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 value is the
433 text address corresponding to the end of the previous include file and
434 the start of this one. To specify the main source file again, use an
435 @code{N_SOL} symbol with the name of the main source file.
440 The @code{N_BINCL} approach works as follows. An @code{N_BINCL} symbol
441 specifies the start of an include file. In an object file, only the
442 string is significant; the Sun linker puts data into some of the
443 other fields. The end of the include file is marked by an
444 @code{N_EINCL} symbol (which has no string field). In an object
445 file, there is no significant data in the @code{N_EINCL} symbol; the Sun
446 linker puts data into some of the fields. @code{N_BINCL} and
447 @code{N_EINCL} can be nested.
449 If the linker detects that two source files have identical stabs between
450 an @code{N_BINCL} and @code{N_EINCL} pair (as will generally be the case
451 for a header file), then it only puts out the stabs once. Each
452 additional occurance is replaced by an @code{N_EXCL} symbol. I believe
453 the Sun (SunOS4, not sure about Solaris) linker is the only one which
454 supports this feature.
455 @c What do the fields of N_EXCL contain? -djm
459 For the start of an include file in XCOFF, use the @file{.bi} assembler
460 directive, which generates a @code{C_BINCL} symbol. A @file{.ei}
461 directive, which generates a @code{C_EINCL} symbol, denotes the end of
462 the include file. Both directives are followed by the name of the
463 source file in quotes, which becomes the string for the symbol.
464 The value of each symbol, produced automatically by the assembler
465 and linker, is the offset into the executable of the beginning
466 (inclusive, as you'd expect) or end (inclusive, as you would not expect)
467 of the portion of the COFF line table that corresponds to this include
468 file. @code{C_BINCL} and @code{C_EINCL} do not nest.
471 @section Line Numbers
474 An @code{N_SLINE} symbol represents the start of a source line. The
475 desc field contains the line number and the value
476 contains the code address for the start of that source line. On most
477 machines the address is absolute; for Sun's stabs-in-ELF, it is relative
478 to the function in which the @code{N_SLINE} symbol occurs.
482 GNU documents @code{N_DSLINE} and @code{N_BSLINE} symbols for line
483 numbers in the data or bss segments, respectively. They are identical
484 to @code{N_SLINE} but are relocated differently by the linker. They
485 were intended to be used to describe the source location of a variable
486 declaration, but I believe that GCC2 actually puts the line number in
487 the desc field of the stab for the variable itself. GDB has been
488 ignoring these symbols (unless they contain a string field) since
491 For single source lines that generate discontiguous code, such as flow
492 of control statements, there may be more than one line number entry for
493 the same source line. In this case there is a line number entry at the
494 start of each code range, each with the same line number.
496 XCOFF does not use stabs for line numbers. Instead, it uses COFF line
497 numbers (which are outside the scope of this document). Standard COFF
498 line numbers cannot deal with include files, but in XCOFF this is fixed
499 with the @code{C_BINCL} method of marking include files (@pxref{Include
505 @findex N_FUN, for functions
507 @findex N_STSYM, for functions (Sun acc)
508 @findex N_GSYM, for functions (Sun acc)
509 All of the following stabs normally use the @code{N_FUN} symbol type.
510 However, Sun's @code{acc} compiler on SunOS4 uses @code{N_GSYM} and
511 @code{N_STSYM}, which means that the value of the stab for the function
512 is useless and the debugger must get the address of the function from
513 the non-stab symbols instead. BSD Fortran is said to use @code{N_FNAME}
514 with the same restriction; the value of the symbol is not useful (I'm
515 not sure it really does use this, because GDB doesn't handle this and no
518 A function is represented by an @samp{F} symbol descriptor for a global
519 (extern) function, and @samp{f} for a static (local) function. The
520 value is the address of the start of the function (absolute
521 for @code{a.out}; relative to the start of the file for Sun's
522 stabs-in-ELF). The type information of the stab represents the return
523 type of the function; thus @samp{foo:f5} means that foo is a function
524 returning type 5. There is no need to try to get the line number of the
525 start of the function from the stab for the function; it is in the next
526 @code{N_SLINE} symbol.
528 @c FIXME: verify whether the "I suspect" below is true or not.
529 Some compilers (such as Sun's Solaris compiler) support an extension for
530 specifying the types of the arguments. I suspect this extension is not
531 used for old (non-prototyped) function definitions in C. If the
532 extension is in use, the type information of the stab for the function
533 is followed by type information for each argument, with each argument
534 preceded by @samp{;}. An argument type of 0 means that additional
535 arguments are being passed, whose types and number may vary (@samp{...}
536 in ANSI C). GDB has tolerated this extension (parsed the syntax, if not
537 necessarily used the information) since at least version 4.8; I don't
538 know whether all versions of dbx tolerate it. The argument types given
539 here are not redundant with the symbols for the formal parameters
540 (@pxref{Parameters}); they are the types of the arguments as they are
541 passed, before any conversions might take place. For example, if a C
542 function which is declared without a prototype takes a @code{float}
543 argument, the value is passed as a @code{double} but then converted to a
544 @code{float}. Debuggers need to use the types given in the arguments
545 when printing values, but when calling the function they need to use the
546 types given in the symbol defining the function.
548 If the return type and types of arguments of a function which is defined
549 in another source file are specified (i.e., a function prototype in ANSI
550 C), traditionally compilers emit no stab; the only way for the debugger
551 to find the information is if the source file where the function is
552 defined was also compiled with debugging symbols. As an extension the
553 Solaris compiler uses symbol descriptor @samp{P} followed by the return
554 type of the function, followed by the arguments, each preceded by
555 @samp{;}, as in a stab with symbol descriptor @samp{f} or @samp{F}.
556 This use of symbol descriptor @samp{P} can be distinguished from its use
557 for register parameters (@pxref{Register Parameters}) by the fact that it has
558 symbol type @code{N_FUN}.
560 The AIX documentation also defines symbol descriptor @samp{J} as an
561 internal function. I assume this means a function nested within another
562 function. It also says symbol descriptor @samp{m} is a module in
563 Modula-2 or extended Pascal.
565 Procedures (functions which do not return values) are represented as
566 functions returning the @code{void} type in C. I don't see why this couldn't
567 be used for all languages (inventing a @code{void} type for this purpose if
568 necessary), but the AIX documentation defines @samp{I}, @samp{P}, and
569 @samp{Q} for internal, global, and static procedures, respectively.
570 These symbol descriptors are unusual in that they are not followed by
573 The following example shows a stab for a function @code{main} which
574 returns type number @code{1}. The @code{_main} specified for the value
575 is a reference to an assembler label which is used to fill in the start
576 address of the function.
579 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
582 The stab representing a procedure is located immediately following the
583 code of the procedure. This stab is in turn directly followed by a
584 group of other stabs describing elements of the procedure. These other
585 stabs describe the procedure's parameters, its block local variables, and
588 @node Nested Procedures
589 @section Nested Procedures
591 For any of the symbol descriptors representing procedures, after the
592 symbol descriptor and the type information is optionally a scope
593 specifier. This consists of a comma, the name of the procedure, another
594 comma, and the name of the enclosing procedure. The first name is local
595 to the scope specified, and seems to be redundant with the name of the
596 symbol (before the @samp{:}). This feature is used by GCC, and
597 presumably Pascal, Modula-2, etc., compilers, for nested functions.
599 If procedures are nested more than one level deep, only the immediately
600 containing scope is specified. For example, this code:
612 return baz (x + 2 * y);
614 return x + bar (3 * x);
622 .stabs "baz:f1,baz,bar",36,0,0,_baz.15 # @r{36 is N_FUN}
623 .stabs "bar:f1,bar,foo",36,0,0,_bar.12
624 .stabs "foo:F1",36,0,0,_foo
627 @node Block Structure
628 @section Block Structure
632 The program's block structure is represented by the @code{N_LBRAC} (left
633 brace) and the @code{N_RBRAC} (right brace) stab types. The variables
634 defined inside a block precede the @code{N_LBRAC} symbol for most
635 compilers, including GCC. Other compilers, such as the Convex, Acorn
636 RISC machine, and Sun @code{acc} compilers, put the variables after the
637 @code{N_LBRAC} symbol. The values of the @code{N_LBRAC} and
638 @code{N_RBRAC} symbols are the start and end addresses of the code of
639 the block, respectively. For most machines, they are relative to the
640 starting address of this source file. For the Gould NP1, they are
641 absolute. For Sun's stabs-in-ELF, they are relative to the function in
644 The @code{N_LBRAC} and @code{N_RBRAC} stabs that describe the block
645 scope of a procedure are located after the @code{N_FUN} stab that
646 represents the procedure itself.
648 Sun documents the desc field of @code{N_LBRAC} and
649 @code{N_RBRAC} symbols as containing the nesting level of the block.
650 However, dbx seems to not care, and GCC always sets desc to
656 The @samp{c} symbol descriptor indicates that this stab represents a
657 constant. This symbol descriptor is an exception to the general rule
658 that symbol descriptors are followed by type information. Instead, it
659 is followed by @samp{=} and one of the following:
663 Boolean constant. @var{value} is a numeric value; I assume it is 0 for
667 Character constant. @var{value} is the numeric value of the constant.
669 @item e @var{type-information} , @var{value}
670 Constant whose value can be represented as integral.
671 @var{type-information} is the type of the constant, as it would appear
672 after a symbol descriptor (@pxref{String Field}). @var{value} is the
673 numeric value of the constant. GDB 4.9 does not actually get the right
674 value if @var{value} does not fit in a host @code{int}, but it does not
675 do anything violent, and future debuggers could be extended to accept
676 integers of any size (whether unsigned or not). This constant type is
677 usually documented as being only for enumeration constants, but GDB has
678 never imposed that restriction; I don't know about other debuggers.
681 Integer constant. @var{value} is the numeric value. The type is some
682 sort of generic integer type (for GDB, a host @code{int}); to specify
683 the type explicitly, use @samp{e} instead.
686 Real constant. @var{value} is the real value, which can be @samp{INF}
687 (optionally preceded by a sign) for infinity, @samp{QNAN} for a quiet
688 NaN (not-a-number), or @samp{SNAN} for a signalling NaN. If it is a
689 normal number the format is that accepted by the C library function
693 String constant. @var{string} is a string enclosed in either @samp{'}
694 (in which case @samp{'} characters within the string are represented as
695 @samp{\'} or @samp{"} (in which case @samp{"} characters within the
696 string are represented as @samp{\"}).
698 @item S @var{type-information} , @var{elements} , @var{bits} , @var{pattern}
699 Set constant. @var{type-information} is the type of the constant, as it
700 would appear after a symbol descriptor (@pxref{String Field}).
701 @var{elements} is the number of elements in the set (does this means
702 how many bits of @var{pattern} are actually used, which would be
703 redundant with the type, or perhaps the number of bits set in
704 @var{pattern}? I don't get it), @var{bits} is the number of bits in the
705 constant (meaning it specifies the length of @var{pattern}, I think),
706 and @var{pattern} is a hexadecimal representation of the set. AIX
707 documentation refers to a limit of 32 bytes, but I see no reason why
708 this limit should exist. This form could probably be used for arbitrary
709 constants, not just sets; the only catch is that @var{pattern} should be
710 understood to be target, not host, byte order and format.
713 The boolean, character, string, and set constants are not supported by
714 GDB 4.9, but it ignores them. GDB 4.8 and earlier gave an error
715 message and refused to read symbols from the file containing the
718 The above information is followed by @samp{;}.
723 Different types of stabs describe the various ways that variables can be
724 allocated: on the stack, globally, in registers, in common blocks,
725 statically, or as arguments to a function.
728 * Stack Variables:: Variables allocated on the stack.
729 * Global Variables:: Variables used by more than one source file.
730 * Register Variables:: Variables in registers.
731 * Common Blocks:: Variables statically allocated together.
732 * Statics:: Variables local to one source file.
733 * Based Variables:: Fortran pointer based variables.
734 * Parameters:: Variables for arguments to functions.
737 @node Stack Variables
738 @section Automatic Variables Allocated on the Stack
740 If a variable's scope is local to a function and its lifetime is only as
741 long as that function executes (C calls such variables
742 @dfn{automatic}), it can be allocated in a register (@pxref{Register
743 Variables}) or on the stack.
746 Each variable allocated on the stack has a stab with the symbol
747 descriptor omitted. Since type information should begin with a digit,
748 @samp{-}, or @samp{(}, only those characters precluded from being used
749 for symbol descriptors. However, the Acorn RISC machine (ARM) is said
750 to get this wrong: it puts out a mere type definition here, without the
751 preceding @samp{@var{type-number}=}. This is a bad idea; there is no
752 guarantee that type descriptors are distinct from symbol descriptors.
753 Stabs for stack variables use the @code{N_LSYM} stab type.
755 The value of the stab is the offset of the variable within the
756 local variables. On most machines this is an offset from the frame
757 pointer and is negative. The location of the stab specifies which block
758 it is defined in; see @ref{Block Structure}.
760 For example, the following C code:
770 produces the following stabs:
773 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
774 .stabs "x:1",128,0,0,-12 # @r{128 is N_LSYM}
775 .stabn 192,0,0,LBB2 # @r{192 is N_LBRAC}
776 .stabn 224,0,0,LBE2 # @r{224 is N_RBRAC}
779 @xref{Procedures} for more information on the @code{N_FUN} stab, and
780 @ref{Block Structure} for more information on the @code{N_LBRAC} and
781 @code{N_RBRAC} stabs.
783 @node Global Variables
784 @section Global Variables
787 A variable whose scope is not specific to just one source file is
788 represented by the @samp{G} symbol descriptor. These stabs use the
789 @code{N_GSYM} stab type. The type information for the stab
790 (@pxref{String Field}) gives the type of the variable.
792 For example, the following source code:
799 yields the following assembly code:
802 .stabs "g_foo:G2",32,0,0,0 # @r{32 is N_GSYM}
809 The address of the variable represented by the @code{N_GSYM} is not
810 contained in the @code{N_GSYM} stab. The debugger gets this information
811 from the external symbol for the global variable. In the example above,
812 the @code{.global _g_foo} and @code{_g_foo:} lines tell the assembler to
813 produce an external symbol.
815 @node Register Variables
816 @section Register Variables
819 @c According to an old version of this manual, AIX uses C_RPSYM instead
820 @c of C_RSYM. I am skeptical; this should be verified.
821 Register variables have their own stab type, @code{N_RSYM}, and their
822 own symbol descriptor, @samp{r}. The stab's value is the
823 number of the register where the variable data will be stored.
824 @c .stabs "name:type",N_RSYM,0,RegSize,RegNumber (Sun doc)
826 AIX defines a separate symbol descriptor @samp{d} for floating point
827 registers. This seems unnecessary; why not just just give floating
828 point registers different register numbers? I have not verified whether
829 the compiler actually uses @samp{d}.
831 If the register is explicitly allocated to a global variable, but not
835 register int g_bar asm ("%g5");
839 then the stab may be emitted at the end of the object file, with
840 the other bss symbols.
843 @section Common Blocks
845 A common block is a statically allocated section of memory which can be
846 referred to by several source files. It may contain several variables.
847 I believe Fortran is the only language with this feature.
851 A @code{N_BCOMM} stab begins a common block and an @code{N_ECOMM} stab
852 ends it. The only field that is significant in these two stabs is the
853 string, which names a normal (non-debugging) symbol that gives the
854 address of the common block.
857 Each stab between the @code{N_BCOMM} and the @code{N_ECOMM} specifies a
858 member of that common block; its value is the offset within the
859 common block of that variable. The @code{N_ECOML} stab type is
860 documented for this purpose, but Sun's Fortran compiler uses
861 @code{N_GSYM} instead. The test case I looked at had a common block
862 local to a function and it used the @samp{V} symbol descriptor; I assume
863 one would use @samp{S} if not local to a function (that is, if a common
864 block @emph{can} be anything other than local to a function).
867 @section Static Variables
869 Initialized static variables are represented by the @samp{S} and
870 @samp{V} symbol descriptors. @samp{S} means file scope static, and
871 @samp{V} means procedure scope static.
873 @c This is probably not worth mentioning; it is only true on the sparc
874 @c for `double' variables which although declared const are actually in
875 @c the data segment (the text segment can't guarantee 8 byte alignment).
877 @c 2.4.5 has a bug in that it uses @code{N_FUN}, so neither dbx nor GDB can
878 @c find the variables)
881 @findex N_FUN, for variables
883 In a.out files, @code{N_STSYM} means the data section, @code{N_FUN}
884 means the text section, and @code{N_LCSYM} means the bss section. For
885 those systems with a read-only data section separate from the text
886 section (Solaris), @code{N_ROSYM} means the read-only data section.
888 For example, the source lines:
891 static const int var_const = 5;
892 static int var_init = 2;
893 static int var_noinit;
897 yield the following stabs:
900 .stabs "var_const:S1",36,0,0,_var_const # @r{36 is N_FUN}
902 .stabs "var_init:S1",38,0,0,_var_init # @r{38 is N_STSYM}
904 .stabs "var_noinit:S1",40,0,0,_var_noinit # @r{40 is N_LCSYM}
907 In XCOFF files, each symbol has a section number, so the stab type
908 need not indicate the section.
910 In ECOFF files, the storage class is used to specify the section, so the
911 stab type need not indicate the section.
913 In ELF files, for Solaris 2.1, symbol descriptor @samp{S} means that the
914 address is absolute (ld relocates it) and symbol descriptor @samp{V}
915 means that the address is relative to the start of the relevant section
916 for that compilation unit. I don't know what it does for @samp{S} stabs
917 on Solaris 2.3 (in which ld no longer relocates stabs). For more
918 information on ld stab relocation, @xref{Stabs In ELF}.
920 @node Based Variables
921 @section Fortran Based Variables
923 Fortran (at least, the Sun and SGI dialects of FORTRAN-77) has a feature
924 which allows allocating arrays with @code{malloc}, but which avoids
925 blurring the line between arrays and pointers the way that C does. In
926 stabs such a variable uses the @samp{b} symbol descriptor.
928 For example, the Fortran declarations
931 real foo, foo10(10), foo10_5(10,5)
933 pointer (foo10p, foo10)
934 pointer (foo105p, foo10_5)
942 foo10_5:bar3;1;5;ar3;1;10;6
945 In this example, @code{real} is type 6 and type 3 is an integral type
946 which is the type of the subscripts of the array (probably
949 The @samp{b} symbol descriptor is like @samp{V} in that it denotes a
950 statically allocated symbol whose scope is local to a function; see
951 @xref{Statics}. The value of the symbol, instead of being the address
952 of the variable itself, is the address of a pointer to that variable.
953 So in the above example, the value of the @code{foo} stab is the address
954 of a pointer to a real, the value of the @code{foo10} stab is the
955 address of a pointer to a 10-element array of reals, and the value of
956 the @code{foo10_5} stab is the address of a pointer to a 5-element array
957 of 10-element arrays of reals.
962 Formal parameters to a function are represented by a stab (or sometimes
963 two; see below) for each parameter. The stabs are in the order in which
964 the debugger should print the parameters (i.e., the order in which the
965 parameters are declared in the source file). The exact form of the stab
966 depends on how the parameter is being passed.
969 Parameters passed on the stack use the symbol descriptor @samp{p} and
970 the @code{N_PSYM} symbol type. The value of the symbol is an offset
971 used to locate the parameter on the stack; its exact meaning is
972 machine-dependent, but on most machines it is an offset from the frame
975 As a simple example, the code:
986 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
987 .stabs "argc:p1",160,0,0,68 # @r{160 is N_PSYM}
988 .stabs "argv:p20=*21=*2",160,0,0,72
991 The type definition of @code{argv} is interesting because it contains
992 several type definitions. Type 21 is pointer to type 2 (char) and
993 @code{argv} (type 20) is pointer to type 21.
995 @c FIXME: figure out what these mean and describe them coherently.
996 The following symbol descriptors are also said to go with @code{N_PSYM}.
997 The value of the symbol is said to be an offset from the argument
998 pointer (I'm not sure whether this is true or not).
1002 pF Fortran function parameter
1003 X (function result variable)
1007 * Register Parameters::
1008 * Local Variable Parameters::
1009 * Reference Parameters::
1010 * Conformant Arrays::
1013 @node Register Parameters
1014 @subsection Passing Parameters in Registers
1016 If the parameter is passed in a register, then traditionally there are
1017 two symbols for each argument:
1020 .stabs "arg:p1" . . . ; N_PSYM
1021 .stabs "arg:r1" . . . ; N_RSYM
1024 Debuggers use the second one to find the value, and the first one to
1025 know that it is an argument.
1028 @findex N_RSYM, for parameters
1029 Because that approach is kind of ugly, some compilers use symbol
1030 descriptor @samp{P} or @samp{R} to indicate an argument which is in a
1031 register. Symbol type @code{C_RPSYM} is used with @samp{R} and
1032 @code{N_RSYM} is used with @samp{P}. The symbol's value is
1033 the register number. @samp{P} and @samp{R} mean the same thing; the
1034 difference is that @samp{P} is a GNU invention and @samp{R} is an IBM
1035 (XCOFF) invention. As of version 4.9, GDB should handle either one.
1037 There is at least one case where GCC uses a @samp{p} and @samp{r} pair
1038 rather than @samp{P}; this is where the argument is passed in the
1039 argument list and then loaded into a register.
1041 According to the AIX documentation, symbol descriptor @samp{D} is for a
1042 parameter passed in a floating point register. This seems
1043 unnecessary---why not just use @samp{R} with a register number which
1044 indicates that it's a floating point register? I haven't verified
1045 whether the system actually does what the documentation indicates.
1047 @c FIXME: On the hppa this is for any type > 8 bytes, I think, and not
1048 @c for small structures (investigate).
1049 On the sparc and hppa, for a @samp{P} symbol whose type is a structure
1050 or union, the register contains the address of the structure. On the
1051 sparc, this is also true of a @samp{p} and @samp{r} pair (using Sun
1052 @code{cc}) or a @samp{p} symbol. However, if a (small) structure is
1053 really in a register, @samp{r} is used. And, to top it all off, on the
1054 hppa it might be a structure which was passed on the stack and loaded
1055 into a register and for which there is a @samp{p} and @samp{r} pair! I
1056 believe that symbol descriptor @samp{i} is supposed to deal with this
1057 case (it is said to mean "value parameter by reference, indirect
1058 access"; I don't know the source for this information), but I don't know
1059 details or what compilers or debuggers use it, if any (not GDB or GCC).
1060 It is not clear to me whether this case needs to be dealt with
1061 differently than parameters passed by reference (@pxref{Reference Parameters}).
1063 @node Local Variable Parameters
1064 @subsection Storing Parameters as Local Variables
1066 There is a case similar to an argument in a register, which is an
1067 argument that is actually stored as a local variable. Sometimes this
1068 happens when the argument was passed in a register and then the compiler
1069 stores it as a local variable. If possible, the compiler should claim
1070 that it's in a register, but this isn't always done.
1072 @findex N_LSYM, for parameter
1073 Some compilers use the pair of symbols approach described above
1074 (@samp{@var{arg}:p} followed by @samp{@var{arg}:}); this includes GCC1
1075 (not GCC2) on the sparc when passing a small structure and GCC2
1076 (sometimes) when the argument type is @code{float} and it is passed as a
1077 @code{double} and converted to @code{float} by the prologue (in the
1078 latter case the type of the @samp{@var{arg}:p} symbol is @code{double}
1079 and the type of the @samp{@var{arg}:} symbol is @code{float}).
1081 GCC, at least on the 960, has another solution to the same problem. It
1082 uses a single @samp{p} symbol descriptor for an argument which is stored
1083 as a local variable but uses @code{N_LSYM} instead of @code{N_PSYM}. In
1084 this case, the value of the symbol is an offset relative to the local
1085 variables for that function, not relative to the arguments; on some
1086 machines those are the same thing, but not on all.
1088 @c This is mostly just background info; the part that logically belongs
1089 @c here is the last sentence.
1090 On the VAX or on other machines in which the calling convention includes
1091 the number of words of arguments actually passed, the debugger (GDB at
1092 least) uses the parameter symbols to keep track of whether it needs to
1093 print nameless arguments in addition to the formal parameters which it
1094 has printed because each one has a stab. For example, in
1097 extern int fprintf (FILE *stream, char *format, @dots{});
1099 fprintf (stdout, "%d\n", x);
1102 there are stabs for @code{stream} and @code{format}. On most machines,
1103 the debugger can only print those two arguments (because it has no way
1104 of knowing that additional arguments were passed), but on the VAX or
1105 other machines with a calling convention which indicates the number of
1106 words of arguments, the debugger can print all three arguments. To do
1107 so, the parameter symbol (symbol descriptor @samp{p}) (not necessarily
1108 @samp{r} or symbol descriptor omitted symbols) needs to contain the
1109 actual type as passed (for example, @code{double} not @code{float} if it
1110 is passed as a double and converted to a float).
1112 @node Reference Parameters
1113 @subsection Passing Parameters by Reference
1115 If the parameter is passed by reference (e.g., Pascal @code{VAR}
1116 parameters), then the symbol descriptor is @samp{v} if it is in the
1117 argument list, or @samp{a} if it in a register. Other than the fact
1118 that these contain the address of the parameter rather than the
1119 parameter itself, they are identical to @samp{p} and @samp{R},
1120 respectively. I believe @samp{a} is an AIX invention; @samp{v} is
1121 supported by all stabs-using systems as far as I know.
1123 @node Conformant Arrays
1124 @subsection Passing Conformant Array Parameters
1126 @c Is this paragraph correct? It is based on piecing together patchy
1127 @c information and some guesswork
1128 Conformant arrays are a feature of Modula-2, and perhaps other
1129 languages, in which the size of an array parameter is not known to the
1130 called function until run-time. Such parameters have two stabs: a
1131 @samp{x} for the array itself, and a @samp{C}, which represents the size
1132 of the array. The value of the @samp{x} stab is the offset in the
1133 argument list where the address of the array is stored (it this right?
1134 it is a guess); the value of the @samp{C} stab is the offset in the
1135 argument list where the size of the array (in elements? in bytes?) is
1139 @chapter Defining Types
1141 The examples so far have described types as references to previously
1142 defined types, or defined in terms of subranges of or pointers to
1143 previously defined types. This chapter describes the other type
1144 descriptors that may follow the @samp{=} in a type definition.
1147 * Builtin Types:: Integers, floating point, void, etc.
1148 * Miscellaneous Types:: Pointers, sets, files, etc.
1149 * Cross-References:: Referring to a type not yet defined.
1150 * Subranges:: A type with a specific range.
1151 * Arrays:: An aggregate type of same-typed elements.
1152 * Strings:: Like an array but also has a length.
1153 * Enumerations:: Like an integer but the values have names.
1154 * Structures:: An aggregate type of different-typed elements.
1155 * Typedefs:: Giving a type a name.
1156 * Unions:: Different types sharing storage.
1161 @section Builtin Types
1163 Certain types are built in (@code{int}, @code{short}, @code{void},
1164 @code{float}, etc.); the debugger recognizes these types and knows how
1165 to handle them. Thus, don't be surprised if some of the following ways
1166 of specifying builtin types do not specify everything that a debugger
1167 would need to know about the type---in some cases they merely specify
1168 enough information to distinguish the type from other types.
1170 The traditional way to define builtin types is convolunted, so new ways
1171 have been invented to describe them. Sun's @code{acc} uses special
1172 builtin type descriptors (@samp{b} and @samp{R}), and IBM uses negative
1173 type numbers. GDB accepts all three ways, as of version 4.8; dbx just
1174 accepts the traditional builtin types and perhaps one of the other two
1175 formats. The following sections describe each of these formats.
1178 * Traditional Builtin Types:: Put on your seatbelts and prepare for kludgery
1179 * Builtin Type Descriptors:: Builtin types with special type descriptors
1180 * Negative Type Numbers:: Builtin types using negative type numbers
1183 @node Traditional Builtin Types
1184 @subsection Traditional Builtin Types
1186 This is the traditional, convoluted method for defining builtin types.
1187 There are several classes of such type definitions: integer, floating
1188 point, and @code{void}.
1191 * Traditional Integer Types::
1192 * Traditional Other Types::
1195 @node Traditional Integer Types
1196 @subsubsection Traditional Integer Types
1198 Often types are defined as subranges of themselves. If the bounding values
1199 fit within an @code{int}, then they are given normally. For example:
1202 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0 # @r{128 is N_LSYM}
1203 .stabs "char:t2=r2;0;127;",128,0,0,0
1206 Builtin types can also be described as subranges of @code{int}:
1209 .stabs "unsigned short:t6=r1;0;65535;",128,0,0,0
1212 If the lower bound of a subrange is 0 and the upper bound is -1,
1213 the type is an unsigned integral type whose bounds are too
1214 big to describe in an @code{int}. Traditionally this is only used for
1215 @code{unsigned int} and @code{unsigned long}:
1218 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
1221 For larger types, GCC 2.4.5 puts out bounds in octal, with a leading 0.
1222 In this case a negative bound consists of a number which is a 1 bit
1223 followed by a bunch of 0 bits, and a positive bound is one in which a
1224 bunch of bits are 1. All known versions of dbx and GDB version 4 accept
1225 this, but GDB 3.5 refuses to read the whole file containing such
1226 symbols. So GCC 2.3.3 did not output the proper size for these types.
1227 @c FIXME: How about an example?
1229 If the lower bound of a subrange is 0 and the upper bound is negative,
1230 the type is an unsigned integral type whose size in bytes is the
1231 absolute value of the upper bound. I believe this is a Convex
1232 convention for @code{unsigned long long}.
1234 If the lower bound of a subrange is negative and the upper bound is 0,
1235 the type is a signed integral type whose size in bytes is
1236 the absolute value of the lower bound. I believe this is a Convex
1237 convention for @code{long long}. To distinguish this from a legitimate
1238 subrange, the type should be a subrange of itself. I'm not sure whether
1239 this is the case for Convex.
1241 @node Traditional Other Types
1242 @subsubsection Traditional Other Types
1244 If the upper bound of a subrange is 0 and the lower bound is positive,
1245 the type is a floating point type, and the lower bound of the subrange
1246 indicates the number of bytes in the type:
1249 .stabs "float:t12=r1;4;0;",128,0,0,0
1250 .stabs "double:t13=r1;8;0;",128,0,0,0
1253 However, GCC writes @code{long double} the same way it writes
1254 @code{double}, so there is no way to distinguish.
1257 .stabs "long double:t14=r1;8;0;",128,0,0,0
1260 Complex types are defined the same way as floating-point types; there is
1261 no way to distinguish a single-precision complex from a double-precision
1262 floating-point type.
1264 The C @code{void} type is defined as itself:
1267 .stabs "void:t15=15",128,0,0,0
1270 I'm not sure how a boolean type is represented.
1272 @node Builtin Type Descriptors
1273 @subsection Defining Builtin Types Using Builtin Type Descriptors
1275 This is the method used by Sun's @code{acc} for defining builtin types.
1276 These are the type descriptors to define builtin types:
1279 @c FIXME: clean up description of width and offset, once we figure out
1281 @item b @var{signed} @var{char-flag} @var{width} ; @var{offset} ; @var{nbits} ;
1282 Define an integral type. @var{signed} is @samp{u} for unsigned or
1283 @samp{s} for signed. @var{char-flag} is @samp{c} which indicates this
1284 is a character type, or is omitted. I assume this is to distinguish an
1285 integral type from a character type of the same size, for example it
1286 might make sense to set it for the C type @code{wchar_t} so the debugger
1287 can print such variables differently (Solaris does not do this). Sun
1288 sets it on the C types @code{signed char} and @code{unsigned char} which
1289 arguably is wrong. @var{width} and @var{offset} appear to be for small
1290 objects stored in larger ones, for example a @code{short} in an
1291 @code{int} register. @var{width} is normally the number of bytes in the
1292 type. @var{offset} seems to always be zero. @var{nbits} is the number
1293 of bits in the type.
1295 Note that type descriptor @samp{b} used for builtin types conflicts with
1296 its use for Pascal space types (@pxref{Miscellaneous Types}); they can
1297 be distinguished because the character following the type descriptor
1298 will be a digit, @samp{(}, or @samp{-} for a Pascal space type, or
1299 @samp{u} or @samp{s} for a builtin type.
1302 Documented by AIX to define a wide character type, but their compiler
1303 actually uses negative type numbers (@pxref{Negative Type Numbers}).
1305 @item R @var{fp-type} ; @var{bytes} ;
1306 Define a floating point type. @var{fp-type} has one of the following values:
1310 IEEE 32-bit (single precision) floating point format.
1313 IEEE 64-bit (double precision) floating point format.
1315 @item 3 (NF_COMPLEX)
1316 @item 4 (NF_COMPLEX16)
1317 @item 5 (NF_COMPLEX32)
1318 @c "GDB source" really means @file{include/aout/stab_gnu.h}, but trying
1319 @c to put that here got an overfull hbox.
1320 These are for complex numbers. A comment in the GDB source describes
1321 them as Fortran @code{complex}, @code{double complex}, and
1322 @code{complex*16}, respectively, but what does that mean? (i.e., Single
1323 precision? Double precison?).
1325 @item 6 (NF_LDOUBLE)
1326 Long double. This should probably only be used for Sun format
1327 @code{long double}, and new codes should be used for other floating
1328 point formats (@code{NF_DOUBLE} can be used if a @code{long double} is
1329 really just an IEEE double, of course).
1332 @var{bytes} is the number of bytes occupied by the type. This allows a
1333 debugger to perform some operations with the type even if it doesn't
1334 understand @var{fp-type}.
1336 @item g @var{type-information} ; @var{nbits}
1337 Documented by AIX to define a floating type, but their compiler actually
1338 uses negative type numbers (@pxref{Negative Type Numbers}).
1340 @item c @var{type-information} ; @var{nbits}
1341 Documented by AIX to define a complex type, but their compiler actually
1342 uses negative type numbers (@pxref{Negative Type Numbers}).
1345 The C @code{void} type is defined as a signed integral type 0 bits long:
1347 .stabs "void:t19=bs0;0;0",128,0,0,0
1349 The Solaris compiler seems to omit the trailing semicolon in this case.
1350 Getting sloppy in this way is not a swift move because if a type is
1351 embedded in a more complex expression it is necessary to be able to tell
1354 I'm not sure how a boolean type is represented.
1356 @node Negative Type Numbers
1357 @subsection Negative Type Numbers
1359 This is the method used in XCOFF for defining builtin types.
1360 Since the debugger knows about the builtin types anyway, the idea of
1361 negative type numbers is simply to give a special type number which
1362 indicates the builtin type. There is no stab defining these types.
1364 There are several subtle issues with negative type numbers.
1366 One is the size of the type. A builtin type (for example the C types
1367 @code{int} or @code{long}) might have different sizes depending on
1368 compiler options, the target architecture, the ABI, etc. This issue
1369 doesn't come up for IBM tools since (so far) they just target the
1370 RS/6000; the sizes indicated below for each size are what the IBM
1371 RS/6000 tools use. To deal with differing sizes, either define separate
1372 negative type numbers for each size (which works but requires changing
1373 the debugger, and, unless you get both AIX dbx and GDB to accept the
1374 change, introduces an incompatibility), or use a type attribute
1375 (@pxref{String Field}) to define a new type with the appropriate size
1376 (which merely requires a debugger which understands type attributes,
1377 like AIX dbx). For example,
1380 .stabs "boolean:t10=@@s8;-16",128,0,0,0
1383 defines an 8-bit boolean type, and
1386 .stabs "boolean:t10=@@s64;-16",128,0,0,0
1389 defines a 64-bit boolean type.
1391 A similar issue is the format of the type. This comes up most often for
1392 floating-point types, which could have various formats (particularly
1393 extended doubles, which vary quite a bit even among IEEE systems).
1394 Again, it is best to define a new negative type number for each
1395 different format; changing the format based on the target system has
1396 various problems. One such problem is that the Alpha has both VAX and
1397 IEEE floating types. One can easily imagine one library using the VAX
1398 types and another library in the same executable using the IEEE types.
1399 Another example is that the interpretation of whether a boolean is true
1400 or false can be based on the least significant bit, most significant
1401 bit, whether it is zero, etc., and different compilers (or different
1402 options to the same compiler) might provide different kinds of boolean.
1404 The last major issue is the names of the types. The name of a given
1405 type depends @emph{only} on the negative type number given; these do not
1406 vary depending on the language, the target system, or anything else.
1407 One can always define separate type numbers---in the following list you
1408 will see for example separate @code{int} and @code{integer*4} types
1409 which are identical except for the name. But compatibility can be
1410 maintained by not inventing new negative type numbers and instead just
1411 defining a new type with a new name. For example:
1414 .stabs "CARDINAL:t10=-8",128,0,0,0
1417 Here is the list of negative type numbers. The phrase @dfn{integral
1418 type} is used to mean twos-complement (I strongly suspect that all
1419 machines which use stabs use twos-complement; most machines use
1420 twos-complement these days).
1424 @code{int}, 32 bit signed integral type.
1427 @code{char}, 8 bit type holding a character. Both GDB and dbx on AIX
1428 treat this as signed. GCC uses this type whether @code{char} is signed
1429 or not, which seems like a bad idea. The AIX compiler (@code{xlc}) seems to
1430 avoid this type; it uses -5 instead for @code{char}.
1433 @code{short}, 16 bit signed integral type.
1436 @code{long}, 32 bit signed integral type.
1439 @code{unsigned char}, 8 bit unsigned integral type.
1442 @code{signed char}, 8 bit signed integral type.
1445 @code{unsigned short}, 16 bit unsigned integral type.
1448 @code{unsigned int}, 32 bit unsigned integral type.
1451 @code{unsigned}, 32 bit unsigned integral type.
1454 @code{unsigned long}, 32 bit unsigned integral type.
1457 @code{void}, type indicating the lack of a value.
1460 @code{float}, IEEE single precision.
1463 @code{double}, IEEE double precision.
1466 @code{long double}, IEEE double precision. The compiler claims the size
1467 will increase in a future release, and for binary compatibility you have
1468 to avoid using @code{long double}. I hope when they increase it they
1469 use a new negative type number.
1472 @code{integer}. 32 bit signed integral type.
1475 @code{boolean}. 32 bit type. How is the truth value encoded? Is it
1476 the least significant bit or is it a question of whether the whole value
1477 is zero or non-zero?
1480 @code{short real}. IEEE single precision.
1483 @code{real}. IEEE double precision.
1486 @code{stringptr}. @xref{Strings}.
1489 @code{character}, 8 bit unsigned character type.
1492 @code{logical*1}, 8 bit type. This Fortran type has a split
1493 personality in that it is used for boolean variables, but can also be
1494 used for unsigned integers. 0 is false, 1 is true, and other values are
1498 @code{logical*2}, 16 bit type. This Fortran type has a split
1499 personality in that it is used for boolean variables, but can also be
1500 used for unsigned integers. 0 is false, 1 is true, and other values are
1504 @code{logical*4}, 32 bit type. This Fortran type has a split
1505 personality in that it is used for boolean variables, but can also be
1506 used for unsigned integers. 0 is false, 1 is true, and other values are
1510 @code{logical}, 32 bit type. This Fortran type has a split
1511 personality in that it is used for boolean variables, but can also be
1512 used for unsigned integers. 0 is false, 1 is true, and other values are
1516 @code{complex}. A complex type consisting of two IEEE single-precision
1517 floating point values.
1520 @code{complex}. A complex type consisting of two IEEE double-precision
1521 floating point values.
1524 @code{integer*1}, 8 bit signed integral type.
1527 @code{integer*2}, 16 bit signed integral type.
1530 @code{integer*4}, 32 bit signed integral type.
1533 @code{wchar}. Wide character, 16 bits wide, unsigned (what format?
1537 @node Miscellaneous Types
1538 @section Miscellaneous Types
1541 @item b @var{type-information} ; @var{bytes}
1542 Pascal space type. This is documented by IBM; what does it mean?
1544 This use of the @samp{b} type descriptor can be distinguished
1545 from its use for builtin integral types (@pxref{Builtin Type
1546 Descriptors}) because the character following the type descriptor is
1547 always a digit, @samp{(}, or @samp{-}.
1549 @item B @var{type-information}
1550 A volatile-qualified version of @var{type-information}. This is
1551 a Sun extension. References and stores to a variable with a
1552 volatile-qualified type must not be optimized or cached; they
1553 must occur as the user specifies them.
1555 @item d @var{type-information}
1556 File of type @var{type-information}. As far as I know this is only used
1559 @item k @var{type-information}
1560 A const-qualified version of @var{type-information}. This is a Sun
1561 extension. A variable with a const-qualified type cannot be modified.
1563 @item M @var{type-information} ; @var{length}
1564 Multiple instance type. The type seems to composed of @var{length}
1565 repetitions of @var{type-information}, for example @code{character*3} is
1566 represented by @samp{M-2;3}, where @samp{-2} is a reference to a
1567 character type (@pxref{Negative Type Numbers}). I'm not sure how this
1568 differs from an array. This appears to be a Fortran feature.
1569 @var{length} is a bound, like those in range types; see @ref{Subranges}.
1571 @item S @var{type-information}
1572 Pascal set type. @var{type-information} must be a small type such as an
1573 enumeration or a subrange, and the type is a bitmask whose length is
1574 specified by the number of elements in @var{type-information}.
1576 @item * @var{type-information}
1577 Pointer to @var{type-information}.
1580 @node Cross-References
1581 @section Cross-References to Other Types
1583 A type can be used before it is defined; one common way to deal with
1584 that situation is just to use a type reference to a type which has not
1587 Another way is with the @samp{x} type descriptor, which is followed by
1588 @samp{s} for a structure tag, @samp{u} for a union tag, or @samp{e} for
1589 a enumerator tag, followed by the name of the tag, followed by @samp{:}.
1590 For example, the following C declarations:
1601 .stabs "bar:G16=*17=xsfoo:",32,0,0,0
1604 Not all debuggers support the @samp{x} type descriptor, so on some
1605 machines GCC does not use it. I believe that for the above example it
1606 would just emit a reference to type 17 and never define it, but I
1607 haven't verified that.
1609 Modula-2 imported types, at least on AIX, use the @samp{i} type
1610 descriptor, which is followed by the name of the module from which the
1611 type is imported, followed by @samp{:}, followed by the name of the
1612 type. There is then optionally a comma followed by type information for
1613 the type. This differs from merely naming the type (@pxref{Typedefs}) in
1614 that it identifies the module; I don't understand whether the name of
1615 the type given here is always just the same as the name we are giving
1616 it, or whether this type descriptor is used with a nameless stab
1617 (@pxref{String Field}), or what. The symbol ends with @samp{;}.
1620 @section Subrange Types
1622 The @samp{r} type descriptor defines a type as a subrange of another
1623 type. It is followed by type information for the type of which it is a
1624 subrange, a semicolon, an integral lower bound, a semicolon, an
1625 integral upper bound, and a semicolon. The AIX documentation does not
1626 specify the trailing semicolon, in an effort to specify array indexes
1627 more cleanly, but a subrange which is not an array index has always
1628 included a trailing semicolon (@pxref{Arrays}).
1630 Instead of an integer, either bound can be one of the following:
1633 @item A @var{offset}
1634 The bound is passed by reference on the stack at offset @var{offset}
1635 from the argument list. @xref{Parameters}, for more information on such
1638 @item T @var{offset}
1639 The bound is passed by value on the stack at offset @var{offset} from
1642 @item a @var{register-number}
1643 The bound is pased by reference in register number
1644 @var{register-number}.
1646 @item t @var{register-number}
1647 The bound is passed by value in register number @var{register-number}.
1653 Subranges are also used for builtin types; see @ref{Traditional Builtin Types}.
1656 @section Array Types
1658 Arrays use the @samp{a} type descriptor. Following the type descriptor
1659 is the type of the index and the type of the array elements. If the
1660 index type is a range type, it ends in a semicolon; otherwise
1661 (for example, if it is a type reference), there does not
1662 appear to be any way to tell where the types are separated. In an
1663 effort to clean up this mess, IBM documents the two types as being
1664 separated by a semicolon, and a range type as not ending in a semicolon
1665 (but this is not right for range types which are not array indexes,
1666 @pxref{Subranges}). I think probably the best solution is to specify
1667 that a semicolon ends a range type, and that the index type and element
1668 type of an array are separated by a semicolon, but that if the index
1669 type is a range type, the extra semicolon can be omitted. GDB (at least
1670 through version 4.9) doesn't support any kind of index type other than a
1671 range anyway; I'm not sure about dbx.
1673 It is well established, and widely used, that the type of the index,
1674 unlike most types found in the stabs, is merely a type definition, not
1675 type information (@pxref{String Field}) (that is, it need not start with
1676 @samp{@var{type-number}=} if it is defining a new type). According to a
1677 comment in GDB, this is also true of the type of the array elements; it
1678 gives @samp{ar1;1;10;ar1;1;10;4} as a legitimate way to express a two
1679 dimensional array. According to AIX documentation, the element type
1680 must be type information. GDB accepts either.
1682 The type of the index is often a range type, expressed as the type
1683 descriptor @samp{r} and some parameters. It defines the size of the
1684 array. In the example below, the range @samp{r1;0;2;} defines an index
1685 type which is a subrange of type 1 (integer), with a lower bound of 0
1686 and an upper bound of 2. This defines the valid range of subscripts of
1687 a three-element C array.
1689 For example, the definition:
1692 char char_vec[3] = @{'a','b','c'@};
1696 produces the output:
1699 .stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
1708 If an array is @dfn{packed}, the elements are spaced more
1709 closely than normal, saving memory at the expense of speed. For
1710 example, an array of 3-byte objects might, if unpacked, have each
1711 element aligned on a 4-byte boundary, but if packed, have no padding.
1712 One way to specify that something is packed is with type attributes
1713 (@pxref{String Field}). In the case of arrays, another is to use the
1714 @samp{P} type descriptor instead of @samp{a}. Other than specifying a
1715 packed array, @samp{P} is identical to @samp{a}.
1717 @c FIXME-what is it? A pointer?
1718 An open array is represented by the @samp{A} type descriptor followed by
1719 type information specifying the type of the array elements.
1721 @c FIXME: what is the format of this type? A pointer to a vector of pointers?
1722 An N-dimensional dynamic array is represented by
1725 D @var{dimensions} ; @var{type-information}
1728 @c Does dimensions really have this meaning? The AIX documentation
1730 @var{dimensions} is the number of dimensions; @var{type-information}
1731 specifies the type of the array elements.
1733 @c FIXME: what is the format of this type? A pointer to some offsets in
1735 A subarray of an N-dimensional array is represented by
1738 E @var{dimensions} ; @var{type-information}
1741 @c Does dimensions really have this meaning? The AIX documentation
1743 @var{dimensions} is the number of dimensions; @var{type-information}
1744 specifies the type of the array elements.
1749 Some languages, like C or the original Pascal, do not have string types,
1750 they just have related things like arrays of characters. But most
1751 Pascals and various other languages have string types, which are
1752 indicated as follows:
1755 @item n @var{type-information} ; @var{bytes}
1756 @var{bytes} is the maximum length. I'm not sure what
1757 @var{type-information} is; I suspect that it means that this is a string
1758 of @var{type-information} (thus allowing a string of integers, a string
1759 of wide characters, etc., as well as a string of characters). Not sure
1760 what the format of this type is. This is an AIX feature.
1762 @item z @var{type-information} ; @var{bytes}
1763 Just like @samp{n} except that this is a gstring, not an ordinary
1764 string. I don't know the difference.
1767 Pascal Stringptr. What is this? This is an AIX feature.
1771 @section Enumerations
1773 Enumerations are defined with the @samp{e} type descriptor.
1775 @c FIXME: Where does this information properly go? Perhaps it is
1776 @c redundant with something we already explain.
1777 The source line below declares an enumeration type at file scope.
1778 The type definition is located after the @code{N_RBRAC} that marks the end of
1779 the previous procedure's block scope, and before the @code{N_FUN} that marks
1780 the beginning of the next procedure's block scope. Therefore it does not
1781 describe a block local symbol, but a file local one.
1786 enum e_places @{first,second=3,last@};
1790 generates the following stab:
1793 .stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0
1796 The symbol descriptor (@samp{T}) says that the stab describes a
1797 structure, enumeration, or union tag. The type descriptor @samp{e},
1798 following the @samp{22=} of the type definition narrows it down to an
1799 enumeration type. Following the @samp{e} is a list of the elements of
1800 the enumeration. The format is @samp{@var{name}:@var{value},}. The
1801 list of elements ends with @samp{;}.
1803 There is no standard way to specify the size of an enumeration type; it
1804 is determined by the architecture (normally all enumerations types are
1805 32 bits). There should be a way to specify an enumeration type of
1806 another size; type attributes would be one way to do this. @xref{Stabs
1812 The encoding of structures in stabs can be shown with an example.
1814 The following source code declares a structure tag and defines an
1815 instance of the structure in global scope. Then a @code{typedef} equates the
1816 structure tag with a new type. Seperate stabs are generated for the
1817 structure tag, the structure @code{typedef}, and the structure instance. The
1818 stabs for the tag and the @code{typedef} are emited when the definitions are
1819 encountered. Since the structure elements are not initialized, the
1820 stab and code for the structure variable itself is located at the end
1821 of the program in the bss section.
1828 struct s_tag* s_next;
1831 typedef struct s_tag s_typedef;
1834 The structure tag has an @code{N_LSYM} stab type because, like the
1835 enumeration, the symbol has file scope. Like the enumeration, the
1836 symbol descriptor is @samp{T}, for enumeration, structure, or tag type.
1837 The type descriptor @samp{s} following the @samp{16=} of the type
1838 definition narrows the symbol type to structure.
1840 Following the @samp{s} type descriptor is the number of bytes the
1841 structure occupies, followed by a description of each structure element.
1842 The structure element descriptions are of the form @var{name:type, bit
1843 offset from the start of the struct, number of bits in the element}.
1845 @c FIXME: phony line break. Can probably be fixed by using an example
1846 @c with fewer fields.
1849 .stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;
1850 s_char_vec:17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0
1853 In this example, the first two structure elements are previously defined
1854 types. For these, the type following the @samp{@var{name}:} part of the
1855 element description is a simple type reference. The other two structure
1856 elements are new types. In this case there is a type definition
1857 embedded after the @samp{@var{name}:}. The type definition for the
1858 array element looks just like a type definition for a standalone array.
1859 The @code{s_next} field is a pointer to the same kind of structure that
1860 the field is an element of. So the definition of structure type 16
1861 contains a type definition for an element which is a pointer to type 16.
1864 @section Giving a Type a Name
1866 To give a type a name, use the @samp{t} symbol descriptor. The type
1867 is specified by the type information (@pxref{String Field}) for the stab.
1871 .stabs "s_typedef:t16",128,0,0,0 # @r{128 is N_LSYM}
1874 specifies that @code{s_typedef} refers to type number 16. Such stabs
1875 have symbol type @code{N_LSYM} (or @code{C_DECL} for XCOFF).
1877 If you are specifying the tag name for a structure, union, or
1878 enumeration, use the @samp{T} symbol descriptor instead. I believe C is
1879 the only language with this feature.
1881 If the type is an opaque type (I believe this is a Modula-2 feature),
1882 AIX provides a type descriptor to specify it. The type descriptor is
1883 @samp{o} and is followed by a name. I don't know what the name
1884 means---is it always the same as the name of the type, or is this type
1885 descriptor used with a nameless stab (@pxref{String Field})? There
1886 optionally follows a comma followed by type information which defines
1887 the type of this type. If omitted, a semicolon is used in place of the
1888 comma and the type information, and the type is much like a generic
1889 pointer type---it has a known size but little else about it is
1903 This code generates a stab for a union tag and a stab for a union
1904 variable. Both use the @code{N_LSYM} stab type. If a union variable is
1905 scoped locally to the procedure in which it is defined, its stab is
1906 located immediately preceding the @code{N_LBRAC} for the procedure's block
1909 The stab for the union tag, however, is located preceding the code for
1910 the procedure in which it is defined. The stab type is @code{N_LSYM}. This
1911 would seem to imply that the union type is file scope, like the struct
1912 type @code{s_tag}. This is not true. The contents and position of the stab
1913 for @code{u_type} do not convey any infomation about its procedure local
1916 @c FIXME: phony line break. Can probably be fixed by using an example
1917 @c with fewer fields.
1920 .stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
1924 The symbol descriptor @samp{T}, following the @samp{name:} means that
1925 the stab describes an enumeration, structure, or union tag. The type
1926 descriptor @samp{u}, following the @samp{23=} of the type definition,
1927 narrows it down to a union type definition. Following the @samp{u} is
1928 the number of bytes in the union. After that is a list of union element
1929 descriptions. Their format is @var{name:type, bit offset into the
1930 union, number of bytes for the element;}.
1932 The stab for the union variable is:
1935 .stabs "an_u:23",128,0,0,-20 # @r{128 is N_LSYM}
1938 @samp{-20} specifies where the variable is stored (@pxref{Stack
1941 @node Function Types
1942 @section Function Types
1944 Various types can be defined for function variables. These types are
1945 not used in defining functions (@pxref{Procedures}); they are used for
1946 things like pointers to functions.
1948 The simple, traditional, type is type descriptor @samp{f} is followed by
1949 type information for the return type of the function, followed by a
1952 This does not deal with functions for which the number and types of the
1953 parameters are part of the type, as in Modula-2 or ANSI C. AIX provides
1954 extensions to specify these, using the @samp{f}, @samp{F}, @samp{p}, and
1955 @samp{R} type descriptors.
1957 First comes the type descriptor. If it is @samp{f} or @samp{F}, this
1958 type involves a function rather than a procedure, and the type
1959 information for the return type of the function follows, followed by a
1960 comma. Then comes the number of parameters to the function and a
1961 semicolon. Then, for each parameter, there is the name of the parameter
1962 followed by a colon (this is only present for type descriptors @samp{R}
1963 and @samp{F} which represent Pascal function or procedure parameters),
1964 type information for the parameter, a comma, 0 if passed by reference or
1965 1 if passed by value, and a semicolon. The type definition ends with a
1968 For example, this variable definition:
1975 generates the following code:
1978 .stabs "g_pf:G24=*25=f1",32,0,0,0
1979 .common _g_pf,4,"bss"
1982 The variable defines a new type, 24, which is a pointer to another new
1983 type, 25, which is a function returning @code{int}.
1986 @chapter Symbol Information in Symbol Tables
1988 This chapter describes the format of symbol table entries
1989 and how stab assembler directives map to them. It also describes the
1990 transformations that the assembler and linker make on data from stabs.
1993 * Symbol Table Format::
1994 * Transformations On Symbol Tables::
1997 @node Symbol Table Format
1998 @section Symbol Table Format
2000 Each time the assembler encounters a stab directive, it puts
2001 each field of the stab into a corresponding field in a symbol table
2002 entry of its output file. If the stab contains a string field, the
2003 symbol table entry for that stab points to a string table entry
2004 containing the string data from the stab. Assembler labels become
2005 relocatable addresses. Symbol table entries in a.out have the format:
2007 @c FIXME: should refer to external, not internal.
2009 struct internal_nlist @{
2010 unsigned long n_strx; /* index into string table of name */
2011 unsigned char n_type; /* type of symbol */
2012 unsigned char n_other; /* misc info (usually empty) */
2013 unsigned short n_desc; /* description field */
2014 bfd_vma n_value; /* value of symbol */
2018 If the stab has a string, the @code{n_strx} field holds the offset in
2019 bytes of the string within the string table. The string is terminated
2020 by a NUL character. If the stab lacks a string (for example, it was
2021 produced by a @code{.stabn} or @code{.stabd} directive), the
2022 @code{n_strx} field is zero.
2024 Symbol table entries with @code{n_type} field values greater than 0x1f
2025 originated as stabs generated by the compiler (with one random
2026 exception). The other entries were placed in the symbol table of the
2027 executable by the assembler or the linker.
2029 @node Transformations On Symbol Tables
2030 @section Transformations on Symbol Tables
2032 The linker concatenates object files and does fixups of externally
2035 You can see the transformations made on stab data by the assembler and
2036 linker by examining the symbol table after each pass of the build. To
2037 do this, use @samp{nm -ap}, which dumps the symbol table, including
2038 debugging information, unsorted. For stab entries the columns are:
2039 @var{value}, @var{other}, @var{desc}, @var{type}, @var{string}. For
2040 assembler and linker symbols, the columns are: @var{value}, @var{type},
2043 The low 5 bits of the stab type tell the linker how to relocate the
2044 value of the stab. Thus for stab types like @code{N_RSYM} and
2045 @code{N_LSYM}, where the value is an offset or a register number, the
2046 low 5 bits are @code{N_ABS}, which tells the linker not to relocate the
2049 Where the value of a stab contains an assembly language label,
2050 it is transformed by each build step. The assembler turns it into a
2051 relocatable address and the linker turns it into an absolute address.
2054 * Transformations On Static Variables::
2055 * Transformations On Global Variables::
2056 * ELF Transformations:: In ELF, things are a bit different.
2059 @node Transformations On Static Variables
2060 @subsection Transformations on Static Variables
2062 This source line defines a static variable at file scope:
2065 static int s_g_repeat
2069 The following stab describes the symbol:
2072 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
2076 The assembler transforms the stab into this symbol table entry in the
2077 @file{.o} file. The location is expressed as a data segment offset.
2080 00000084 - 00 0000 STSYM s_g_repeat:S1
2084 In the symbol table entry from the executable, the linker has made the
2085 relocatable address absolute.
2088 0000e00c - 00 0000 STSYM s_g_repeat:S1
2091 @node Transformations On Global Variables
2092 @subsection Transformations on Global Variables
2094 Stabs for global variables do not contain location information. In
2095 this case, the debugger finds location information in the assembler or
2096 linker symbol table entry describing the variable. The source line:
2106 .stabs "g_foo:G2",32,0,0,0
2109 The variable is represented by two symbol table entries in the object
2110 file (see below). The first one originated as a stab. The second one
2111 is an external symbol. The upper case @samp{D} signifies that the
2112 @code{n_type} field of the symbol table contains 7, @code{N_DATA} with
2113 local linkage. The stab's value is zero since the value is not used for
2114 @code{N_GSYM} stabs. The value of the linker symbol is the relocatable
2115 address corresponding to the variable.
2118 00000000 - 00 0000 GSYM g_foo:G2
2123 These entries as transformed by the linker. The linker symbol table
2124 entry now holds an absolute address:
2127 00000000 - 00 0000 GSYM g_foo:G2
2132 @node ELF Transformations
2133 @subsection Transformations of Stabs in ELF Files
2135 For ELF files, use @code{objdump --stabs} instead of @code{nm} to show
2136 the stabs in an object or executable file. @code{objdump} is a GNU
2137 utility; Sun does not provide any equivalent.
2139 The following example is for a stab whose value is an address is
2140 relative to the compilation unit (@pxref{Stabs In ELF}). For example,
2147 appears within a function, then the assembly language output from the
2153 .stabs "ld:V(0,3)",0x26,0,4,.L18-Ddata.data # @r{0x26 is N_STSYM}
2160 Because the value is formed by subtracting one symbol from another, the
2161 value is absolute, not relocatable, and so the object file contains
2164 Symnum n_type n_othr n_desc n_value n_strx String
2165 31 STSYM 0 4 00000004 680 ld:V(0,3)
2168 without any relocations, and the executable file also contains
2171 Symnum n_type n_othr n_desc n_value n_strx String
2172 31 STSYM 0 4 00000004 680 ld:V(0,3)
2176 @chapter GNU C++ Stabs
2179 * Basic Cplusplus Types::
2182 * Methods:: Method definition
2184 * Method Modifiers::
2187 * Virtual Base Classes::
2191 Type descriptors added for C++ descriptions:
2195 method type (@code{##} if minimal debug)
2198 Member (class and variable) type. It is followed by type information
2199 for the offset basetype, a comma, and type information for the type of
2200 the field being pointed to. (FIXME: this is acknowledged to be
2201 gibberish. Can anyone say what really goes here?).
2203 Note that there is a conflict between this and type attributes
2204 (@pxref{String Field}); both use type descriptor @samp{@@}.
2205 Fortunately, the @samp{@@} type descriptor used in this C++ sense always
2206 will be followed by a digit, @samp{(}, or @samp{-}, and type attributes
2207 never start with those things.
2210 @node Basic Cplusplus Types
2211 @section Basic Types For C++
2213 << the examples that follow are based on a01.C >>
2216 C++ adds two more builtin types to the set defined for C. These are
2217 the unknown type and the vtable record type. The unknown type, type
2218 16, is defined in terms of itself like the void type.
2220 The vtable record type, type 17, is defined as a structure type and
2221 then as a structure tag. The structure has four fields: delta, index,
2222 pfn, and delta2. pfn is the function pointer.
2224 << In boilerplate $vtbl_ptr_type, what are the fields delta,
2225 index, and delta2 used for? >>
2227 This basic type is present in all C++ programs even if there are no
2228 virtual methods defined.
2231 .stabs "struct_name:sym_desc(type)type_def(17)=type_desc(struct)struct_bytes(8)
2232 elem_name(delta):type_ref(short int),bit_offset(0),field_bits(16);
2233 elem_name(index):type_ref(short int),bit_offset(16),field_bits(16);
2234 elem_name(pfn):type_def(18)=type_desc(ptr to)type_ref(void),
2235 bit_offset(32),field_bits(32);
2236 elem_name(delta2):type_def(short int);bit_offset(32),field_bits(16);;"
2241 .stabs "$vtbl_ptr_type:t17=s8
2242 delta:6,0,16;index:6,16,16;pfn:18=*15,32,32;delta2:6,32,16;;"
2247 .stabs "name:sym_dec(struct tag)type_ref($vtbl_ptr_type)",N_LSYM,NIL,NIL,NIL
2251 .stabs "$vtbl_ptr_type:T17",128,0,0,0
2254 @node Simple Classes
2255 @section Simple Class Definition
2257 The stabs describing C++ language features are an extension of the
2258 stabs describing C. Stabs representing C++ class types elaborate
2259 extensively on the stab format used to describe structure types in C.
2260 Stabs representing class type variables look just like stabs
2261 representing C language variables.
2263 Consider the following very simple class definition.
2269 int Ameth(int in, char other);
2273 The class @code{baseA} is represented by two stabs. The first stab describes
2274 the class as a structure type. The second stab describes a structure
2275 tag of the class type. Both stabs are of stab type @code{N_LSYM}. Since the
2276 stab is not located between an @code{N_FUN} and an @code{N_LBRAC} stab this indicates
2277 that the class is defined at file scope. If it were, then the @code{N_LSYM}
2278 would signify a local variable.
2280 A stab describing a C++ class type is similar in format to a stab
2281 describing a C struct, with each class member shown as a field in the
2282 structure. The part of the struct format describing fields is
2283 expanded to include extra information relevent to C++ class members.
2284 In addition, if the class has multiple base classes or virtual
2285 functions the struct format outside of the field parts is also
2288 In this simple example the field part of the C++ class stab
2289 representing member data looks just like the field part of a C struct
2290 stab. The section on protections describes how its format is
2291 sometimes extended for member data.
2293 The field part of a C++ class stab representing a member function
2294 differs substantially from the field part of a C struct stab. It
2295 still begins with @samp{name:} but then goes on to define a new type number
2296 for the member function, describe its return type, its argument types,
2297 its protection level, any qualifiers applied to the method definition,
2298 and whether the method is virtual or not. If the method is virtual
2299 then the method description goes on to give the vtable index of the
2300 method, and the type number of the first base class defining the
2303 When the field name is a method name it is followed by two colons rather
2304 than one. This is followed by a new type definition for the method.
2305 This is a number followed by an equal sign and the type descriptor
2306 @samp{#}, indicating a method type, and a second @samp{#}, indicating
2307 that this is the @dfn{minimal} type of method definition used by GCC2,
2308 not larger method definitions used by earlier versions of GCC. This is
2309 followed by a type reference showing the return type of the method and a
2312 The format of an overloaded operator method name differs from that of
2313 other methods. It is @samp{op$::@var{operator-name}.} where
2314 @var{operator-name} is the operator name such as @samp{+} or @samp{+=}.
2315 The name ends with a period, and any characters except the period can
2316 occur in the @var{operator-name} string.
2318 The next part of the method description represents the arguments to the
2319 method, preceeded by a colon and ending with a semi-colon. The types of
2320 the arguments are expressed in the same way argument types are expressed
2321 in C++ name mangling. In this example an @code{int} and a @code{char}
2324 This is followed by a number, a letter, and an asterisk or period,
2325 followed by another semicolon. The number indicates the protections
2326 that apply to the member function. Here the 2 means public. The
2327 letter encodes any qualifier applied to the method definition. In
2328 this case, @samp{A} means that it is a normal function definition. The dot
2329 shows that the method is not virtual. The sections that follow
2330 elaborate further on these fields and describe the additional
2331 information present for virtual methods.
2335 .stabs "class_name:sym_desc(type)type_def(20)=type_desc(struct)struct_bytes(4)
2336 field_name(Adat):type(int),bit_offset(0),field_bits(32);
2338 method_name(Ameth)::type_def(21)=type_desc(method)return_type(int);
2339 :arg_types(int char);
2340 protection(public)qualifier(normal)virtual(no);;"
2345 .stabs "baseA:t20=s4Adat:1,0,32;Ameth::21=##1;:ic;2A.;;",128,0,0,0
2347 .stabs "class_name:sym_desc(struct tag)",N_LSYM,NIL,NIL,NIL
2349 .stabs "baseA:T20",128,0,0,0
2352 @node Class Instance
2353 @section Class Instance
2355 As shown above, describing even a simple C++ class definition is
2356 accomplished by massively extending the stab format used in C to
2357 describe structure types. However, once the class is defined, C stabs
2358 with no modifications can be used to describe class instances. The
2368 yields the following stab describing the class instance. It looks no
2369 different from a standard C stab describing a local variable.
2372 .stabs "name:type_ref(baseA)", N_LSYM, NIL, NIL, frame_ptr_offset
2376 .stabs "AbaseA:20",128,0,0,-20
2380 @section Method Definition
2382 The class definition shown above declares Ameth. The C++ source below
2387 baseA::Ameth(int in, char other)
2394 This method definition yields three stabs following the code of the
2395 method. One stab describes the method itself and following two describe
2396 its parameters. Although there is only one formal argument all methods
2397 have an implicit argument which is the @code{this} pointer. The @code{this}
2398 pointer is a pointer to the object on which the method was called. Note
2399 that the method name is mangled to encode the class name and argument
2400 types. Name mangling is described in the @sc{arm} (@cite{The Annotated
2401 C++ Reference Manual}, by Ellis and Stroustrup, @sc{isbn}
2402 0-201-51459-1); @file{gpcompare.texi} in Cygnus GCC distributions
2403 describes the differences between GNU mangling and @sc{arm}
2405 @c FIXME: Use @xref, especially if this is generally installed in the
2407 @c FIXME: This information should be in a net release, either of GCC or
2408 @c GDB. But gpcompare.texi doesn't seem to be in the FSF GCC.
2411 .stabs "name:symbol_desriptor(global function)return_type(int)",
2412 N_FUN, NIL, NIL, code_addr_of_method_start
2414 .stabs "Ameth__5baseAic:F1",36,0,0,_Ameth__5baseAic
2417 Here is the stab for the @code{this} pointer implicit argument. The
2418 name of the @code{this} pointer is always @code{this}. Type 19, the
2419 @code{this} pointer is defined as a pointer to type 20, @code{baseA},
2420 but a stab defining @code{baseA} has not yet been emited. Since the
2421 compiler knows it will be emited shortly, here it just outputs a cross
2422 reference to the undefined symbol, by prefixing the symbol name with
2426 .stabs "name:sym_desc(register param)type_def(19)=
2427 type_desc(ptr to)type_ref(baseA)=
2428 type_desc(cross-reference to)baseA:",N_RSYM,NIL,NIL,register_number
2430 .stabs "this:P19=*20=xsbaseA:",64,0,0,8
2433 The stab for the explicit integer argument looks just like a parameter
2434 to a C function. The last field of the stab is the offset from the
2435 argument pointer, which in most systems is the same as the frame
2439 .stabs "name:sym_desc(value parameter)type_ref(int)",
2440 N_PSYM,NIL,NIL,offset_from_arg_ptr
2442 .stabs "in:p1",160,0,0,72
2445 << The examples that follow are based on A1.C >>
2448 @section Protections
2451 In the simple class definition shown above all member data and
2452 functions were publicly accessable. The example that follows
2453 contrasts public, protected and privately accessable fields and shows
2454 how these protections are encoded in C++ stabs.
2456 @c FIXME: What does "part of the string" mean?
2457 Protections for class member data are signified by two characters
2458 embedded in the stab defining the class type. These characters are
2459 located after the name: part of the string. @samp{/0} means private,
2460 @samp{/1} means protected, and @samp{/2} means public. If these
2461 characters are omited this means that the member is public. The
2462 following C++ source:
2476 generates the following stab to describe the class type all_data.
2479 .stabs "class_name:sym_desc(type)type_def(19)=type_desc(struct)struct_bytes
2480 data_name:/protection(private)type_ref(int),bit_offset,num_bits;
2481 data_name:/protection(protected)type_ref(char),bit_offset,num_bits;
2482 data_name:(/num omited, private)type_ref(float),bit_offset,num_bits;;"
2487 .stabs "all_data:t19=s12
2488 priv_dat:/01,0,32;prot_dat:/12,32,8;pub_dat:12,64,32;;",128,0,0,0
2491 Protections for member functions are signified by one digit embeded in
2492 the field part of the stab describing the method. The digit is 0 if
2493 private, 1 if protected and 2 if public. Consider the C++ class
2497 class all_methods @{
2499 int priv_meth(int in)@{return in;@};
2501 char protMeth(char in)@{return in;@};
2503 float pubMeth(float in)@{return in;@};
2507 It generates the following stab. The digit in question is to the left
2508 of an @samp{A} in each case. Notice also that in this case two symbol
2509 descriptors apply to the class name struct tag and struct type.
2512 .stabs "class_name:sym_desc(struct tag&type)type_def(21)=
2513 sym_desc(struct)struct_bytes(1)
2514 meth_name::type_def(22)=sym_desc(method)returning(int);
2515 :args(int);protection(private)modifier(normal)virtual(no);
2516 meth_name::type_def(23)=sym_desc(method)returning(char);
2517 :args(char);protection(protected)modifier(normal)virual(no);
2518 meth_name::type_def(24)=sym_desc(method)returning(float);
2519 :args(float);protection(public)modifier(normal)virtual(no);;",
2524 .stabs "all_methods:Tt21=s1priv_meth::22=##1;:i;0A.;protMeth::23=##2;:c;1A.;
2525 pubMeth::24=##12;:f;2A.;;",128,0,0,0
2528 @node Method Modifiers
2529 @section Method Modifiers (@code{const}, @code{volatile}, @code{const volatile})
2533 In the class example described above all the methods have the normal
2534 modifier. This method modifier information is located just after the
2535 protection information for the method. This field has four possible
2536 character values. Normal methods use @samp{A}, const methods use
2537 @samp{B}, volatile methods use @samp{C}, and const volatile methods use
2538 @samp{D}. Consider the class definition below:
2543 int ConstMeth (int arg) const @{ return arg; @};
2544 char VolatileMeth (char arg) volatile @{ return arg; @};
2545 float ConstVolMeth (float arg) const volatile @{return arg; @};
2549 This class is described by the following stab:
2552 .stabs "class(A):sym_desc(struct)type_def(20)=type_desc(struct)struct_bytes(1)
2553 meth_name(ConstMeth)::type_def(21)sym_desc(method)
2554 returning(int);:arg(int);protection(public)modifier(const)virtual(no);
2555 meth_name(VolatileMeth)::type_def(22)=sym_desc(method)
2556 returning(char);:arg(char);protection(public)modifier(volatile)virt(no)
2557 meth_name(ConstVolMeth)::type_def(23)=sym_desc(method)
2558 returning(float);:arg(float);protection(public)modifer(const volatile)
2559 virtual(no);;", @dots{}
2563 .stabs "A:T20=s1ConstMeth::21=##1;:i;2B.;VolatileMeth::22=##2;:c;2C.;
2564 ConstVolMeth::23=##12;:f;2D.;;",128,0,0,0
2567 @node Virtual Methods
2568 @section Virtual Methods
2570 << The following examples are based on a4.C >>
2572 The presence of virtual methods in a class definition adds additional
2573 data to the class description. The extra data is appended to the
2574 description of the virtual method and to the end of the class
2575 description. Consider the class definition below:
2581 virtual int A_virt (int arg) @{ return arg; @};
2585 This results in the stab below describing class A. It defines a new
2586 type (20) which is an 8 byte structure. The first field of the class
2587 struct is @samp{Adat}, an integer, starting at structure offset 0 and
2590 The second field in the class struct is not explicitly defined by the
2591 C++ class definition but is implied by the fact that the class
2592 contains a virtual method. This field is the vtable pointer. The
2593 name of the vtable pointer field starts with @samp{$vf} and continues with a
2594 type reference to the class it is part of. In this example the type
2595 reference for class A is 20 so the name of its vtable pointer field is
2596 @samp{$vf20}, followed by the usual colon.
2598 Next there is a type definition for the vtable pointer type (21).
2599 This is in turn defined as a pointer to another new type (22).
2601 Type 22 is the vtable itself, which is defined as an array, indexed by
2602 a range of integers between 0 and 1, and whose elements are of type
2603 17. Type 17 was the vtable record type defined by the boilerplate C++
2604 type definitions, as shown earlier.
2606 The bit offset of the vtable pointer field is 32. The number of bits
2607 in the field are not specified when the field is a vtable pointer.
2609 Next is the method definition for the virtual member function @code{A_virt}.
2610 Its description starts out using the same format as the non-virtual
2611 member functions described above, except instead of a dot after the
2612 @samp{A} there is an asterisk, indicating that the function is virtual.
2613 Since is is virtual some addition information is appended to the end
2614 of the method description.
2616 The first number represents the vtable index of the method. This is a
2617 32 bit unsigned number with the high bit set, followed by a
2620 The second number is a type reference to the first base class in the
2621 inheritence hierarchy defining the virtual member function. In this
2622 case the class stab describes a base class so the virtual function is
2623 not overriding any other definition of the method. Therefore the
2624 reference is to the type number of the class that the stab is
2627 This is followed by three semi-colons. One marks the end of the
2628 current sub-section, one marks the end of the method field, and the
2629 third marks the end of the struct definition.
2631 For classes containing virtual functions the very last section of the
2632 string part of the stab holds a type reference to the first base
2633 class. This is preceeded by @samp{~%} and followed by a final semi-colon.
2636 .stabs "class_name(A):type_def(20)=sym_desc(struct)struct_bytes(8)
2637 field_name(Adat):type_ref(int),bit_offset(0),field_bits(32);
2638 field_name(A virt func ptr):type_def(21)=type_desc(ptr to)type_def(22)=
2639 sym_desc(array)index_type_ref(range of int from 0 to 1);
2640 elem_type_ref(vtbl elem type),
2642 meth_name(A_virt)::typedef(23)=sym_desc(method)returning(int);
2643 :arg_type(int),protection(public)normal(yes)virtual(yes)
2644 vtable_index(1);class_first_defining(A);;;~%first_base(A);",
2648 @c FIXME: bogus line break.
2650 .stabs "A:t20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2651 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2655 @section Inheritence
2657 Stabs describing C++ derived classes include additional sections that
2658 describe the inheritence hierarchy of the class. A derived class stab
2659 also encodes the number of base classes. For each base class it tells
2660 if the base class is virtual or not, and if the inheritence is private
2661 or public. It also gives the offset into the object of the portion of
2662 the object corresponding to each base class.
2664 This additional information is embeded in the class stab following the
2665 number of bytes in the struct. First the number of base classes
2666 appears bracketed by an exclamation point and a comma.
2668 Then for each base type there repeats a series: two digits, a number,
2669 a comma, another number, and a semi-colon.
2671 The first of the two digits is 1 if the base class is virtual and 0 if
2672 not. The second digit is 2 if the derivation is public and 0 if not.
2674 The number following the first two digits is the offset from the start
2675 of the object to the part of the object pertaining to the base class.
2677 After the comma, the second number is a type_descriptor for the base
2678 type. Finally a semi-colon ends the series, which repeats for each
2681 The source below defines three base classes @code{A}, @code{B}, and
2682 @code{C} and the derived class @code{D}.
2689 virtual int A_virt (int arg) @{ return arg; @};
2695 virtual int B_virt (int arg) @{return arg; @};
2701 virtual int C_virt (int arg) @{return arg; @};
2704 class D : A, virtual B, public C @{
2707 virtual int A_virt (int arg ) @{ return arg+1; @};
2708 virtual int B_virt (int arg) @{ return arg+2; @};
2709 virtual int C_virt (int arg) @{ return arg+3; @};
2710 virtual int D_virt (int arg) @{ return arg; @};
2714 Class stabs similar to the ones described earlier are generated for
2717 @c FIXME!!! the linebreaks in the following example probably make the
2718 @c examples literally unusable, but I don't know any other way to get
2719 @c them on the page.
2720 @c One solution would be to put some of the type definitions into
2721 @c separate stabs, even if that's not exactly what the compiler actually
2724 .stabs "A:T20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2725 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2727 .stabs "B:Tt25=s8Bdat:1,0,32;$vf25:21,32;B_virt::26=##1;
2728 :i;2A*-2147483647;25;;;~%25;",128,0,0,0
2730 .stabs "C:Tt28=s8Cdat:1,0,32;$vf28:21,32;C_virt::29=##1;
2731 :i;2A*-2147483647;28;;;~%28;",128,0,0,0
2734 In the stab describing derived class @code{D} below, the information about
2735 the derivation of this class is encoded as follows.
2738 .stabs "derived_class_name:symbol_descriptors(struct tag&type)=
2739 type_descriptor(struct)struct_bytes(32)!num_bases(3),
2740 base_virtual(no)inheritence_public(no)base_offset(0),
2741 base_class_type_ref(A);
2742 base_virtual(yes)inheritence_public(no)base_offset(NIL),
2743 base_class_type_ref(B);
2744 base_virtual(no)inheritence_public(yes)base_offset(64),
2745 base_class_type_ref(C); @dots{}
2748 @c FIXME! fake linebreaks.
2750 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:
2751 1,160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt:
2752 :32:i;2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;
2753 28;;D_virt::32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2756 @node Virtual Base Classes
2757 @section Virtual Base Classes
2759 A derived class object consists of a concatination in memory of the data
2760 areas defined by each base class, starting with the leftmost and ending
2761 with the rightmost in the list of base classes. The exception to this
2762 rule is for virtual inheritence. In the example above, class @code{D}
2763 inherits virtually from base class @code{B}. This means that an
2764 instance of a @code{D} object will not contain its own @code{B} part but
2765 merely a pointer to a @code{B} part, known as a virtual base pointer.
2767 In a derived class stab, the base offset part of the derivation
2768 information, described above, shows how the base class parts are
2769 ordered. The base offset for a virtual base class is always given as 0.
2770 Notice that the base offset for @code{B} is given as 0 even though
2771 @code{B} is not the first base class. The first base class @code{A}
2774 The field information part of the stab for class @code{D} describes the field
2775 which is the pointer to the virtual base class @code{B}. The vbase pointer
2776 name is @samp{$vb} followed by a type reference to the virtual base class.
2777 Since the type id for @code{B} in this example is 25, the vbase pointer name
2780 @c FIXME!! fake linebreaks below
2782 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:1,
2783 160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt::32:i;
2784 2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;28;;D_virt:
2785 :32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2788 Following the name and a semicolon is a type reference describing the
2789 type of the virtual base class pointer, in this case 24. Type 24 was
2790 defined earlier as the type of the @code{B} class @code{this} pointer. The
2791 @code{this} pointer for a class is a pointer to the class type.
2794 .stabs "this:P24=*25=xsB:",64,0,0,8
2797 Finally the field offset part of the vbase pointer field description
2798 shows that the vbase pointer is the first field in the @code{D} object,
2799 before any data fields defined by the class. The layout of a @code{D}
2800 class object is a follows, @code{Adat} at 0, the vtable pointer for
2801 @code{A} at 32, @code{Cdat} at 64, the vtable pointer for C at 96, the
2802 virtual base pointer for @code{B} at 128, and @code{Ddat} at 160.
2805 @node Static Members
2806 @section Static Members
2808 The data area for a class is a concatenation of the space used by the
2809 data members of the class. If the class has virtual methods, a vtable
2810 pointer follows the class data. The field offset part of each field
2811 description in the class stab shows this ordering.
2813 << How is this reflected in stabs? See Cygnus bug #677 for some info. >>
2816 @appendix Table of Stab Types
2818 The following are all the possible values for the stab type field, for
2819 @code{a.out} files, in numeric order. This does not apply to XCOFF, but
2820 it does apply to stabs in ELF. Stabs in ECOFF use these values but add
2821 0x8f300 to distinguish them from non-stab symbols.
2823 The symbolic names are defined in the file @file{include/aout/stabs.def}.
2826 * Non-Stab Symbol Types:: Types from 0 to 0x1f
2827 * Stab Symbol Types:: Types from 0x20 to 0xff
2830 @node Non-Stab Symbol Types
2831 @appendixsec Non-Stab Symbol Types
2833 The following types are used by the linker and assembler, not by stab
2834 directives. Since this document does not attempt to describe aspects of
2835 object file format other than the debugging format, no details are
2838 @c Try to get most of these to fit on a single line.
2848 File scope absolute symbol
2850 @item 0x3 N_ABS | N_EXT
2851 External absolute symbol
2854 File scope text symbol
2856 @item 0x5 N_TEXT | N_EXT
2857 External text symbol
2860 File scope data symbol
2862 @item 0x7 N_DATA | N_EXT
2863 External data symbol
2866 File scope BSS symbol
2868 @item 0x9 N_BSS | N_EXT
2872 Same as @code{N_FN}, for Sequent compilers
2875 Symbol is indirected to another symbol
2878 Common---visible after shared library dynamic link
2881 Absolute set element
2884 Text segment set element
2887 Data segment set element
2890 BSS segment set element
2893 Pointer to set vector
2895 @item 0x1e N_WARNING
2896 Print a warning message during linking
2899 File name of a @file{.o} file
2902 @node Stab Symbol Types
2903 @appendixsec Stab Symbol Types
2905 The following symbol types indicate that this is a stab. This is the
2906 full list of stab numbers, including stab types that are used in
2907 languages other than C.
2911 Global symbol; see @ref{Global Variables}.
2914 Function name (for BSD Fortran); see @ref{Procedures}.
2917 Function name (@pxref{Procedures}) or text segment variable
2921 Data segment file-scope variable; see @ref{Statics}.
2924 BSS segment file-scope variable; see @ref{Statics}.
2927 Name of main routine; see @ref{Main Program}.
2930 Variable in @code{.rodata} section; see @ref{Statics}.
2933 Global symbol (for Pascal); see @ref{N_PC}.
2936 Number of symbols (according to Ultrix V4.0); see @ref{N_NSYMS}.
2939 No DST map; see @ref{N_NOMAP}.
2941 @c FIXME: describe this solaris feature in the body of the text (see
2942 @c comments in include/aout/stab.def).
2944 Object file (Solaris2).
2946 @c See include/aout/stab.def for (a little) more info.
2948 Debugger options (Solaris2).
2951 Register variable; see @ref{Register Variables}.
2954 Modula-2 compilation unit; see @ref{N_M2C}.
2957 Line number in text segment; see @ref{Line Numbers}.
2960 Line number in data segment; see @ref{Line Numbers}.
2963 Line number in bss segment; see @ref{Line Numbers}.
2966 Sun source code browser, path to @file{.cb} file; see @ref{N_BROWS}.
2969 GNU Modula2 definition module dependency; see @ref{N_DEFD}.
2972 Function start/body/end line numbers (Solaris2).
2975 GNU C++ exception variable; see @ref{N_EHDECL}.
2978 Modula2 info "for imc" (according to Ultrix V4.0); see @ref{N_MOD2}.
2981 GNU C++ @code{catch} clause; see @ref{N_CATCH}.
2984 Structure of union element; see @ref{N_SSYM}.
2987 Last stab for module (Solaris2).
2990 Path and name of source file; see @ref{Source Files}.
2993 Stack variable (@pxref{Stack Variables}) or type (@pxref{Typedefs}).
2996 Beginning of an include file (Sun only); see @ref{Include Files}.
2999 Name of include file; see @ref{Include Files}.
3002 Parameter variable; see @ref{Parameters}.
3005 End of an include file; see @ref{Include Files}.
3008 Alternate entry point; see @ref{N_ENTRY}.
3011 Beginning of a lexical block; see @ref{Block Structure}.
3014 Place holder for a deleted include file; see @ref{Include Files}.
3017 Modula2 scope information (Sun linker); see @ref{N_SCOPE}.
3020 End of a lexical block; see @ref{Block Structure}.
3023 Begin named common block; see @ref{Common Blocks}.
3026 End named common block; see @ref{Common Blocks}.
3029 Member of a common block; see @ref{Common Blocks}.
3031 @c FIXME: How does this really work? Move it to main body of document.
3033 Pascal @code{with} statement: type,,0,0,offset (Solaris2).
3036 Gould non-base registers; see @ref{Gould}.
3039 Gould non-base registers; see @ref{Gould}.
3042 Gould non-base registers; see @ref{Gould}.
3045 Gould non-base registers; see @ref{Gould}.
3048 Gould non-base registers; see @ref{Gould}.
3051 @c Restore the default table indent
3056 @node Symbol Descriptors
3057 @appendix Table of Symbol Descriptors
3059 The symbol descriptor is the character which follows the colon in many
3060 stabs, and which tells what kind of stab it is. @xref{String Field},
3061 for more information about their use.
3063 @c Please keep this alphabetical
3065 @c In TeX, this looks great, digit is in italics. But makeinfo insists
3066 @c on putting it in `', not realizing that @var should override @code.
3067 @c I don't know of any way to make makeinfo do the right thing. Seems
3068 @c like a makeinfo bug to me.
3072 Variable on the stack; see @ref{Stack Variables}.
3075 Parameter passed by reference in register; see @ref{Reference Parameters}.
3078 Based variable; see @ref{Based Variables}.
3081 Constant; see @ref{Constants}.
3084 Conformant array bound (Pascal, maybe other languages); @ref{Conformant
3085 Arrays}. Name of a caught exception (GNU C++). These can be
3086 distinguished because the latter uses @code{N_CATCH} and the former uses
3087 another symbol type.
3090 Floating point register variable; see @ref{Register Variables}.
3093 Parameter in floating point register; see @ref{Register Parameters}.
3096 File scope function; see @ref{Procedures}.
3099 Global function; see @ref{Procedures}.
3102 Global variable; see @ref{Global Variables}.
3105 @xref{Register Parameters}.
3108 Internal (nested) procedure; see @ref{Nested Procedures}.
3111 Internal (nested) function; see @ref{Nested Procedures}.
3114 Label name (documented by AIX, no further information known).
3117 Module; see @ref{Procedures}.
3120 Argument list parameter; see @ref{Parameters}.
3126 Fortran Function parameter; see @ref{Parameters}.
3129 Unfortunately, three separate meanings have been independently invented
3130 for this symbol descriptor. At least the GNU and Sun uses can be
3131 distinguished by the symbol type. Global Procedure (AIX) (symbol type
3132 used unknown); see @ref{Procedures}. Register parameter (GNU) (symbol
3133 type @code{N_PSYM}); see @ref{Parameters}. Prototype of function
3134 referenced by this file (Sun @code{acc}) (symbol type @code{N_FUN}).
3137 Static Procedure; see @ref{Procedures}.
3140 Register parameter; see @ref{Register Parameters}.
3143 Register variable; see @ref{Register Variables}.
3146 File scope variable; see @ref{Statics}.
3149 Type name; see @ref{Typedefs}.
3152 Enumeration, structure, or union tag; see @ref{Typedefs}.
3155 Parameter passed by reference; see @ref{Reference Parameters}.
3158 Procedure scope static variable; see @ref{Statics}.
3161 Conformant array; see @ref{Conformant Arrays}.
3164 Function return variable; see @ref{Parameters}.
3167 @node Type Descriptors
3168 @appendix Table of Type Descriptors
3170 The type descriptor is the character which follows the type number and
3171 an equals sign. It specifies what kind of type is being defined.
3172 @xref{String Field}, for more information about their use.
3177 Type reference; see @ref{String Field}.
3180 Reference to builtin type; see @ref{Negative Type Numbers}.
3183 Method (C++); see @ref{Cplusplus}.
3186 Pointer; see @ref{Miscellaneous Types}.
3192 Type Attributes (AIX); see @ref{String Field}. Member (class and variable)
3193 type (GNU C++); see @ref{Cplusplus}.
3196 Array; see @ref{Arrays}.
3199 Open array; see @ref{Arrays}.
3202 Pascal space type (AIX); see @ref{Miscellaneous Types}. Builtin integer
3203 type (Sun); see @ref{Builtin Type Descriptors}.
3206 Volatile-qualified type; see @ref{Miscellaneous Types}.
3209 Complex builtin type; see @ref{Builtin Type Descriptors}.
3212 COBOL Picture type. See AIX documentation for details.
3215 File type; see @ref{Miscellaneous Types}.
3218 N-dimensional dynamic array; see @ref{Arrays}.
3221 Enumeration type; see @ref{Enumerations}.
3224 N-dimensional subarray; see @ref{Arrays}.
3227 Function type; see @ref{Function Types}.
3230 Pascal function parameter; see @ref{Function Types}
3233 Builtin floating point type; see @ref{Builtin Type Descriptors}.
3236 COBOL Group. See AIX documentation for details.
3239 Imported type; see @ref{Cross-References}.
3242 Const-qualified type; see @ref{Miscellaneous Types}.
3245 COBOL File Descriptor. See AIX documentation for details.
3248 Multiple instance type; see @ref{Miscellaneous Types}.
3251 String type; see @ref{Strings}.
3254 Stringptr; see @ref{Strings}.
3257 Opaque type; see @ref{Typedefs}.
3260 Procedure; see @ref{Function Types}.
3263 Packed array; see @ref{Arrays}.
3266 Range type; see @ref{Subranges}.
3269 Builtin floating type; see @ref{Builtin Type Descriptors} (Sun). Pascal
3270 subroutine parameter; see @ref{Function Types} (AIX). Detecting this
3271 conflict is possible with careful parsing (hint: a Pascal subroutine
3272 parameter type will always contain a comma, and a builtin type
3273 descriptor never will).
3276 Structure type; see @ref{Structures}.
3279 Set type; see @ref{Miscellaneous Types}.
3282 Union; see @ref{Unions}.
3285 Variant record. This is a Pascal and Modula-2 feature which is like a
3286 union within a struct in C. See AIX documentation for details.
3289 Wide character; see @ref{Builtin Type Descriptors}.
3292 Cross-reference; see @ref{Cross-References}.
3295 gstring; see @ref{Strings}.
3298 @node Expanded Reference
3299 @appendix Expanded Reference by Stab Type
3301 @c FIXME: This appendix should go away; see N_PSYM or N_SO for an example.
3303 For a full list of stab types, and cross-references to where they are
3304 described, see @ref{Stab Types}. This appendix just duplicates certain
3305 information from the main body of this document; eventually the
3306 information will all be in one place.
3310 The first line is the symbol type (see @file{include/aout/stab.def}).
3312 The second line describes the language constructs the symbol type
3315 The third line is the stab format with the significant stab fields
3316 named and the rest NIL.
3318 Subsequent lines expand upon the meaning and possible values for each
3319 significant stab field. @samp{#} stands in for the type descriptor.
3321 Finally, any further information.
3324 * N_PC:: Pascal global symbol
3325 * N_NSYMS:: Number of symbols
3326 * N_NOMAP:: No DST map
3327 * N_M2C:: Modula-2 compilation unit
3328 * N_BROWS:: Path to .cb file for Sun source code browser
3329 * N_DEFD:: GNU Modula2 definition module dependency
3330 * N_EHDECL:: GNU C++ exception variable
3331 * N_MOD2:: Modula2 information "for imc"
3332 * N_CATCH:: GNU C++ "catch" clause
3333 * N_SSYM:: Structure or union element
3334 * N_ENTRY:: Alternate entry point
3335 * N_SCOPE:: Modula2 scope information (Sun only)
3336 * Gould:: non-base register symbols used on Gould systems
3337 * N_LENG:: Length of preceding entry
3343 @deffn @code{.stabs} N_PC
3345 Global symbol (for Pascal).
3348 "name" -> "symbol_name" <<?>>
3349 value -> supposedly the line number (stab.def is skeptical)
3353 @file{stabdump.c} says:
3355 global pascal symbol: name,,0,subtype,line
3363 @deffn @code{.stabn} N_NSYMS
3365 Number of symbols (according to Ultrix V4.0).
3368 0, files,,funcs,lines (stab.def)
3375 @deffn @code{.stabs} N_NOMAP
3377 No DST map for symbol (according to Ultrix V4.0). I think this means a
3378 variable has been optimized out.
3381 name, ,0,type,ignored (stab.def)
3388 @deffn @code{.stabs} N_M2C
3390 Modula-2 compilation unit.
3393 "string" -> "unit_name,unit_time_stamp[,code_time_stamp]"
3395 value -> 0 (main unit)
3403 @deffn @code{.stabs} N_BROWS
3405 Sun source code browser, path to @file{.cb} file
3408 "path to associated @file{.cb} file"
3410 Note: N_BROWS has the same value as N_BSLINE.
3416 @deffn @code{.stabn} N_DEFD
3418 GNU Modula2 definition module dependency.
3420 GNU Modula-2 definition module dependency. The value is the
3421 modification time of the definition file. The other field is non-zero
3422 if it is imported with the GNU M2 keyword @code{%INITIALIZE}. Perhaps
3423 @code{N_M2C} can be used if there are enough empty fields?
3429 @deffn @code{.stabs} N_EHDECL
3431 GNU C++ exception variable <<?>>.
3433 "@var{string} is variable name"
3435 Note: conflicts with @code{N_MOD2}.
3441 @deffn @code{.stab?} N_MOD2
3443 Modula2 info "for imc" (according to Ultrix V4.0)
3445 Note: conflicts with @code{N_EHDECL} <<?>>
3451 @deffn @code{.stabn} N_CATCH
3453 GNU C++ @code{catch} clause
3455 GNU C++ @code{catch} clause. The value is its address. The desc field
3456 is nonzero if this entry is immediately followed by a @code{CAUGHT} stab
3457 saying what exception was caught. Multiple @code{CAUGHT} stabs means
3458 that multiple exceptions can be caught here. If desc is 0, it means all
3459 exceptions are caught here.
3465 @deffn @code{.stabn} N_SSYM
3467 Structure or union element.
3469 The value is the offset in the structure.
3471 <<?looking at structs and unions in C I didn't see these>>
3477 @deffn @code{.stabn} N_ENTRY
3479 Alternate entry point.
3480 The value is its address.
3487 @deffn @code{.stab?} N_SCOPE
3489 Modula2 scope information (Sun linker)
3494 @section Non-base registers on Gould systems
3496 @deffn @code{.stab?} N_NBTEXT
3497 @deffnx @code{.stab?} N_NBDATA
3498 @deffnx @code{.stab?} N_NBBSS
3499 @deffnx @code{.stab?} N_NBSTS
3500 @deffnx @code{.stab?} N_NBLCS
3506 These are used on Gould systems for non-base registers syms.
3508 However, the following values are not the values used by Gould; they are
3509 the values which GNU has been documenting for these values for a long
3510 time, without actually checking what Gould uses. I include these values
3511 only because perhaps some someone actually did something with the GNU
3512 information (I hope not, why GNU knowingly assigned wrong values to
3513 these in the header file is a complete mystery to me).
3516 240 0xf0 N_NBTEXT ??
3517 242 0xf2 N_NBDATA ??
3527 @deffn @code{.stabn} N_LENG
3529 Second symbol entry containing a length-value for the preceding entry.
3530 The value is the length.
3534 @appendix Questions and Anomalies
3538 @c I think this is changed in GCC 2.4.5 to put the line number there.
3539 For GNU C stabs defining local and global variables (@code{N_LSYM} and
3540 @code{N_GSYM}), the desc field is supposed to contain the source
3541 line number on which the variable is defined. In reality the desc
3542 field is always 0. (This behavior is defined in @file{dbxout.c} and
3543 putting a line number in desc is controlled by @samp{#ifdef
3544 WINNING_GDB}, which defaults to false). GDB supposedly uses this
3545 information if you say @samp{list @var{var}}. In reality, @var{var} can
3546 be a variable defined in the program and GDB says @samp{function
3547 @var{var} not defined}.
3550 In GNU C stabs, there seems to be no way to differentiate tag types:
3551 structures, unions, and enums (symbol descriptor @samp{T}) and typedefs
3552 (symbol descriptor @samp{t}) defined at file scope from types defined locally
3553 to a procedure or other more local scope. They all use the @code{N_LSYM}
3554 stab type. Types defined at procedure scope are emited after the
3555 @code{N_RBRAC} of the preceding function and before the code of the
3556 procedure in which they are defined. This is exactly the same as
3557 types defined in the source file between the two procedure bodies.
3558 GDB overcompensates by placing all types in block #1, the block for
3559 symbols of file scope. This is true for default, @samp{-ansi} and
3560 @samp{-traditional} compiler options. (Bugs gcc/1063, gdb/1066.)
3563 What ends the procedure scope? Is it the proc block's @code{N_RBRAC} or the
3564 next @code{N_FUN}? (I believe its the first.)
3567 @c FIXME: This should go with the other stuff about global variables.
3568 Global variable stabs don't have location information. This comes
3569 from the external symbol for the same variable. The external symbol
3570 has a leading underbar on the _name of the variable and the stab does
3571 not. How do we know these two symbol table entries are talking about
3572 the same symbol when their names are different? (Answer: the debugger
3573 knows that external symbols have leading underbars).
3575 @c FIXME: This is absurdly vague; there all kinds of differences, some
3576 @c of which are the same between gnu & sun, and some of which aren't.
3577 @c In particular, I'm pretty sure GCC works with Sun dbx by default.
3579 @c Can GCC be configured to output stabs the way the Sun compiler
3580 @c does, so that their native debugging tools work? <NO?> It doesn't by
3581 @c default. GDB reads either format of stab. (GCC or SunC). How about
3585 @node XCOFF Differences
3586 @appendix Differences Between GNU Stabs in a.out and GNU Stabs in XCOFF
3588 @c FIXME: Merge *all* these into the main body of the document.
3589 The AIX/RS6000 native object file format is XCOFF with stabs. This
3590 appendix only covers those differences which are not covered in the main
3591 body of this document.
3595 BSD a.out stab types correspond to AIX XCOFF storage classes. In general
3596 the mapping is @code{N_@var{stabtype}} becomes @code{C_@var{stabtype}}.
3597 Some stab types in a.out are not supported in XCOFF; most of these use
3600 @c FIXME: Get C_* types for the block, figure out whether it is always
3601 @c used (I suspect not), explain clearly, and move to node Statics.
3602 Exception: initialised static @code{N_STSYM} and un-initialized static
3603 @code{N_LCSYM} both map to the @code{C_STSYM} storage class. But the
3604 distinction is preserved because in XCOFF @code{N_STSYM} and
3605 @code{N_LCSYM} must be emited in a named static block. Begin the block
3606 with @samp{.bs s[RW] data_section_name} for @code{N_STSYM} or @samp{.bs
3607 s bss_section_name} for @code{N_LCSYM}. End the block with @samp{.es}.
3609 @c FIXME: I think they are trying to say something about whether the
3610 @c assembler defaults the value to the location counter.
3612 If the XCOFF stab is an @code{N_FUN} (@code{C_FUN}) then follow the
3613 string field with @samp{,.} instead of just @samp{,}.
3616 I think that's it for @file{.s} file differences. They could stand to be
3617 better presented. This is just a list of what I have noticed so far.
3618 There are a @emph{lot} of differences in the information in the symbol
3619 tables of the executable and object files.
3621 Mapping of a.out stab types to XCOFF storage classes:
3624 stab type storage class
3625 -------------------------------
3661 @node Sun Differences
3662 @appendix Differences Between GNU Stabs and Sun Native Stabs
3664 @c FIXME: Merge all this stuff into the main body of the document.
3668 GNU C stabs define @emph{all} types, file or procedure scope, as
3669 @code{N_LSYM}. Sun doc talks about using @code{N_GSYM} too.
3672 Sun C stabs use type number pairs in the format
3673 (@var{file-number},@var{type-number}) where @var{file-number} is a
3674 number starting with 1 and incremented for each sub-source file in the
3675 compilation. @var{type-number} is a number starting with 1 and
3676 incremented for each new type defined in the compilation. GNU C stabs
3677 use the type number alone, with no source file number.
3681 @appendix Using Stabs With The ELF Object File Format
3683 The ELF object file format allows tools to create object files with
3684 custom sections containing any arbitrary data. To use stabs in ELF
3685 object files, the tools create two custom sections, a section named
3686 @code{.stab} which contains an array of fixed length structures, one
3687 struct per stab, and a section named @code{.stabstr} containing all the
3688 variable length strings that are referenced by stabs in the @code{.stab}
3689 section. The byte order of the stabs binary data matches the byte order
3690 of the ELF file itself, as determined from the @code{EI_DATA} field in
3691 the @code{e_ident} member of the ELF header.
3693 The first stab in the @code{.stab} section for each compilation unit is
3694 synthetic, generated entirely by the assembler, with no corresponding
3695 @code{.stab} directive as input to the assembler. This stab contains
3696 the following fields:
3700 Offset in the @code{.stabstr} section to the source filename.
3706 Unused field, always zero.
3709 Count of upcoming symbols, i.e., the number of remaining stabs for this
3713 Size of the string table fragment associated with this source file, in
3717 The @code{.stabstr} section always starts with a null byte (so that string
3718 offsets of zero reference a null string), followed by random length strings,
3719 each of which is null byte terminated.
3721 The ELF section header for the @code{.stab} section has its
3722 @code{sh_link} member set to the section number of the @code{.stabstr}
3723 section, and the @code{.stabstr} section has its ELF section
3724 header @code{sh_type} member set to @code{SHT_STRTAB} to mark it as a
3727 To keep linking fast, it is a bad idea to have the linker relocating
3728 stabs, so (except for Solaris 2.2 and earlier, see below) none of the
3729 addresses in the @code{n_value} field of the stabs are relocated by the
3730 linker. Instead they are relative to the source file (or some entity
3731 smaller than a source file, like a function). To find the address of
3732 each section corresponding to a given source file, the compiler puts out
3733 symbols giving the address of each section for a given source file.
3734 Since these are ELF (not stab) symbols, the linker can relocate them
3735 correctly. They are named @code{Bbss.bss} for the bss section,
3736 @code{Ddata.data} for the data section, and @code{Drodata.rodata} for
3737 the rodata section. For the text section, there is no such symbol. For
3738 an example of how these symbols work, @xref{ELF Transformations}. GCC
3739 does not provide these symbols; it instead relies on the stabs getting
3740 relocated, which loses for Solaris 2.3 (see below). Thus address which
3741 would normally be relative to @code{Bbss.bss}, etc., are absolute. The
3742 linker provided with Solaris 2.2 and earlier relocates stabs using
3743 relocation information from a @code{.rela.stab} section, which means
3744 that the value of an @code{N_FUN} stab in an executable is the actual
3745 address. I think this is just standard ELF relocations, as it would do
3746 for any section, rather than a special-purpose stabs hack. For Solaris
3747 2.3 and later, the linker ignores relocations for the stabs section.
3748 The value of a @code{N_FUN} stab is zero and the address of a function
3749 can be obtained from the ELF (non-stab) symbols. Sun, in reference to
3750 bug 1142109, has verified that this is intentional. Because looking
3751 things up in the ELF symbols would probably be slow, and this doesn't
3752 provide any way to deal with nested functions, it would probably be
3753 better to use a @code{Ttext.text} symbol for stabs-in-elf on non-Solaris
3754 machines, and make the address in the @code{N_FUN} relative to the
3755 @code{Ttext.text} symbol. In addition to @code{N_FUN} symbols, whether
3756 the linker relocates stabs also affects some @code{N_ROSYM},
3757 @code{N_STSYM}, and @code{N_LCSYM} symbols; see @ref{Statics}.
3759 @node Symbol Types Index
3760 @unnumbered Symbol Types Index