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 * Stab Sections:: In some object file formats, stabs are
89 * Symbol Types Index:: Index of symbolic stab symbol type names.
95 @chapter Overview of Stabs
97 @dfn{Stabs} refers to a format for information that describes a program
98 to a debugger. This format was apparently invented by
100 the University of California at Berkeley, for the @code{pdx} Pascal
101 debugger; the format has spread widely since then.
103 This document is one of the few published sources of documentation on
104 stabs. It is believed to be comprehensive for stabs used by C. The
105 lists of symbol descriptors (@pxref{Symbol Descriptors}) and type
106 descriptors (@pxref{Type Descriptors}) are believed to be completely
107 comprehensive. Stabs for COBOL-specific features and for variant
108 records (used by Pascal and Modula-2) are poorly documented here.
110 @c FIXME: Need to document all OS9000 stuff in GDB; see all references
111 @c to os9k_stabs in stabsread.c.
113 Other sources of information on stabs are @cite{Dbx and Dbxtool
114 Interfaces}, 2nd edition, by Sun, 1988, and @cite{AIX Version 3.2 Files
115 Reference}, Fourth Edition, September 1992, "dbx Stabstring Grammar" in
116 the a.out section, page 2-31. This document is believed to incorporate
117 the information from those two sources except where it explicitly directs
118 you to them for more information.
121 * Flow:: Overview of debugging information flow
122 * Stabs Format:: Overview of stab format
123 * String Field:: The string field
124 * C Example:: A simple example in C source
125 * Assembly Code:: The simple example at the assembly level
129 @section Overview of Debugging Information Flow
131 The GNU C compiler compiles C source in a @file{.c} file into assembly
132 language in a @file{.s} file, which the assembler translates into
133 a @file{.o} file, which the linker combines with other @file{.o} files and
134 libraries to produce an executable file.
136 With the @samp{-g} option, GCC puts in the @file{.s} file additional
137 debugging information, which is slightly transformed by the assembler
138 and linker, and carried through into the final executable. This
139 debugging information describes features of the source file like line
140 numbers, the types and scopes of variables, and function names,
141 parameters, and scopes.
143 For some object file formats, the debugging information is encapsulated
144 in assembler directives known collectively as @dfn{stab} (symbol table)
145 directives, which are interspersed with the generated code. Stabs are
146 the native format for debugging information in the a.out and XCOFF
147 object file formats. The GNU tools can also emit stabs in the COFF and
148 ECOFF object file formats.
150 The assembler adds the information from stabs to the symbol information
151 it places by default in the symbol table and the string table of the
152 @file{.o} file it is building. The linker consolidates the @file{.o}
153 files into one executable file, with one symbol table and one string
154 table. Debuggers use the symbol and string tables in the executable as
155 a source of debugging information about the program.
158 @section Overview of Stab Format
160 There are three overall formats for stab assembler directives,
161 differentiated by the first word of the stab. The name of the directive
162 describes which combination of four possible data fields follows. It is
163 either @code{.stabs} (string), @code{.stabn} (number), or @code{.stabd}
164 (dot). IBM's XCOFF assembler uses @code{.stabx} (and some other
165 directives such as @code{.file} and @code{.bi}) instead of
166 @code{.stabs}, @code{.stabn} or @code{.stabd}.
168 The overall format of each class of stab is:
171 .stabs "@var{string}",@var{type},@var{other},@var{desc},@var{value}
172 .stabn @var{type},@var{other},@var{desc},@var{value}
173 .stabd @var{type},@var{other},@var{desc}
174 .stabx "@var{string}",@var{value},@var{type},@var{sdb-type}
177 @c what is the correct term for "current file location"? My AIX
178 @c assembler manual calls it "the value of the current location counter".
179 For @code{.stabn} and @code{.stabd}, there is no @var{string} (the
180 @code{n_strx} field is zero; see @ref{Symbol Tables}). For
181 @code{.stabd}, the @var{value} field is implicit and has the value of
182 the current file location. For @code{.stabx}, the @var{sdb-type} field
183 is unused for stabs and can always be set to zero. The @var{other}
184 field is almost always unused and can be set to zero.
186 The number in the @var{type} field gives some basic information about
187 which type of stab this is (or whether it @emph{is} a stab, as opposed
188 to an ordinary symbol). Each valid type number defines a different stab
189 type; further, the stab type defines the exact interpretation of, and
190 possible values for, any remaining @var{string}, @var{desc}, or
191 @var{value} fields present in the stab. @xref{Stab Types}, for a list
192 in numeric order of the valid @var{type} field values for stab directives.
195 @section The String Field
197 For most stabs the string field holds the meat of the
198 debugging information. The flexible nature of this field
199 is what makes stabs extensible. For some stab types the string field
200 contains only a name. For other stab types the contents can be a great
203 The overall format of the string field for most stab types is:
206 "@var{name}:@var{symbol-descriptor} @var{type-information}"
209 @var{name} is the name of the symbol represented by the stab; it can
210 contain a pair of colons (@pxref{Nested Symbols}). @var{name} can be
211 omitted, which means the stab represents an unnamed object. For
212 example, @samp{:t10=*2} defines type 10 as a pointer to type 2, but does
213 not give the type a name. Omitting the @var{name} field is supported by
214 AIX dbx and GDB after about version 4.8, but not other debuggers. GCC
215 sometimes uses a single space as the name instead of omitting the name
216 altogether; apparently that is supported by most debuggers.
218 The @var{symbol-descriptor} following the @samp{:} is an alphabetic
219 character that tells more specifically what kind of symbol the stab
220 represents. If the @var{symbol-descriptor} is omitted, but type
221 information follows, then the stab represents a local variable. For a
222 list of symbol descriptors, see @ref{Symbol Descriptors}. The @samp{c}
223 symbol descriptor is an exception in that it is not followed by type
224 information. @xref{Constants}.
226 @var{type-information} is either a @var{type-number}, or
227 @samp{@var{type-number}=}. A @var{type-number} alone is a type
228 reference, referring directly to a type that has already been defined.
230 The @samp{@var{type-number}=} form is a type definition, where the
231 number represents a new type which is about to be defined. The type
232 definition may refer to other types by number, and those type numbers
233 may be followed by @samp{=} and nested definitions. Also, the Lucid
234 compiler will repeat @samp{@var{type-number}=} more than once if it
235 wants to define several type numbers at once.
237 In a type definition, if the character that follows the equals sign is
238 non-numeric then it is a @var{type-descriptor}, and tells what kind of
239 type is about to be defined. Any other values following the
240 @var{type-descriptor} vary, depending on the @var{type-descriptor}.
241 @xref{Type Descriptors}, for a list of @var{type-descriptor} values. If
242 a number follows the @samp{=} then the number is a @var{type-reference}.
243 For a full description of types, @ref{Types}.
245 There is an AIX extension for type attributes. Following the @samp{=}
246 are any number of type attributes. Each one starts with @samp{@@} and
247 ends with @samp{;}. Debuggers, including AIX's dbx and GDB 4.10, skip
248 any type attributes they do not recognize. GDB 4.9 and other versions
249 of dbx may not do this. Because of a conflict with C++
250 (@pxref{Cplusplus}), new attributes should not be defined which begin
251 with a digit, @samp{(}, or @samp{-}; GDB may be unable to distinguish
252 those from the C++ type descriptor @samp{@@}. The attributes are:
255 @item a@var{boundary}
256 @var{boundary} is an integer specifying the alignment. I assume it
257 applies to all variables of this type.
260 Pointer class (for checking). Not sure what this means, or how
261 @var{integer} is interpreted.
264 Indicate this is a packed type, meaning that structure fields or array
265 elements are placed more closely in memory, to save memory at the
269 Size in bits of a variable of this type. This is fully supported by GDB
273 Indicate that this type is a string instead of an array of characters,
274 or a bitstring instead of a set. It doesn't change the layout of the
275 data being represented, but does enable the debugger to know which type
279 All of this can make the string field quite long. All versions of GDB,
280 and some versions of dbx, can handle arbitrarily long strings. But many
281 versions of dbx (or assemblers or linkers, I'm not sure which)
282 cretinously limit the strings to about 80 characters, so compilers which
283 must work with such systems need to split the @code{.stabs} directive
284 into several @code{.stabs} directives. Each stab duplicates every field
285 except the string field. The string field of every stab except the last
286 is marked as continued with a backslash at the end (in the assembly code
287 this may be written as a double backslash, depending on the assembler).
288 Removing the backslashes and concatenating the string fields of each
289 stab produces the original, long string. Just to be incompatible (or so
290 they don't have to worry about what the assembler does with
291 backslashes), AIX can use @samp{?} instead of backslash.
294 @section A Simple Example in C Source
296 To get the flavor of how stabs describe source information for a C
297 program, let's look at the simple program:
302 printf("Hello world");
306 When compiled with @samp{-g}, the program above yields the following
307 @file{.s} file. Line numbers have been added to make it easier to refer
308 to parts of the @file{.s} file in the description of the stabs that
312 @section The Simple Example at the Assembly Level
314 This simple ``hello world'' example demonstrates several of the stab
315 types used to describe C language source files.
319 2 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0
320 3 .stabs "hello.c",100,0,0,Ltext0
323 6 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
324 7 .stabs "char:t2=r2;0;127;",128,0,0,0
325 8 .stabs "long int:t3=r1;-2147483648;2147483647;",128,0,0,0
326 9 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
327 10 .stabs "long unsigned int:t5=r1;0;-1;",128,0,0,0
328 11 .stabs "short int:t6=r1;-32768;32767;",128,0,0,0
329 12 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
330 13 .stabs "short unsigned int:t8=r1;0;65535;",128,0,0,0
331 14 .stabs "long long unsigned int:t9=r1;0;-1;",128,0,0,0
332 15 .stabs "signed char:t10=r1;-128;127;",128,0,0,0
333 16 .stabs "unsigned char:t11=r1;0;255;",128,0,0,0
334 17 .stabs "float:t12=r1;4;0;",128,0,0,0
335 18 .stabs "double:t13=r1;8;0;",128,0,0,0
336 19 .stabs "long double:t14=r1;8;0;",128,0,0,0
337 20 .stabs "void:t15=15",128,0,0,0
340 23 .ascii "Hello, world!\12\0"
355 38 sethi %hi(LC0),%o1
356 39 or %o1,%lo(LC0),%o0
367 50 .stabs "main:F1",36,0,0,_main
368 51 .stabn 192,0,0,LBB2
369 52 .stabn 224,0,0,LBE2
372 @node Program Structure
373 @chapter Encoding the Structure of the Program
375 The elements of the program structure that stabs encode include the name
376 of the main function, the names of the source and include files, the
377 line numbers, procedure names and types, and the beginnings and ends of
381 * Main Program:: Indicate what the main program is
382 * Source Files:: The path and name of the source file
383 * Include Files:: Names of include files
386 * Nested Procedures::
388 * Alternate Entry Points:: Entering procedures except at the beginning.
392 @section Main Program
395 Most languages allow the main program to have any name. The
396 @code{N_MAIN} stab type tells the debugger the name that is used in this
397 program. Only the string field is significant; it is the name of
398 a function which is the main program. Most C compilers do not use this
399 stab (they expect the debugger to assume that the name is @code{main}),
400 but some C compilers emit an @code{N_MAIN} stab for the @code{main}
404 @section Paths and Names of the Source Files
407 Before any other stabs occur, there must be a stab specifying the source
408 file. This information is contained in a symbol of stab type
409 @code{N_SO}; the string field contains the name of the file. The
410 value of the symbol is the start address of the portion of the
411 text section corresponding to that file.
413 With the Sun Solaris2 compiler, the desc field contains a
414 source-language code.
415 @c Do the debuggers use it? What are the codes? -djm
417 Some compilers (for example, GCC2 and SunOS4 @file{/bin/cc}) also
418 include the directory in which the source was compiled, in a second
419 @code{N_SO} symbol preceding the one containing the file name. This
420 symbol can be distinguished by the fact that it ends in a slash. Code
421 from the @code{cfront} C++ compiler can have additional @code{N_SO} symbols for
422 nonexistent source files after the @code{N_SO} for the real source file;
423 these are believed to contain no useful information.
428 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0 # @r{100 is N_SO}
429 .stabs "hello.c",100,0,0,Ltext0
434 Instead of @code{N_SO} symbols, XCOFF uses a @code{.file} assembler
435 directive which assembles to a standard COFF @code{.file} symbol;
436 explaining this in detail is outside the scope of this document.
439 @section Names of Include Files
441 There are several schemes for dealing with include files: the
442 traditional @code{N_SOL} approach, Sun's @code{N_BINCL} approach, and the
443 XCOFF @code{C_BINCL} approach (which despite the similar name has little in
444 common with @code{N_BINCL}).
447 An @code{N_SOL} symbol specifies which include file subsequent symbols
448 refer to. The string field is the name of the file and the value is the
449 text address corresponding to the end of the previous include file and
450 the start of this one. To specify the main source file again, use an
451 @code{N_SOL} symbol with the name of the main source file.
456 The @code{N_BINCL} approach works as follows. An @code{N_BINCL} symbol
457 specifies the start of an include file. In an object file, only the
458 string is significant; the Sun linker puts data into some of the
459 other fields. The end of the include file is marked by an
460 @code{N_EINCL} symbol (which has no string field). In an object
461 file, there is no significant data in the @code{N_EINCL} symbol; the Sun
462 linker puts data into some of the fields. @code{N_BINCL} and
463 @code{N_EINCL} can be nested.
465 If the linker detects that two source files have identical stabs between
466 an @code{N_BINCL} and @code{N_EINCL} pair (as will generally be the case
467 for a header file), then it only puts out the stabs once. Each
468 additional occurance is replaced by an @code{N_EXCL} symbol. I believe
469 the Sun (SunOS4, not sure about Solaris) linker is the only one which
470 supports this feature.
471 @c What do the fields of N_EXCL contain? -djm
475 For the start of an include file in XCOFF, use the @file{.bi} assembler
476 directive, which generates a @code{C_BINCL} symbol. A @file{.ei}
477 directive, which generates a @code{C_EINCL} symbol, denotes the end of
478 the include file. Both directives are followed by the name of the
479 source file in quotes, which becomes the string for the symbol.
480 The value of each symbol, produced automatically by the assembler
481 and linker, is the offset into the executable of the beginning
482 (inclusive, as you'd expect) or end (inclusive, as you would not expect)
483 of the portion of the COFF line table that corresponds to this include
484 file. @code{C_BINCL} and @code{C_EINCL} do not nest.
487 @section Line Numbers
490 An @code{N_SLINE} symbol represents the start of a source line. The
491 desc field contains the line number and the value contains the code
492 address for the start of that source line. On most machines the address
493 is absolute; for stabs in sections (@pxref{Stab Sections}), it is
494 relative to the function in which the @code{N_SLINE} symbol occurs.
498 GNU documents @code{N_DSLINE} and @code{N_BSLINE} symbols for line
499 numbers in the data or bss segments, respectively. They are identical
500 to @code{N_SLINE} but are relocated differently by the linker. They
501 were intended to be used to describe the source location of a variable
502 declaration, but I believe that GCC2 actually puts the line number in
503 the desc field of the stab for the variable itself. GDB has been
504 ignoring these symbols (unless they contain a string field) since
507 For single source lines that generate discontiguous code, such as flow
508 of control statements, there may be more than one line number entry for
509 the same source line. In this case there is a line number entry at the
510 start of each code range, each with the same line number.
512 XCOFF does not use stabs for line numbers. Instead, it uses COFF line
513 numbers (which are outside the scope of this document). Standard COFF
514 line numbers cannot deal with include files, but in XCOFF this is fixed
515 with the @code{C_BINCL} method of marking include files (@pxref{Include
521 @findex N_FUN, for functions
523 @findex N_STSYM, for functions (Sun acc)
524 @findex N_GSYM, for functions (Sun acc)
525 All of the following stabs normally use the @code{N_FUN} symbol type.
526 However, Sun's @code{acc} compiler on SunOS4 uses @code{N_GSYM} and
527 @code{N_STSYM}, which means that the value of the stab for the function
528 is useless and the debugger must get the address of the function from
529 the non-stab symbols instead. BSD Fortran is said to use @code{N_FNAME}
530 with the same restriction; the value of the symbol is not useful (I'm
531 not sure it really does use this, because GDB doesn't handle this and no
534 A function is represented by an @samp{F} symbol descriptor for a global
535 (extern) function, and @samp{f} for a static (local) function. The
536 value is the address of the start of the function. For @code{a.out}, it
537 is already relocated. For stabs in ELF, the SunPRO compiler version
538 2.0.1 and GCC put out an address which gets relocated by the linker. In
539 a future release SunPRO is planning to put out zero, in which case the
540 address can be found from the ELF (non-stab) symbol. Because looking
541 things up in the ELF symbols would probably be slow, I'm not sure how to
542 find which symbol of that name is the right one, and this doesn't
543 provide any way to deal with nested functions, it would probably be
544 better to make the value of the stab an address relative to the start of
545 the file, or just absolute. See @ref{ELF Linker Relocation} for more
546 information on linker relocation of stabs in ELF files.
548 The type information of the stab represents the return type of the
549 function; thus @samp{foo:f5} means that foo is a function returning type
550 5. There is no need to try to get the line number of the start of the
551 function from the stab for the function; it is in the next
552 @code{N_SLINE} symbol.
554 @c FIXME: verify whether the "I suspect" below is true or not.
555 Some compilers (such as Sun's Solaris compiler) support an extension for
556 specifying the types of the arguments. I suspect this extension is not
557 used for old (non-prototyped) function definitions in C. If the
558 extension is in use, the type information of the stab for the function
559 is followed by type information for each argument, with each argument
560 preceded by @samp{;}. An argument type of 0 means that additional
561 arguments are being passed, whose types and number may vary (@samp{...}
562 in ANSI C). GDB has tolerated this extension (parsed the syntax, if not
563 necessarily used the information) since at least version 4.8; I don't
564 know whether all versions of dbx tolerate it. The argument types given
565 here are not redundant with the symbols for the formal parameters
566 (@pxref{Parameters}); they are the types of the arguments as they are
567 passed, before any conversions might take place. For example, if a C
568 function which is declared without a prototype takes a @code{float}
569 argument, the value is passed as a @code{double} but then converted to a
570 @code{float}. Debuggers need to use the types given in the arguments
571 when printing values, but when calling the function they need to use the
572 types given in the symbol defining the function.
574 If the return type and types of arguments of a function which is defined
575 in another source file are specified (i.e., a function prototype in ANSI
576 C), traditionally compilers emit no stab; the only way for the debugger
577 to find the information is if the source file where the function is
578 defined was also compiled with debugging symbols. As an extension the
579 Solaris compiler uses symbol descriptor @samp{P} followed by the return
580 type of the function, followed by the arguments, each preceded by
581 @samp{;}, as in a stab with symbol descriptor @samp{f} or @samp{F}.
582 This use of symbol descriptor @samp{P} can be distinguished from its use
583 for register parameters (@pxref{Register Parameters}) by the fact that it has
584 symbol type @code{N_FUN}.
586 The AIX documentation also defines symbol descriptor @samp{J} as an
587 internal function. I assume this means a function nested within another
588 function. It also says symbol descriptor @samp{m} is a module in
589 Modula-2 or extended Pascal.
591 Procedures (functions which do not return values) are represented as
592 functions returning the @code{void} type in C. I don't see why this couldn't
593 be used for all languages (inventing a @code{void} type for this purpose if
594 necessary), but the AIX documentation defines @samp{I}, @samp{P}, and
595 @samp{Q} for internal, global, and static procedures, respectively.
596 These symbol descriptors are unusual in that they are not followed by
599 The following example shows a stab for a function @code{main} which
600 returns type number @code{1}. The @code{_main} specified for the value
601 is a reference to an assembler label which is used to fill in the start
602 address of the function.
605 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
608 The stab representing a procedure is located immediately following the
609 code of the procedure. This stab is in turn directly followed by a
610 group of other stabs describing elements of the procedure. These other
611 stabs describe the procedure's parameters, its block local variables, and
614 @node Nested Procedures
615 @section Nested Procedures
617 For any of the symbol descriptors representing procedures, after the
618 symbol descriptor and the type information is optionally a scope
619 specifier. This consists of a comma, the name of the procedure, another
620 comma, and the name of the enclosing procedure. The first name is local
621 to the scope specified, and seems to be redundant with the name of the
622 symbol (before the @samp{:}). This feature is used by GCC, and
623 presumably Pascal, Modula-2, etc., compilers, for nested functions.
625 If procedures are nested more than one level deep, only the immediately
626 containing scope is specified. For example, this code:
638 return baz (x + 2 * y);
640 return x + bar (3 * x);
648 .stabs "baz:f1,baz,bar",36,0,0,_baz.15 # @r{36 is N_FUN}
649 .stabs "bar:f1,bar,foo",36,0,0,_bar.12
650 .stabs "foo:F1",36,0,0,_foo
653 @node Block Structure
654 @section Block Structure
658 @c For GCC 2.5.8 or so stabs-in-coff, these are absolute instead of
659 @c function relative (as documented below). But GDB has never been able
660 @c to deal with that (it had wanted them to be relative to the file, but
661 @c I just fixed that (between GDB 4.12 and 4.13)), so it is function
662 @c relative just like ELF and SOM and the below documentation.
663 The program's block structure is represented by the @code{N_LBRAC} (left
664 brace) and the @code{N_RBRAC} (right brace) stab types. The variables
665 defined inside a block precede the @code{N_LBRAC} symbol for most
666 compilers, including GCC. Other compilers, such as the Convex, Acorn
667 RISC machine, and Sun @code{acc} compilers, put the variables after the
668 @code{N_LBRAC} symbol. The values of the @code{N_LBRAC} and
669 @code{N_RBRAC} symbols are the start and end addresses of the code of
670 the block, respectively. For most machines, they are relative to the
671 starting address of this source file. For the Gould NP1, they are
672 absolute. For stabs in sections (@pxref{Stab Sections}), they are
673 relative to the function in which they occur.
675 The @code{N_LBRAC} and @code{N_RBRAC} stabs that describe the block
676 scope of a procedure are located after the @code{N_FUN} stab that
677 represents the procedure itself.
679 Sun documents the desc field of @code{N_LBRAC} and
680 @code{N_RBRAC} symbols as containing the nesting level of the block.
681 However, dbx seems to not care, and GCC always sets desc to
684 @node Alternate Entry Points
685 @section Alternate Entry Points
687 Some languages, like Fortran, have the ability to enter procedures at
688 some place other than the beginning. One can declare an alternate entry
689 point. The @code{N_ENTRY} stab is for this; however, the Sun FORTRAN compiler
690 doesn't use it. According to AIX documentation, only the name of a
691 @code{C_ENTRY} stab is significant; the address of the alternate entry
692 point comes from the corresponding external symbol. A previous revision
693 of this document said that the value of an @code{N_ENTRY} stab was the
694 address of the alternate entry point, but I don't know the source for
700 The @samp{c} symbol descriptor indicates that this stab represents a
701 constant. This symbol descriptor is an exception to the general rule
702 that symbol descriptors are followed by type information. Instead, it
703 is followed by @samp{=} and one of the following:
707 Boolean constant. @var{value} is a numeric value; I assume it is 0 for
711 Character constant. @var{value} is the numeric value of the constant.
713 @item e @var{type-information} , @var{value}
714 Constant whose value can be represented as integral.
715 @var{type-information} is the type of the constant, as it would appear
716 after a symbol descriptor (@pxref{String Field}). @var{value} is the
717 numeric value of the constant. GDB 4.9 does not actually get the right
718 value if @var{value} does not fit in a host @code{int}, but it does not
719 do anything violent, and future debuggers could be extended to accept
720 integers of any size (whether unsigned or not). This constant type is
721 usually documented as being only for enumeration constants, but GDB has
722 never imposed that restriction; I don't know about other debuggers.
725 Integer constant. @var{value} is the numeric value. The type is some
726 sort of generic integer type (for GDB, a host @code{int}); to specify
727 the type explicitly, use @samp{e} instead.
730 Real constant. @var{value} is the real value, which can be @samp{INF}
731 (optionally preceded by a sign) for infinity, @samp{QNAN} for a quiet
732 NaN (not-a-number), or @samp{SNAN} for a signalling NaN. If it is a
733 normal number the format is that accepted by the C library function
737 String constant. @var{string} is a string enclosed in either @samp{'}
738 (in which case @samp{'} characters within the string are represented as
739 @samp{\'} or @samp{"} (in which case @samp{"} characters within the
740 string are represented as @samp{\"}).
742 @item S @var{type-information} , @var{elements} , @var{bits} , @var{pattern}
743 Set constant. @var{type-information} is the type of the constant, as it
744 would appear after a symbol descriptor (@pxref{String Field}).
745 @var{elements} is the number of elements in the set (does this means
746 how many bits of @var{pattern} are actually used, which would be
747 redundant with the type, or perhaps the number of bits set in
748 @var{pattern}? I don't get it), @var{bits} is the number of bits in the
749 constant (meaning it specifies the length of @var{pattern}, I think),
750 and @var{pattern} is a hexadecimal representation of the set. AIX
751 documentation refers to a limit of 32 bytes, but I see no reason why
752 this limit should exist. This form could probably be used for arbitrary
753 constants, not just sets; the only catch is that @var{pattern} should be
754 understood to be target, not host, byte order and format.
757 The boolean, character, string, and set constants are not supported by
758 GDB 4.9, but it ignores them. GDB 4.8 and earlier gave an error
759 message and refused to read symbols from the file containing the
762 The above information is followed by @samp{;}.
767 Different types of stabs describe the various ways that variables can be
768 allocated: on the stack, globally, in registers, in common blocks,
769 statically, or as arguments to a function.
772 * Stack Variables:: Variables allocated on the stack.
773 * Global Variables:: Variables used by more than one source file.
774 * Register Variables:: Variables in registers.
775 * Common Blocks:: Variables statically allocated together.
776 * Statics:: Variables local to one source file.
777 * Based Variables:: Fortran pointer based variables.
778 * Parameters:: Variables for arguments to functions.
781 @node Stack Variables
782 @section Automatic Variables Allocated on the Stack
784 If a variable's scope is local to a function and its lifetime is only as
785 long as that function executes (C calls such variables
786 @dfn{automatic}), it can be allocated in a register (@pxref{Register
787 Variables}) or on the stack.
790 Each variable allocated on the stack has a stab with the symbol
791 descriptor omitted. Since type information should begin with a digit,
792 @samp{-}, or @samp{(}, only those characters precluded from being used
793 for symbol descriptors. However, the Acorn RISC machine (ARM) is said
794 to get this wrong: it puts out a mere type definition here, without the
795 preceding @samp{@var{type-number}=}. This is a bad idea; there is no
796 guarantee that type descriptors are distinct from symbol descriptors.
797 Stabs for stack variables use the @code{N_LSYM} stab type.
799 The value of the stab is the offset of the variable within the
800 local variables. On most machines this is an offset from the frame
801 pointer and is negative. The location of the stab specifies which block
802 it is defined in; see @ref{Block Structure}.
804 For example, the following C code:
814 produces the following stabs:
817 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
818 .stabs "x:1",128,0,0,-12 # @r{128 is N_LSYM}
819 .stabn 192,0,0,LBB2 # @r{192 is N_LBRAC}
820 .stabn 224,0,0,LBE2 # @r{224 is N_RBRAC}
823 @xref{Procedures} for more information on the @code{N_FUN} stab, and
824 @ref{Block Structure} for more information on the @code{N_LBRAC} and
825 @code{N_RBRAC} stabs.
827 @node Global Variables
828 @section Global Variables
831 A variable whose scope is not specific to just one source file is
832 represented by the @samp{G} symbol descriptor. These stabs use the
833 @code{N_GSYM} stab type. The type information for the stab
834 (@pxref{String Field}) gives the type of the variable.
836 For example, the following source code:
843 yields the following assembly code:
846 .stabs "g_foo:G2",32,0,0,0 # @r{32 is N_GSYM}
853 The address of the variable represented by the @code{N_GSYM} is not
854 contained in the @code{N_GSYM} stab. The debugger gets this information
855 from the external symbol for the global variable. In the example above,
856 the @code{.global _g_foo} and @code{_g_foo:} lines tell the assembler to
857 produce an external symbol.
859 @node Register Variables
860 @section Register Variables
863 @c According to an old version of this manual, AIX uses C_RPSYM instead
864 @c of C_RSYM. I am skeptical; this should be verified.
865 Register variables have their own stab type, @code{N_RSYM}, and their
866 own symbol descriptor, @samp{r}. The stab's value is the
867 number of the register where the variable data will be stored.
868 @c .stabs "name:type",N_RSYM,0,RegSize,RegNumber (Sun doc)
870 AIX defines a separate symbol descriptor @samp{d} for floating point
871 registers. This seems unnecessary; why not just just give floating
872 point registers different register numbers? I have not verified whether
873 the compiler actually uses @samp{d}.
875 If the register is explicitly allocated to a global variable, but not
879 register int g_bar asm ("%g5");
883 then the stab may be emitted at the end of the object file, with
884 the other bss symbols.
887 @section Common Blocks
889 A common block is a statically allocated section of memory which can be
890 referred to by several source files. It may contain several variables.
891 I believe Fortran is the only language with this feature.
897 A @code{N_BCOMM} stab begins a common block and an @code{N_ECOMM} stab
898 ends it. The only field that is significant in these two stabs is the
899 string, which names a normal (non-debugging) symbol that gives the
900 address of the common block. According to IBM documentation, only the
901 @code{N_BCOMM} has the name of the common block (even though their
902 compiler actually puts it both places).
906 The stabs for the members of the common block are between the
907 @code{N_BCOMM} and the @code{N_ECOMM}; the value of each stab is the
908 offset within the common block of that variable. IBM uses the
909 @code{C_ECOML} stab type, and there is a corresponding @code{N_ECOML}
910 stab type, but Sun's Fortran compiler uses @code{N_GSYM} instead. The
911 variables within a common block use the @samp{V} symbol descriptor (I
912 believe this is true of all Fortran variables). Other stabs (at least
913 type declarations using @code{C_DECL}) can also be between the
914 @code{N_BCOMM} and the @code{N_ECOMM}.
917 @section Static Variables
919 Initialized static variables are represented by the @samp{S} and
920 @samp{V} symbol descriptors. @samp{S} means file scope static, and
921 @samp{V} means procedure scope static. One exception: in XCOFF, IBM's
922 xlc compiler always uses @samp{V}, and whether it is file scope or not
923 is distinguished by whether the stab is located within a function.
925 @c This is probably not worth mentioning; it is only true on the sparc
926 @c for `double' variables which although declared const are actually in
927 @c the data segment (the text segment can't guarantee 8 byte alignment).
929 @c 2.4.5 has a bug in that it uses @code{N_FUN}, so neither dbx nor GDB can
930 @c find the variables)
933 @findex N_FUN, for variables
935 In a.out files, @code{N_STSYM} means the data section, @code{N_FUN}
936 means the text section, and @code{N_LCSYM} means the bss section. For
937 those systems with a read-only data section separate from the text
938 section (Solaris), @code{N_ROSYM} means the read-only data section.
940 For example, the source lines:
943 static const int var_const = 5;
944 static int var_init = 2;
945 static int var_noinit;
949 yield the following stabs:
952 .stabs "var_const:S1",36,0,0,_var_const # @r{36 is N_FUN}
954 .stabs "var_init:S1",38,0,0,_var_init # @r{38 is N_STSYM}
956 .stabs "var_noinit:S1",40,0,0,_var_noinit # @r{40 is N_LCSYM}
959 In XCOFF files, the stab type need not indicate the section;
960 @code{C_STSYM} can be used for all statics. Also, each static variable
961 is enclosed in a static block. A @code{C_BSTAT} (emitted with a
962 @samp{.bs} assembler directive) symbol begins the static block; its
963 value is the address of the static block, its section is the section of
964 the variables in that static block, and its name is @samp{.bs}. A
965 @code{C_ESTAT} (emitted with a @samp{.es} assembler directive) symbol
966 ends the static block; its name is @samp{.es} and its value and section
969 In ECOFF files, the storage class is used to specify the section, so the
970 stab type need not indicate the section.
972 In ELF files, for the SunPRO compiler version 2.0.1, symbol descriptor
973 @samp{S} means that the address is absolute (the linker relocates it)
974 and symbol descriptor @samp{V} means that the address is relative to the
975 start of the relevant section for that compilation unit. SunPRO has
976 plans to have the linker stop relocating stabs; I suspect that their the
977 debugger gets the address from the corresponding ELF (not stab) symbol.
978 I'm not sure how to find which symbol of that name is the right one.
979 The clean way to do all this would be to have a the value of a symbol
980 descriptor @samp{S} symbol be an offset relative to the start of the
981 file, just like everything else, but that introduces obvious
982 compatibility problems. For more information on linker stab relocation,
983 @xref{ELF Linker Relocation}.
985 @node Based Variables
986 @section Fortran Based Variables
988 Fortran (at least, the Sun and SGI dialects of FORTRAN-77) has a feature
989 which allows allocating arrays with @code{malloc}, but which avoids
990 blurring the line between arrays and pointers the way that C does. In
991 stabs such a variable uses the @samp{b} symbol descriptor.
993 For example, the Fortran declarations
996 real foo, foo10(10), foo10_5(10,5)
998 pointer (foo10p, foo10)
999 pointer (foo105p, foo10_5)
1007 foo10_5:bar3;1;5;ar3;1;10;6
1010 In this example, @code{real} is type 6 and type 3 is an integral type
1011 which is the type of the subscripts of the array (probably
1014 The @samp{b} symbol descriptor is like @samp{V} in that it denotes a
1015 statically allocated symbol whose scope is local to a function; see
1016 @xref{Statics}. The value of the symbol, instead of being the address
1017 of the variable itself, is the address of a pointer to that variable.
1018 So in the above example, the value of the @code{foo} stab is the address
1019 of a pointer to a real, the value of the @code{foo10} stab is the
1020 address of a pointer to a 10-element array of reals, and the value of
1021 the @code{foo10_5} stab is the address of a pointer to a 5-element array
1022 of 10-element arrays of reals.
1027 Formal parameters to a function are represented by a stab (or sometimes
1028 two; see below) for each parameter. The stabs are in the order in which
1029 the debugger should print the parameters (i.e., the order in which the
1030 parameters are declared in the source file). The exact form of the stab
1031 depends on how the parameter is being passed.
1034 Parameters passed on the stack use the symbol descriptor @samp{p} and
1035 the @code{N_PSYM} symbol type. The value of the symbol is an offset
1036 used to locate the parameter on the stack; its exact meaning is
1037 machine-dependent, but on most machines it is an offset from the frame
1040 As a simple example, the code:
1051 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
1052 .stabs "argc:p1",160,0,0,68 # @r{160 is N_PSYM}
1053 .stabs "argv:p20=*21=*2",160,0,0,72
1056 The type definition of @code{argv} is interesting because it contains
1057 several type definitions. Type 21 is pointer to type 2 (char) and
1058 @code{argv} (type 20) is pointer to type 21.
1060 @c FIXME: figure out what these mean and describe them coherently.
1061 The following symbol descriptors are also said to go with @code{N_PSYM}.
1062 The value of the symbol is said to be an offset from the argument
1063 pointer (I'm not sure whether this is true or not).
1067 pF Fortran function parameter
1068 X (function result variable)
1072 * Register Parameters::
1073 * Local Variable Parameters::
1074 * Reference Parameters::
1075 * Conformant Arrays::
1078 @node Register Parameters
1079 @subsection Passing Parameters in Registers
1081 If the parameter is passed in a register, then traditionally there are
1082 two symbols for each argument:
1085 .stabs "arg:p1" . . . ; N_PSYM
1086 .stabs "arg:r1" . . . ; N_RSYM
1089 Debuggers use the second one to find the value, and the first one to
1090 know that it is an argument.
1093 @findex N_RSYM, for parameters
1094 Because that approach is kind of ugly, some compilers use symbol
1095 descriptor @samp{P} or @samp{R} to indicate an argument which is in a
1096 register. Symbol type @code{C_RPSYM} is used with @samp{R} and
1097 @code{N_RSYM} is used with @samp{P}. The symbol's value is
1098 the register number. @samp{P} and @samp{R} mean the same thing; the
1099 difference is that @samp{P} is a GNU invention and @samp{R} is an IBM
1100 (XCOFF) invention. As of version 4.9, GDB should handle either one.
1102 There is at least one case where GCC uses a @samp{p} and @samp{r} pair
1103 rather than @samp{P}; this is where the argument is passed in the
1104 argument list and then loaded into a register.
1106 According to the AIX documentation, symbol descriptor @samp{D} is for a
1107 parameter passed in a floating point register. This seems
1108 unnecessary---why not just use @samp{R} with a register number which
1109 indicates that it's a floating point register? I haven't verified
1110 whether the system actually does what the documentation indicates.
1112 @c FIXME: On the hppa this is for any type > 8 bytes, I think, and not
1113 @c for small structures (investigate).
1114 On the sparc and hppa, for a @samp{P} symbol whose type is a structure
1115 or union, the register contains the address of the structure. On the
1116 sparc, this is also true of a @samp{p} and @samp{r} pair (using Sun
1117 @code{cc}) or a @samp{p} symbol. However, if a (small) structure is
1118 really in a register, @samp{r} is used. And, to top it all off, on the
1119 hppa it might be a structure which was passed on the stack and loaded
1120 into a register and for which there is a @samp{p} and @samp{r} pair! I
1121 believe that symbol descriptor @samp{i} is supposed to deal with this
1122 case (it is said to mean "value parameter by reference, indirect
1123 access"; I don't know the source for this information), but I don't know
1124 details or what compilers or debuggers use it, if any (not GDB or GCC).
1125 It is not clear to me whether this case needs to be dealt with
1126 differently than parameters passed by reference (@pxref{Reference Parameters}).
1128 @node Local Variable Parameters
1129 @subsection Storing Parameters as Local Variables
1131 There is a case similar to an argument in a register, which is an
1132 argument that is actually stored as a local variable. Sometimes this
1133 happens when the argument was passed in a register and then the compiler
1134 stores it as a local variable. If possible, the compiler should claim
1135 that it's in a register, but this isn't always done.
1137 If a parameter is passed as one type and converted to a smaller type by
1138 the prologue (for example, the parameter is declared as a @code{float},
1139 but the calling conventions specify that it is passed as a
1140 @code{double}), then GCC2 (sometimes) uses a pair of symbols. The first
1141 symbol uses symbol descriptor @samp{p} and the type which is passed.
1142 The second symbol has the type and location which the parameter actually
1143 has after the prologue. For example, suppose the following C code
1144 appears with no prototypes involved:
1153 if @code{f} is passed as a double at stack offset 8, and the prologue
1154 converts it to a float in register number 0, then the stabs look like:
1157 .stabs "f:p13",160,0,3,8 # @r{160 is @code{N_PSYM}, here 13 is @code{double}}
1158 .stabs "f:r12",64,0,3,0 # @r{64 is @code{N_RSYM}, here 12 is @code{float}}
1161 In both stabs 3 is the line number where @code{f} is declared
1162 (@pxref{Line Numbers}).
1164 @findex N_LSYM, for parameter
1165 GCC, at least on the 960, has another solution to the same problem. It
1166 uses a single @samp{p} symbol descriptor for an argument which is stored
1167 as a local variable but uses @code{N_LSYM} instead of @code{N_PSYM}. In
1168 this case, the value of the symbol is an offset relative to the local
1169 variables for that function, not relative to the arguments; on some
1170 machines those are the same thing, but not on all.
1172 @c This is mostly just background info; the part that logically belongs
1173 @c here is the last sentence.
1174 On the VAX or on other machines in which the calling convention includes
1175 the number of words of arguments actually passed, the debugger (GDB at
1176 least) uses the parameter symbols to keep track of whether it needs to
1177 print nameless arguments in addition to the formal parameters which it
1178 has printed because each one has a stab. For example, in
1181 extern int fprintf (FILE *stream, char *format, @dots{});
1183 fprintf (stdout, "%d\n", x);
1186 there are stabs for @code{stream} and @code{format}. On most machines,
1187 the debugger can only print those two arguments (because it has no way
1188 of knowing that additional arguments were passed), but on the VAX or
1189 other machines with a calling convention which indicates the number of
1190 words of arguments, the debugger can print all three arguments. To do
1191 so, the parameter symbol (symbol descriptor @samp{p}) (not necessarily
1192 @samp{r} or symbol descriptor omitted symbols) needs to contain the
1193 actual type as passed (for example, @code{double} not @code{float} if it
1194 is passed as a double and converted to a float).
1196 @node Reference Parameters
1197 @subsection Passing Parameters by Reference
1199 If the parameter is passed by reference (e.g., Pascal @code{VAR}
1200 parameters), then the symbol descriptor is @samp{v} if it is in the
1201 argument list, or @samp{a} if it in a register. Other than the fact
1202 that these contain the address of the parameter rather than the
1203 parameter itself, they are identical to @samp{p} and @samp{R},
1204 respectively. I believe @samp{a} is an AIX invention; @samp{v} is
1205 supported by all stabs-using systems as far as I know.
1207 @node Conformant Arrays
1208 @subsection Passing Conformant Array Parameters
1210 @c Is this paragraph correct? It is based on piecing together patchy
1211 @c information and some guesswork
1212 Conformant arrays are a feature of Modula-2, and perhaps other
1213 languages, in which the size of an array parameter is not known to the
1214 called function until run-time. Such parameters have two stabs: a
1215 @samp{x} for the array itself, and a @samp{C}, which represents the size
1216 of the array. The value of the @samp{x} stab is the offset in the
1217 argument list where the address of the array is stored (it this right?
1218 it is a guess); the value of the @samp{C} stab is the offset in the
1219 argument list where the size of the array (in elements? in bytes?) is
1223 @chapter Defining Types
1225 The examples so far have described types as references to previously
1226 defined types, or defined in terms of subranges of or pointers to
1227 previously defined types. This chapter describes the other type
1228 descriptors that may follow the @samp{=} in a type definition.
1231 * Builtin Types:: Integers, floating point, void, etc.
1232 * Miscellaneous Types:: Pointers, sets, files, etc.
1233 * Cross-References:: Referring to a type not yet defined.
1234 * Subranges:: A type with a specific range.
1235 * Arrays:: An aggregate type of same-typed elements.
1236 * Strings:: Like an array but also has a length.
1237 * Enumerations:: Like an integer but the values have names.
1238 * Structures:: An aggregate type of different-typed elements.
1239 * Typedefs:: Giving a type a name.
1240 * Unions:: Different types sharing storage.
1245 @section Builtin Types
1247 Certain types are built in (@code{int}, @code{short}, @code{void},
1248 @code{float}, etc.); the debugger recognizes these types and knows how
1249 to handle them. Thus, don't be surprised if some of the following ways
1250 of specifying builtin types do not specify everything that a debugger
1251 would need to know about the type---in some cases they merely specify
1252 enough information to distinguish the type from other types.
1254 The traditional way to define builtin types is convolunted, so new ways
1255 have been invented to describe them. Sun's @code{acc} uses special
1256 builtin type descriptors (@samp{b} and @samp{R}), and IBM uses negative
1257 type numbers. GDB accepts all three ways, as of version 4.8; dbx just
1258 accepts the traditional builtin types and perhaps one of the other two
1259 formats. The following sections describe each of these formats.
1262 * Traditional Builtin Types:: Put on your seatbelts and prepare for kludgery
1263 * Builtin Type Descriptors:: Builtin types with special type descriptors
1264 * Negative Type Numbers:: Builtin types using negative type numbers
1267 @node Traditional Builtin Types
1268 @subsection Traditional Builtin Types
1270 This is the traditional, convoluted method for defining builtin types.
1271 There are several classes of such type definitions: integer, floating
1272 point, and @code{void}.
1275 * Traditional Integer Types::
1276 * Traditional Other Types::
1279 @node Traditional Integer Types
1280 @subsubsection Traditional Integer Types
1282 Often types are defined as subranges of themselves. If the bounding values
1283 fit within an @code{int}, then they are given normally. For example:
1286 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0 # @r{128 is N_LSYM}
1287 .stabs "char:t2=r2;0;127;",128,0,0,0
1290 Builtin types can also be described as subranges of @code{int}:
1293 .stabs "unsigned short:t6=r1;0;65535;",128,0,0,0
1296 If the lower bound of a subrange is 0 and the upper bound is -1,
1297 the type is an unsigned integral type whose bounds are too
1298 big to describe in an @code{int}. Traditionally this is only used for
1299 @code{unsigned int} and @code{unsigned long}:
1302 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
1305 For larger types, GCC 2.4.5 puts out bounds in octal, with one or more
1306 leading zeroes. In this case a negative bound consists of a number
1307 which is a 1 bit (for the sign bit) followed by a 0 bit for each bit in
1308 the number (except the sign bit), and a positive bound is one which is a
1309 1 bit for each bit in the number (except possibly the sign bit). All
1310 known versions of dbx and GDB version 4 accept this (at least in the
1311 sense of not refusing to process the file), but GDB 3.5 refuses to read
1312 the whole file containing such symbols. So GCC 2.3.3 did not output the
1313 proper size for these types. As an example of octal bounds, the string
1314 fields of the stabs for 64 bit integer types look like:
1316 @c .stabs directives, etc., omitted to make it fit on the page.
1318 long int:t3=r1;001000000000000000000000;000777777777777777777777;
1319 long unsigned int:t5=r1;000000000000000000000000;001777777777777777777777;
1322 If the lower bound of a subrange is 0 and the upper bound is negative,
1323 the type is an unsigned integral type whose size in bytes is the
1324 absolute value of the upper bound. I believe this is a Convex
1325 convention for @code{unsigned long long}.
1327 If the lower bound of a subrange is negative and the upper bound is 0,
1328 the type is a signed integral type whose size in bytes is
1329 the absolute value of the lower bound. I believe this is a Convex
1330 convention for @code{long long}. To distinguish this from a legitimate
1331 subrange, the type should be a subrange of itself. I'm not sure whether
1332 this is the case for Convex.
1334 @node Traditional Other Types
1335 @subsubsection Traditional Other Types
1337 If the upper bound of a subrange is 0 and the lower bound is positive,
1338 the type is a floating point type, and the lower bound of the subrange
1339 indicates the number of bytes in the type:
1342 .stabs "float:t12=r1;4;0;",128,0,0,0
1343 .stabs "double:t13=r1;8;0;",128,0,0,0
1346 However, GCC writes @code{long double} the same way it writes
1347 @code{double}, so there is no way to distinguish.
1350 .stabs "long double:t14=r1;8;0;",128,0,0,0
1353 Complex types are defined the same way as floating-point types; there is
1354 no way to distinguish a single-precision complex from a double-precision
1355 floating-point type.
1357 The C @code{void} type is defined as itself:
1360 .stabs "void:t15=15",128,0,0,0
1363 I'm not sure how a boolean type is represented.
1365 @node Builtin Type Descriptors
1366 @subsection Defining Builtin Types Using Builtin Type Descriptors
1368 This is the method used by Sun's @code{acc} for defining builtin types.
1369 These are the type descriptors to define builtin types:
1372 @c FIXME: clean up description of width and offset, once we figure out
1374 @item b @var{signed} @var{char-flag} @var{width} ; @var{offset} ; @var{nbits} ;
1375 Define an integral type. @var{signed} is @samp{u} for unsigned or
1376 @samp{s} for signed. @var{char-flag} is @samp{c} which indicates this
1377 is a character type, or is omitted. I assume this is to distinguish an
1378 integral type from a character type of the same size, for example it
1379 might make sense to set it for the C type @code{wchar_t} so the debugger
1380 can print such variables differently (Solaris does not do this). Sun
1381 sets it on the C types @code{signed char} and @code{unsigned char} which
1382 arguably is wrong. @var{width} and @var{offset} appear to be for small
1383 objects stored in larger ones, for example a @code{short} in an
1384 @code{int} register. @var{width} is normally the number of bytes in the
1385 type. @var{offset} seems to always be zero. @var{nbits} is the number
1386 of bits in the type.
1388 Note that type descriptor @samp{b} used for builtin types conflicts with
1389 its use for Pascal space types (@pxref{Miscellaneous Types}); they can
1390 be distinguished because the character following the type descriptor
1391 will be a digit, @samp{(}, or @samp{-} for a Pascal space type, or
1392 @samp{u} or @samp{s} for a builtin type.
1395 Documented by AIX to define a wide character type, but their compiler
1396 actually uses negative type numbers (@pxref{Negative Type Numbers}).
1398 @item R @var{fp-type} ; @var{bytes} ;
1399 Define a floating point type. @var{fp-type} has one of the following values:
1403 IEEE 32-bit (single precision) floating point format.
1406 IEEE 64-bit (double precision) floating point format.
1408 @item 3 (NF_COMPLEX)
1409 @item 4 (NF_COMPLEX16)
1410 @item 5 (NF_COMPLEX32)
1411 @c "GDB source" really means @file{include/aout/stab_gnu.h}, but trying
1412 @c to put that here got an overfull hbox.
1413 These are for complex numbers. A comment in the GDB source describes
1414 them as Fortran @code{complex}, @code{double complex}, and
1415 @code{complex*16}, respectively, but what does that mean? (i.e., Single
1416 precision? Double precison?).
1418 @item 6 (NF_LDOUBLE)
1419 Long double. This should probably only be used for Sun format
1420 @code{long double}, and new codes should be used for other floating
1421 point formats (@code{NF_DOUBLE} can be used if a @code{long double} is
1422 really just an IEEE double, of course).
1425 @var{bytes} is the number of bytes occupied by the type. This allows a
1426 debugger to perform some operations with the type even if it doesn't
1427 understand @var{fp-type}.
1429 @item g @var{type-information} ; @var{nbits}
1430 Documented by AIX to define a floating type, but their compiler actually
1431 uses negative type numbers (@pxref{Negative Type Numbers}).
1433 @item c @var{type-information} ; @var{nbits}
1434 Documented by AIX to define a complex type, but their compiler actually
1435 uses negative type numbers (@pxref{Negative Type Numbers}).
1438 The C @code{void} type is defined as a signed integral type 0 bits long:
1440 .stabs "void:t19=bs0;0;0",128,0,0,0
1442 The Solaris compiler seems to omit the trailing semicolon in this case.
1443 Getting sloppy in this way is not a swift move because if a type is
1444 embedded in a more complex expression it is necessary to be able to tell
1447 I'm not sure how a boolean type is represented.
1449 @node Negative Type Numbers
1450 @subsection Negative Type Numbers
1452 This is the method used in XCOFF for defining builtin types.
1453 Since the debugger knows about the builtin types anyway, the idea of
1454 negative type numbers is simply to give a special type number which
1455 indicates the builtin type. There is no stab defining these types.
1457 There are several subtle issues with negative type numbers.
1459 One is the size of the type. A builtin type (for example the C types
1460 @code{int} or @code{long}) might have different sizes depending on
1461 compiler options, the target architecture, the ABI, etc. This issue
1462 doesn't come up for IBM tools since (so far) they just target the
1463 RS/6000; the sizes indicated below for each size are what the IBM
1464 RS/6000 tools use. To deal with differing sizes, either define separate
1465 negative type numbers for each size (which works but requires changing
1466 the debugger, and, unless you get both AIX dbx and GDB to accept the
1467 change, introduces an incompatibility), or use a type attribute
1468 (@pxref{String Field}) to define a new type with the appropriate size
1469 (which merely requires a debugger which understands type attributes,
1470 like AIX dbx). For example,
1473 .stabs "boolean:t10=@@s8;-16",128,0,0,0
1476 defines an 8-bit boolean type, and
1479 .stabs "boolean:t10=@@s64;-16",128,0,0,0
1482 defines a 64-bit boolean type.
1484 A similar issue is the format of the type. This comes up most often for
1485 floating-point types, which could have various formats (particularly
1486 extended doubles, which vary quite a bit even among IEEE systems).
1487 Again, it is best to define a new negative type number for each
1488 different format; changing the format based on the target system has
1489 various problems. One such problem is that the Alpha has both VAX and
1490 IEEE floating types. One can easily imagine one library using the VAX
1491 types and another library in the same executable using the IEEE types.
1492 Another example is that the interpretation of whether a boolean is true
1493 or false can be based on the least significant bit, most significant
1494 bit, whether it is zero, etc., and different compilers (or different
1495 options to the same compiler) might provide different kinds of boolean.
1497 The last major issue is the names of the types. The name of a given
1498 type depends @emph{only} on the negative type number given; these do not
1499 vary depending on the language, the target system, or anything else.
1500 One can always define separate type numbers---in the following list you
1501 will see for example separate @code{int} and @code{integer*4} types
1502 which are identical except for the name. But compatibility can be
1503 maintained by not inventing new negative type numbers and instead just
1504 defining a new type with a new name. For example:
1507 .stabs "CARDINAL:t10=-8",128,0,0,0
1510 Here is the list of negative type numbers. The phrase @dfn{integral
1511 type} is used to mean twos-complement (I strongly suspect that all
1512 machines which use stabs use twos-complement; most machines use
1513 twos-complement these days).
1517 @code{int}, 32 bit signed integral type.
1520 @code{char}, 8 bit type holding a character. Both GDB and dbx on AIX
1521 treat this as signed. GCC uses this type whether @code{char} is signed
1522 or not, which seems like a bad idea. The AIX compiler (@code{xlc}) seems to
1523 avoid this type; it uses -5 instead for @code{char}.
1526 @code{short}, 16 bit signed integral type.
1529 @code{long}, 32 bit signed integral type.
1532 @code{unsigned char}, 8 bit unsigned integral type.
1535 @code{signed char}, 8 bit signed integral type.
1538 @code{unsigned short}, 16 bit unsigned integral type.
1541 @code{unsigned int}, 32 bit unsigned integral type.
1544 @code{unsigned}, 32 bit unsigned integral type.
1547 @code{unsigned long}, 32 bit unsigned integral type.
1550 @code{void}, type indicating the lack of a value.
1553 @code{float}, IEEE single precision.
1556 @code{double}, IEEE double precision.
1559 @code{long double}, IEEE double precision. The compiler claims the size
1560 will increase in a future release, and for binary compatibility you have
1561 to avoid using @code{long double}. I hope when they increase it they
1562 use a new negative type number.
1565 @code{integer}. 32 bit signed integral type.
1568 @code{boolean}. 32 bit type. GDB and GCC assume that zero is false,
1569 one is true, and other values have unspecified meaning. I hope this
1570 agrees with how the IBM tools use the type.
1573 @code{short real}. IEEE single precision.
1576 @code{real}. IEEE double precision.
1579 @code{stringptr}. @xref{Strings}.
1582 @code{character}, 8 bit unsigned character type.
1585 @code{logical*1}, 8 bit type. This Fortran type has a split
1586 personality in that it is used for boolean variables, but can also be
1587 used for unsigned integers. 0 is false, 1 is true, and other values are
1591 @code{logical*2}, 16 bit type. This Fortran type has a split
1592 personality in that it is used for boolean variables, but can also be
1593 used for unsigned integers. 0 is false, 1 is true, and other values are
1597 @code{logical*4}, 32 bit type. This Fortran type has a split
1598 personality in that it is used for boolean variables, but can also be
1599 used for unsigned integers. 0 is false, 1 is true, and other values are
1603 @code{logical}, 32 bit type. This Fortran type has a split
1604 personality in that it is used for boolean variables, but can also be
1605 used for unsigned integers. 0 is false, 1 is true, and other values are
1609 @code{complex}. A complex type consisting of two IEEE single-precision
1610 floating point values.
1613 @code{complex}. A complex type consisting of two IEEE double-precision
1614 floating point values.
1617 @code{integer*1}, 8 bit signed integral type.
1620 @code{integer*2}, 16 bit signed integral type.
1623 @code{integer*4}, 32 bit signed integral type.
1626 @code{wchar}. Wide character, 16 bits wide, unsigned (what format?
1630 @node Miscellaneous Types
1631 @section Miscellaneous Types
1634 @item b @var{type-information} ; @var{bytes}
1635 Pascal space type. This is documented by IBM; what does it mean?
1637 This use of the @samp{b} type descriptor can be distinguished
1638 from its use for builtin integral types (@pxref{Builtin Type
1639 Descriptors}) because the character following the type descriptor is
1640 always a digit, @samp{(}, or @samp{-}.
1642 @item B @var{type-information}
1643 A volatile-qualified version of @var{type-information}. This is
1644 a Sun extension. References and stores to a variable with a
1645 volatile-qualified type must not be optimized or cached; they
1646 must occur as the user specifies them.
1648 @item d @var{type-information}
1649 File of type @var{type-information}. As far as I know this is only used
1652 @item k @var{type-information}
1653 A const-qualified version of @var{type-information}. This is a Sun
1654 extension. A variable with a const-qualified type cannot be modified.
1656 @item M @var{type-information} ; @var{length}
1657 Multiple instance type. The type seems to composed of @var{length}
1658 repetitions of @var{type-information}, for example @code{character*3} is
1659 represented by @samp{M-2;3}, where @samp{-2} is a reference to a
1660 character type (@pxref{Negative Type Numbers}). I'm not sure how this
1661 differs from an array. This appears to be a Fortran feature.
1662 @var{length} is a bound, like those in range types; see @ref{Subranges}.
1664 @item S @var{type-information}
1665 Pascal set type. @var{type-information} must be a small type such as an
1666 enumeration or a subrange, and the type is a bitmask whose length is
1667 specified by the number of elements in @var{type-information}.
1669 In CHILL, if it is a bitstring instead of a set, also use the @samp{S}
1670 type attribute (@pxref{String Field}).
1672 @item * @var{type-information}
1673 Pointer to @var{type-information}.
1676 @node Cross-References
1677 @section Cross-References to Other Types
1679 A type can be used before it is defined; one common way to deal with
1680 that situation is just to use a type reference to a type which has not
1683 Another way is with the @samp{x} type descriptor, which is followed by
1684 @samp{s} for a structure tag, @samp{u} for a union tag, or @samp{e} for
1685 a enumerator tag, followed by the name of the tag, followed by @samp{:}.
1686 If the name contains @samp{::} between a @samp{<} and @samp{>} pair (for
1687 C++ templates), such a @samp{::} does not end the name---only a single
1688 @samp{:} ends the name; see @ref{Nested Symbols}.
1690 For example, the following C declarations:
1701 .stabs "bar:G16=*17=xsfoo:",32,0,0,0
1704 Not all debuggers support the @samp{x} type descriptor, so on some
1705 machines GCC does not use it. I believe that for the above example it
1706 would just emit a reference to type 17 and never define it, but I
1707 haven't verified that.
1709 Modula-2 imported types, at least on AIX, use the @samp{i} type
1710 descriptor, which is followed by the name of the module from which the
1711 type is imported, followed by @samp{:}, followed by the name of the
1712 type. There is then optionally a comma followed by type information for
1713 the type. This differs from merely naming the type (@pxref{Typedefs}) in
1714 that it identifies the module; I don't understand whether the name of
1715 the type given here is always just the same as the name we are giving
1716 it, or whether this type descriptor is used with a nameless stab
1717 (@pxref{String Field}), or what. The symbol ends with @samp{;}.
1720 @section Subrange Types
1722 The @samp{r} type descriptor defines a type as a subrange of another
1723 type. It is followed by type information for the type of which it is a
1724 subrange, a semicolon, an integral lower bound, a semicolon, an
1725 integral upper bound, and a semicolon. The AIX documentation does not
1726 specify the trailing semicolon, in an effort to specify array indexes
1727 more cleanly, but a subrange which is not an array index has always
1728 included a trailing semicolon (@pxref{Arrays}).
1730 Instead of an integer, either bound can be one of the following:
1733 @item A @var{offset}
1734 The bound is passed by reference on the stack at offset @var{offset}
1735 from the argument list. @xref{Parameters}, for more information on such
1738 @item T @var{offset}
1739 The bound is passed by value on the stack at offset @var{offset} from
1742 @item a @var{register-number}
1743 The bound is pased by reference in register number
1744 @var{register-number}.
1746 @item t @var{register-number}
1747 The bound is passed by value in register number @var{register-number}.
1753 Subranges are also used for builtin types; see @ref{Traditional Builtin Types}.
1756 @section Array Types
1758 Arrays use the @samp{a} type descriptor. Following the type descriptor
1759 is the type of the index and the type of the array elements. If the
1760 index type is a range type, it ends in a semicolon; otherwise
1761 (for example, if it is a type reference), there does not
1762 appear to be any way to tell where the types are separated. In an
1763 effort to clean up this mess, IBM documents the two types as being
1764 separated by a semicolon, and a range type as not ending in a semicolon
1765 (but this is not right for range types which are not array indexes,
1766 @pxref{Subranges}). I think probably the best solution is to specify
1767 that a semicolon ends a range type, and that the index type and element
1768 type of an array are separated by a semicolon, but that if the index
1769 type is a range type, the extra semicolon can be omitted. GDB (at least
1770 through version 4.9) doesn't support any kind of index type other than a
1771 range anyway; I'm not sure about dbx.
1773 It is well established, and widely used, that the type of the index,
1774 unlike most types found in the stabs, is merely a type definition, not
1775 type information (@pxref{String Field}) (that is, it need not start with
1776 @samp{@var{type-number}=} if it is defining a new type). According to a
1777 comment in GDB, this is also true of the type of the array elements; it
1778 gives @samp{ar1;1;10;ar1;1;10;4} as a legitimate way to express a two
1779 dimensional array. According to AIX documentation, the element type
1780 must be type information. GDB accepts either.
1782 The type of the index is often a range type, expressed as the type
1783 descriptor @samp{r} and some parameters. It defines the size of the
1784 array. In the example below, the range @samp{r1;0;2;} defines an index
1785 type which is a subrange of type 1 (integer), with a lower bound of 0
1786 and an upper bound of 2. This defines the valid range of subscripts of
1787 a three-element C array.
1789 For example, the definition:
1792 char char_vec[3] = @{'a','b','c'@};
1796 produces the output:
1799 .stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
1808 If an array is @dfn{packed}, the elements are spaced more
1809 closely than normal, saving memory at the expense of speed. For
1810 example, an array of 3-byte objects might, if unpacked, have each
1811 element aligned on a 4-byte boundary, but if packed, have no padding.
1812 One way to specify that something is packed is with type attributes
1813 (@pxref{String Field}). In the case of arrays, another is to use the
1814 @samp{P} type descriptor instead of @samp{a}. Other than specifying a
1815 packed array, @samp{P} is identical to @samp{a}.
1817 @c FIXME-what is it? A pointer?
1818 An open array is represented by the @samp{A} type descriptor followed by
1819 type information specifying the type of the array elements.
1821 @c FIXME: what is the format of this type? A pointer to a vector of pointers?
1822 An N-dimensional dynamic array is represented by
1825 D @var{dimensions} ; @var{type-information}
1828 @c Does dimensions really have this meaning? The AIX documentation
1830 @var{dimensions} is the number of dimensions; @var{type-information}
1831 specifies the type of the array elements.
1833 @c FIXME: what is the format of this type? A pointer to some offsets in
1835 A subarray of an N-dimensional array is represented by
1838 E @var{dimensions} ; @var{type-information}
1841 @c Does dimensions really have this meaning? The AIX documentation
1843 @var{dimensions} is the number of dimensions; @var{type-information}
1844 specifies the type of the array elements.
1849 Some languages, like C or the original Pascal, do not have string types,
1850 they just have related things like arrays of characters. But most
1851 Pascals and various other languages have string types, which are
1852 indicated as follows:
1855 @item n @var{type-information} ; @var{bytes}
1856 @var{bytes} is the maximum length. I'm not sure what
1857 @var{type-information} is; I suspect that it means that this is a string
1858 of @var{type-information} (thus allowing a string of integers, a string
1859 of wide characters, etc., as well as a string of characters). Not sure
1860 what the format of this type is. This is an AIX feature.
1862 @item z @var{type-information} ; @var{bytes}
1863 Just like @samp{n} except that this is a gstring, not an ordinary
1864 string. I don't know the difference.
1867 Pascal Stringptr. What is this? This is an AIX feature.
1870 Languages, such as CHILL which have a string type which is basically
1871 just an array of characters use the @samp{S} type attribute
1872 (@pxref{String Field}).
1875 @section Enumerations
1877 Enumerations are defined with the @samp{e} type descriptor.
1879 @c FIXME: Where does this information properly go? Perhaps it is
1880 @c redundant with something we already explain.
1881 The source line below declares an enumeration type at file scope.
1882 The type definition is located after the @code{N_RBRAC} that marks the end of
1883 the previous procedure's block scope, and before the @code{N_FUN} that marks
1884 the beginning of the next procedure's block scope. Therefore it does not
1885 describe a block local symbol, but a file local one.
1890 enum e_places @{first,second=3,last@};
1894 generates the following stab:
1897 .stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0
1900 The symbol descriptor (@samp{T}) says that the stab describes a
1901 structure, enumeration, or union tag. The type descriptor @samp{e},
1902 following the @samp{22=} of the type definition narrows it down to an
1903 enumeration type. Following the @samp{e} is a list of the elements of
1904 the enumeration. The format is @samp{@var{name}:@var{value},}. The
1905 list of elements ends with @samp{;}. The fact that @var{value} is
1906 specified as an integer can cause problems if the value is large. GCC
1907 2.5.2 tries to output it in octal in that case with a leading zero,
1908 which is probably a good thing, although GDB 4.11 supports octal only in
1909 cases where decimal is perfectly good. Negative decimal values are
1910 supported by both GDB and dbx.
1912 There is no standard way to specify the size of an enumeration type; it
1913 is determined by the architecture (normally all enumerations types are
1914 32 bits). Type attributes can be used to specify an enumeration type of
1915 another size for debuggers which support them; see @ref{String Field}.
1917 Enumeration types are unusual in that they define symbols for the
1918 enumeration values (@code{first}, @code{second}, and @code{third} in the
1919 above example), and even though these symbols are visible in the file as
1920 a whole (rather than being in a more local namespace like structure
1921 member names), they are defined in the type definition for the
1922 enumeration type rather than each having their own symbol. In order to
1923 be fast, GDB will only get symbols from such types (in its initial scan
1924 of the stabs) if the type is the first thing defined after a @samp{T} or
1925 @samp{t} symbol descriptor (the above example fulfills this
1926 requirement). If the type does not have a name, the compiler should
1927 emit it in a nameless stab (@pxref{String Field}); GCC does this.
1932 The encoding of structures in stabs can be shown with an example.
1934 The following source code declares a structure tag and defines an
1935 instance of the structure in global scope. Then a @code{typedef} equates the
1936 structure tag with a new type. Seperate stabs are generated for the
1937 structure tag, the structure @code{typedef}, and the structure instance. The
1938 stabs for the tag and the @code{typedef} are emited when the definitions are
1939 encountered. Since the structure elements are not initialized, the
1940 stab and code for the structure variable itself is located at the end
1941 of the program in the bss section.
1948 struct s_tag* s_next;
1951 typedef struct s_tag s_typedef;
1954 The structure tag has an @code{N_LSYM} stab type because, like the
1955 enumeration, the symbol has file scope. Like the enumeration, the
1956 symbol descriptor is @samp{T}, for enumeration, structure, or tag type.
1957 The type descriptor @samp{s} following the @samp{16=} of the type
1958 definition narrows the symbol type to structure.
1960 Following the @samp{s} type descriptor is the number of bytes the
1961 structure occupies, followed by a description of each structure element.
1962 The structure element descriptions are of the form @var{name:type, bit
1963 offset from the start of the struct, number of bits in the element}.
1965 @c FIXME: phony line break. Can probably be fixed by using an example
1966 @c with fewer fields.
1969 .stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;
1970 s_char_vec:17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0
1973 In this example, the first two structure elements are previously defined
1974 types. For these, the type following the @samp{@var{name}:} part of the
1975 element description is a simple type reference. The other two structure
1976 elements are new types. In this case there is a type definition
1977 embedded after the @samp{@var{name}:}. The type definition for the
1978 array element looks just like a type definition for a standalone array.
1979 The @code{s_next} field is a pointer to the same kind of structure that
1980 the field is an element of. So the definition of structure type 16
1981 contains a type definition for an element which is a pointer to type 16.
1983 If a field is a static member (this is a C++ feature in which a single
1984 variable appears to be a field of every structure of a given type) it
1985 still starts out with the field name, a colon, and the type, but then
1986 instead of a comma, bit position, comma, and bit size, there is a colon
1987 followed by the name of the variable which each such field refers to.
1989 If the structure has methods (a C++ feature), they follow the non-method
1990 fields; see @ref{Cplusplus}.
1993 @section Giving a Type a Name
1995 To give a type a name, use the @samp{t} symbol descriptor. The type
1996 is specified by the type information (@pxref{String Field}) for the stab.
2000 .stabs "s_typedef:t16",128,0,0,0 # @r{128 is N_LSYM}
2003 specifies that @code{s_typedef} refers to type number 16. Such stabs
2004 have symbol type @code{N_LSYM} (or @code{C_DECL} for XCOFF).
2006 If you are specifying the tag name for a structure, union, or
2007 enumeration, use the @samp{T} symbol descriptor instead. I believe C is
2008 the only language with this feature.
2010 If the type is an opaque type (I believe this is a Modula-2 feature),
2011 AIX provides a type descriptor to specify it. The type descriptor is
2012 @samp{o} and is followed by a name. I don't know what the name
2013 means---is it always the same as the name of the type, or is this type
2014 descriptor used with a nameless stab (@pxref{String Field})? There
2015 optionally follows a comma followed by type information which defines
2016 the type of this type. If omitted, a semicolon is used in place of the
2017 comma and the type information, and the type is much like a generic
2018 pointer type---it has a known size but little else about it is
2032 This code generates a stab for a union tag and a stab for a union
2033 variable. Both use the @code{N_LSYM} stab type. If a union variable is
2034 scoped locally to the procedure in which it is defined, its stab is
2035 located immediately preceding the @code{N_LBRAC} for the procedure's block
2038 The stab for the union tag, however, is located preceding the code for
2039 the procedure in which it is defined. The stab type is @code{N_LSYM}. This
2040 would seem to imply that the union type is file scope, like the struct
2041 type @code{s_tag}. This is not true. The contents and position of the stab
2042 for @code{u_type} do not convey any infomation about its procedure local
2045 @c FIXME: phony line break. Can probably be fixed by using an example
2046 @c with fewer fields.
2049 .stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
2053 The symbol descriptor @samp{T}, following the @samp{name:} means that
2054 the stab describes an enumeration, structure, or union tag. The type
2055 descriptor @samp{u}, following the @samp{23=} of the type definition,
2056 narrows it down to a union type definition. Following the @samp{u} is
2057 the number of bytes in the union. After that is a list of union element
2058 descriptions. Their format is @var{name:type, bit offset into the
2059 union, number of bytes for the element;}.
2061 The stab for the union variable is:
2064 .stabs "an_u:23",128,0,0,-20 # @r{128 is N_LSYM}
2067 @samp{-20} specifies where the variable is stored (@pxref{Stack
2070 @node Function Types
2071 @section Function Types
2073 Various types can be defined for function variables. These types are
2074 not used in defining functions (@pxref{Procedures}); they are used for
2075 things like pointers to functions.
2077 The simple, traditional, type is type descriptor @samp{f} is followed by
2078 type information for the return type of the function, followed by a
2081 This does not deal with functions for which the number and types of the
2082 parameters are part of the type, as in Modula-2 or ANSI C. AIX provides
2083 extensions to specify these, using the @samp{f}, @samp{F}, @samp{p}, and
2084 @samp{R} type descriptors.
2086 First comes the type descriptor. If it is @samp{f} or @samp{F}, this
2087 type involves a function rather than a procedure, and the type
2088 information for the return type of the function follows, followed by a
2089 comma. Then comes the number of parameters to the function and a
2090 semicolon. Then, for each parameter, there is the name of the parameter
2091 followed by a colon (this is only present for type descriptors @samp{R}
2092 and @samp{F} which represent Pascal function or procedure parameters),
2093 type information for the parameter, a comma, 0 if passed by reference or
2094 1 if passed by value, and a semicolon. The type definition ends with a
2097 For example, this variable definition:
2104 generates the following code:
2107 .stabs "g_pf:G24=*25=f1",32,0,0,0
2108 .common _g_pf,4,"bss"
2111 The variable defines a new type, 24, which is a pointer to another new
2112 type, 25, which is a function returning @code{int}.
2115 @chapter Symbol Information in Symbol Tables
2117 This chapter describes the format of symbol table entries
2118 and how stab assembler directives map to them. It also describes the
2119 transformations that the assembler and linker make on data from stabs.
2122 * Symbol Table Format::
2123 * Transformations On Symbol Tables::
2126 @node Symbol Table Format
2127 @section Symbol Table Format
2129 Each time the assembler encounters a stab directive, it puts
2130 each field of the stab into a corresponding field in a symbol table
2131 entry of its output file. If the stab contains a string field, the
2132 symbol table entry for that stab points to a string table entry
2133 containing the string data from the stab. Assembler labels become
2134 relocatable addresses. Symbol table entries in a.out have the format:
2136 @c FIXME: should refer to external, not internal.
2138 struct internal_nlist @{
2139 unsigned long n_strx; /* index into string table of name */
2140 unsigned char n_type; /* type of symbol */
2141 unsigned char n_other; /* misc info (usually empty) */
2142 unsigned short n_desc; /* description field */
2143 bfd_vma n_value; /* value of symbol */
2147 If the stab has a string, the @code{n_strx} field holds the offset in
2148 bytes of the string within the string table. The string is terminated
2149 by a NUL character. If the stab lacks a string (for example, it was
2150 produced by a @code{.stabn} or @code{.stabd} directive), the
2151 @code{n_strx} field is zero.
2153 Symbol table entries with @code{n_type} field values greater than 0x1f
2154 originated as stabs generated by the compiler (with one random
2155 exception). The other entries were placed in the symbol table of the
2156 executable by the assembler or the linker.
2158 @node Transformations On Symbol Tables
2159 @section Transformations on Symbol Tables
2161 The linker concatenates object files and does fixups of externally
2164 You can see the transformations made on stab data by the assembler and
2165 linker by examining the symbol table after each pass of the build. To
2166 do this, use @samp{nm -ap}, which dumps the symbol table, including
2167 debugging information, unsorted. For stab entries the columns are:
2168 @var{value}, @var{other}, @var{desc}, @var{type}, @var{string}. For
2169 assembler and linker symbols, the columns are: @var{value}, @var{type},
2172 The low 5 bits of the stab type tell the linker how to relocate the
2173 value of the stab. Thus for stab types like @code{N_RSYM} and
2174 @code{N_LSYM}, where the value is an offset or a register number, the
2175 low 5 bits are @code{N_ABS}, which tells the linker not to relocate the
2178 Where the value of a stab contains an assembly language label,
2179 it is transformed by each build step. The assembler turns it into a
2180 relocatable address and the linker turns it into an absolute address.
2183 * Transformations On Static Variables::
2184 * Transformations On Global Variables::
2185 * Stab Section Transformations:: For some object file formats,
2186 things are a bit different.
2189 @node Transformations On Static Variables
2190 @subsection Transformations on Static Variables
2192 This source line defines a static variable at file scope:
2195 static int s_g_repeat
2199 The following stab describes the symbol:
2202 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
2206 The assembler transforms the stab into this symbol table entry in the
2207 @file{.o} file. The location is expressed as a data segment offset.
2210 00000084 - 00 0000 STSYM s_g_repeat:S1
2214 In the symbol table entry from the executable, the linker has made the
2215 relocatable address absolute.
2218 0000e00c - 00 0000 STSYM s_g_repeat:S1
2221 @node Transformations On Global Variables
2222 @subsection Transformations on Global Variables
2224 Stabs for global variables do not contain location information. In
2225 this case, the debugger finds location information in the assembler or
2226 linker symbol table entry describing the variable. The source line:
2236 .stabs "g_foo:G2",32,0,0,0
2239 The variable is represented by two symbol table entries in the object
2240 file (see below). The first one originated as a stab. The second one
2241 is an external symbol. The upper case @samp{D} signifies that the
2242 @code{n_type} field of the symbol table contains 7, @code{N_DATA} with
2243 local linkage. The stab's value is zero since the value is not used for
2244 @code{N_GSYM} stabs. The value of the linker symbol is the relocatable
2245 address corresponding to the variable.
2248 00000000 - 00 0000 GSYM g_foo:G2
2253 These entries as transformed by the linker. The linker symbol table
2254 entry now holds an absolute address:
2257 00000000 - 00 0000 GSYM g_foo:G2
2262 @node Stab Section Transformations
2263 @subsection Transformations of Stabs in separate sections
2265 For object file formats using stabs in separate sections (@pxref{Stab
2266 Sections}), use @code{objdump --stabs} instead of @code{nm} to show the
2267 stabs in an object or executable file. @code{objdump} is a GNU utility;
2268 Sun does not provide any equivalent.
2270 The following example is for a stab whose value is an address is
2271 relative to the compilation unit (@pxref{ELF Linker Relocation}). For
2272 example, if the source line
2278 appears within a function, then the assembly language output from the
2284 .stabs "ld:V(0,3)",0x26,0,4,.L18-Ddata.data # @r{0x26 is N_STSYM}
2291 Because the value is formed by subtracting one symbol from another, the
2292 value is absolute, not relocatable, and so the object file contains
2295 Symnum n_type n_othr n_desc n_value n_strx String
2296 31 STSYM 0 4 00000004 680 ld:V(0,3)
2299 without any relocations, and the executable file also contains
2302 Symnum n_type n_othr n_desc n_value n_strx String
2303 31 STSYM 0 4 00000004 680 ld:V(0,3)
2307 @chapter GNU C++ Stabs
2310 * Class Names:: C++ class names are both tags and typedefs.
2311 * Nested Symbols:: C++ symbol names can be within other types.
2312 * Basic Cplusplus Types::
2315 * Methods:: Method definition
2317 * Method Modifiers::
2320 * Virtual Base Classes::
2324 Type descriptors added for C++ descriptions:
2328 method type (@code{##} if minimal debug)
2331 Member (class and variable) type. It is followed by type information
2332 for the offset basetype, a comma, and type information for the type of
2333 the field being pointed to. (FIXME: this is acknowledged to be
2334 gibberish. Can anyone say what really goes here?).
2336 Note that there is a conflict between this and type attributes
2337 (@pxref{String Field}); both use type descriptor @samp{@@}.
2338 Fortunately, the @samp{@@} type descriptor used in this C++ sense always
2339 will be followed by a digit, @samp{(}, or @samp{-}, and type attributes
2340 never start with those things.
2344 @section C++ Class Names
2346 In C++, a class name which is declared with @code{class}, @code{struct},
2347 or @code{union}, is not only a tag, as in C, but also a type name. Thus
2348 there should be stabs with both @samp{t} and @samp{T} symbol descriptors
2351 To save space, there is a special abbreviation for this case. If the
2352 @samp{T} symbol descriptor is followed by @samp{t}, then the stab
2353 defines both a type name and a tag.
2355 For example, the C++ code
2358 struct foo @{int x;@};
2361 can be represented as either
2364 .stabs "foo:T19=s4x:1,0,32;;",128,0,0,0 # @r{128 is N_LSYM}
2365 .stabs "foo:t19",128,0,0,0
2371 .stabs "foo:Tt19=s4x:1,0,32;;",128,0,0,0
2374 @node Nested Symbols
2375 @section Defining a Symbol Within Another Type
2377 In C++, a symbol (such as a type name) can be defined within another type.
2378 @c FIXME: Needs example.
2380 In stabs, this is sometimes represented by making the name of a symbol
2381 which contains @samp{::}. Such a pair of colons does not end the name
2382 of the symbol, the way a single colon would (@pxref{String Field}). I'm
2383 not sure how consistently used or well thought out this mechanism is.
2384 So that a pair of colons in this position always has this meaning,
2385 @samp{:} cannot be used as a symbol descriptor.
2387 For example, if the string for a stab is @samp{foo::bar::baz:t5=*6},
2388 then @code{foo::bar::baz} is the name of the symbol, @samp{t} is the
2389 symbol descriptor, and @samp{5=*6} is the type information.
2391 @node Basic Cplusplus Types
2392 @section Basic Types For C++
2394 << the examples that follow are based on a01.C >>
2397 C++ adds two more builtin types to the set defined for C. These are
2398 the unknown type and the vtable record type. The unknown type, type
2399 16, is defined in terms of itself like the void type.
2401 The vtable record type, type 17, is defined as a structure type and
2402 then as a structure tag. The structure has four fields: delta, index,
2403 pfn, and delta2. pfn is the function pointer.
2405 << In boilerplate $vtbl_ptr_type, what are the fields delta,
2406 index, and delta2 used for? >>
2408 This basic type is present in all C++ programs even if there are no
2409 virtual methods defined.
2412 .stabs "struct_name:sym_desc(type)type_def(17)=type_desc(struct)struct_bytes(8)
2413 elem_name(delta):type_ref(short int),bit_offset(0),field_bits(16);
2414 elem_name(index):type_ref(short int),bit_offset(16),field_bits(16);
2415 elem_name(pfn):type_def(18)=type_desc(ptr to)type_ref(void),
2416 bit_offset(32),field_bits(32);
2417 elem_name(delta2):type_def(short int);bit_offset(32),field_bits(16);;"
2422 .stabs "$vtbl_ptr_type:t17=s8
2423 delta:6,0,16;index:6,16,16;pfn:18=*15,32,32;delta2:6,32,16;;"
2428 .stabs "name:sym_dec(struct tag)type_ref($vtbl_ptr_type)",N_LSYM,NIL,NIL,NIL
2432 .stabs "$vtbl_ptr_type:T17",128,0,0,0
2435 @node Simple Classes
2436 @section Simple Class Definition
2438 The stabs describing C++ language features are an extension of the
2439 stabs describing C. Stabs representing C++ class types elaborate
2440 extensively on the stab format used to describe structure types in C.
2441 Stabs representing class type variables look just like stabs
2442 representing C language variables.
2444 Consider the following very simple class definition.
2450 int Ameth(int in, char other);
2454 The class @code{baseA} is represented by two stabs. The first stab describes
2455 the class as a structure type. The second stab describes a structure
2456 tag of the class type. Both stabs are of stab type @code{N_LSYM}. Since the
2457 stab is not located between an @code{N_FUN} and an @code{N_LBRAC} stab this indicates
2458 that the class is defined at file scope. If it were, then the @code{N_LSYM}
2459 would signify a local variable.
2461 A stab describing a C++ class type is similar in format to a stab
2462 describing a C struct, with each class member shown as a field in the
2463 structure. The part of the struct format describing fields is
2464 expanded to include extra information relevent to C++ class members.
2465 In addition, if the class has multiple base classes or virtual
2466 functions the struct format outside of the field parts is also
2469 In this simple example the field part of the C++ class stab
2470 representing member data looks just like the field part of a C struct
2471 stab. The section on protections describes how its format is
2472 sometimes extended for member data.
2474 The field part of a C++ class stab representing a member function
2475 differs substantially from the field part of a C struct stab. It
2476 still begins with @samp{name:} but then goes on to define a new type number
2477 for the member function, describe its return type, its argument types,
2478 its protection level, any qualifiers applied to the method definition,
2479 and whether the method is virtual or not. If the method is virtual
2480 then the method description goes on to give the vtable index of the
2481 method, and the type number of the first base class defining the
2484 When the field name is a method name it is followed by two colons rather
2485 than one. This is followed by a new type definition for the method.
2486 This is a number followed by an equal sign and the type descriptor
2487 @samp{#}, indicating a method type, and a second @samp{#}, indicating
2488 that this is the @dfn{minimal} type of method definition used by GCC2,
2489 not larger method definitions used by earlier versions of GCC. This is
2490 followed by a type reference showing the return type of the method and a
2493 The format of an overloaded operator method name differs from that of
2494 other methods. It is @samp{op$::@var{operator-name}.} where
2495 @var{operator-name} is the operator name such as @samp{+} or @samp{+=}.
2496 The name ends with a period, and any characters except the period can
2497 occur in the @var{operator-name} string.
2499 The next part of the method description represents the arguments to the
2500 method, preceeded by a colon and ending with a semi-colon. The types of
2501 the arguments are expressed in the same way argument types are expressed
2502 in C++ name mangling. In this example an @code{int} and a @code{char}
2505 This is followed by a number, a letter, and an asterisk or period,
2506 followed by another semicolon. The number indicates the protections
2507 that apply to the member function. Here the 2 means public. The
2508 letter encodes any qualifier applied to the method definition. In
2509 this case, @samp{A} means that it is a normal function definition. The dot
2510 shows that the method is not virtual. The sections that follow
2511 elaborate further on these fields and describe the additional
2512 information present for virtual methods.
2516 .stabs "class_name:sym_desc(type)type_def(20)=type_desc(struct)struct_bytes(4)
2517 field_name(Adat):type(int),bit_offset(0),field_bits(32);
2519 method_name(Ameth)::type_def(21)=type_desc(method)return_type(int);
2520 :arg_types(int char);
2521 protection(public)qualifier(normal)virtual(no);;"
2526 .stabs "baseA:t20=s4Adat:1,0,32;Ameth::21=##1;:ic;2A.;;",128,0,0,0
2528 .stabs "class_name:sym_desc(struct tag)",N_LSYM,NIL,NIL,NIL
2530 .stabs "baseA:T20",128,0,0,0
2533 @node Class Instance
2534 @section Class Instance
2536 As shown above, describing even a simple C++ class definition is
2537 accomplished by massively extending the stab format used in C to
2538 describe structure types. However, once the class is defined, C stabs
2539 with no modifications can be used to describe class instances. The
2549 yields the following stab describing the class instance. It looks no
2550 different from a standard C stab describing a local variable.
2553 .stabs "name:type_ref(baseA)", N_LSYM, NIL, NIL, frame_ptr_offset
2557 .stabs "AbaseA:20",128,0,0,-20
2561 @section Method Definition
2563 The class definition shown above declares Ameth. The C++ source below
2568 baseA::Ameth(int in, char other)
2575 This method definition yields three stabs following the code of the
2576 method. One stab describes the method itself and following two describe
2577 its parameters. Although there is only one formal argument all methods
2578 have an implicit argument which is the @code{this} pointer. The @code{this}
2579 pointer is a pointer to the object on which the method was called. Note
2580 that the method name is mangled to encode the class name and argument
2581 types. Name mangling is described in the @sc{arm} (@cite{The Annotated
2582 C++ Reference Manual}, by Ellis and Stroustrup, @sc{isbn}
2583 0-201-51459-1); @file{gpcompare.texi} in Cygnus GCC distributions
2584 describes the differences between GNU mangling and @sc{arm}
2586 @c FIXME: Use @xref, especially if this is generally installed in the
2588 @c FIXME: This information should be in a net release, either of GCC or
2589 @c GDB. But gpcompare.texi doesn't seem to be in the FSF GCC.
2592 .stabs "name:symbol_desriptor(global function)return_type(int)",
2593 N_FUN, NIL, NIL, code_addr_of_method_start
2595 .stabs "Ameth__5baseAic:F1",36,0,0,_Ameth__5baseAic
2598 Here is the stab for the @code{this} pointer implicit argument. The
2599 name of the @code{this} pointer is always @code{this}. Type 19, the
2600 @code{this} pointer is defined as a pointer to type 20, @code{baseA},
2601 but a stab defining @code{baseA} has not yet been emited. Since the
2602 compiler knows it will be emited shortly, here it just outputs a cross
2603 reference to the undefined symbol, by prefixing the symbol name with
2607 .stabs "name:sym_desc(register param)type_def(19)=
2608 type_desc(ptr to)type_ref(baseA)=
2609 type_desc(cross-reference to)baseA:",N_RSYM,NIL,NIL,register_number
2611 .stabs "this:P19=*20=xsbaseA:",64,0,0,8
2614 The stab for the explicit integer argument looks just like a parameter
2615 to a C function. The last field of the stab is the offset from the
2616 argument pointer, which in most systems is the same as the frame
2620 .stabs "name:sym_desc(value parameter)type_ref(int)",
2621 N_PSYM,NIL,NIL,offset_from_arg_ptr
2623 .stabs "in:p1",160,0,0,72
2626 << The examples that follow are based on A1.C >>
2629 @section Protections
2631 In the simple class definition shown above all member data and
2632 functions were publicly accessable. The example that follows
2633 contrasts public, protected and privately accessable fields and shows
2634 how these protections are encoded in C++ stabs.
2636 If the character following the @samp{@var{field-name}:} part of the
2637 string is @samp{/}, then the next character is the visibility. @samp{0}
2638 means private, @samp{1} means protected, and @samp{2} means public.
2639 Debuggers should ignore visibility characters they do not recognize, and
2640 assume a reasonable default (such as public) (GDB 4.11 does not, but
2641 this should be fixed in the next GDB release). If no visibility is
2642 specified the field is public. The visibility @samp{9} means that the
2643 field has been optimized out and is public (there is no way to specify
2644 an optimized out field with a private or protected visibility).
2645 Visibility @samp{9} is not supported by GDB 4.11; this should be fixed
2646 in the next GDB release.
2648 The following C++ source:
2662 generates the following stab:
2666 .stabs "vis:T19=s12priv:/01,0,32;prot:/12,32,8;pub:12,64,32;;",128,0,0,0
2669 @samp{vis:T19=s12} indicates that type number 19 is a 12 byte structure
2670 named @code{vis} The @code{priv} field has public visibility
2671 (@samp{/0}), type int (@samp{1}), and offset and size @samp{,0,32;}.
2672 The @code{prot} field has protected visibility (@samp{/1}), type char
2673 (@samp{2}) and offset and size @samp{,32,8;}. The @code{pub} field has
2674 type float (@samp{12}), and offset and size @samp{,64,32;}.
2676 Protections for member functions are signified by one digit embeded in
2677 the field part of the stab describing the method. The digit is 0 if
2678 private, 1 if protected and 2 if public. Consider the C++ class
2682 class all_methods @{
2684 int priv_meth(int in)@{return in;@};
2686 char protMeth(char in)@{return in;@};
2688 float pubMeth(float in)@{return in;@};
2692 It generates the following stab. The digit in question is to the left
2693 of an @samp{A} in each case. Notice also that in this case two symbol
2694 descriptors apply to the class name struct tag and struct type.
2697 .stabs "class_name:sym_desc(struct tag&type)type_def(21)=
2698 sym_desc(struct)struct_bytes(1)
2699 meth_name::type_def(22)=sym_desc(method)returning(int);
2700 :args(int);protection(private)modifier(normal)virtual(no);
2701 meth_name::type_def(23)=sym_desc(method)returning(char);
2702 :args(char);protection(protected)modifier(normal)virual(no);
2703 meth_name::type_def(24)=sym_desc(method)returning(float);
2704 :args(float);protection(public)modifier(normal)virtual(no);;",
2709 .stabs "all_methods:Tt21=s1priv_meth::22=##1;:i;0A.;protMeth::23=##2;:c;1A.;
2710 pubMeth::24=##12;:f;2A.;;",128,0,0,0
2713 @node Method Modifiers
2714 @section Method Modifiers (@code{const}, @code{volatile}, @code{const volatile})
2718 In the class example described above all the methods have the normal
2719 modifier. This method modifier information is located just after the
2720 protection information for the method. This field has four possible
2721 character values. Normal methods use @samp{A}, const methods use
2722 @samp{B}, volatile methods use @samp{C}, and const volatile methods use
2723 @samp{D}. Consider the class definition below:
2728 int ConstMeth (int arg) const @{ return arg; @};
2729 char VolatileMeth (char arg) volatile @{ return arg; @};
2730 float ConstVolMeth (float arg) const volatile @{return arg; @};
2734 This class is described by the following stab:
2737 .stabs "class(A):sym_desc(struct)type_def(20)=type_desc(struct)struct_bytes(1)
2738 meth_name(ConstMeth)::type_def(21)sym_desc(method)
2739 returning(int);:arg(int);protection(public)modifier(const)virtual(no);
2740 meth_name(VolatileMeth)::type_def(22)=sym_desc(method)
2741 returning(char);:arg(char);protection(public)modifier(volatile)virt(no)
2742 meth_name(ConstVolMeth)::type_def(23)=sym_desc(method)
2743 returning(float);:arg(float);protection(public)modifer(const volatile)
2744 virtual(no);;", @dots{}
2748 .stabs "A:T20=s1ConstMeth::21=##1;:i;2B.;VolatileMeth::22=##2;:c;2C.;
2749 ConstVolMeth::23=##12;:f;2D.;;",128,0,0,0
2752 @node Virtual Methods
2753 @section Virtual Methods
2755 << The following examples are based on a4.C >>
2757 The presence of virtual methods in a class definition adds additional
2758 data to the class description. The extra data is appended to the
2759 description of the virtual method and to the end of the class
2760 description. Consider the class definition below:
2766 virtual int A_virt (int arg) @{ return arg; @};
2770 This results in the stab below describing class A. It defines a new
2771 type (20) which is an 8 byte structure. The first field of the class
2772 struct is @samp{Adat}, an integer, starting at structure offset 0 and
2775 The second field in the class struct is not explicitly defined by the
2776 C++ class definition but is implied by the fact that the class
2777 contains a virtual method. This field is the vtable pointer. The
2778 name of the vtable pointer field starts with @samp{$vf} and continues with a
2779 type reference to the class it is part of. In this example the type
2780 reference for class A is 20 so the name of its vtable pointer field is
2781 @samp{$vf20}, followed by the usual colon.
2783 Next there is a type definition for the vtable pointer type (21).
2784 This is in turn defined as a pointer to another new type (22).
2786 Type 22 is the vtable itself, which is defined as an array, indexed by
2787 a range of integers between 0 and 1, and whose elements are of type
2788 17. Type 17 was the vtable record type defined by the boilerplate C++
2789 type definitions, as shown earlier.
2791 The bit offset of the vtable pointer field is 32. The number of bits
2792 in the field are not specified when the field is a vtable pointer.
2794 Next is the method definition for the virtual member function @code{A_virt}.
2795 Its description starts out using the same format as the non-virtual
2796 member functions described above, except instead of a dot after the
2797 @samp{A} there is an asterisk, indicating that the function is virtual.
2798 Since is is virtual some addition information is appended to the end
2799 of the method description.
2801 The first number represents the vtable index of the method. This is a
2802 32 bit unsigned number with the high bit set, followed by a
2805 The second number is a type reference to the first base class in the
2806 inheritence hierarchy defining the virtual member function. In this
2807 case the class stab describes a base class so the virtual function is
2808 not overriding any other definition of the method. Therefore the
2809 reference is to the type number of the class that the stab is
2812 This is followed by three semi-colons. One marks the end of the
2813 current sub-section, one marks the end of the method field, and the
2814 third marks the end of the struct definition.
2816 For classes containing virtual functions the very last section of the
2817 string part of the stab holds a type reference to the first base
2818 class. This is preceeded by @samp{~%} and followed by a final semi-colon.
2821 .stabs "class_name(A):type_def(20)=sym_desc(struct)struct_bytes(8)
2822 field_name(Adat):type_ref(int),bit_offset(0),field_bits(32);
2823 field_name(A virt func ptr):type_def(21)=type_desc(ptr to)type_def(22)=
2824 sym_desc(array)index_type_ref(range of int from 0 to 1);
2825 elem_type_ref(vtbl elem type),
2827 meth_name(A_virt)::typedef(23)=sym_desc(method)returning(int);
2828 :arg_type(int),protection(public)normal(yes)virtual(yes)
2829 vtable_index(1);class_first_defining(A);;;~%first_base(A);",
2833 @c FIXME: bogus line break.
2835 .stabs "A:t20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2836 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2840 @section Inheritence
2842 Stabs describing C++ derived classes include additional sections that
2843 describe the inheritence hierarchy of the class. A derived class stab
2844 also encodes the number of base classes. For each base class it tells
2845 if the base class is virtual or not, and if the inheritence is private
2846 or public. It also gives the offset into the object of the portion of
2847 the object corresponding to each base class.
2849 This additional information is embeded in the class stab following the
2850 number of bytes in the struct. First the number of base classes
2851 appears bracketed by an exclamation point and a comma.
2853 Then for each base type there repeats a series: a virtual character, a
2854 visibilty character, a number, a comma, another number, and a
2857 The virtual character is @samp{1} if the base class is virtual and
2858 @samp{0} if not. The visibility character is @samp{2} if the derivation
2859 is public, @samp{1} if it is protected, and @samp{0} if it is private.
2860 Debuggers should ignore virtual or visibility characters they do not
2861 recognize, and assume a reasonable default (such as public and
2862 non-virtual) (GDB 4.11 does not, but this should be fixed in the next
2865 The number following the virtual and visibility characters is the offset
2866 from the start of the object to the part of the object pertaining to the
2869 After the comma, the second number is a type_descriptor for the base
2870 type. Finally a semi-colon ends the series, which repeats for each
2873 The source below defines three base classes @code{A}, @code{B}, and
2874 @code{C} and the derived class @code{D}.
2881 virtual int A_virt (int arg) @{ return arg; @};
2887 virtual int B_virt (int arg) @{return arg; @};
2893 virtual int C_virt (int arg) @{return arg; @};
2896 class D : A, virtual B, public C @{
2899 virtual int A_virt (int arg ) @{ return arg+1; @};
2900 virtual int B_virt (int arg) @{ return arg+2; @};
2901 virtual int C_virt (int arg) @{ return arg+3; @};
2902 virtual int D_virt (int arg) @{ return arg; @};
2906 Class stabs similar to the ones described earlier are generated for
2909 @c FIXME!!! the linebreaks in the following example probably make the
2910 @c examples literally unusable, but I don't know any other way to get
2911 @c them on the page.
2912 @c One solution would be to put some of the type definitions into
2913 @c separate stabs, even if that's not exactly what the compiler actually
2916 .stabs "A:T20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2917 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2919 .stabs "B:Tt25=s8Bdat:1,0,32;$vf25:21,32;B_virt::26=##1;
2920 :i;2A*-2147483647;25;;;~%25;",128,0,0,0
2922 .stabs "C:Tt28=s8Cdat:1,0,32;$vf28:21,32;C_virt::29=##1;
2923 :i;2A*-2147483647;28;;;~%28;",128,0,0,0
2926 In the stab describing derived class @code{D} below, the information about
2927 the derivation of this class is encoded as follows.
2930 .stabs "derived_class_name:symbol_descriptors(struct tag&type)=
2931 type_descriptor(struct)struct_bytes(32)!num_bases(3),
2932 base_virtual(no)inheritence_public(no)base_offset(0),
2933 base_class_type_ref(A);
2934 base_virtual(yes)inheritence_public(no)base_offset(NIL),
2935 base_class_type_ref(B);
2936 base_virtual(no)inheritence_public(yes)base_offset(64),
2937 base_class_type_ref(C); @dots{}
2940 @c FIXME! fake linebreaks.
2942 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:
2943 1,160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt:
2944 :32:i;2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;
2945 28;;D_virt::32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2948 @node Virtual Base Classes
2949 @section Virtual Base Classes
2951 A derived class object consists of a concatination in memory of the data
2952 areas defined by each base class, starting with the leftmost and ending
2953 with the rightmost in the list of base classes. The exception to this
2954 rule is for virtual inheritence. In the example above, class @code{D}
2955 inherits virtually from base class @code{B}. This means that an
2956 instance of a @code{D} object will not contain its own @code{B} part but
2957 merely a pointer to a @code{B} part, known as a virtual base pointer.
2959 In a derived class stab, the base offset part of the derivation
2960 information, described above, shows how the base class parts are
2961 ordered. The base offset for a virtual base class is always given as 0.
2962 Notice that the base offset for @code{B} is given as 0 even though
2963 @code{B} is not the first base class. The first base class @code{A}
2966 The field information part of the stab for class @code{D} describes the field
2967 which is the pointer to the virtual base class @code{B}. The vbase pointer
2968 name is @samp{$vb} followed by a type reference to the virtual base class.
2969 Since the type id for @code{B} in this example is 25, the vbase pointer name
2972 @c FIXME!! fake linebreaks below
2974 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:1,
2975 160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt::32:i;
2976 2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;28;;D_virt:
2977 :32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2980 Following the name and a semicolon is a type reference describing the
2981 type of the virtual base class pointer, in this case 24. Type 24 was
2982 defined earlier as the type of the @code{B} class @code{this} pointer. The
2983 @code{this} pointer for a class is a pointer to the class type.
2986 .stabs "this:P24=*25=xsB:",64,0,0,8
2989 Finally the field offset part of the vbase pointer field description
2990 shows that the vbase pointer is the first field in the @code{D} object,
2991 before any data fields defined by the class. The layout of a @code{D}
2992 class object is a follows, @code{Adat} at 0, the vtable pointer for
2993 @code{A} at 32, @code{Cdat} at 64, the vtable pointer for C at 96, the
2994 virtual base pointer for @code{B} at 128, and @code{Ddat} at 160.
2997 @node Static Members
2998 @section Static Members
3000 The data area for a class is a concatenation of the space used by the
3001 data members of the class. If the class has virtual methods, a vtable
3002 pointer follows the class data. The field offset part of each field
3003 description in the class stab shows this ordering.
3005 << How is this reflected in stabs? See Cygnus bug #677 for some info. >>
3008 @appendix Table of Stab Types
3010 The following are all the possible values for the stab type field, for
3011 @code{a.out} files, in numeric order. This does not apply to XCOFF, but
3012 it does apply to stabs in sections (@pxref{Stab Sections}). Stabs in
3013 ECOFF use these values but add 0x8f300 to distinguish them from non-stab
3016 The symbolic names are defined in the file @file{include/aout/stabs.def}.
3019 * Non-Stab Symbol Types:: Types from 0 to 0x1f
3020 * Stab Symbol Types:: Types from 0x20 to 0xff
3023 @node Non-Stab Symbol Types
3024 @appendixsec Non-Stab Symbol Types
3026 The following types are used by the linker and assembler, not by stab
3027 directives. Since this document does not attempt to describe aspects of
3028 object file format other than the debugging format, no details are
3031 @c Try to get most of these to fit on a single line.
3041 File scope absolute symbol
3043 @item 0x3 N_ABS | N_EXT
3044 External absolute symbol
3047 File scope text symbol
3049 @item 0x5 N_TEXT | N_EXT
3050 External text symbol
3053 File scope data symbol
3055 @item 0x7 N_DATA | N_EXT
3056 External data symbol
3059 File scope BSS symbol
3061 @item 0x9 N_BSS | N_EXT
3065 Same as @code{N_FN}, for Sequent compilers
3068 Symbol is indirected to another symbol
3071 Common---visible after shared library dynamic link
3074 @itemx 0x15 N_SETA | N_EXT
3075 Absolute set element
3078 @itemx 0x17 N_SETT | N_EXT
3079 Text segment set element
3082 @itemx 0x19 N_SETD | N_EXT
3083 Data segment set element
3086 @itemx 0x1b N_SETB | N_EXT
3087 BSS segment set element
3090 @itemx 0x1d N_SETV | N_EXT
3091 Pointer to set vector
3093 @item 0x1e N_WARNING
3094 Print a warning message during linking
3097 File name of a @file{.o} file
3100 @node Stab Symbol Types
3101 @appendixsec Stab Symbol Types
3103 The following symbol types indicate that this is a stab. This is the
3104 full list of stab numbers, including stab types that are used in
3105 languages other than C.
3109 Global symbol; see @ref{Global Variables}.
3112 Function name (for BSD Fortran); see @ref{Procedures}.
3115 Function name (@pxref{Procedures}) or text segment variable
3119 Data segment file-scope variable; see @ref{Statics}.
3122 BSS segment file-scope variable; see @ref{Statics}.
3125 Name of main routine; see @ref{Main Program}.
3128 Variable in @code{.rodata} section; see @ref{Statics}.
3131 Global symbol (for Pascal); see @ref{N_PC}.
3134 Number of symbols (according to Ultrix V4.0); see @ref{N_NSYMS}.
3137 No DST map; see @ref{N_NOMAP}.
3139 @c FIXME: describe this solaris feature in the body of the text (see
3140 @c comments in include/aout/stab.def).
3142 Object file (Solaris2).
3144 @c See include/aout/stab.def for (a little) more info.
3146 Debugger options (Solaris2).
3149 Register variable; see @ref{Register Variables}.
3152 Modula-2 compilation unit; see @ref{N_M2C}.
3155 Line number in text segment; see @ref{Line Numbers}.
3158 Line number in data segment; see @ref{Line Numbers}.
3161 Line number in bss segment; see @ref{Line Numbers}.
3164 Sun source code browser, path to @file{.cb} file; see @ref{N_BROWS}.
3167 GNU Modula2 definition module dependency; see @ref{N_DEFD}.
3170 Function start/body/end line numbers (Solaris2).
3173 GNU C++ exception variable; see @ref{N_EHDECL}.
3176 Modula2 info "for imc" (according to Ultrix V4.0); see @ref{N_MOD2}.
3179 GNU C++ @code{catch} clause; see @ref{N_CATCH}.
3182 Structure of union element; see @ref{N_SSYM}.
3185 Last stab for module (Solaris2).
3188 Path and name of source file; see @ref{Source Files}.
3191 Stack variable (@pxref{Stack Variables}) or type (@pxref{Typedefs}).
3194 Beginning of an include file (Sun only); see @ref{Include Files}.
3197 Name of include file; see @ref{Include Files}.
3200 Parameter variable; see @ref{Parameters}.
3203 End of an include file; see @ref{Include Files}.
3206 Alternate entry point; see @ref{Alternate Entry Points}.
3209 Beginning of a lexical block; see @ref{Block Structure}.
3212 Place holder for a deleted include file; see @ref{Include Files}.
3215 Modula2 scope information (Sun linker); see @ref{N_SCOPE}.
3218 End of a lexical block; see @ref{Block Structure}.
3221 Begin named common block; see @ref{Common Blocks}.
3224 End named common block; see @ref{Common Blocks}.
3227 Member of a common block; see @ref{Common Blocks}.
3229 @c FIXME: How does this really work? Move it to main body of document.
3231 Pascal @code{with} statement: type,,0,0,offset (Solaris2).
3234 Gould non-base registers; see @ref{Gould}.
3237 Gould non-base registers; see @ref{Gould}.
3240 Gould non-base registers; see @ref{Gould}.
3243 Gould non-base registers; see @ref{Gould}.
3246 Gould non-base registers; see @ref{Gould}.
3249 @c Restore the default table indent
3254 @node Symbol Descriptors
3255 @appendix Table of Symbol Descriptors
3257 The symbol descriptor is the character which follows the colon in many
3258 stabs, and which tells what kind of stab it is. @xref{String Field},
3259 for more information about their use.
3261 @c Please keep this alphabetical
3263 @c In TeX, this looks great, digit is in italics. But makeinfo insists
3264 @c on putting it in `', not realizing that @var should override @code.
3265 @c I don't know of any way to make makeinfo do the right thing. Seems
3266 @c like a makeinfo bug to me.
3270 Variable on the stack; see @ref{Stack Variables}.
3273 C++ nested symbol; see @xref{Nested Symbols}
3276 Parameter passed by reference in register; see @ref{Reference Parameters}.
3279 Based variable; see @ref{Based Variables}.
3282 Constant; see @ref{Constants}.
3285 Conformant array bound (Pascal, maybe other languages); @ref{Conformant
3286 Arrays}. Name of a caught exception (GNU C++). These can be
3287 distinguished because the latter uses @code{N_CATCH} and the former uses
3288 another symbol type.
3291 Floating point register variable; see @ref{Register Variables}.
3294 Parameter in floating point register; see @ref{Register Parameters}.
3297 File scope function; see @ref{Procedures}.
3300 Global function; see @ref{Procedures}.
3303 Global variable; see @ref{Global Variables}.
3306 @xref{Register Parameters}.
3309 Internal (nested) procedure; see @ref{Nested Procedures}.
3312 Internal (nested) function; see @ref{Nested Procedures}.
3315 Label name (documented by AIX, no further information known).
3318 Module; see @ref{Procedures}.
3321 Argument list parameter; see @ref{Parameters}.
3327 Fortran Function parameter; see @ref{Parameters}.
3330 Unfortunately, three separate meanings have been independently invented
3331 for this symbol descriptor. At least the GNU and Sun uses can be
3332 distinguished by the symbol type. Global Procedure (AIX) (symbol type
3333 used unknown); see @ref{Procedures}. Register parameter (GNU) (symbol
3334 type @code{N_PSYM}); see @ref{Parameters}. Prototype of function
3335 referenced by this file (Sun @code{acc}) (symbol type @code{N_FUN}).
3338 Static Procedure; see @ref{Procedures}.
3341 Register parameter; see @ref{Register Parameters}.
3344 Register variable; see @ref{Register Variables}.
3347 File scope variable; see @ref{Statics}.
3350 Local variable (OS9000).
3353 Type name; see @ref{Typedefs}.
3356 Enumeration, structure, or union tag; see @ref{Typedefs}.
3359 Parameter passed by reference; see @ref{Reference Parameters}.
3362 Procedure scope static variable; see @ref{Statics}.
3365 Conformant array; see @ref{Conformant Arrays}.
3368 Function return variable; see @ref{Parameters}.
3371 @node Type Descriptors
3372 @appendix Table of Type Descriptors
3374 The type descriptor is the character which follows the type number and
3375 an equals sign. It specifies what kind of type is being defined.
3376 @xref{String Field}, for more information about their use.
3381 Type reference; see @ref{String Field}.
3384 Reference to builtin type; see @ref{Negative Type Numbers}.
3387 Method (C++); see @ref{Cplusplus}.
3390 Pointer; see @ref{Miscellaneous Types}.
3396 Type Attributes (AIX); see @ref{String Field}. Member (class and variable)
3397 type (GNU C++); see @ref{Cplusplus}.
3400 Array; see @ref{Arrays}.
3403 Open array; see @ref{Arrays}.
3406 Pascal space type (AIX); see @ref{Miscellaneous Types}. Builtin integer
3407 type (Sun); see @ref{Builtin Type Descriptors}. Const and volatile
3408 qualfied type (OS9000).
3411 Volatile-qualified type; see @ref{Miscellaneous Types}.
3414 Complex builtin type (AIX); see @ref{Builtin Type Descriptors}.
3415 Const-qualified type (OS9000).
3418 COBOL Picture type. See AIX documentation for details.
3421 File type; see @ref{Miscellaneous Types}.
3424 N-dimensional dynamic array; see @ref{Arrays}.
3427 Enumeration type; see @ref{Enumerations}.
3430 N-dimensional subarray; see @ref{Arrays}.
3433 Function type; see @ref{Function Types}.
3436 Pascal function parameter; see @ref{Function Types}
3439 Builtin floating point type; see @ref{Builtin Type Descriptors}.
3442 COBOL Group. See AIX documentation for details.
3445 Imported type (AIX); see @ref{Cross-References}. Volatile-qualified
3449 Const-qualified type; see @ref{Miscellaneous Types}.
3452 COBOL File Descriptor. See AIX documentation for details.
3455 Multiple instance type; see @ref{Miscellaneous Types}.
3458 String type; see @ref{Strings}.
3461 Stringptr; see @ref{Strings}.
3464 Opaque type; see @ref{Typedefs}.
3467 Procedure; see @ref{Function Types}.
3470 Packed array; see @ref{Arrays}.
3473 Range type; see @ref{Subranges}.
3476 Builtin floating type; see @ref{Builtin Type Descriptors} (Sun). Pascal
3477 subroutine parameter; see @ref{Function Types} (AIX). Detecting this
3478 conflict is possible with careful parsing (hint: a Pascal subroutine
3479 parameter type will always contain a comma, and a builtin type
3480 descriptor never will).
3483 Structure type; see @ref{Structures}.
3486 Set type; see @ref{Miscellaneous Types}.
3489 Union; see @ref{Unions}.
3492 Variant record. This is a Pascal and Modula-2 feature which is like a
3493 union within a struct in C. See AIX documentation for details.
3496 Wide character; see @ref{Builtin Type Descriptors}.
3499 Cross-reference; see @ref{Cross-References}.
3502 Used by IBM's xlC C++ compiler (for structures, I think).
3505 gstring; see @ref{Strings}.
3508 @node Expanded Reference
3509 @appendix Expanded Reference by Stab Type
3511 @c FIXME: This appendix should go away; see N_PSYM or N_SO for an example.
3513 For a full list of stab types, and cross-references to where they are
3514 described, see @ref{Stab Types}. This appendix just duplicates certain
3515 information from the main body of this document; eventually the
3516 information will all be in one place.
3520 The first line is the symbol type (see @file{include/aout/stab.def}).
3522 The second line describes the language constructs the symbol type
3525 The third line is the stab format with the significant stab fields
3526 named and the rest NIL.
3528 Subsequent lines expand upon the meaning and possible values for each
3529 significant stab field. @samp{#} stands in for the type descriptor.
3531 Finally, any further information.
3534 * N_PC:: Pascal global symbol
3535 * N_NSYMS:: Number of symbols
3536 * N_NOMAP:: No DST map
3537 * N_M2C:: Modula-2 compilation unit
3538 * N_BROWS:: Path to .cb file for Sun source code browser
3539 * N_DEFD:: GNU Modula2 definition module dependency
3540 * N_EHDECL:: GNU C++ exception variable
3541 * N_MOD2:: Modula2 information "for imc"
3542 * N_CATCH:: GNU C++ "catch" clause
3543 * N_SSYM:: Structure or union element
3544 * N_SCOPE:: Modula2 scope information (Sun only)
3545 * Gould:: non-base register symbols used on Gould systems
3546 * N_LENG:: Length of preceding entry
3552 @deffn @code{.stabs} N_PC
3554 Global symbol (for Pascal).
3557 "name" -> "symbol_name" <<?>>
3558 value -> supposedly the line number (stab.def is skeptical)
3562 @file{stabdump.c} says:
3564 global pascal symbol: name,,0,subtype,line
3572 @deffn @code{.stabn} N_NSYMS
3574 Number of symbols (according to Ultrix V4.0).
3577 0, files,,funcs,lines (stab.def)
3584 @deffn @code{.stabs} N_NOMAP
3586 No DST map for symbol (according to Ultrix V4.0). I think this means a
3587 variable has been optimized out.
3590 name, ,0,type,ignored (stab.def)
3597 @deffn @code{.stabs} N_M2C
3599 Modula-2 compilation unit.
3602 "string" -> "unit_name,unit_time_stamp[,code_time_stamp]"
3604 value -> 0 (main unit)
3608 See @cite{Dbx and Dbxtool Interfaces}, 2nd edition, by Sun, 1988, for
3616 @deffn @code{.stabs} N_BROWS
3618 Sun source code browser, path to @file{.cb} file
3621 "path to associated @file{.cb} file"
3623 Note: N_BROWS has the same value as N_BSLINE.
3629 @deffn @code{.stabn} N_DEFD
3631 GNU Modula2 definition module dependency.
3633 GNU Modula-2 definition module dependency. The value is the
3634 modification time of the definition file. The other field is non-zero
3635 if it is imported with the GNU M2 keyword @code{%INITIALIZE}. Perhaps
3636 @code{N_M2C} can be used if there are enough empty fields?
3642 @deffn @code{.stabs} N_EHDECL
3644 GNU C++ exception variable <<?>>.
3646 "@var{string} is variable name"
3648 Note: conflicts with @code{N_MOD2}.
3654 @deffn @code{.stab?} N_MOD2
3656 Modula2 info "for imc" (according to Ultrix V4.0)
3658 Note: conflicts with @code{N_EHDECL} <<?>>
3664 @deffn @code{.stabn} N_CATCH
3666 GNU C++ @code{catch} clause
3668 GNU C++ @code{catch} clause. The value is its address. The desc field
3669 is nonzero if this entry is immediately followed by a @code{CAUGHT} stab
3670 saying what exception was caught. Multiple @code{CAUGHT} stabs means
3671 that multiple exceptions can be caught here. If desc is 0, it means all
3672 exceptions are caught here.
3678 @deffn @code{.stabn} N_SSYM
3680 Structure or union element.
3682 The value is the offset in the structure.
3684 <<?looking at structs and unions in C I didn't see these>>
3690 @deffn @code{.stab?} N_SCOPE
3692 Modula2 scope information (Sun linker)
3697 @section Non-base registers on Gould systems
3699 @deffn @code{.stab?} N_NBTEXT
3700 @deffnx @code{.stab?} N_NBDATA
3701 @deffnx @code{.stab?} N_NBBSS
3702 @deffnx @code{.stab?} N_NBSTS
3703 @deffnx @code{.stab?} N_NBLCS
3709 These are used on Gould systems for non-base registers syms.
3711 However, the following values are not the values used by Gould; they are
3712 the values which GNU has been documenting for these values for a long
3713 time, without actually checking what Gould uses. I include these values
3714 only because perhaps some someone actually did something with the GNU
3715 information (I hope not, why GNU knowingly assigned wrong values to
3716 these in the header file is a complete mystery to me).
3719 240 0xf0 N_NBTEXT ??
3720 242 0xf2 N_NBDATA ??
3730 @deffn @code{.stabn} N_LENG
3732 Second symbol entry containing a length-value for the preceding entry.
3733 The value is the length.
3737 @appendix Questions and Anomalies
3741 @c I think this is changed in GCC 2.4.5 to put the line number there.
3742 For GNU C stabs defining local and global variables (@code{N_LSYM} and
3743 @code{N_GSYM}), the desc field is supposed to contain the source
3744 line number on which the variable is defined. In reality the desc
3745 field is always 0. (This behavior is defined in @file{dbxout.c} and
3746 putting a line number in desc is controlled by @samp{#ifdef
3747 WINNING_GDB}, which defaults to false). GDB supposedly uses this
3748 information if you say @samp{list @var{var}}. In reality, @var{var} can
3749 be a variable defined in the program and GDB says @samp{function
3750 @var{var} not defined}.
3753 In GNU C stabs, there seems to be no way to differentiate tag types:
3754 structures, unions, and enums (symbol descriptor @samp{T}) and typedefs
3755 (symbol descriptor @samp{t}) defined at file scope from types defined locally
3756 to a procedure or other more local scope. They all use the @code{N_LSYM}
3757 stab type. Types defined at procedure scope are emited after the
3758 @code{N_RBRAC} of the preceding function and before the code of the
3759 procedure in which they are defined. This is exactly the same as
3760 types defined in the source file between the two procedure bodies.
3761 GDB overcompensates by placing all types in block #1, the block for
3762 symbols of file scope. This is true for default, @samp{-ansi} and
3763 @samp{-traditional} compiler options. (Bugs gcc/1063, gdb/1066.)
3766 What ends the procedure scope? Is it the proc block's @code{N_RBRAC} or the
3767 next @code{N_FUN}? (I believe its the first.)
3770 @c FIXME: This should go with the other stuff about global variables.
3771 Global variable stabs don't have location information. This comes
3772 from the external symbol for the same variable. The external symbol
3773 has a leading underbar on the _name of the variable and the stab does
3774 not. How do we know these two symbol table entries are talking about
3775 the same symbol when their names are different? (Answer: the debugger
3776 knows that external symbols have leading underbars).
3778 @c FIXME: This is absurdly vague; there all kinds of differences, some
3779 @c of which are the same between gnu & sun, and some of which aren't.
3780 @c In particular, I'm pretty sure GCC works with Sun dbx by default.
3782 @c Can GCC be configured to output stabs the way the Sun compiler
3783 @c does, so that their native debugging tools work? <NO?> It doesn't by
3784 @c default. GDB reads either format of stab. (GCC or SunC). How about
3788 @node XCOFF Differences
3789 @appendix Differences Between GNU Stabs in a.out and GNU Stabs in XCOFF
3791 @c FIXME: Merge *all* these into the main body of the document.
3792 The AIX/RS6000 native object file format is XCOFF with stabs. This
3793 appendix only covers those differences which are not covered in the main
3794 body of this document.
3798 BSD a.out stab types correspond to AIX XCOFF storage classes. In general
3799 the mapping is @code{N_@var{stabtype}} becomes @code{C_@var{stabtype}}.
3800 Some stab types in a.out are not supported in XCOFF; most of these use
3803 @c FIXME: I think they are trying to say something about whether the
3804 @c assembler defaults the value to the location counter.
3806 If the XCOFF stab is an @code{N_FUN} (@code{C_FUN}) then follow the
3807 string field with @samp{,.} instead of just @samp{,}.
3810 I think that's it for @file{.s} file differences. They could stand to be
3811 better presented. This is just a list of what I have noticed so far.
3812 There are a @emph{lot} of differences in the information in the symbol
3813 tables of the executable and object files.
3815 Mapping of a.out stab types to XCOFF storage classes:
3818 stab type storage class
3819 -------------------------------
3855 @node Sun Differences
3856 @appendix Differences Between GNU Stabs and Sun Native Stabs
3858 @c FIXME: Merge all this stuff into the main body of the document.
3862 GNU C stabs define @emph{all} types, file or procedure scope, as
3863 @code{N_LSYM}. Sun doc talks about using @code{N_GSYM} too.
3866 Sun C stabs use type number pairs in the format
3867 (@var{file-number},@var{type-number}) where @var{file-number} is a
3868 number starting with 1 and incremented for each sub-source file in the
3869 compilation. @var{type-number} is a number starting with 1 and
3870 incremented for each new type defined in the compilation. GNU C stabs
3871 use the type number alone, with no source file number.
3875 @appendix Using Stabs in Their Own Sections
3877 Many object file formats allow tools to create object files with custom
3878 sections containing any arbitrary data. For any such object file
3879 format, stabs can be embedded in special sections. This is how stabs
3880 are used with ELF and SOM, and aside from ECOFF and XCOFF, is how stabs
3884 * Stab Section Basics:: How to embed stabs in sections
3885 * ELF Linker Relocation:: Sun ELF hacks
3888 @node Stab Section Basics
3889 @appendixsec How to Embed Stabs in Sections
3891 The assembler creates two custom sections, a section named @code{.stab}
3892 which contains an array of fixed length structures, one struct per stab,
3893 and a section named @code{.stabstr} containing all the variable length
3894 strings that are referenced by stabs in the @code{.stab} section. The
3895 byte order of the stabs binary data depends on the object file format.
3896 For ELF, it matches the byte order of the ELF file itself, as determined
3897 from the @code{EI_DATA} field in the @code{e_ident} member of the ELF
3898 header. For SOM, it is always big-endian (is this true??? FIXME). For
3899 COFF, it matches the byte order of the COFF headers.
3901 The first stab in the @code{.stab} section for each compilation unit is
3902 synthetic, generated entirely by the assembler, with no corresponding
3903 @code{.stab} directive as input to the assembler. This stab contains
3904 the following fields:
3908 Offset in the @code{.stabstr} section to the source filename.
3914 Unused field, always zero.
3917 Count of upcoming symbols, i.e., the number of remaining stabs for this
3921 Size of the string table fragment associated with this source file, in
3925 The @code{.stabstr} section always starts with a null byte (so that string
3926 offsets of zero reference a null string), followed by random length strings,
3927 each of which is null byte terminated.
3929 The ELF section header for the @code{.stab} section has its
3930 @code{sh_link} member set to the section number of the @code{.stabstr}
3931 section, and the @code{.stabstr} section has its ELF section
3932 header @code{sh_type} member set to @code{SHT_STRTAB} to mark it as a
3933 string table. SOM and COFF have no way of linking the sections together
3934 or marking them as string tables.
3936 @node ELF Linker Relocation
3937 @appendixsec Having the Linker Relocate Stabs in ELF
3939 This section describes some Sun hacks for Stabs in ELF; it does not
3940 apply to COFF or SOM.
3942 To keep linking fast, you don't want the linker to have to relocate very
3943 many stabs. Making sure this is done for @code{N_SLINE},
3944 @code{N_RBRAC}, and @code{N_LBRAC} stabs is the most important thing
3945 (see the descriptions of those stabs for more information). But Sun's
3946 stabs in ELF has taken this further, to make all addresses in the
3947 @code{n_value} field (functions and static variables) relative to the
3948 source file. For the @code{N_SO} symbol itself, Sun simply omits the
3949 address. To find the address of each section corresponding to a given
3950 source file, the compiler puts out symbols giving the address of each
3951 section for a given source file. Since these are ELF (not stab)
3952 symbols, the linker relocates them correctly without having to touch the
3953 stabs section. They are named @code{Bbss.bss} for the bss section,
3954 @code{Ddata.data} for the data section, and @code{Drodata.rodata} for
3955 the rodata section. For the text section, there is no such symbol (but
3956 there should be, see below). For an example of how these symbols work,
3957 @xref{Stab Section Transformations}. GCC does not provide these symbols;
3958 it instead relies on the stabs getting relocated. Thus addresses which
3959 would normally be relative to @code{Bbss.bss}, etc., are already
3960 relocated. The Sun linker provided with Solaris 2.2 and earlier
3961 relocates stabs using normal ELF relocation information, as it would do
3962 for any section. Sun has been threatening to kludge their linker to not
3963 do this (to speed up linking), even though the correct way to avoid
3964 having the linker do these relocations is to have the compiler no longer
3965 output relocatable values. Last I heard they had been talked out of the
3966 linker kludge. See Sun point patch 101052-01 and Sun bug 1142109. With
3967 the Sun compiler this affects @samp{S} symbol descriptor stabs
3968 (@pxref{Statics}) and functions (@pxref{Procedures}). In the latter
3969 case, to adopt the clean solution (making the value of the stab relative
3970 to the start of the compilation unit), it would be necessary to invent a
3971 @code{Ttext.text} symbol, analogous to the @code{Bbss.bss}, etc.,
3972 symbols. I recommend this rather than using a zero value and getting
3973 the address from the ELF symbols.
3975 Finding the correct @code{Bbss.bss}, etc., symbol is difficult, because
3976 the linker simply concatenates the @code{.stab} sections from each
3977 @file{.o} file without including any information about which part of a
3978 @code{.stab} section comes from which @file{.o} file. The way GDB does
3979 this is to look for an ELF @code{STT_FILE} symbol which has the same
3980 name as the last component of the file name from the @code{N_SO} symbol
3981 in the stabs (for example, if the file name is @file{../../gdb/main.c},
3982 it looks for an ELF @code{STT_FILE} symbol named @code{main.c}). This
3983 loses if different files have the same name (they could be in different
3984 directories, a library could have been copied from one system to
3985 another, etc.). It would be much cleaner to have the @code{Bbss.bss}
3986 symbols in the stabs themselves. Having the linker relocate them there
3987 is no more work than having the linker relocate ELF symbols, and it
3988 solves the problem of having to associate the ELF and stab symbols.
3989 However, no one has yet designed or implemented such a scheme.
3991 @node Symbol Types Index
3992 @unnumbered Symbol Types Index