods use
2538 @samp{B}, volatile methods use @samp{C}, and const volatile methods use
2539 @samp{D}. Consider the class definition below:
2544 int ConstMeth (int arg) const @{ return arg; @};
2545 char VolatileMeth (char arg) volatile @{ return arg; @};
2546 float ConstVolMeth (float arg) const volatile @{return arg; @};
2550 This class is described by the following stab:
2553 .stabs "class(A):sym_desc(struct)type_def(20)=type_desc(struct)struct_bytes(1)
2554 meth_name(ConstMeth)::type_def(21)sym_desc(method)
2555 returning(int);:arg(int);protection(public)modifier(const)virtual(no);
2556 meth_name(VolatileMeth)::type_def(22)=sym_desc(method)
2557 returning(char);:arg(char);protection(public)modifier(volatile)virt(no)
2558 meth_name(ConstVolMeth)::type_def(23)=sym_desc(method)
2559 returning(float);:arg(float);protection(public)modifer(const volatile)
2560 virtual(no);;", @dots{}
2564 .stabs "A:T20=s1ConstMeth::21=##1;:i;2B.;VolatileMeth::22=##2;:c;2C.;
2565 ConstVolMeth::23=##12;:f;2D.;;",128,0,0,0
2568 @node Virtual Methods
2569 @section Virtual Methods
2571 << The following examples are based on a4.C >>
2573 The presence of virtual methods in a class definition adds additional
2574 data to the class description. The extra data is appended to the
2575 description of the virtual method and to the end of the class
2576 description. Consider the class definition below:
2582 virtual int A_virt (int arg) @{ return arg; @};
2586 This results in the stab below describing class A. It defines a new
2587 type (20) which is an 8 byte structure. The first field of the class
2588 struct is @samp{Adat}, an integer, starting at structure offset 0 and
2591 The second field in the class struct is not explicitly defined by the
2592 C++ class definition but is implied by the fact that the class
2593 contains a virtual method. This field is the vtable pointer. The
2594 name of the vtable pointer field starts with @samp{$vf} and continues with a
2595 type reference to the class it is part of. In this example the type
2596 reference for class A is 20 so the name of its vtable pointer field is
2597 @samp{$vf20}, followed by the usual colon.
2599 Next there is a type definition for the vtable pointer type (21).
2600 This is in turn defined as a pointer to another new type (22).
2602 Type 22 is the vtable itself, which is defined as an array, indexed by
2603 a range of integers between 0 and 1, and whose elements are of type
2604 17. Type 17 was the vtable record type defined by the boilerplate C++
2605 type definitions, as shown earlier.
2607 The bit offset of the vtable pointer field is 32. The number of bits
2608 in the field are not specified when the field is a vtable pointer.
2610 Next is the method definition for the virtual member function @code{A_virt}.
2611 Its description starts out using the same format as the non-virtual
2612 member functions described above, except instead of a dot after the
2613 @samp{A} there is an asterisk, indicating that the function is virtual.
2614 Since is is virtual some addition information is appended to the end
2615 of the method description.
2617 The first number represents the vtable index of the method. This is a
2618 32 bit unsigned number with the high bit set, followed by a
2621 The second number is a type reference to the first base class in the
2622 inheritence hierarchy defining the virtual member function. In this
2623 case the class stab describes a base class so the virtual function is
2624 not overriding any other definition of the method. Therefore the
2625 reference is to the type number of the class that the stab is
2628 This is followed by three semi-colons. One marks the end of the
2629 current sub-section, one marks the end of the method field, and the
2630 third marks the end of the struct definition.
2632 For classes containing virtual functions the very last section of the
2633 string part of the stab holds a type reference to the first base
2634 class. This is preceeded by @samp{~%} and followed by a final semi-colon.
2637 .stabs "class_name(A):type_def(20)=sym_desc(struct)struct_bytes(8)
2638 field_name(Adat):type_ref(int),bit_offset(0),field_bits(32);
2639 field_name(A virt func ptr):type_def(21)=type_desc(ptr to)type_def(22)=
2640 sym_desc(array)index_type_ref(range of int from 0 to 1);
2641 elem_type_ref(vtbl elem type),
2643 meth_name(A_virt)::typedef(23)=sym_desc(method)returning(int);
2644 :arg_type(int),protection(public)normal(yes)virtual(yes)
2645 vtable_index(1);class_first_defining(A);;;~%first_base(A);",
2649 @c FIXME: bogus line break.
2651 .stabs "A:t20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2652 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2656 @section Inheritence
2658 Stabs describing C++ derived classes include additional sections that
2659 describe the inheritence hierarchy of the class. A derived class stab
2660 also encodes the number of base classes. For each base class it tells
2661 if the base class is virtual or not, and if the inheritence is private
2662 or public. It also gives the offset into the object of the portion of
2663 the object corresponding to each base class.
2665 This additional information is embeded in the class stab following the
2666 number of bytes in the struct. First the number of base classes
2667 appears bracketed by an exclamation point and a comma.
2669 Then for each base type there repeats a series: two digits, a number,
2670 a comma, another number, and a semi-colon.
2672 The first of the two digits is 1 if the base class is virtual and 0 if
2673 not. The second digit is 2 if the derivation is public and 0 if not.
2675 The number following the first two digits is the offset from the start
2676 of the object to the part of the object pertaining to the base class.
2678 After the comma, the second number is a type_descriptor for the base
2679 type. Finally a semi-colon ends the series, which repeats for each
2682 The source below defines three base classes @code{A}, @code{B}, and
2683 @code{C} and the derived class @code{D}.
2690 virtual int A_virt (int arg) @{ return arg; @};
2696 virtual int B_virt (int arg) @{return arg; @};
2702 virtual int C_virt (int arg) @{return arg; @};
2705 class D : A, virtual B, public C @{
2708 virtual int A_virt (int arg ) @{ return arg+1; @};
2709 virtual int B_virt (int arg) @{ return arg+2; @};
2710 virtual int C_virt (int arg) @{ return arg+3; @};
2711 virtual int D_virt (int arg) @{ return arg; @};
2715 Class stabs similar to the ones described earlier are generated for
2718 @c FIXME!!! the linebreaks in the following example probably make the
2719 @c examples literally unusable, but I don't know any other way to get
2720 @c them on the page.
2721 @c One solution would be to put some of the type definitions into
2722 @c separate stabs, even if that's not exactly what the compiler actually
2725 .stabs "A:T20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2726 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2728 .stabs "B:Tt25=s8Bdat:1,0,32;$vf25:21,32;B_virt::26=##1;
2729 :i;2A*-2147483647;25;;;~%25;",128,0,0,0
2731 .stabs "C:Tt28=s8Cdat:1,0,32;$vf28:21,32;C_virt::29=##1;
2732 :i;2A*-2147483647;28;;;~%28;",128,0,0,0
2735 In the stab describing derived class @code{D} below, the information about
2736 the derivation of this class is encoded as follows.
2739 .stabs "derived_class_name:symbol_descriptors(struct tag&type)=
2740 type_descriptor(struct)struct_bytes(32)!num_bases(3),
2741 base_virtual(no)inheritence_public(no)base_offset(0),
2742 base_class_type_ref(A);
2743 base_virtual(yes)inheritence_public(no)base_offset(NIL),
2744 base_class_type_ref(B);
2745 base_virtual(no)inheritence_public(yes)base_offset(64),
2746 base_class_type_ref(C); @dots{}
2749 @c FIXME! fake linebreaks.
2751 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:
2752 1,160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt:
2753 :32:i;2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;
2754 28;;D_virt::32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2757 @node Virtual Base Classes
2758 @section Virtual Base Classes
2760 A derived class object consists of a concatination in memory of the data
2761 areas defined by each base class, starting with the leftmost and ending
2762 with the rightmost in the list of base classes. The exception to this
2763 rule is for virtual inheritence. In the example above, class @code{D}
2764 inherits virtually from base class @code{B}. This means that an
2765 instance of a @code{D} object will not contain its own @code{B} part but
2766 merely a pointer to a @code{B} part, known as a virtual base pointer.
2768 In a derived class stab, the base offset part of the derivation
2769 information, described above, shows how the base class parts are
2770 ordered. The base offset for a virtual base class is always given as 0.
2771 Notice that the base offset for @code{B} is given as 0 even though
2772 @code{B} is not the first base class. The first base class @code{A}
2775 The field information part of the stab for class @code{D} describes the field
2776 which is the pointer to the virtual base class @code{B}. The vbase pointer
2777 name is @samp{$vb} followed by a type reference to the virtual base class.
2778 Since the type id for @code{B} in this example is 25, the vbase pointer name
2781 @c FIXME!! fake linebreaks below
2783 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:1,
2784 160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt::32:i;
2785 2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;28;;D_virt:
2786 :32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2789 Following the name and a semicolon is a type reference describing the
2790 type of the virtual base class pointer, in this case 24. Type 24 was
2791 defined earlier as the type of the @code{B} class @code{this} pointer. The
2792 @code{this} pointer for a class is a pointer to the class type.
2795 .stabs "this:P24=*25=xsB:",64,0,0,8
2798 Finally the field offset part of the vbase pointer field description
2799 shows that the vbase pointer is the first field in the @code{D} object,
2800 before any data fields defined by the class. The layout of a @code{D}
2801 class object is a follows, @code{Adat} at 0, the vtable pointer for
2802 @code{A} at 32, @code{Cdat} at 64, the vtable pointer for C at 96, the
2803 virtual base pointer for @code{B} at 128, and @code{Ddat} at 160.
2806 @node Static Members
2807 @section Static Members
2809 The data area for a class is a concatenation of the space used by the
2810 data members of the class. If the class has virtual methods, a vtable
2811 pointer follows the class data. The field offset part of each field
2812 description in the class stab shows this ordering.
2814 << How is this reflected in stabs? See Cygnus bug #677 for some info. >>
2817 @appendix Table of Stab Types
2819 The following are all the possible values for the stab type field, for
2820 @code{a.out} files, in numeric order. This does not apply to XCOFF, but
2821 it does apply to stabs in ELF. Stabs in ECOFF use these values but add
2822 0x8f300 to distinguish them from non-stab symbols.
2824 The symbolic names are defined in the file @file{include/aout/stabs.def}.
2827 * Non-Stab Symbol Types:: Types from 0 to 0x1f
2828 * Stab Symbol Types:: Types from 0x20 to 0xff
2831 @node Non-Stab Symbol Types
2832 @appendixsec Non-Stab Symbol Types
2834 The following types are used by the linker and assembler, not by stab
2835 directives. Since this document does not attempt to describe aspects of
2836 object file format other than the debugging format, no details are
2839 @c Try to get most of these to fit on a single line.
2849 File scope absolute symbol
2851 @item 0x3 N_ABS | N_EXT
2852 External absolute symbol
2855 File scope text symbol
2857 @item 0x5 N_TEXT | N_EXT
2858 External text symbol
2861 File scope data symbol
2863 @item 0x7 N_DATA | N_EXT
2864 External data symbol
2867 File scope BSS symbol
2869 @item 0x9 N_BSS | N_EXT
2873 Same as @code{N_FN}, for Sequent compilers
2876 Symbol is indirected to another symbol
2879 Common---visible after shared library dynamic link
2882 Absolute set element
2885 Text segment set element
2888 Data segment set element
2891 BSS segment set element
2894 Pointer to set vector
2896 @item 0x1e N_WARNING
2897 Print a warning message during linking
2900 File name of a @file{.o} file
2903 @node Stab Symbol Types
2904 @appendixsec Stab Symbol Types
2906 The following symbol types indicate that this is a stab. This is the
2907 full list of stab numbers, including stab types that are used in
2908 languages other than C.
2912 Global symbol; see @ref{Global Variables}.
2915 Function name (for BSD Fortran); see @ref{Procedures}.
2918 Function name (@pxref{Procedures}) or text segment variable
2922 Data segment file-scope variable; see @ref{Statics}.
2925 BSS segment file-scope variable; see @ref{Statics}.
2928 Name of main routine; see @ref{Main Program}.
2931 Variable in @code{.rodata} section; see @ref{Statics}.
2934 Global symbol (for Pascal); see @ref{N_PC}.
2937 Number of symbols (according to Ultrix V4.0); see @ref{N_NSYMS}.
2940 No DST map; see @ref{N_NOMAP}.
2942 @c FIXME: describe this solaris feature in the body of the text (see
2943 @c comments in include/aout/stab.def).
2945 Object file (Solaris2).
2947 @c See include/aout/stab.def for (a little) more info.
2949 Debugger options (Solaris2).
2952 Register variable; see @ref{Register Variables}.
2955 Modula-2 compilation unit; see @ref{N_M2C}.
2958 Line number in text segment; see @ref{Line Numbers}.
2961 Line number in data segment; see @ref{Line Numbers}.
2964 Line number in bss segment; see @ref{Line Numbers}.
2967 Sun source code browser, path to @file{.cb} file; see @ref{N_BROWS}.
2970 GNU Modula2 definition module dependency; see @ref{N_DEFD}.
2973 Function start/body/end line numbers (Solaris2).
2976 GNU C++ exception variable; see @ref{N_EHDECL}.
2979 Modula2 info "for imc" (according to Ultrix V4.0); see @ref{N_MOD2}.
2982 GNU C++ @code{catch} clause; see @ref{N_CATCH}.
2985 Structure of union element; see @ref{N_SSYM}.
2988 Last stab for module (Solaris2).
2991 Path and name of source file; see @ref{Source Files}.
2994 Stack variable (@pxref{Stack Variables}) or type (@pxref{Typedefs}).
2997 Beginning of an include file (Sun only); see @ref{Include Files}.
3000 Name of include file; see @ref{Include Files}.
3003 Parameter variable; see @ref{Parameters}.
3006 End of an include file; see @ref{Include Files}.
3009 Alternate entry point; see @ref{N_ENTRY}.
3012 Beginning of a lexical block; see @ref{Block Structure}.
3015 Place holder for a deleted include file; see @ref{Include Files}.
3018 Modula2 scope information (Sun linker); see @ref{N_SCOPE}.
3021 End of a lexical block; see @ref{Block Structure}.
3024 Begin named common block; see @ref{Common Blocks}.
3027 End named common block; see @ref{Common Blocks}.
3030 Member of a common block; see @ref{Common Blocks}.
3032 @c FIXME: How does this really work? Move it to main body of document.
3034 Pascal @code{with} statement: type,,0,0,offset (Solaris2).
3037 Gould non-base registers; see @ref{Gould}.
3040 Gould non-base registers; see @ref{Gould}.
3043 Gould non-base registers; see @ref{Gould}.
3046 Gould non-base registers; see @ref{Gould}.
3049 Gould non-base registers; see @ref{Gould}.
3052 @c Restore the default table indent
3057 @node Symbol Descriptors
3058 @appendix Table of Symbol Descriptors
3060 The symbol descriptor is the character which follows the colon in many
3061 stabs, and which tells what kind of stab it is. @xref{String Field},
3062 for more information about their use.
3064 @c Please keep this alphabetical
3066 @c In TeX, this looks great, digit is in italics. But makeinfo insists
3067 @c on putting it in `', not realizing that @var should override @code.
3068 @c I don't know of any way to make makeinfo do the right thing. Seems
3069 @c like a makeinfo bug to me.
3073 Variable on the stack; see @ref{Stack Variables}.
3076 Parameter passed by reference in register; see @ref{Reference Parameters}.
3079 Based variable; see @ref{Based Variables}.
3082 Constant; see @ref{Constants}.
3085 Conformant array bound (Pascal, maybe other languages); @ref{Conformant
3086 Arrays}. Name of a caught exception (GNU C++). These can be
3087 distinguished because the latter uses @code{N_CATCH} and the former uses
3088 another symbol type.
3091 Floating point register variable; see @ref{Register Variables}.
3094 Parameter in floating point register; see @ref{Register Parameters}.
3097 File scope function; see @ref{Procedures}.
3100 Global function; see @ref{Procedures}.
3103 Global variable; see @ref{Global Variables}.
3106 @xref{Register Parameters}.
3109 Internal (nested) procedure; see @ref{Nested Procedures}.
3112 Internal (nested) function; see @ref{Nested Procedures}.
3115 Label name (documented by AIX, no further information known).
3118 Module; see @ref{Procedures}.
3121 Argument list parameter; see @ref{Parameters}.
3127 Fortran Function parameter; see @ref{Parameters}.
3130 Unfortunately, three separate meanings have been independently invented
3131 for this symbol descriptor. At least the GNU and Sun uses can be
3132 distinguished by the symbol type. Global Procedure (AIX) (symbol type
3133 used unknown); see @ref{Procedures}. Register parameter (GNU) (symbol
3134 type @code{N_PSYM}); see @ref{Parameters}. Prototype of function
3135 referenced by this file (Sun @code{acc}) (symbol type @code{N_FUN}).
3138 Static Procedure; see @ref{Procedures}.
3141 Register parameter; see @ref{Register Parameters}.
3144 Register variable; see @ref{Register Variables}.
3147 File scope variable; see @ref{Statics}.
3150 Type name; see @ref{Typedefs}.
3153 Enumeration, structure, or union tag; see @ref{Typedefs}.
3156 Parameter passed by reference; see @ref{Reference Parameters}.
3159 Procedure scope static variable; see @ref{Statics}.
3162 Conformant array; see @ref{Conformant Arrays}.
3165 Function return variable; see @ref{Parameters}.
3168 @node Type Descriptors
3169 @appendix Table of Type Descriptors
3171 The type descriptor is the character which follows the type number and
3172 an equals sign. It specifies what kind of type is being defined.
3173 @xref{String Field}, for more information about their use.
3178 Type reference; see @ref{String Field}.
3181 Reference to builtin type; see @ref{Negative Type Numbers}.
3184 Method (C++); see @ref{Cplusplus}.
3187 Pointer; see @ref{Miscellaneous Types}.
3193 Type Attributes (AIX); see @ref{String Field}. Member (class and variable)
3194 type (GNU C++); see @ref{Cplusplus}.
3197 Array; see @ref{Arrays}.
3200 Open array; see @ref{Arrays}.
3203 Pascal space type (AIX); see @ref{Miscellaneous Types}. Builtin integer
3204 type (Sun); see @ref{Builtin Type Descriptors}.
3207 Volatile-qualified type; see @ref{Miscellaneous Types}.
3210 Complex builtin type; see @ref{Builtin Type Descriptors}.
3213 COBOL Picture type. See AIX documentation for details.
3216 File type; see @ref{Miscellaneous Types}.
3219 N-dimensional dynamic array; see @ref{Arrays}.
3222 Enumeration type; see @ref{Enumerations}.
3225 N-dimensional subarray; see @ref{Arrays}.
3228 Function type; see @ref{Function Types}.
3231 Pascal function parameter; see @ref{Function Types}
3234 Builtin floating point type; see @ref{Builtin Type Descriptors}.
3237 COBOL Group. See AIX documentation for details.
3240 Imported type; see @ref{Cross-References}.
3243 Const-qualified type; see @ref{Miscellaneous Types}.
3246 COBOL File Descriptor. See AIX documentation for details.
3249 Multiple instance type; see @ref{Miscellaneous Types}.
3252 String type; see @ref{Strings}.
3255 Stringptr; see @ref{Strings}.
3258 Opaque type; see @ref{Typedefs}.
3261 Procedure; see @ref{Function Types}.
3264 Packed array; see @ref{Arrays}.
3267 Range type; see @ref{Subranges}.
3270 Builtin floating type; see @ref{Builtin Type Descriptors} (Sun). Pascal
3271 subroutine parameter; see @ref{Function Types} (AIX). Detecting this
3272 conflict is possible with careful parsing (hint: a Pascal subroutine
3273 parameter type will always contain a comma, and a builtin type
3274 descriptor never will).
3277 Structure type; see @ref{Structures}.
3280 Set type; see @ref{Miscellaneous Types}.
3283 Union; see @ref{Unions}.
3286 Variant record. This is a Pascal and Modula-2 feature which is like a
3287 union within a struct in C. See AIX documentation for details.
3290 Wide character; see @ref{Builtin Type Descriptors}.
3293 Cross-reference; see @ref{Cross-References}.
3296 gstring; see @ref{Strings}.
3299 @node Expanded Reference
3300 @appendix Expanded Reference by Stab Type
3302 @c FIXME: This appendix should go away; see N_PSYM or N_SO for an example.
3304 For a full list of stab types, and cross-references to where they are
3305 described, see @ref{Stab Types}. This appendix just duplicates certain
3306 information from the main body of this document; eventually the
3307 information will all be in one place.
3311 The first line is the symbol type (see @file{include/aout/stab.def}).
3313 The second line describes the language constructs the symbol type
3316 The third line is the stab format with the significant stab fields
3317 named and the rest NIL.
3319 Subsequent lines expand upon the meaning and possible values for each
3320 significant stab field. @samp{#} stands in for the type descriptor.
3322 Finally, any further information.
3325 * N_PC:: Pascal global symbol
3326 * N_NSYMS:: Number of symbols
3327 * N_NOMAP:: No DST map
3328 * N_M2C:: Modula-2 compilation unit
3329 * N_BROWS:: Path to .cb file for Sun source code browser
3330 * N_DEFD:: GNU Modula2 definition module dependency
3331 * N_EHDECL:: GNU C++ exception variable
3332 * N_MOD2:: Modula2 information "for imc"
3333 * N_CATCH:: GNU C++ "catch" clause
3334 * N_SSYM:: Structure or union element
3335 * N_ENTRY:: Alternate entry point
3336 * N_SCOPE:: Modula2 scope information (Sun only)
3337 * Gould:: non-base register symbols used on Gould systems
3338 * N_LENG:: Length of preceding entry
3344 @deffn @code{.stabs} N_PC
3346 Global symbol (for Pascal).
3349 "name" -> "symbol_name" <<?>>
3350 value -> supposedly the line number (stab.def is skeptical)
3354 @file{stabdump.c} says:
3356 global pascal symbol: name,,0,subtype,line
3364 @deffn @code{.stabn} N_NSYMS
3366 Number of symbols (according to Ultrix V4.0).
3369 0, files,,funcs,lines (stab.def)
3376 @deffn @code{.stabs} N_NOMAP
3378 No DST map for symbol (according to Ultrix V4.0). I think this means a
3379 variable has been optimized out.
3382 name, ,0,type,ignored (stab.def)
3389 @deffn @code{.stabs} N_M2C
3391 Modula-2 compilation unit.
3394 "string" -> "unit_name,unit_time_stamp[,code_time_stamp]"
3396 value -> 0 (main unit)
3404 @deffn @code{.stabs} N_BROWS
3406 Sun source code browser, path to @file{.cb} file
3409 "path to associated @file{.cb} file"
3411 Note: N_BROWS has the same value as N_BSLINE.
3417 @deffn @code{.stabn} N_DEFD
3419 GNU Modula2 definition module dependency.
3421 GNU Modula-2 definition module dependency. The value is the
3422 modification time of the definition file. The other field is non-zero
3423 if it is imported with the GNU M2 keyword @code{%INITIALIZE}. Perhaps
3424 @code{N_M2C} can be used if there are enough empty fields?
3430 @deffn @code{.stabs} N_EHDECL
3432 GNU C++ exception variable <<?>>.
3434 "@var{string} is variable name"
3436 Note: conflicts with @code{N_MOD2}.
3442 @deffn @code{.stab?} N_MOD2
3444 Modula2 info "for imc" (according to Ultrix V4.0)
3446 Note: conflicts with @code{N_EHDECL} <<?>>
3452 @deffn @code{.stabn} N_CATCH
3454 GNU C++ @code{catch} clause
3456 GNU C++ @code{catch} clause. The value is its address. The desc field
3457 is nonzero if this entry is immediately followed by a @code{CAUGHT} stab
3458 saying what exception was caught. Multiple @code{CAUGHT} stabs means
3459 that multiple exceptions can be caught here. If desc is 0, it means all
3460 exceptions are caught here.
3466 @deffn @code{.stabn} N_SSYM
3468 Structure or union element.
3470 The value is the offset in the structure.
3472 <<?looking at structs and unions in C I didn't see these>>
3478 @deffn @code{.stabn} N_ENTRY
3480 Alternate entry point.
3481 The value is its address.
3488 @deffn @code{.stab?} N_SCOPE
3490 Modula2 scope information (Sun linker)
3495 @section Non-base registers on Gould systems
3497 @deffn @code{.stab?} N_NBTEXT
3498 @deffnx @code{.stab?} N_NBDATA
3499 @deffnx @code{.stab?} N_NBBSS
3500 @deffnx @code{.stab?} N_NBSTS
3501 @deffnx @code{.stab?} N_NBLCS
3507 These are used on Gould systems for non-base registers syms.
3509 However, the following values are not the values used by Gould; they are
3510 the values which GNU has been documenting for these values for a long
3511 time, without actually checking what Gould uses. I include these values
3512 only because perhaps some someone actually did something with the GNU
3513 information (I hope not, why GNU knowingly assigned wrong values to
3514 these in the header file is a complete mystery to me).
3517 240 0xf0 N_NBTEXT ??
3518 242 0xf2 N_NBDATA ??
3528 @deffn @code{.stabn} N_LENG
3530 Second symbol entry containing a length-value for the preceding entry.
3531 The value is the length.
3535 @appendix Questions and Anomalies
3539 @c I think this is changed in GCC 2.4.5 to put the line number there.
3540 For GNU C stabs defining local and global variables (@code{N_LSYM} and
3541 @code{N_GSYM}), the desc field is supposed to contain the source
3542 line number on which the variable is defined. In reality the desc
3543 field is always 0. (This behavior is defined in @file{dbxout.c} and
3544 putting a line number in desc is controlled by @samp{#ifdef
3545 WINNING_GDB}, which defaults to false). GDB supposedly uses this
3546 information if you say @samp{list @var{var}}. In reality, @var{var} can
3547 be a variable defined in the program and GDB says @samp{function
3548 @var{var} not defined}.
3551 In GNU C stabs, there seems to be no way to differentiate tag types:
3552 structures, unions, and enums (symbol descriptor @samp{T}) and typedefs
3553 (symbol descriptor @samp{t}) defined at file scope from types defined locally
3554 to a procedure or other more local scope. They all use the @code{N_LSYM}
3555 stab type. Types defined at procedure scope are emited after the
3556 @code{N_RBRAC} of the preceding function and before the code of the
3557 procedure in which they are defined. This is exactly the same as
3558 types defined in the source file between the two procedure bodies.
3559 GDB overcompensates by placing all types in block #1, the block for
3560 symbols of file scope. This is true for default, @samp{-ansi} and
3561 @samp{-traditional} compiler options. (Bugs gcc/1063, gdb/1066.)
3564 What ends the procedure scope? Is it the proc block's @code{N_RBRAC} or the
3565 next @code{N_FUN}? (I believe its the first.)
3568 @c FIXME: This should go with the other stuff about global variables.
3569 Global variable stabs don't have location information. This comes
3570 from the external symbol for the same variable. The external symbol
3571 has a leading underbar on the _name of the variable and the stab does
3572 not. How do we know these two symbol table entries are talking about
3573 the same symbol when their names are different? (Answer: the debugger
3574 knows that external symbols have leading underbars).
3576 @c FIXME: This is absurdly vague; there all kinds of differences, some
3577 @c of which are the same between gnu & sun, and some of which aren't.
3578 @c In particular, I'm pretty sure GCC works with Sun dbx by default.
3580 @c Can GCC be configured to output stabs the way the Sun compiler
3581 @c does, so that their native debugging tools work? <NO?> It doesn't by
3582 @c default. GDB reads either format of stab. (GCC or SunC). How about
3586 @node XCOFF Differences
3587 @appendix Differences Between GNU Stabs in a.out and GNU Stabs in XCOFF
3589 @c FIXME: Merge *all* these into the main body of the document.
3590 The AIX/RS6000 native object file format is XCOFF with stabs. This
3591 appendix only covers those differences which are not covered in the main
3592 body of this document.
3596 BSD a.out stab types correspond to AIX XCOFF storage classes. In general
3597 the mapping is @code{N_@var{stabtype}} becomes @code{C_@var{stabtype}}.
3598 Some stab types in a.out are not supported in XCOFF; most of these use
3601 @c FIXME: Get C_* types for the block, figure out whether it is always
3602 @c used (I suspect not), explain clearly, and move to node Statics.
3603 Exception: initialised static @code{N_STSYM} and un-initialized static
3604 @code{N_LCSYM} both map to the @code{C_STSYM} storage class. But the
3605 distinction is preserved because in XCOFF @code{N_STSYM} and
3606 @code{N_LCSYM} must be emited in a named static block. Begin the block
3607 with @samp{.bs s[RW] data_section_name} for @code{N_STSYM} or @samp{.bs
3608 s bss_section_name} for @code{N_LCSYM}. End the block with @samp{.es}.
3610 @c FIXME: I think they are trying to say something about whether the
3611 @c assembler defaults the value to the location counter.
3613 If the XCOFF stab is an @code{N_FUN} (@code{C_FUN}) then follow the
3614 string field with @samp{,.} instead of just @samp{,}.
3617 I think that's it for @file{.s} file differences. They could stand to be
3618 better presented. This is just a list of what I have noticed so far.
3619 There are a @emph{lot} of differences in the information in the symbol
3620 tables of the executable and object files.
3622 Mapping of a.out stab types to XCOFF storage classes:
3625 stab type storage class
3626 -------------------------------
3662 @node Sun Differences
3663 @appendix Differences Between GNU Stabs and Sun Native Stabs
3665 @c FIXME: Merge all this stuff into the main body of the document.
3669 GNU C stabs define @emph{all} types, file or procedure scope, as
3670 @code{N_LSYM}. Sun doc talks about using @code{N_GSYM} too.
3673 Sun C stabs use type number pairs in the format
3674 (@var{file-number},@var{type-number}) where @var{file-number} is a
3675 number starting with 1 and incremented for each sub-source file in the
3676 compilation. @var{type-number} is a number starting with 1 and
3677 incremented for each new type defined in the compilation. GNU C stabs
3678 use the type number alone, with no source file number.
3682 @appendix Using Stabs With The ELF Object File Format
3684 The ELF object file format allows tools to create object files with
3685 custom sections containing any arbitrary data. To use stabs in ELF
3686 object files, the tools create two custom sections, a section named
3687 @code{.stab} which contains an array of fixed length structures, one
3688 struct per stab, and a section named @code{.stabstr} containing all the
3689 variable length strings that are referenced by stabs in the @code{.stab}
3690 section. The byte order of the stabs binary data matches the byte order
3691 of the ELF file itself, as determined from the @code{EI_DATA} field in
3692 the @code{e_ident} member of the ELF header.
3694 The first stab in the @code{.stab} section for each compilation unit is
3695 synthetic, generated entirely by the assembler, with no corresponding
3696 @code{.stab} directive as input to the assembler. This stab contains
3697 the following fields:
3701 Offset in the @code{.stabstr} section to the source filename.
3707 Unused field, always zero.
3710 Count of upcoming symbols, i.e., the number of remaining stabs for this
3714 Size of the string table fragment associated with this source file, in
3718 The @code{.stabstr} section always starts with a null byte (so that string
3719 offsets of zero reference a null string), followed by random length strings,
3720 each of which is null byte terminated.
3722 The ELF section header for the @code{.stab} section has its
3723 @code{sh_link} member set to the section number of the @code{.stabstr}
3724 section, and the @code{.stabstr} section has its ELF section
3725 header @code{sh_type} member set to @code{SHT_STRTAB} to mark it as a
3728 To keep linking fast, it is a bad idea to have the linker relocating
3729 stabs, so (except for Solaris 2.2 and earlier, see below) none of the
3730 addresses in the @code{n_value} field of the stabs are relocated by the
3731 linker. Instead they are relative to the source file (or some entity
3732 smaller than a source file, like a function). To find the address of
3733 each section corresponding to a given source file, the compiler puts out
3734 symbols giving the address of each section for a given source file.
3735 Since these are ELF (not stab) symbols, the linker can relocate them
3736 correctly. They are named @code{Bbss.bss} for the bss section,
3737 @code{Ddata.data} for the data section, and @code{Drodata.rodata} for
3738 the rodata section. For the text section, there is no such symbol. For
3739 an example of how these symbols work, @xref{ELF Transformations}. GCC
3740 does not provide these symbols; it instead relies on the stabs getting
3741 relocated, which loses for Solaris 2.3 (see below). Thus address which
3742 would normally be relative to @code{Bbss.bss}, etc., are absolute. The
3743 linker provided with Solaris 2.2 and earlier relocates stabs using
3744 relocation information from a @code{.rela.stab} section, which means
3745 that the value of an @code{N_FUN} stab in an executable is the actual
3746 address. I think this is just standard ELF relocations, as it would do
3747 for any section, rather than a special-purpose stabs hack. For Solaris
3748 2.3 and later, the linker ignores relocations for the stabs section.
3749 The value of a @code{N_FUN} stab is zero and the address of a function
3750 can be obtained from the ELF (non-stab) symbols. Sun, in reference to
3751 bug 1142109, has verified that this is intentional. Because looking
3752 things up in the ELF symbols would probably be slow, and this doesn't
3753 provide any way to deal with nested functions, it would probably be
3754 better to use a @code{Ttext.text} symbol for stabs-in-elf on non-Solaris
3755 machines, and make the address in the @code{N_FUN} relative to the
3756 @code{Ttext.text} symbol. In addition to @code{N_FUN} symbols, whether
3757 the linker relocates stabs also affects some @code{N_ROSYM},
3758 @code{N_STSYM}, and @code{N_LCSYM} symbols; see @ref{Statics}.
3760 @node Symbol Types Index
3761 @unnumbered Symbol Types Index