2 @setfilename stabs.info
9 * Stabs:: The "stabs" debugging information format.
15 This document describes the stabs debugging symbol tables.
17 Copyright 1992, 1993 Free Software Foundation, Inc.
18 Contributed by Cygnus Support. Written by Julia Menapace.
20 Permission is granted to make and distribute verbatim copies of
21 this manual provided the copyright notice and this permission notice
22 are preserved on all copies.
25 Permission is granted to process this file through Tex and print the
26 results, provided the printed document carries copying permission
27 notice identical to this one except for the removal of this paragraph
28 (this paragraph not being relevant to the printed manual).
31 Permission is granted to copy or distribute modified versions of this
32 manual under the terms of the GPL (for which purpose this text may be
33 regarded as a program in the language TeX).
36 @setchapternewpage odd
39 @title The ``stabs'' debug format
40 @author Julia Menapace
41 @author Cygnus Support
44 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
45 \xdef\manvers{\$Revision$} % For use in headers, footers too
47 \hfill Cygnus Support\par
49 \hfill \TeX{}info \texinfoversion\par
53 @vskip 0pt plus 1filll
54 Copyright @copyright{} 1992, 1993 Free Software Foundation, Inc.
55 Contributed by Cygnus Support.
57 Permission is granted to make and distribute verbatim copies of
58 this manual provided the copyright notice and this permission notice
59 are preserved on all copies.
65 @top The "stabs" representation of debugging information
67 This document describes the stabs debugging format.
70 * Overview:: Overview of stabs
71 * Program structure:: Encoding of the structure of the program
72 * Constants:: Constants
73 * Example:: A comprehensive example in C
75 * Types:: Type definitions
76 * Symbol Tables:: Symbol information in symbol tables
77 * Cplusplus:: Appendixes:
78 * Example2.c:: Source code for extended example
79 * Example2.s:: Assembly code for extended example
80 * Stab Types:: Symbol types in a.out files
81 * Symbol Descriptors:: Table of Symbol Descriptors
82 * Type Descriptors:: Table of Symbol Descriptors
83 * Expanded reference:: Reference information by stab type
84 * Questions:: Questions and anomolies
85 * xcoff-differences:: Differences between GNU stabs in a.out
86 and GNU stabs in xcoff
87 * Sun-differences:: Differences between GNU stabs and Sun
89 * Stabs-in-ELF:: Stabs in an ELF file.
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
99 @c FIXME! <<name of inventor>> at
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 completely comprehensive for stabs used by
105 C. The lists of symbol descriptors (@pxref{Symbol Descriptors}) and
106 type descriptors (@pxref{Type Descriptors}) are believed to be completely
107 comprehensive. There are known to be stabs for C++ and COBOL which are
108 poorly documented here. Stabs specific to other languages (e.g., Pascal,
109 Modula-2) are probably not as well documented as they should be.
111 Other sources of information on stabs are @cite{dbx and dbxtool
112 interfaces}, 2nd edition, by Sun, circa 1988, and @cite{AIX Version 3.2
113 Files Reference}, Fourth Edition, September 1992, "dbx Stabstring
114 Grammar" in the a.out section, page 2-31. This document is believed to
115 incorporate the information from those two sources except where it
116 explictly directs you to them for more information.
119 * Flow:: Overview of debugging information flow
120 * Stabs Format:: Overview of stab format
121 * String Field:: The @code{.stabs} @var{string} field
122 * C example:: A simple example in C source
123 * Assembly code:: The simple example at the assembly level
127 @section Overview of debugging information flow
129 The GNU C compiler compiles C source in a @file{.c} file into assembly
130 language in a @file{.s} file, which the assembler translates into
131 a @file{.o} file, which the linker combines with other @file{.o} files and
132 libraries to produce an executable file.
134 With the @samp{-g} option, GCC puts in the @file{.s} file additional
135 debugging information, which is slightly transformed by the assembler
136 and linker, and carried through into the final executable. This
137 debugging information describes features of the source file like line
138 numbers, the types and scopes of variables, and function names,
139 parameters, and scopes.
141 For some object file formats, the debugging information is encapsulated
142 in assembler directives known collectively as @dfn{stab} (symbol table)
143 directives, which are interspersed with the generated code. Stabs are
144 the native format for debugging information in the a.out and xcoff
145 object file formats. The GNU tools can also emit stabs in the coff and
146 ecoff object file formats.
148 The assembler adds the information from stabs to the symbol information
149 it places by default in the symbol table and the string table of the
150 @file{.o} file it is building. The linker consolidates the @file{.o}
151 files into one executable file, with one symbol table and one string
152 table. Debuggers use the symbol and string tables in the executable as
153 a source of debugging information about the program.
156 @section Overview of stab format
158 There are three overall formats for stab assembler directives,
159 differentiated by the first word of the stab. The name of the directive
160 describes which combination of four possible data fields follows. It is
161 either @code{.stabs} (string), @code{.stabn} (number), or @code{.stabd}
162 (dot). IBM's xcoff assembler uses @code{.stabx} (and some other
163 directives such as @code{.file} and @code{.bi}) instead of
164 @code{.stabs}, @code{.stabn} or @code{.stabd}.
166 The overall format of each class of stab is:
169 .stabs "@var{string}",@var{type},0,@var{desc},@var{value}
170 .stabn @var{type},0,@var{desc},@var{value}
171 .stabd @var{type},0,@var{desc}
172 .stabx "@var{string}",@var{value},@var{type},@var{sdb-type}
175 @c what is the correct term for "current file location"? My AIX
176 @c assembler manual calls it "the value of the current location counter".
177 For @code{.stabn} and @code{.stabd}, there is no @var{string} (the
178 @code{n_strx} field is zero; see @xref{Symbol Tables}). For
179 @code{.stabd}, the @var{value} field is implicit and has the value of
180 the current file location. For @code{.stabx}, the @var{sdb-type} field
181 is unused for stabs and can always be set to 0.
183 The number in the @var{type} field gives some basic information about
184 which type of stab this is (or whether it @emph{is} a stab, as opposed
185 to an ordinary symbol). Each valid type number defines a different stab
186 type. Further, the stab type defines the exact interpretation of, and
187 possible values for, any remaining @var{string}, @var{desc}, or
188 @var{value} fields present in the stab. @xref{Stab Types}, for a list
189 in numeric order of the valid type field values for stab directives.
192 @section The @code{.stabs} @var{string} field
194 For @code{.stabs} the @var{string} field holds the meat of the
195 debugging information. The generally unstructured nature of this field
196 is what makes stabs extensible. For some stab types the string field
197 contains only a name. For other stab types the contents can be a great
200 The overall format is of the @var{string} field is:
203 "@var{name}:@var{symbol-descriptor} @var{type-information}"
206 @var{name} is the name of the symbol represented by the stab.
207 @var{name} can be omitted, which means the stab represents an unnamed
208 object. For example, @samp{:t10=*2} defines type 10 as a pointer to
209 type 2, but does not give the type a name. Omitting the @var{name}
210 field is supported by AIX DBX and GDB after about version 4.8, but not
211 other debuggers. GCC sometimes uses a single space as the name instead
212 of omitting the name altogether; apparently that is supported by most
215 The @var{symbol_descriptor} following the @samp{:} is an alphabetic
216 character that tells more specifically what kind of symbol the stab
217 represents. If the @var{symbol_descriptor} is omitted, but type
218 information follows, then the stab represents a local variable. For a
219 list of symbol descriptors, see @ref{Symbol Descriptors}. The @samp{c}
220 symbol descriptor is an exception in that it is not followed by type
221 information. @xref{Constants}.
223 @var{type-information} is either a @var{type_number}, or
224 @samp{@var{type_number}=}. The @var{type_number} alone is a type
225 reference, referring directly to a type that has already been defined.
227 The @samp{@var{type_number}=} form is a type definition, where the
228 number represents a new type which is about to be defined. The type
229 definition may refer to other types by number, and those type numbers
230 may be followed by @samp{=} and nested definitions.
232 In a type definition, if the character that follows the equals sign is
233 non-numeric then it is a @var{type_descriptor}, and tells what kind of
234 type is about to be defined. Any other values following the
235 @var{type_descriptor} vary, depending on the @var{type_descriptor}. If
236 a number follows the @samp{=} then the number is a @var{type_reference}.
237 For a full description of types, @ref{Types}. @xref{Type
238 Descriptors}, for a list of
239 @var{type_descriptor} values.
241 There is an AIX extension for type attributes. Following the @samp{=}
242 is any number of type attributes. Each one starts with @samp{@@} and
243 ends with @samp{;}. Debuggers, including AIX's DBX, skip any type
244 attributes they do not recognize. GDB 4.9 does not do this---it will
245 ignore the entire symbol containing a type attribute. Hopefully this
246 will be fixed in the next GDB release. Because of a conflict with C++
247 (@pxref{Cplusplus}), new attributes should not be defined which begin
248 with a digit, @samp{(}, or @samp{-}; GDB may be unable to distinguish
249 those from the C++ type descriptor @samp{@@}. The attributes are:
252 @item a@var{boundary}
253 @var{boundary} is an integer specifying the alignment. I assume it
254 applies to all variables of this type.
257 Size in bits of a variable 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 All this can make the @var{string} field quite long. All
270 versions of GDB, and some versions of DBX, can handle arbitrarily long
271 strings. But many versions of DBX cretinously limit the strings to
272 about 80 characters, so compilers which must work with such DBX's need
273 to split the @code{.stabs} directive into several @code{.stabs}
274 directives. Each stab duplicates exactly all but the
275 @var{string} field. The @var{string} field of
276 every stab except the last is marked as continued with a
277 double-backslash at the end. Removing the backslashes and concatenating
278 the @var{string} fields of each stab produces the original,
282 @section A simple example in C source
284 To get the flavor of how stabs describe source information for a C
285 program, let's look at the simple program:
290 printf("Hello world");
294 When compiled with @samp{-g}, the program above yields the following
295 @file{.s} file. Line numbers have been added to make it easier to refer
296 to parts of the @file{.s} file in the description of the stabs that
300 @section The simple example at the assembly level
302 This simple ``hello world'' example demonstrates several of the stab
303 types used to describe C language source files.
307 2 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0
308 3 .stabs "hello.c",100,0,0,Ltext0
311 6 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
312 7 .stabs "char:t2=r2;0;127;",128,0,0,0
313 8 .stabs "long int:t3=r1;-2147483648;2147483647;",128,0,0,0
314 9 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
315 10 .stabs "long unsigned int:t5=r1;0;-1;",128,0,0,0
316 11 .stabs "short int:t6=r1;-32768;32767;",128,0,0,0
317 12 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
318 13 .stabs "short unsigned int:t8=r1;0;65535;",128,0,0,0
319 14 .stabs "long long unsigned int:t9=r1;0;-1;",128,0,0,0
320 15 .stabs "signed char:t10=r1;-128;127;",128,0,0,0
321 16 .stabs "unsigned char:t11=r1;0;255;",128,0,0,0
322 17 .stabs "float:t12=r1;4;0;",128,0,0,0
323 18 .stabs "double:t13=r1;8;0;",128,0,0,0
324 19 .stabs "long double:t14=r1;8;0;",128,0,0,0
325 20 .stabs "void:t15=15",128,0,0,0
328 23 .ascii "Hello, world!\12\0"
343 38 sethi %hi(LC0),%o1
344 39 or %o1,%lo(LC0),%o0
355 50 .stabs "main:F1",36,0,0,_main
356 51 .stabn 192,0,0,LBB2
357 52 .stabn 224,0,0,LBE2
360 @node Program structure
361 @chapter Encoding for the structure of the program
363 For the numeric values of the symbolic stab types, see @ref{Stab Types}.
364 For a reference to them, see @ref{Expanded reference}.
367 * Main Program:: Indicate what the main program is
368 * Source Files:: The path and name of the source file
369 * Include Files:: Names of include files
376 @section Main Program
378 Most languages allow the main program to have any name. The
379 @code{N_MAIN} stab type is used for a stab telling the debugger what
380 name is used in this program. Only the name is significant; it will be
381 the name of a function which is the main program. Most C compilers do
382 not use this stab; they expect the debugger to simply assume that the
383 name is @samp{main}, but some C compilers emit an @code{N_MAIN} stab for
384 the @samp{main} function.
387 @section Paths and names of the source files
389 Before any other stabs occur, there must be a stab specifying the source
390 file. This information is contained in a symbol of stab type
391 @code{N_SO}; the string contains the name of the file. The value of the
392 symbol is the start address of portion of the text section corresponding
395 With the Sun Solaris2 compiler, the @code{desc} field contains a
396 source-language code.
398 Some compilers (for example, GCC2 and SunOS4 @file{/bin/cc}) also
399 include the directory in which the source was compiled, in a second
400 @code{N_SO} symbol preceding the one containing the file name. This
401 symbol can be distinguished by the fact that it ends in a slash. Code
402 from the cfront C++ compiler can have additional @code{N_SO} symbols for
403 nonexistent source files after the @code{N_SO} for the real source file;
404 these are believed to contain no useful information.
409 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0 # @r{100 is N_SO}
410 .stabs "hello.c",100,0,0,Ltext0
415 Instead of @code{N_SO} symbols, XCOFF uses a @code{.file} assembler
416 directive which assembles to a standard COFF @code{.file} symbol;
417 explaining this in detail is outside the scope of this document.
420 @section Names of include files
422 There are several different schemes for dealing with include files: the
423 traditional @code{N_SOL} approach, Sun's @code{N_BINCL} approach, and the
424 XCOFF @code{C_BINCL} approach (which despite the similar name has little in
425 common with @code{N_BINCL}).
427 An @code{N_SOL} symbol specifies which include file subsequent symbols
428 refer to. The string field is the name of the file and the value is the
429 text address corresponding to the start of the previous include file and
430 the start of this one. To specify the main source file again, use an
431 @code{N_SOL} symbol with the name of the main source file.
433 A @code{N_BINCL} symbol specifies the start of an include file. In an
434 object file, only the name is significant. The Sun linker puts data
435 into some of the other fields. The end of the include file is marked by
436 a @code{N_EINCL} symbol (which has no name field). In an ojbect file,
437 there is no significant data in the @code{N_EINCL} symbol; the Sun
438 linker puts data into some of the fields. @code{N_BINCL} and
439 @code{N_EINCL} can be nested. If the linker detects that two source
440 files have identical stabs with a @code{N_BINCL} and @code{N_EINCL} pair
441 (as will generally be the case for a header file), then it only puts out
442 the stabs once. Each additional occurance is replaced by an
443 @code{N_EXCL} symbol. I believe the Sun (SunOS4, not sure about
444 Solaris) linker is the only one which supports this feature.
446 For the start of an include file in XCOFF, use the @file{.bi} assembler
447 directive, which generates a @code{C_BINCL} symbol. A @file{.ei}
448 directive, which generates a @code{C_EINCL} symbol, denotes the end of
449 the include file. Both directives are followed by the name of the
450 source file in quotes, which becomes the string for the symbol. The
451 value of each symbol, produced automatically by the assembler and
452 linker, is an offset into the executable which points to the beginning
453 (inclusive, as you'd expect) and end (inclusive, as you would not
454 expect) of the portion of the COFF linetable which corresponds to this
455 include file. @code{C_BINCL} and @code{C_EINCL} do not nest.
458 @section Line Numbers
460 A @code{N_SLINE} symbol represents the start of a source line. The
461 @var{desc} field contains the line number and the @var{value} field
462 contains the code address for the start of that source line. On most
463 machines the address is absolute; for Sun's stabs-in-ELF, it is relative
464 to the function in which the @code{N_SLINE} symbol occurs.
466 GNU documents @code{N_DSLINE} and @code{N_BSLINE} symbols for line
467 numbers in the data or bss segments, respectively. They are identical
468 to @code{N_SLINE} but are relocated differently by the linker. They
469 were intended to be used to describe the source location of a variable
470 declaration, but I believe that GCC2 actually puts the line number in
471 the desc field of the stab for the variable itself. GDB has been
472 ignoring these symbols (unless they contain a string field) at least
475 XCOFF uses COFF line numbers instead, which are outside the scope of
476 this document, ammeliorated by adequate marking of include files
477 (@pxref{Source Files}).
479 For single source lines that generate discontiguous code, such as flow
480 of control statements, there may be more than one line number entry for
481 the same source line. In this case there is a line number entry at the
482 start of each code range, each with the same line number.
487 All of the following stabs use the @code{N_FUN} symbol type.
489 A function is represented by an @samp{F} symbol descriptor (@var{desc}
490 field) for a global (extern) function, and @samp{f} for a static (local)
491 function. The next @code{N_SLINE} symbol can be used to find the line
492 number of the start of the function. The value field is the address of
493 the start of the function (absolute for @code{a.out}; relative to the
494 start of the file for Sun's stabs-in-ELF). The type information of the
495 stab represents the return type of the function; thus @samp{foo:f5}
496 means that foo is a function returning type 5.
498 The type information of the stab is optionally followed by type
499 information for each argument, with each argument preceded by @samp{;}.
500 An argument type of 0 means that additional arguments are being passed,
501 whose types and number may vary (@samp{...} in ANSI C). This extension
502 is used by Sun's Solaris compiler. GDB has tolerated it (i.e., at least
503 parsed the syntax, if not necessarily used the information) at least
504 since version 4.8; I don't know whether all versions of DBX will
505 tolerate it. The argument types given here are not redundant
506 with the symbols for the arguments themselves (@pxref{Parameters}), they
507 are the types of the arguments as they are passed, before any
508 conversions might take place. For example, if a C function which is
509 declared without a prototype takes a @code{float} argument, the value is
510 passed as a @code{double} but then converted to a @code{float}.
511 Debuggers need to use the types given in the arguments when printing
512 values, but if calling the function they need to use the types given in
513 the symbol defining the function.
515 If the return type and types of arguments of a function which is defined
516 in another source file are specified (i.e., a function prototype in ANSI
517 C), traditionally compilers emit no stab; the only way for the debugger
518 to find the information is if the source file where the function is
519 defined was also compiled with debugging symbols. As an extension the
520 Solaris compiler uses symbol descriptor @samp{P} followed by the return
521 type of the function, followed by the arguments, each preceded by
522 @samp{;}, as in a stab with symbol descriptor @samp{f} or @samp{F}.
523 This use of symbol descriptor @samp{P} can be distinguished from its use
524 for register parameters (@pxref{Parameters}) by the fact that it has
525 symbol type @code{N_FUN}.
527 The AIX documentation also defines symbol descriptor @samp{J} as an
528 internal function. I assume this means a function nested within another
529 function. It also says symbol descriptor @samp{m} is a module in
530 Modula-2 or extended Pascal.
532 Procedures (functions which do not return values) are represented as
533 functions returning the @code{void} type in C. I don't see why this couldn't
534 be used for all languages (inventing a @code{void} type for this purpose if
535 necessary), but the AIX documentation defines @samp{I}, @samp{P}, and
536 @samp{Q} for internal, global, and static procedures, respectively.
537 These symbol descriptors are unusual in that they are not followed by
540 For any of the above symbol descriptors, after the symbol descriptor and
541 the type information, there is optionally a comma, followed by the name
542 of the procedure, followed by a comma, followed by a name specifying the
543 scope. The first name is local to the scope specified, and seems to be
544 redundant with the name of the symbol (before the @samp{:}). The name
545 specifying the scope is the name of a procedure specifying that scope.
546 This feature is used by GCC, and presumably Pascal, Modula-2, etc.,
547 compilers, for nested functions.
549 If procedures are nested more than one level deep, only the immediately
550 containing scope is specified, for example:
562 return baz (x + 2 * y);
564 return x + bar (3 * x);
572 .stabs "baz:f1,baz,bar",36,0,0,_baz.15 # @r{36 is N_FUN}
573 .stabs "bar:f1,bar,foo",36,0,0,_bar.12
574 .stabs "foo:F1",36,0,0,_foo
577 The stab representing a procedure is located immediately following the
578 code of the procedure. This stab is in turn directly followed by a
579 group of other stabs describing elements of the procedure. These other
580 stabs describe the procedure's parameters, its block local variables, and
583 Going back to our "hello world" example program,
591 The @code{.stabs} entry after this code fragment shows the @var{name} of
592 the procedure (@code{main}); the type descriptor @var{desc} (@code{F},
593 for a global procedure); a reference to the predefined type @code{int}
594 for the return type; and the starting @var{address} of the procedure.
596 Here is an exploded summary (with whitespace introduced for clarity),
597 followed by line 50 of our sample assembly output, which has this form:
601 @var{desc} @r{(global proc @samp{F})}
602 @var{return_type_ref} @r{(int)}
608 50 .stabs "main:F1",36,0,0,_main
611 @node Block Structure
612 @section Block Structure
614 The program's block structure is represented by the @code{N_LBRAC} (left
615 brace) and the @code{N_RBRAC} (right brace) stab types. The variables
616 defined inside a block preceded the @code{N_LBRAC} symbol for most
617 compilers, including GCC. Other compilers, such as the Convex, Acorn
618 RISC machine, and Sun acc compilers, put the variables after the
619 @code{N_LBRAC} symbol. The values of the @code{N_LBRAC} and
620 @code{N_RBRAC} symbols are the start and end addresses of the code of
621 the block, respectively. For most machines, they are relative to the
622 starting address of this source file. For the Gould NP1, they are
623 absolute. For Sun's stabs-in-ELF, they are relative to the function in
626 The @code{N_LBRAC} and @code{N_RBRAC} stabs that describe the block
627 scope of a procedure are located after the @code{N_FUN} stab that
628 represents the procedure itself.
630 Sun documents the @code{desc} field of @code{N_LBRAC} and
631 @code{N_RBRAC} symbols as containing the nesting level of the block.
632 However, DBX seems not to care, and GCC always sets @code{desc} to
638 The @samp{c} symbol descriptor indicates that this stab represents a
639 constant. This symbol descriptor is an exception to the general rule
640 that symbol descriptors are followed by type information. Instead, it
641 is followed by @samp{=} and one of the following:
645 Boolean constant. @var{value} is a numeric value; I assume it is 0 for
649 Character constant. @var{value} is the numeric value of the constant.
651 @item e @var{type-information} , @var{value}
652 Constant whose value can be represented as integral.
653 @var{type-information} is the type of the constant, as it would appear
654 after a symbol descriptor (@pxref{Stabs Format}). @var{value} is the
655 numeric value of the constant. GDB 4.9 does not actually get the right
656 value if @var{value} does not fit in a host @code{int}, but it does not
657 do anything violent, and future debuggers could be extended to accept
658 integers of any size (whether unsigned or not). This constant type is
659 usually documented as being only for enumeration constants, but GDB has
660 never imposed that restriction; I don't know about other debuggers.
663 Integer constant. @var{value} is the numeric value. The type is some
664 sort of generic integer type (for GDB, a host @code{int}); to specify
665 the type explicitly, use @samp{e} instead.
668 Real constant. @var{value} is the real value, which can be @samp{INF}
669 (optionally preceded by a sign) for infinity, @samp{QNAN} for a quiet
670 NaN (not-a-number), or @samp{SNAN} for a signalling NaN. If it is a
671 normal number the format is that accepted by the C library function
675 String constant. @var{string} is a string enclosed in either @samp{'}
676 (in which case @samp{'} characters within the string are represented as
677 @samp{\'} or @samp{"} (in which case @samp{"} characters within the
678 string are represented as @samp{\"}).
680 @item S @var{type-information} , @var{elements} , @var{bits} , @var{pattern}
681 Set constant. @var{type-information} is the type of the constant, as it
682 would appear after a symbol descriptor (@pxref{Stabs Format}).
683 @var{elements} is the number of elements in the set (Does this means
684 how many bits of @var{pattern} are actually used, which would be
685 redundant with the type, or perhaps the number of bits set in
686 @var{pattern}? I don't get it), @var{bits} is the number of bits in the
687 constant (meaning it specifies the length of @var{pattern}, I think),
688 and @var{pattern} is a hexadecimal representation of the set. AIX
689 documentation refers to a limit of 32 bytes, but I see no reason why
690 this limit should exist. This form could probably be used for arbitrary
691 constants, not just sets; the only catch is that @var{pattern} should be
692 understood to be target, not host, byte order and format.
695 The boolean, character, string, and set constants are not supported by
696 GDB 4.9, but it will ignore them. GDB 4.8 and earlier gave an error
697 message and refused to read symbols from the file containing the
700 This information is followed by @samp{;}.
703 @chapter A Comprehensive Example in C
705 To describe the other stab types,
706 we'll examine a second program, @code{example2}, which builds on the
707 first example to introduce the rest of the stab types, symbol
708 descriptors, and type descriptors used in C.
709 @xref{Example2.c} for the complete @file{.c} source,
710 and @pxref{Example2.s} for the @file{.s} assembly code.
711 This description includes parts of those files.
713 @section Flow of control and nested scopes
719 @code{N_SLINE}, @code{N_LBRAC}, @code{N_RBRAC} (cont.)
722 Consider the body of @code{main}, from @file{example2.c}. It shows more
723 about how @code{N_SLINE}, @code{N_RBRAC}, and @code{N_LBRAC} stabs are used.
727 21 static float s_flap;
729 23 for (times=0; times < s_g_repeat; times++)@{
731 25 printf ("Hello world\n");
736 Here we have a single source line, the @code{for} line, that generates
737 non-linear flow of control, and non-contiguous code. In this case, an
738 @code{N_SLINE} stab with the same line number proceeds each block of
739 non-contiguous code generated from the same source line.
741 The example also shows nested scopes. The @code{N_LBRAC} and
742 @code{N_LBRAC} stabs that describe block structure are nested in the
743 same order as the corresponding code blocks, those of the for loop
744 inside those for the body of main.
747 This is the label for the @code{N_LBRAC} (left brace) stab marking the
748 start of @code{main}.
755 In the first code range for C source line 23, the @code{for} loop
756 initialize and test, @code{N_SLINE} (68) records the line number:
763 58 .stabn 68,0,23,LM2
767 62 sethi %hi(_s_g_repeat),%o0
769 64 ld [%o0+%lo(_s_g_repeat)],%o0
774 @exdent label for the @code{N_LBRAC} (start block) marking the start of @code{for} loop
777 69 .stabn 68,0,25,LM3
779 71 sethi %hi(LC0),%o1
780 72 or %o1,%lo(LC0),%o0
783 75 .stabn 68,0,26,LM4
786 @exdent label for the @code{N_RBRAC} (end block) stab marking the end of the @code{for} loop
792 Now we come to the second code range for source line 23, the @code{for}
793 loop increment and return. Once again, @code{N_SLINE} (68) records the
797 .stabn, N_SLINE, NIL,
801 78 .stabn 68,0,23,LM5
809 86 .stabn 68,0,27,LM6
812 @exdent label for the @code{N_RBRAC} (end block) stab marking the end of the @code{for} loop
815 89 .stabn 68,0,27,LM7
820 94 .stabs "main:F1",36,0,0,_main
821 95 .stabs "argc:p1",160,0,0,68
822 96 .stabs "argv:p20=*21=*2",160,0,0,72
823 97 .stabs "s_flap:V12",40,0,0,_s_flap.0
824 98 .stabs "times:1",128,0,0,-20
828 Here is an illustration of stabs describing nested scopes. The scope
829 nesting is reflected in the nested bracketing stabs (@code{N_LBRAC},
833 .stabn N_LBRAC,NIL,NIL,
834 @var{block-start-address}
836 99 .stabn 192,0,0,LBB2 ## begin proc label
837 100 .stabs "inner:1",128,0,0,-24
838 101 .stabn 192,0,0,LBB3 ## begin for label
842 @code{N_RBRAC} (224), ``right brace'' ends a lexical block (scope).
845 .stabn N_RBRAC,NIL,NIL,
846 @var{block-end-address}
848 102 .stabn 224,0,0,LBE3 ## end for label
849 103 .stabn 224,0,0,LBE2 ## end proc label
856 * Stack Variables:: Variables allocated on the stack.
857 * Global Variables:: Variables used by more than one source file.
858 * Register variables:: Variables in registers.
859 * Common Blocks:: Variables statically allocated together.
860 * Statics:: Variables local to one source file.
861 * Parameters:: Variables for arguments to functions.
864 @node Stack Variables
865 @section Automatic Variables Allocated on the Stack
867 If a variable is declared whose scope is local to a function and whose
868 lifetime is only as long as that function executes (C calls such
869 variables automatic), they can be allocated in a register
870 (@pxref{Register variables}) or on the stack.
872 For variables allocated on the stack, each variable has a stab with the
873 symbol descriptor omitted. Since type information should being with a
874 digit, @samp{-}, or @samp{(}, only digits, @samp{-}, and @samp{(} are
875 precluded from being used for symbol descriptors by this fact. However,
876 the Acorn RISC machine (ARM) is said to get this wrong: it puts out a
877 mere type definition here, without the preceding
878 @code{@var{typenumber}=}. This is a bad idea; there is no guarantee
879 that type descriptors are distinct from symbol descriptors.
881 These stabs have the @code{N_LSYM} stab type. The value of the stab is
882 the offset of the variable within the local variables. On most machines
883 this is an offset from the frame pointer and is negative.
885 The stab for an automatic variable is located just before the
886 @code{N_LBRAC} stab describing the open brace of the block to which it
887 is scoped, except for some compilers which put the automatic variables
888 after the @code{N_LBRAC} (see @code{VARIABLES_INSIDE_BLOCK} in GDB).
890 For example, the following C code
900 produces the following stabs
903 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
904 .stabs "x:1",128,0,0,-12 # @r{128 is N_LSYM}
905 .stabn 192,0,0,LBB2 # @r{192 is N_LBRAC}
906 .stabn 224,0,0,LBE2 # @r{224 is N_RBRAC}
909 @xref{Procedures} for more information on the first stab, and @ref{Block
910 Structure} for more information on the @code{N_LBRAC} and @code{N_RBRAC}
913 @node Global Variables
914 @section Global Variables
916 A variable whose scope which is not specific to just one source file is
917 represented by the @samp{G} symbol descriptor. These stabs use the
918 @code{N_GSYM} stab type. The type information for the stab
919 (@pxref{Stabs Format}) gives the type of the variable.
921 For example, the following source code:
928 yields the following assembly code:
931 .stabs "g_foo:G2",32,0,0,0 # @r{32 is N_GSYM}
938 The address of the variable represented by the @code{N_GSYM} is not
939 contained in the @code{N_GSYM} stab. The debugger gets this information
940 from the external symbol for the global variable. In the example above,
941 the @code{.global _g_foo} and @code{_g_foo:} lines tell the assembler to
942 produce an external symbol.
944 @node Register variables
945 @section Register variables
947 @c According to an old version of this manual, AIX uses C_RPSYM instead
948 @c of C_RSYM. I am skeptical; this should be verified.
949 Register variables have their own stab type, @code{N_RSYM}, and their
950 own symbol descriptor, @code{r}. The stab's value field contains the
951 number of the register where the variable data will be stored.
953 The value is the register number.
955 AIX defines a separate symbol descriptor @samp{d} for floating point
956 registers. This seems unnecessary; why not just just give floating
957 point registers different register numbers? I have not verified whether
958 the compiler actually uses @samp{d}.
960 If the register is explicitly allocated to a global variable, but not
964 register int g_bar asm ("%g5");
967 the stab may be emitted at the end of the object file, with
968 the other bss symbols.
971 @section Common Blocks
973 A common block is a statically allocated section of memory which can be
974 referred to by several source files. It may contain several variables.
975 I believe Fortran is the only language with this feature. A
976 @code{N_BCOMM} stab begins a common block and an @code{N_ECOMM} stab
977 ends it. The only thing which is significant about these two stabs is
978 their name, which can be used to look up a normal (non-debugging) symbol
979 which gives the address of the common block. Then each stab between the
980 @code{N_BCOMM} and the @code{N_ECOMM} specifies a member of that common
981 block; its value is the offset within the common block of that variable.
982 The @code{N_ECOML} stab type is documented for this purpose, but Sun's
983 Fortran compiler uses @code{N_GSYM} instead. The test case I
984 looked at had a common block local to a function and it used the
985 @samp{V} symbol descriptor; I assume one would use @samp{S} if not local
986 to a function (that is, if a common block @emph{can} be anything other
987 than local to a function).
990 @section Static Variables
992 Initialized static variables are represented by the @samp{S} and
993 @samp{V} symbol descriptors. @samp{S} means file scope static, and
994 @samp{V} means procedure scope static.
996 @c This is probably not worth mentioning; it is only true on the sparc
997 @c for `double' variables which although declared const are actually in
998 @c the data segment (the text segment can't guarantee 8 byte alignment).
1000 @c 2.4.5 has a bug in that it uses @code{N_FUN}, so neither DBX nor GDB can
1001 @c find the variables)
1002 In a.out files, @code{N_STSYM} means the data segment, @code{N_FUN}
1003 means the text segment, and @code{N_LCSYM} means the bss segment.
1005 In xcoff files, each symbol has a section number, so the stab type
1006 need not indicate the segment.
1008 In ecoff files, the storage class is used to specify the section, so the
1009 stab type need not indicate the segment.
1011 @c In ELF files, it apparently is a big mess. See kludge in dbxread.c
1012 @c in GDB. FIXME: Investigate where this kludge comes from.
1014 @c This is the place to mention N_ROSYM; I'd rather do so once I can
1015 @c coherently explain how this stuff works for stabs-in-ELF.
1017 For example, the source lines
1020 static const int var_const = 5;
1021 static int var_init = 2;
1022 static int var_noinit;
1026 yield the following stabs:
1029 .stabs "var_const:S1",36,0,0,_var_const # @r{36 is N_FUN}
1031 .stabs "var_init:S1",38,0,0,_var_init # @r{38 is N_STSYM}
1033 .stabs "var_noinit:S1",40,0,0,_var_noinit # @r{40 is N_LCSYM}
1039 Parameters to a function are represented by a stab (or sometimes two,
1040 see below) for each parameter. The stabs are in the order in which the
1041 debugger should print the parameters (i.e., the order in which the
1042 parameters are declared in the source file).
1044 The symbol descriptor @samp{p} is used to refer to parameters which are
1045 in the arglist. Symbols have symbol type @code{N_PSYM}. The value of
1046 the symbol is the offset relative to the argument list.
1048 If the parameter is passed in a register, then the traditional way to do
1049 this is to provide two symbols for each argument:
1052 .stabs "arg:p1" . . . ; N_PSYM
1053 .stabs "arg:r1" . . . ; N_RSYM
1056 Debuggers are expected to use the second one to find the value, and the
1057 first one to know that it is an argument.
1059 Because this is kind of ugly, some compilers use symbol descriptor
1060 @samp{P} or @samp{R} to indicate an argument which is in a register.
1061 The symbol value is the register number. @samp{P} and @samp{R} mean the
1062 same thing, the difference is that @samp{P} is a GNU invention and
1063 @samp{R} is an IBM (xcoff) invention. As of version 4.9, GDB should
1064 handle either one. Symbol type @code{C_RPSYM} is used with @samp{R} and
1065 @code{N_RSYM} is used with @samp{P}.
1067 According to the AIX documentation symbol descriptor @samp{D} is for a
1068 parameter passed in a floating point register. This seems
1069 unnecessary---why not just use @samp{R} with a register number which
1070 indicates that it's a floating point register? I haven't verified
1071 whether the system actually does what the documentation indicates.
1073 There is at least one case where GCC uses a @samp{p} and @samp{r} pair
1074 rather than @samp{P}; this is where the argument is passed in the
1075 argument list and then loaded into a register.
1077 On the sparc and hppa, for a @samp{P} symbol whose type is a structure
1078 or union, the register contains the address of the structure. On the
1079 sparc, this is also true of a @samp{p} and @samp{r} pair (using Sun cc) or a
1080 @samp{p} symbol. However, if a (small) structure is really in a
1081 register, @samp{r} is used. And, to top it all off, on the hppa it
1082 might be a structure which was passed on the stack and loaded into a
1083 register and for which there is a @samp{p} and @samp{r} pair! I believe
1084 that symbol descriptor @samp{i} is supposed to deal with this case, (it
1085 is said to mean "value parameter by reference, indirect access", I don't
1086 know the source for this information) but I don't know details or what
1087 compilers or debuggers use it, if any (not GDB or GCC). It is not clear
1088 to me whether this case needs to be dealt with differently than
1089 parameters passed by reference (see below).
1091 There is another case similar to an argument in a register, which is an
1092 argument which is actually stored as a local variable. Sometimes this
1093 happens when the argument was passed in a register and then the compiler
1094 stores it as a local variable. If possible, the compiler should claim
1095 that it's in a register, but this isn't always done. Some compilers use
1096 the pair of symbols approach described above (@samp{@var{arg}:p} followed by
1097 @samp{@var{arg}:}); this includes GCC1 (not GCC2) on the sparc when passing a small
1098 structure and GCC2 (sometimes) when the argument type is float and it is
1099 passed as a double and converted to float by the prologue (in the latter
1100 case the type of the @samp{@var{arg}:p} symbol is double and the type of the @samp{@var{arg}:}
1101 symbol is float). GCC, at least on the 960, uses a single @samp{p}
1102 symbol descriptor for an argument which is stored as a local variable
1103 but uses @code{N_LSYM} instead of @code{N_PSYM}. In this case the value
1104 of the symbol is an offset relative to the local variables for that
1105 function, not relative to the arguments (on some machines those are the
1106 same thing, but not on all).
1108 If the parameter is passed by reference (e.g., Pascal VAR parameters),
1109 then type symbol descriptor is @samp{v} if it is in the argument list,
1110 or @samp{a} if it in a register. Other than the fact that these contain
1111 the address of the parameter other than the parameter itself, they are
1112 identical to @samp{p} and @samp{R}, respectively. I believe @samp{a} is
1113 an AIX invention; @samp{v} is supported by all stabs-using systems as
1116 @c Is this paragraph correct? It is based on piecing together patchy
1117 @c information and some guesswork
1118 Conformant arrays refer to a feature of Modula-2, and perhaps other
1119 languages, in which the size of an array parameter is not known to the
1120 called function until run-time. Such parameters have two stabs, a
1121 @samp{x} for the array itself, and a @samp{C}, which represents the size
1122 of the array. The value of the @samp{x} stab is the offset in the
1123 argument list where the address of the array is stored (it this right?
1124 it is a guess); the value of the @samp{C} stab is the offset in the
1125 argument list where the size of the array (in elements? in bytes?) is
1128 The following are also said to go with @code{N_PSYM}:
1131 "name" -> "param_name:#type"
1133 -> pF Fortran function parameter
1134 -> X (function result variable)
1135 -> b (based variable)
1137 value -> offset from the argument pointer (positive).
1140 As a simple example, the code
1152 .stabs "main:F1",36,0,0,_main # @r{36 is N_FUN}
1153 .stabs "argc:p1",160,0,0,68 # @r{160 is N_PSYM}
1154 .stabs "argv:p20=*21=*2",160,0,0,72
1157 The type definition of argv is interesting because it contains several
1158 type definitions. Type 21 is pointer to type 2 (char) and argv (type 20) is
1162 @chapter Defining types
1164 Now let's look at some variable definitions involving complex types.
1165 This involves understanding better how types are described. In the
1166 examples so far types have been described as references to previously
1167 defined types or defined in terms of subranges of or pointers to
1168 previously defined types. The section that follows discusses
1169 the other type descriptors that may follow the @samp{=} sign in a
1173 * Builtin types:: Integers, floating point, void, etc.
1174 * Miscellaneous Types:: Pointers, sets, files, etc.
1175 * Cross-references:: Referring to a type not yet defined.
1176 * Subranges:: A type with a specific range.
1177 * Arrays:: An aggregate type of same-typed elements.
1178 * Strings:: Like an array but also has a length.
1179 * Enumerations:: Like an integer but the values have names.
1180 * Structures:: An aggregate type of different-typed elements.
1181 * Typedefs:: Giving a type a name.
1182 * Unions:: Different types sharing storage.
1187 @section Builtin types
1189 Certain types are built in (@code{int}, @code{short}, @code{void},
1190 @code{float}, etc.); the debugger recognizes these types and knows how
1191 to handle them. Thus don't be surprised if some of the following ways
1192 of specifying builtin types do not specify everything that a debugger
1193 would need to know about the type---in some cases they merely specify
1194 enough information to distinguish the type from other types.
1196 The traditional way to define builtin types is convolunted, so new ways
1197 have been invented to describe them. Sun's ACC uses the @samp{b} and
1198 @samp{R} type descriptors (@pxref{Builtin Type Descriptors}), and IBM
1199 uses negative type numbers (@pxref{Negative Type Numbers}). GDB can
1200 accept all three, as of version 4.8; DBX just accepts the traditional
1201 builtin types and perhaps one of the other two formats.
1204 * Traditional Builtin Types:: Put on your seatbelts and prepare for kludgery
1205 * Builtin Type Descriptors:: Builtin types with special type descriptors
1206 * Negative Type Numbers:: Builtin types using negative type numbers
1209 @node Traditional Builtin Types
1210 @subsection Traditional Builtin types
1212 Often types are defined as subranges of themselves. If the array bounds
1213 can fit within an @code{int}, then they are given normally. For example:
1216 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0 # @r{128 is N_LSYM}
1217 .stabs "char:t2=r2;0;127;",128,0,0,0
1220 Builtin types can also be described as subranges of @code{int}:
1223 .stabs "unsigned short:t6=r1;0;65535;",128,0,0,0
1226 If the lower bound of a subrange is 0 and the upper bound is -1, it
1227 means that the type is an unsigned integral type whose bounds are too
1228 big to describe in an int. Traditionally this is only used for
1229 @code{unsigned int} and @code{unsigned long}; GCC also sometimes uses it
1230 for @code{long long} and @code{unsigned long long}, and the only way to
1231 tell those types apart is to look at their names. On other machines GCC
1232 puts out bounds in octal, with a leading 0. In this case a negative
1233 bound consists of a number which is a 1 bit followed by a bunch of 0
1234 bits, and a positive bound is one in which a bunch of bits are 1.
1237 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
1238 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
1241 If the lower bound of a subrange is 0 and the upper bound is negative,
1242 it means that it is an unsigned integral type whose size in bytes is the
1243 absolute value of the upper bound. I believe this is a Convex
1244 convention for @code{unsigned long long}.
1246 If the lower bound of a subrange is negative and the upper bound is 0,
1247 it means that the type is a signed integral type whose size in bytes is
1248 the absolute value of the lower bound. I believe this is a Convex
1249 convention for @code{long long}. To distinguish this from a legitimate
1250 subrange, the type should be a subrange of itself. I'm not sure whether
1251 this is the case for Convex.
1253 If the upper bound of a subrange is 0, it means that this is a floating
1254 point type, and the lower bound of the subrange indicates the number of
1258 .stabs "float:t12=r1;4;0;",128,0,0,0
1259 .stabs "double:t13=r1;8;0;",128,0,0,0
1262 However, GCC writes @code{long double} the same way it writes
1263 @code{double}; the only way to distinguish them is by the name:
1266 .stabs "long double:t14=r1;8;0;",128,0,0,0
1269 Complex types are defined the same way as floating-point types; the only
1270 way to distinguish a single-precision complex from a double-precision
1271 floating-point type is by the name.
1273 The C @code{void} type is defined as itself:
1276 .stabs "void:t15=15",128,0,0,0
1279 I'm not sure how a boolean type is represented.
1281 @node Builtin Type Descriptors
1282 @subsection Defining Builtin Types using Builtin Type Descriptors
1284 There are various type descriptors to define builtin types:
1287 @c FIXME: clean up description of width and offset, once we figure out
1289 @item b @var{signed} @var{char-flag} @var{width} ; @var{offset} ; @var{nbits} ;
1290 Define an integral type. @var{signed} is @samp{u} for unsigned or
1291 @samp{s} for signed. @var{char-flag} is @samp{c} which indicates this
1292 is a character type, or is omitted. I assume this is to distinguish an
1293 integral type from a character type of the same size, for example it
1294 might make sense to set it for the C type @code{wchar_t} so the debugger
1295 can print such variables differently (Solaris does not do this). Sun
1296 sets it on the C types @code{signed char} and @code{unsigned char} which
1297 arguably is wrong. @var{width} and @var{offset} appear to be for small
1298 objects stored in larger ones, for example a @code{short} in an
1299 @code{int} register. @var{width} is normally the number of bytes in the
1300 type. @var{offset} seems to always be zero. @var{nbits} is the number
1301 of bits in the type.
1303 Note that type descriptor @samp{b} used for builtin types conflicts with
1304 its use for Pascal space types (@pxref{Miscellaneous Types}); they can
1305 be distinguished because the character following the type descriptor
1306 will be a digit, @samp{(}, or @samp{-} for a Pascal space type, or
1307 @samp{u} or @samp{s} for a builtin type.
1310 Documented by AIX to define a wide character type, but their compiler
1311 actually uses negative type numbers (@pxref{Negative Type Numbers}).
1313 @item R @var{fp_type} ; @var{bytes} ;
1314 Define a floating point type. @var{fp_type} has one of the following values:
1318 IEEE 32-bit (single precision) floating point format.
1321 IEEE 64-bit (double precision) floating point format.
1323 @item 3 (NF_COMPLEX)
1324 @item 4 (NF_COMPLEX16)
1325 @item 5 (NF_COMPLEX32)
1326 @c "GDB source" really means @file{include/aout/stab_gnu.h}, but trying
1327 @c to put that here got an overfull hbox.
1328 These are for complex numbers. A comment in the GDB source describes
1329 them as Fortran complex, double complex, and complex*16, respectively,
1330 but what does that mean? (i.e., Single precision? Double precison?).
1332 @item 6 (NF_LDOUBLE)
1333 Long double. This should probably only be used for Sun format long
1334 double, and new codes should be used for other floating point formats
1335 (@code{NF_DOUBLE} can be used if a long double is really just an IEEE double,
1339 @var{bytes} is the number of bytes occupied by the type. This allows a
1340 debugger to perform some operations with the type even if it doesn't
1341 understand @var{fp_code}.
1343 @item g @var{type-information} ; @var{nbits}
1344 Documented by AIX to define a floating type, but their compiler actually
1345 uses negative type numbers (@pxref{Negative Type Numbers}).
1347 @item c @var{type-information} ; @var{nbits}
1348 Documented by AIX to define a complex type, but their compiler actually
1349 uses negative type numbers (@pxref{Negative Type Numbers}).
1352 The C @code{void} type is defined as a signed integral type 0 bits long:
1354 .stabs "void:t19=bs0;0;0",128,0,0,0
1356 The Solaris compiler seems to omit the trailing semicolon in this case.
1357 Getting sloppy in this way is not a swift move because if a type is
1358 embedded in a more complex expression it is necessary to be able to tell
1361 I'm not sure how a boolean type is represented.
1363 @node Negative Type Numbers
1364 @subsection Negative Type numbers
1366 Since the debugger knows about the builtin types anyway, the idea of
1367 negative type numbers is simply to give a special type number which
1368 indicates the built in type. There is no stab defining these types.
1370 I'm not sure whether anyone has tried to define what this means if
1371 @code{int} can be other than 32 bits (or other types can be other than
1372 their customary size). If @code{int} has exactly one size for each
1373 architecture, then it can be handled easily enough, but if the size of
1374 @code{int} can vary according the compiler options, then it gets hairy.
1375 The best way to do this would be to define separate negative type
1376 numbers for 16-bit @code{int} and 32-bit @code{int}; therefore I have
1377 indicated below the customary size (and other format information) for
1378 each type. The information below is currently correct because AIX on
1379 the RS6000 is the only system which uses these type numbers. If these
1380 type numbers start to get used on other systems, I suspect the correct
1381 thing to do is to define a new number in cases where a type does not
1382 have the size and format indicated below (or avoid negative type numbers
1385 Also note that part of the definition of the negative type number is
1386 the name of the type. Types with identical size and format but
1387 different names have different negative type numbers.
1391 @code{int}, 32 bit signed integral type.
1394 @code{char}, 8 bit type holding a character. Both GDB and DBX on AIX
1395 treat this as signed. GCC uses this type whether @code{char} is signed
1396 or not, which seems like a bad idea. The AIX compiler (xlc) seems to
1397 avoid this type; it uses -5 instead for @code{char}.
1400 @code{short}, 16 bit signed integral type.
1403 @code{long}, 32 bit signed integral type.
1406 @code{unsigned char}, 8 bit unsigned integral type.
1409 @code{signed char}, 8 bit signed integral type.
1412 @code{unsigned short}, 16 bit unsigned integral type.
1415 @code{unsigned int}, 32 bit unsigned integral type.
1418 @code{unsigned}, 32 bit unsigned integral type.
1421 @code{unsigned long}, 32 bit unsigned integral type.
1424 @code{void}, type indicating the lack of a value.
1427 @code{float}, IEEE single precision.
1430 @code{double}, IEEE double precision.
1433 @code{long double}, IEEE double precision. The compiler claims the size
1434 will increase in a future release, and for binary compatibility you have
1435 to avoid using @code{long double}. I hope when they increase it they
1436 use a new negative type number.
1439 @code{integer}. 32 bit signed integral type.
1442 @code{boolean}. 32 bit type. How is the truth value encoded? Is it
1443 the least significant bit or is it a question of whether the whole value
1444 is zero or non-zero?
1447 @code{short real}. IEEE single precision.
1450 @code{real}. IEEE double precision.
1453 @code{stringptr}. @xref{Strings}.
1456 @code{character}, 8 bit unsigned character type.
1459 @code{logical*1}, 8 bit type. This Fortran type has a split
1460 personality in that it is used for boolean variables, but can also be
1461 used for unsigned integers. 0 is false, 1 is true, and other values are
1465 @code{logical*2}, 16 bit type. This Fortran type has a split
1466 personality in that it is used for boolean variables, but can also be
1467 used for unsigned integers. 0 is false, 1 is true, and other values are
1471 @code{logical*4}, 32 bit type. This Fortran type has a split
1472 personality in that it is used for boolean variables, but can also be
1473 used for unsigned integers. 0 is false, 1 is true, and other values are
1477 @code{logical}, 32 bit type. This Fortran type has a split
1478 personality in that it is used for boolean variables, but can also be
1479 used for unsigned integers. 0 is false, 1 is true, and other values are
1483 @code{complex}. A complex type consisting of two IEEE single-precision
1484 floating point values.
1487 @code{complex}. A complex type consisting of two IEEE double-precision
1488 floating point values.
1491 @code{integer*1}, 8 bit signed integral type.
1494 @code{integer*2}, 16 bit signed integral type.
1497 @code{integer*4}, 32 bit signed integral type.
1500 @code{wchar}. Wide character, 16 bits wide, unsigned (what format?
1504 @node Miscellaneous Types
1505 @section Miscellaneous Types
1508 @item b @var{type-information} ; @var{bytes}
1509 Pascal space type. This is documented by IBM; what does it mean?
1511 Note that this use of the @samp{b} type descriptor can be distinguished
1512 from its use for builtin integral types (@pxref{Builtin Type
1513 Descriptors}) because the character following the type descriptor is
1514 always a digit, @samp{(}, or @samp{-}.
1516 @item B @var{type-information}
1517 A volatile-qualified version of @var{type-information}. This is a Sun
1518 extension. A volatile-qualified type means that references and stores
1519 to a variable of that type must not be optimized or cached; they must
1520 occur as the user specifies them.
1522 @item d @var{type-information}
1523 File of type @var{type-information}. As far as I know this is only used
1526 @item k @var{type-information}
1527 A const-qualified version of @var{type-information}. This is a Sun
1528 extension. A const-qualified type means that a variable of this type
1531 @item M @var{type-information} ; @var{length}
1532 Multiple instance type. The type seems to composed of @var{length}
1533 repetitions of @var{type-information}, for example @code{character*3} is
1534 represented by @samp{M-2;3}, where @samp{-2} is a reference to a
1535 character type (@pxref{Negative Type Numbers}). I'm not sure how this
1536 differs from an array. This appears to be a Fortran feature.
1537 @var{length} is a bound, like those in range types; see @ref{Subranges}.
1539 @item S @var{type-information}
1540 Pascal set type. @var{type-information} must be a small type such as an
1541 enumeration or a subrange, and the type is a bitmask whose length is
1542 specified by the number of elements in @var{type-information}.
1544 @item * @var{type-information}
1545 Pointer to @var{type-information}.
1548 @node Cross-references
1549 @section Cross-references to other types
1551 If a type is used before it is defined, one common way to deal with this
1552 is just to use a type reference to a type which has not yet been
1553 defined. The debugger is expected to be able to deal with this.
1555 Another way is with the @samp{x} type descriptor, which is followed by
1556 @samp{s} for a structure tag, @samp{u} for a union tag, or @samp{e} for
1557 a enumerator tag, followed by the name of the tag, followed by @samp{:}.
1558 for example the following C declarations:
1568 .stabs "bar:G16=*17=xsfoo:",32,0,0,0
1571 Not all debuggers support the @samp{x} type descriptor, so on some
1572 machines GCC does not use it. I believe that for the above example it
1573 would just emit a reference to type 17 and never define it, but I
1574 haven't verified that.
1576 Modula-2 imported types, at least on AIX, use the @samp{i} type
1577 descriptor, which is followed by the name of the module from which the
1578 type is imported, followed by @samp{:}, followed by the name of the
1579 type. There is then optionally a comma followed by type information for
1580 the type (This differs from merely naming the type (@pxref{Typedefs}) in
1581 that it identifies the module; I don't understand whether the name of
1582 the type given here is always just the same as the name we are giving
1583 it, or whether this type descriptor is used with a nameless stab
1584 (@pxref{Stabs Format}), or what). The symbol ends with @samp{;}.
1587 @section Subrange types
1589 The @samp{r} type descriptor defines a type as a subrange of another
1590 type. It is followed by type information for the type which it is a
1591 subrange of, a semicolon, an integral lower bound, a semicolon, an
1592 integral upper bound, and a semicolon. The AIX documentation does not
1593 specify the trailing semicolon, in an effort to specify array indexes
1594 more cleanly, but a subrange which is not an array index has always
1595 included a trailing semicolon (@pxref{Arrays}).
1597 Instead of an integer, either bound can be one of the following:
1600 @item A @var{offset}
1601 The bound is passed by reference on the stack at offset @var{offset}
1602 from the argument list. @xref{Parameters}, for more information on such
1605 @item T @var{offset}
1606 The bound is passed by value on the stack at offset @var{offset} from
1609 @item a @var{register-number}
1610 The bound is pased by reference in register number
1611 @var{register-number}.
1613 @item t @var{register-number}
1614 The bound is passed by value in register number @var{register-number}.
1620 Subranges are also used for builtin types; see @ref{Traditional Builtin Types}.
1623 @section Array types
1625 Arrays use the @samp{a} type descriptor. Following the type descriptor
1626 is the type of the index and the type of the array elements. If the
1627 index type is a range type, it will end in a semicolon; if it is not a
1628 range type (for example, if it is a type reference), there does not
1629 appear to be any way to tell where the types are separated. In an
1630 effort to clean up this mess, IBM documents the two types as being
1631 separated by a semicolon, and a range type as not ending in a semicolon
1632 (but this is not right for range types which are not array indexes,
1633 @pxref{Subranges}). I think probably the best solution is to specify
1634 that a semicolon ends a range type, and that the index type and element
1635 type of an array are separated by a semicolon, but that if the index
1636 type is a range type, the extra semicolon can be omitted. GDB (at least
1637 through version 4.9) doesn't support any kind of index type other than a
1638 range anyway; I'm not sure about dbx.
1640 It is well established, and widely used, that the type of the index,
1641 unlike most types found in the stabs, is merely a type definition, not
1642 type information (@pxref{Stabs Format}) (that is, it need not start with
1643 @var{type-number}@code{=} if it is defining a new type). According to a
1644 comment in GDB, this is also true of the type of the array elements; it
1645 gives @samp{ar1;1;10;ar1;1;10;4} as a legitimate way to express a two
1646 dimensional array. According to AIX documentation, the element type
1647 must be type information. GDB accepts either.
1649 The type of the index is often a range type, expressed as the letter r
1650 and some parameters. It defines the size of the array. In the example
1651 below, the range @code{r1;0;2;} defines an index type which is a
1652 subrange of type 1 (integer), with a lower bound of 0 and an upper bound
1653 of 2. This defines the valid range of subscripts of a three-element C
1656 For example, the definition
1659 char char_vec[3] = @{'a','b','c'@};
1666 .stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
1675 If an array is @dfn{packed}, it means that the elements are spaced more
1676 closely than normal, saving memory at the expense of speed. For
1677 example, an array of 3-byte objects might, if unpacked, have each
1678 element aligned on a 4-byte boundary, but if packed, have no padding.
1679 One way to specify that something is packed is with type attributes
1680 (@pxref{Stabs Format}), in the case of arrays another is to use the
1681 @samp{P} type descriptor instead of @samp{a}. Other than specifying a
1682 packed array, @samp{P} is identical to @samp{a}.
1684 @c FIXME-what is it? A pointer?
1685 An open array is represented by the @samp{A} type descriptor followed by
1686 type information specifying the type of the array elements.
1688 @c FIXME: what is the format of this type? A pointer to a vector of pointers?
1689 An N-dimensional dynamic array is represented by
1692 D @var{dimensions} ; @var{type-information}
1695 @c Does dimensions really have this meaning? The AIX documentation
1697 @var{dimensions} is the number of dimensions; @var{type-information}
1698 specifies the type of the array elements.
1700 @c FIXME: what is the format of this type? A pointer to some offsets in
1702 A subarray of an N-dimensional array is represented by
1705 E @var{dimensions} ; @var{type-information}
1708 @c Does dimensions really have this meaning? The AIX documentation
1710 @var{dimensions} is the number of dimensions; @var{type-information}
1711 specifies the type of the array elements.
1716 Some languages, like C or the original Pascal, do not have string types,
1717 they just have related things like arrays of characters. But most
1718 Pascals and various other languages have string types, which are
1719 indicated as follows:
1722 @item n @var{type-information} ; @var{bytes}
1723 @var{bytes} is the maximum length. I'm not sure what
1724 @var{type-information} is; I suspect that it means that this is a string
1725 of @var{type-information} (thus allowing a string of integers, a string
1726 of wide characters, etc., as well as a string of characters). Not sure
1727 what the format of this type is. This is an AIX feature.
1729 @item z @var{type-information} ; @var{bytes}
1730 Just like @samp{n} except that this is a gstring, not an ordinary
1731 string. I don't know the difference.
1734 Pascal Stringptr. What is this? This is an AIX feature.
1738 @section Enumerations
1740 Enumerations are defined with the @samp{e} type descriptor.
1742 @c FIXME: Where does this information properly go? Perhaps it is
1743 @c redundant with something we already explain.
1744 The source line below declares an enumeration type. It is defined at
1745 file scope between the bodies of main and s_proc in example2.c.
1746 The type definition is located after the @code{N_RBRAC} that marks the end of
1747 the previous procedure's block scope, and before the @code{N_FUN} that marks
1748 the beginning of the next procedure's block scope. Therefore it does not
1749 describe a block local symbol, but a file local one.
1754 enum e_places @{first,second=3,last@};
1758 generates the following stab
1761 .stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0
1764 The symbol descriptor (T) says that the stab describes a structure,
1765 enumeration, or type tag. The type descriptor e, following the 22= of
1766 the type definition narrows it down to an enumeration type. Following
1767 the e is a list of the elements of the enumeration. The format is
1768 name:value,. The list of elements ends with a ;.
1770 There is no standard way to specify the size of an enumeration type; it
1771 is determined by the architecture (normally all enumerations types are
1772 32 bits). There should be a way to specify an enumeration type of
1773 another size; type attributes would be one way to do this @xref{Stabs
1783 @code{N_LSYM} or @code{C_DECL}
1784 @item Symbol Descriptor:
1786 @item Type Descriptor:
1790 The following source code declares a structure tag and defines an
1791 instance of the structure in global scope. Then a typedef equates the
1792 structure tag with a new type. A seperate stab is generated for the
1793 structure tag, the structure typedef, and the structure instance. The
1794 stabs for the tag and the typedef are emited when the definitions are
1795 encountered. Since the structure elements are not initialized, the
1796 stab and code for the structure variable itself is located at the end
1797 of the program in .common.
1803 9 char s_char_vec[8];
1804 10 struct s_tag* s_next;
1807 13 typedef struct s_tag s_typedef;
1810 The structure tag is an @code{N_LSYM} stab type because, like the enum, the
1811 symbol is file scope. Like the enum, the symbol descriptor is T, for
1812 enumeration, struct or tag type. The symbol descriptor s following
1813 the 16= of the type definition narrows the symbol type to struct.
1815 Following the struct symbol descriptor is the number of bytes the
1816 struct occupies, followed by a description of each structure element.
1817 The structure element descriptions are of the form name:type, bit
1818 offset from the start of the struct, and number of bits in the
1823 <128> N_LSYM - type definition
1824 .stabs "name:sym_desc(struct tag) Type_def(16)=type_desc(struct type)
1826 elem_name:type_ref(int),bit_offset,field_bits;
1827 elem_name:type_ref(float),bit_offset,field_bits;
1828 elem_name:type_def(17)=type_desc(array)
1829 index_type(range of int from 0 to 7);
1830 element_type(char),bit_offset,field_bits;;",
1833 30 .stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;
1834 s_char_vec:17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0
1837 In this example, two of the structure elements are previously defined
1838 types. For these, the type following the name: part of the element
1839 description is a simple type reference. The other two structure
1840 elements are new types. In this case there is a type definition
1841 embedded after the name:. The type definition for the array element
1842 looks just like a type definition for a standalone array. The s_next
1843 field is a pointer to the same kind of structure that the field is an
1844 element of. So the definition of structure type 16 contains an type
1845 definition for an element which is a pointer to type 16.
1848 @section Giving a Type a Name
1850 To give a type a name, use the @samp{t} symbol descriptor. The type
1851 specified by the type information (@pxref{Stabs Format}) for the stab.
1855 .stabs "s_typedef:t16",128,0,0,0
1858 specifies that @code{s_typedef} refers to type number 16. Such stabs
1859 have symbol type @code{N_LSYM} or @code{C_DECL}.
1861 If instead, you are specifying the tag name for a structure, union, or
1862 enumeration, use the @samp{T} symbol descriptor instead. I believe C is
1863 the only language with this feature.
1865 If the type is an opaque type (I believe this is a Modula-2 feature),
1866 AIX provides a type descriptor to specify it. The type descriptor is
1867 @samp{o} and is followed by a name. I don't know what the name
1868 means---is it always the same as the name of the type, or is this type
1869 descriptor used with a nameless stab (@pxref{Stabs Format})? There
1870 optionally follows a comma followed by type information which defines
1871 the type of this type. If omitted, a semicolon is used in place of the
1872 comma and the type information, and the type is much like a generic
1873 pointer type---it has a known size but little else about it is
1879 Next let's look at unions. In example2 this union type is declared
1880 locally to a procedure and an instance of the union is defined.
1890 This code generates a stab for the union tag and a stab for the union
1891 variable. Both use the @code{N_LSYM} stab type. Since the union variable is
1892 scoped locally to the procedure in which it is defined, its stab is
1893 located immediately preceding the @code{N_LBRAC} for the procedure's block
1896 The stab for the union tag, however is located preceding the code for
1897 the procedure in which it is defined. The stab type is @code{N_LSYM}. This
1898 would seem to imply that the union type is file scope, like the struct
1899 type s_tag. This is not true. The contents and position of the stab
1900 for u_type do not convey any infomation about its procedure local
1905 .stabs "name:sym_desc(union tag)type_def(22)=type_desc(union)
1907 elem_name:type_ref(int),bit_offset(0),bit_size(32);
1908 elem_name:type_ref(float),bit_offset(0),bit_size(32);
1909 elem_name:type_ref(ptr to char),bit_offset(0),bit_size(32);;"
1910 N_LSYM, NIL, NIL, NIL
1914 105 .stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
1918 The symbol descriptor, T, following the name: means that the stab
1919 describes an enumeration, struct or type tag. The type descriptor u,
1920 following the 23= of the type definition, narrows it down to a union
1921 type definition. Following the u is the number of bytes in the union.
1922 After that is a list of union element descriptions. Their format is
1923 name:type, bit offset into the union, and number of bytes for the
1926 The stab for the union variable follows.
1929 <128> N_LSYM - local variable (with no symbol descriptor)
1930 .stabs "name:type_ref(u_tag)", N_LSYM, NIL, NIL, frame_ptr_offset
1934 130 .stabs "an_u:23",128,0,0,-20
1937 @node Function Types
1938 @section Function types
1940 There are various types for function variables. These types are not
1941 used in defining functions; see symbol descriptor @samp{f}; they are
1942 used for things like pointers to functions.
1944 The simple, traditional, type is type descriptor @samp{f} is followed by
1945 type information for the return type of the function, followed by a
1948 This does not deal with functions the number and type of whose
1949 parameters are part of their type, as found in Modula-2 or ANSI C. AIX
1950 provides extensions to specify these, using the @samp{f}, @samp{F},
1951 @samp{p}, and @samp{R} type descriptors.
1953 First comes the type descriptor. Then, if it is @samp{f} or @samp{F},
1954 this is a function, and the type information for the return type of the
1955 function follows, followed by a comma. Then comes the number of
1956 parameters to the function and a semicolon. Then, for each parameter,
1957 there is the name of the parameter followed by a colon (this is only
1958 present for type descriptors @samp{R} and @samp{F} which represent
1959 Pascal function or procedure parameters), type information for the
1960 parameter, a comma, @samp{0} if passed by reference or @samp{1} if
1961 passed by value, and a semicolon. The type definition ends with a
1971 generates the following code:
1974 .stabs "g_pf:G24=*25=f1",32,0,0,0
1975 .common _g_pf,4,"bss"
1978 The variable defines a new type, 24, which is a pointer to another new
1979 type, 25, which is defined as a function returning int.
1982 @chapter Symbol information in symbol tables
1984 This chapter describes the format of symbol table entries
1985 and how stab assembler directives map to them. It also describes the
1986 transformations that the assembler and linker make on data from stabs.
1988 Each time the assembler encounters a stab in its input file it puts
1989 each field of the stab into corresponding fields in a symbol table
1990 entry of its output file. If the stab contains a string field, the
1991 symbol table entry for that stab points to a string table entry
1992 containing the string data from the stab. Assembler labels become
1993 relocatable addresses. Symbol table entries in a.out have the format:
1996 struct internal_nlist @{
1997 unsigned long n_strx; /* index into string table of name */
1998 unsigned char n_type; /* type of symbol */
1999 unsigned char n_other; /* misc info (usually empty) */
2000 unsigned short n_desc; /* description field */
2001 bfd_vma n_value; /* value of symbol */
2005 For @code{.stabs} directives, the @code{n_strx} field holds the character offset
2006 from the start of the string table to the string table entry
2007 containing the @var{string} field. For other classes of stabs (@code{.stabn} and
2008 @code{.stabd}) this field is null.
2010 Symbol table entries with @code{n_type} fields containing a value greater or
2011 equal to 0x20 originated as stabs generated by the compiler (with one
2012 random exception). Those with n_type values less than 0x20 were
2013 placed in the symbol table of the executable by the assembler or the
2016 The linker concatenates object files and does fixups of externally
2017 defined symbols. You can see the transformations made on stab data by
2018 the assembler and linker by examining the symbol table after each pass
2019 of the build, first the assemble and then the link.
2021 To do this, use @samp{nm -ap}. This dumps the symbol table, including
2022 debugging information, unsorted. For stab entries the columns are:
2023 @var{value}, @var{other}, @var{desc}, @var{type}, @var{string}. For
2024 assembler and linker symbols, the columns are: @var{value}, @var{type},
2027 There are a few important things to notice about symbol tables. Where
2028 the value field of a stab contains a frame pointer offset, or a
2029 register number, that value is unchanged by the rest of the build.
2031 Where the value field of a stab contains an assembly language label,
2032 it is transformed by each build step. The assembler turns it into a
2033 relocatable address and the linker turns it into an absolute address.
2034 This source line defines a static variable at file scope:
2037 3 static int s_g_repeat
2041 The following stab describes the symbol:
2044 26 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
2048 The assembler transforms the stab into this symbol table entry in the
2049 @file{.o} file. The location is expressed as a data segment offset.
2052 21 00000084 - 00 0000 STSYM s_g_repeat:S1
2056 in the symbol table entry from the executable, the linker has made the
2057 relocatable address absolute.
2060 22 0000e00c - 00 0000 STSYM s_g_repeat:S1
2063 Stabs for global variables do not contain location information. In
2064 this case the debugger finds location information in the assembler or
2065 linker symbol table entry describing the variable. The source line:
2075 21 .stabs "g_foo:G2",32,0,0,0
2078 The variable is represented by the following two symbol table entries
2079 in the object file. The first one originated as a stab. The second
2080 one is an external symbol. The upper case D signifies that the @code{n_type}
2081 field of the symbol table contains 7, @code{N_DATA} with local linkage.
2082 The value field following the file's line number is empty
2083 for the stab entry. For the linker symbol it contains the
2084 relocatable address corresponding to the variable.
2087 19 00000000 - 00 0000 GSYM g_foo:G2
2088 20 00000080 D _g_foo
2092 These entries as transformed by the linker. The linker symbol table
2093 entry now holds an absolute address.
2096 21 00000000 - 00 0000 GSYM g_foo:G2
2098 215 0000e008 D _g_foo
2102 @chapter GNU C++ stabs
2105 * Basic Cplusplus types::
2108 * Methods:: Method definition
2110 * Method Modifiers::
2113 * Virtual Base Classes::
2117 Type descriptors added for C++ descriptions:
2121 method type (@code{##} if minimal debug)
2124 Member (class and variable) type. It is followed by type information
2125 for the offset basetype, a comma, and type information for the type of
2126 the field being pointed to. (FIXME: this is acknowledged to be
2127 gibberish. Can anyone say what really goes here?).
2129 Note that there is a conflict between this and type attributes
2130 (@pxref{Stabs Format}); both use type descriptor @samp{@@}.
2131 Fortunately, the @samp{@@} type descriptor used in this C++ sense always
2132 will be followed by a digit, @samp{(}, or @samp{-}, and type attributes
2133 never start with those things.
2136 @node Basic Cplusplus types
2137 @section Basic types for C++
2139 << the examples that follow are based on a01.C >>
2142 C++ adds two more builtin types to the set defined for C. These are
2143 the unknown type and the vtable record type. The unknown type, type
2144 16, is defined in terms of itself like the void type.
2146 The vtable record type, type 17, is defined as a structure type and
2147 then as a structure tag. The structure has four fields: delta, index,
2148 pfn, and delta2. pfn is the function pointer.
2150 << In boilerplate $vtbl_ptr_type, what are the fields delta,
2151 index, and delta2 used for? >>
2153 This basic type is present in all C++ programs even if there are no
2154 virtual methods defined.
2157 .stabs "struct_name:sym_desc(type)type_def(17)=type_desc(struct)struct_bytes(8)
2158 elem_name(delta):type_ref(short int),bit_offset(0),field_bits(16);
2159 elem_name(index):type_ref(short int),bit_offset(16),field_bits(16);
2160 elem_name(pfn):type_def(18)=type_desc(ptr to)type_ref(void),
2161 bit_offset(32),field_bits(32);
2162 elem_name(delta2):type_def(short int);bit_offset(32),field_bits(16);;"
2167 .stabs "$vtbl_ptr_type:t17=s8
2168 delta:6,0,16;index:6,16,16;pfn:18=*15,32,32;delta2:6,32,16;;"
2173 .stabs "name:sym_dec(struct tag)type_ref($vtbl_ptr_type)",N_LSYM,NIL,NIL,NIL
2177 .stabs "$vtbl_ptr_type:T17",128,0,0,0
2180 @node Simple classes
2181 @section Simple class definition
2183 The stabs describing C++ language features are an extension of the
2184 stabs describing C. Stabs representing C++ class types elaborate
2185 extensively on the stab format used to describe structure types in C.
2186 Stabs representing class type variables look just like stabs
2187 representing C language variables.
2189 Consider the following very simple class definition.
2195 int Ameth(int in, char other);
2199 The class @code{baseA} is represented by two stabs. The first stab describes
2200 the class as a structure type. The second stab describes a structure
2201 tag of the class type. Both stabs are of stab type @code{N_LSYM}. Since the
2202 stab is not located between an @code{N_FUN} and a @code{N_LBRAC} stab this indicates
2203 that the class is defined at file scope. If it were, then the @code{N_LSYM}
2204 would signify a local variable.
2206 A stab describing a C++ class type is similar in format to a stab
2207 describing a C struct, with each class member shown as a field in the
2208 structure. The part of the struct format describing fields is
2209 expanded to include extra information relevent to C++ class members.
2210 In addition, if the class has multiple base classes or virtual
2211 functions the struct format outside of the field parts is also
2214 In this simple example the field part of the C++ class stab
2215 representing member data looks just like the field part of a C struct
2216 stab. The section on protections describes how its format is
2217 sometimes extended for member data.
2219 The field part of a C++ class stab representing a member function
2220 differs substantially from the field part of a C struct stab. It
2221 still begins with @samp{name:} but then goes on to define a new type number
2222 for the member function, describe its return type, its argument types,
2223 its protection level, any qualifiers applied to the method definition,
2224 and whether the method is virtual or not. If the method is virtual
2225 then the method description goes on to give the vtable index of the
2226 method, and the type number of the first base class defining the
2229 When the field name is a method name it is followed by two colons
2230 rather than one. This is followed by a new type definition for the
2231 method. This is a number followed by an equal sign and then the
2232 symbol descriptor @samp{##}, indicating a method type. This is followed by
2233 a type reference showing the return type of the method and a
2236 The format of an overloaded operator method name differs from that
2237 of other methods. It is @samp{op$::@var{XXXX}.} where @var{XXXX} is the operator name
2238 such as @samp{+} or @samp{+=}. The name ends with a period, and any characters except
2239 the period can occur in the @var{XXXX} string.
2241 The next part of the method description represents the arguments to
2242 the method, preceeded by a colon and ending with a semi-colon. The
2243 types of the arguments are expressed in the same way argument types
2244 are expressed in C++ name mangling. In this example an @code{int} and a @code{char}
2247 This is followed by a number, a letter, and an asterisk or period,
2248 followed by another semicolon. The number indicates the protections
2249 that apply to the member function. Here the 2 means public. The
2250 letter encodes any qualifier applied to the method definition. In
2251 this case, @samp{A} means that it is a normal function definition. The dot
2252 shows that the method is not virtual. The sections that follow
2253 elaborate further on these fields and describe the additional
2254 information present for virtual methods.
2258 .stabs "class_name:sym_desc(type)type_def(20)=type_desc(struct)struct_bytes(4)
2259 field_name(Adat):type(int),bit_offset(0),field_bits(32);
2261 method_name(Ameth)::type_def(21)=type_desc(method)return_type(int);
2262 :arg_types(int char);
2263 protection(public)qualifier(normal)virtual(no);;"
2268 .stabs "baseA:t20=s4Adat:1,0,32;Ameth::21=##1;:ic;2A.;;",128,0,0,0
2270 .stabs "class_name:sym_desc(struct tag)",N_LSYM,NIL,NIL,NIL
2272 .stabs "baseA:T20",128,0,0,0
2275 @node Class instance
2276 @section Class instance
2278 As shown above, describing even a simple C++ class definition is
2279 accomplished by massively extending the stab format used in C to
2280 describe structure types. However, once the class is defined, C stabs
2281 with no modifications can be used to describe class instances. The
2291 yields the following stab describing the class instance. It looks no
2292 different from a standard C stab describing a local variable.
2295 .stabs "name:type_ref(baseA)", N_LSYM, NIL, NIL, frame_ptr_offset
2299 .stabs "AbaseA:20",128,0,0,-20
2303 @section Method defintion
2305 The class definition shown above declares Ameth. The C++ source below
2310 baseA::Ameth(int in, char other)
2317 This method definition yields three stabs following the code of the
2318 method. One stab describes the method itself and following two describe
2319 its parameters. Although there is only one formal argument all methods
2320 have an implicit argument which is the @code{this} pointer. The @code{this}
2321 pointer is a pointer to the object on which the method was called. Note
2322 that the method name is mangled to encode the class name and argument
2323 types. Name mangling is described in the @sc{arm} (@cite{The Annotated
2324 C++ Reference Manual}, by Ellis and Stroustrup, @sc{isbn}
2325 0-201-51459-1); @file{gpcompare.texi} in Cygnus GCC distributions
2326 describes the differences between GNU mangling and @sc{arm}
2328 @c FIXME: Use @xref, especially if this is generally installed in the
2330 @c FIXME: This information should be in a net release, either of GCC or
2331 @c GDB. But gpcompare.texi doesn't seem to be in the FSF GCC.
2334 .stabs "name:symbol_desriptor(global function)return_type(int)",
2335 N_FUN, NIL, NIL, code_addr_of_method_start
2337 .stabs "Ameth__5baseAic:F1",36,0,0,_Ameth__5baseAic
2340 Here is the stab for the @code{this} pointer implicit argument. The
2341 name of the @code{this} pointer is always @code{this}. Type 19, the
2342 @code{this} pointer is defined as a pointer to type 20, @code{baseA},
2343 but a stab defining @code{baseA} has not yet been emited. Since the
2344 compiler knows it will be emited shortly, here it just outputs a cross
2345 reference to the undefined symbol, by prefixing the symbol name with
2349 .stabs "name:sym_desc(register param)type_def(19)=
2350 type_desc(ptr to)type_ref(baseA)=
2351 type_desc(cross-reference to)baseA:",N_RSYM,NIL,NIL,register_number
2353 .stabs "this:P19=*20=xsbaseA:",64,0,0,8
2356 The stab for the explicit integer argument looks just like a parameter
2357 to a C function. The last field of the stab is the offset from the
2358 argument pointer, which in most systems is the same as the frame
2362 .stabs "name:sym_desc(value parameter)type_ref(int)",
2363 N_PSYM,NIL,NIL,offset_from_arg_ptr
2365 .stabs "in:p1",160,0,0,72
2368 << The examples that follow are based on A1.C >>
2371 @section Protections
2374 In the simple class definition shown above all member data and
2375 functions were publicly accessable. The example that follows
2376 contrasts public, protected and privately accessable fields and shows
2377 how these protections are encoded in C++ stabs.
2379 Protections for class member data are signified by two characters
2380 embeded in the stab defining the class type. These characters are
2381 located after the name: part of the string. @samp{/0} means private, @samp{/1}
2382 means protected, and @samp{/2} means public. If these characters are omited
2383 this means that the member is public. The following C++ source:
2397 generates the following stab to describe the class type all_data.
2400 .stabs "class_name:sym_desc(type)type_def(19)=type_desc(struct)struct_bytes
2401 data_name:/protection(private)type_ref(int),bit_offset,num_bits;
2402 data_name:/protection(protected)type_ref(char),bit_offset,num_bits;
2403 data_name:(/num omited, private)type_ref(float),bit_offset,num_bits;;"
2408 .stabs "all_data:t19=s12
2409 priv_dat:/01,0,32;prot_dat:/12,32,8;pub_dat:12,64,32;;",128,0,0,0
2412 Protections for member functions are signified by one digit embeded in
2413 the field part of the stab describing the method. The digit is 0 if
2414 private, 1 if protected and 2 if public. Consider the C++ class
2418 class all_methods @{
2420 int priv_meth(int in)@{return in;@};
2422 char protMeth(char in)@{return in;@};
2424 float pubMeth(float in)@{return in;@};
2428 It generates the following stab. The digit in question is to the left
2429 of an @samp{A} in each case. Notice also that in this case two symbol
2430 descriptors apply to the class name struct tag and struct type.
2433 .stabs "class_name:sym_desc(struct tag&type)type_def(21)=
2434 sym_desc(struct)struct_bytes(1)
2435 meth_name::type_def(22)=sym_desc(method)returning(int);
2436 :args(int);protection(private)modifier(normal)virtual(no);
2437 meth_name::type_def(23)=sym_desc(method)returning(char);
2438 :args(char);protection(protected)modifier(normal)virual(no);
2439 meth_name::type_def(24)=sym_desc(method)returning(float);
2440 :args(float);protection(public)modifier(normal)virtual(no);;",
2445 .stabs "all_methods:Tt21=s1priv_meth::22=##1;:i;0A.;protMeth::23=##2;:c;1A.;
2446 pubMeth::24=##12;:f;2A.;;",128,0,0,0
2449 @node Method Modifiers
2450 @section Method Modifiers (const, volatile, const volatile)
2454 In the class example described above all the methods have the normal
2455 modifier. This method modifier information is located just after the
2456 protection information for the method. This field has four possible
2457 character values. Normal methods use @samp{A}, const methods use
2458 @samp{B}, volatile methods use @samp{C}, and const volatile methods use
2459 @samp{D}. Consider the class definition below:
2464 int ConstMeth (int arg) const @{ return arg; @};
2465 char VolatileMeth (char arg) volatile @{ return arg; @};
2466 float ConstVolMeth (float arg) const volatile @{return arg; @};
2470 This class is described by the following stab:
2473 .stabs "class(A):sym_desc(struct)type_def(20)=type_desc(struct)struct_bytes(1)
2474 meth_name(ConstMeth)::type_def(21)sym_desc(method)
2475 returning(int);:arg(int);protection(public)modifier(const)virtual(no);
2476 meth_name(VolatileMeth)::type_def(22)=sym_desc(method)
2477 returning(char);:arg(char);protection(public)modifier(volatile)virt(no)
2478 meth_name(ConstVolMeth)::type_def(23)=sym_desc(method)
2479 returning(float);:arg(float);protection(public)modifer(const volatile)
2480 virtual(no);;", @dots{}
2484 .stabs "A:T20=s1ConstMeth::21=##1;:i;2B.;VolatileMeth::22=##2;:c;2C.;
2485 ConstVolMeth::23=##12;:f;2D.;;",128,0,0,0
2488 @node Virtual Methods
2489 @section Virtual Methods
2491 << The following examples are based on a4.C >>
2493 The presence of virtual methods in a class definition adds additional
2494 data to the class description. The extra data is appended to the
2495 description of the virtual method and to the end of the class
2496 description. Consider the class definition below:
2502 virtual int A_virt (int arg) @{ return arg; @};
2506 This results in the stab below describing class A. It defines a new
2507 type (20) which is an 8 byte structure. The first field of the class
2508 struct is @samp{Adat}, an integer, starting at structure offset 0 and
2511 The second field in the class struct is not explicitly defined by the
2512 C++ class definition but is implied by the fact that the class
2513 contains a virtual method. This field is the vtable pointer. The
2514 name of the vtable pointer field starts with @samp{$vf} and continues with a
2515 type reference to the class it is part of. In this example the type
2516 reference for class A is 20 so the name of its vtable pointer field is
2517 @samp{$vf20}, followed by the usual colon.
2519 Next there is a type definition for the vtable pointer type (21).
2520 This is in turn defined as a pointer to another new type (22).
2522 Type 22 is the vtable itself, which is defined as an array, indexed by
2523 a range of integers between 0 and 1, and whose elements are of type
2524 17. Type 17 was the vtable record type defined by the boilerplate C++
2525 type definitions, as shown earlier.
2527 The bit offset of the vtable pointer field is 32. The number of bits
2528 in the field are not specified when the field is a vtable pointer.
2530 Next is the method definition for the virtual member function @code{A_virt}.
2531 Its description starts out using the same format as the non-virtual
2532 member functions described above, except instead of a dot after the
2533 @samp{A} there is an asterisk, indicating that the function is virtual.
2534 Since is is virtual some addition information is appended to the end
2535 of the method description.
2537 The first number represents the vtable index of the method. This is a
2538 32 bit unsigned number with the high bit set, followed by a
2541 The second number is a type reference to the first base class in the
2542 inheritence hierarchy defining the virtual member function. In this
2543 case the class stab describes a base class so the virtual function is
2544 not overriding any other definition of the method. Therefore the
2545 reference is to the type number of the class that the stab is
2548 This is followed by three semi-colons. One marks the end of the
2549 current sub-section, one marks the end of the method field, and the
2550 third marks the end of the struct definition.
2552 For classes containing virtual functions the very last section of the
2553 string part of the stab holds a type reference to the first base
2554 class. This is preceeded by @samp{~%} and followed by a final semi-colon.
2557 .stabs "class_name(A):type_def(20)=sym_desc(struct)struct_bytes(8)
2558 field_name(Adat):type_ref(int),bit_offset(0),field_bits(32);
2559 field_name(A virt func ptr):type_def(21)=type_desc(ptr to)type_def(22)=
2560 sym_desc(array)index_type_ref(range of int from 0 to 1);
2561 elem_type_ref(vtbl elem type),
2563 meth_name(A_virt)::typedef(23)=sym_desc(method)returning(int);
2564 :arg_type(int),protection(public)normal(yes)virtual(yes)
2565 vtable_index(1);class_first_defining(A);;;~%first_base(A);",
2569 @c FIXME: bogus line break.
2571 .stabs "A:t20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2572 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2576 @section Inheritence
2578 Stabs describing C++ derived classes include additional sections that
2579 describe the inheritence hierarchy of the class. A derived class stab
2580 also encodes the number of base classes. For each base class it tells
2581 if the base class is virtual or not, and if the inheritence is private
2582 or public. It also gives the offset into the object of the portion of
2583 the object corresponding to each base class.
2585 This additional information is embeded in the class stab following the
2586 number of bytes in the struct. First the number of base classes
2587 appears bracketed by an exclamation point and a comma.
2589 Then for each base type there repeats a series: two digits, a number,
2590 a comma, another number, and a semi-colon.
2592 The first of the two digits is 1 if the base class is virtual and 0 if
2593 not. The second digit is 2 if the derivation is public and 0 if not.
2595 The number following the first two digits is the offset from the start
2596 of the object to the part of the object pertaining to the base class.
2598 After the comma, the second number is a type_descriptor for the base
2599 type. Finally a semi-colon ends the series, which repeats for each
2602 The source below defines three base classes @code{A}, @code{B}, and
2603 @code{C} and the derived class @code{D}.
2610 virtual int A_virt (int arg) @{ return arg; @};
2616 virtual int B_virt (int arg) @{return arg; @};
2622 virtual int C_virt (int arg) @{return arg; @};
2625 class D : A, virtual B, public C @{
2628 virtual int A_virt (int arg ) @{ return arg+1; @};
2629 virtual int B_virt (int arg) @{ return arg+2; @};
2630 virtual int C_virt (int arg) @{ return arg+3; @};
2631 virtual int D_virt (int arg) @{ return arg; @};
2635 Class stabs similar to the ones described earlier are generated for
2638 @c FIXME!!! the linebreaks in the following example probably make the
2639 @c examples literally unusable, but I don't know any other way to get
2640 @c them on the page.
2641 @c One solution would be to put some of the type definitions into
2642 @c separate stabs, even if that's not exactly what the compiler actually
2645 .stabs "A:T20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2646 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2648 .stabs "B:Tt25=s8Bdat:1,0,32;$vf25:21,32;B_virt::26=##1;
2649 :i;2A*-2147483647;25;;;~%25;",128,0,0,0
2651 .stabs "C:Tt28=s8Cdat:1,0,32;$vf28:21,32;C_virt::29=##1;
2652 :i;2A*-2147483647;28;;;~%28;",128,0,0,0
2655 In the stab describing derived class @code{D} below, the information about
2656 the derivation of this class is encoded as follows.
2659 .stabs "derived_class_name:symbol_descriptors(struct tag&type)=
2660 type_descriptor(struct)struct_bytes(32)!num_bases(3),
2661 base_virtual(no)inheritence_public(no)base_offset(0),
2662 base_class_type_ref(A);
2663 base_virtual(yes)inheritence_public(no)base_offset(NIL),
2664 base_class_type_ref(B);
2665 base_virtual(no)inheritence_public(yes)base_offset(64),
2666 base_class_type_ref(C); @dots{}
2669 @c FIXME! fake linebreaks.
2671 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:
2672 1,160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt:
2673 :32:i;2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;
2674 28;;D_virt::32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2677 @node Virtual Base Classes
2678 @section Virtual Base Classes
2680 A derived class object consists of a concatination in memory of the
2681 data areas defined by each base class, starting with the leftmost and
2682 ending with the rightmost in the list of base classes. The exception
2683 to this rule is for virtual inheritence. In the example above, class
2684 @code{D} inherits virtually from base class @code{B}. This means that an instance
2685 of a @code{D} object will not contain its own @code{B} part but merely a pointer to
2686 a @code{B} part, known as a virtual base pointer.
2688 In a derived class stab, the base offset part of the derivation
2689 information, described above, shows how the base class parts are
2690 ordered. The base offset for a virtual base class is always given as
2691 0. Notice that the base offset for @code{B} is given as 0 even though @code{B} is
2692 not the first base class. The first base class @code{A} starts at offset 0.
2694 The field information part of the stab for class @code{D} describes the field
2695 which is the pointer to the virtual base class @code{B}. The vbase pointer
2696 name is @samp{$vb} followed by a type reference to the virtual base class.
2697 Since the type id for @code{B} in this example is 25, the vbase pointer name
2700 @c FIXME!! fake linebreaks below
2702 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:1,
2703 160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt::32:i;
2704 2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;28;;D_virt:
2705 :32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2708 Following the name and a semicolon is a type reference describing the
2709 type of the virtual base class pointer, in this case 24. Type 24 was
2710 defined earlier as the type of the @code{B} class @code{this} pointer. The
2711 @code{this} pointer for a class is a pointer to the class type.
2714 .stabs "this:P24=*25=xsB:",64,0,0,8
2717 Finally the field offset part of the vbase pointer field description
2718 shows that the vbase pointer is the first field in the @code{D} object,
2719 before any data fields defined by the class. The layout of a @code{D}
2720 class object is a follows, @code{Adat} at 0, the vtable pointer for
2721 @code{A} at 32, @code{Cdat} at 64, the vtable pointer for C at 96, the
2722 virtual base pointer for @code{B} at 128, and @code{Ddat} at 160.
2725 @node Static Members
2726 @section Static Members
2728 The data area for a class is a concatenation of the space used by the
2729 data members of the class. If the class has virtual methods, a vtable
2730 pointer follows the class data. The field offset part of each field
2731 description in the class stab shows this ordering.
2733 << How is this reflected in stabs? See Cygnus bug #677 for some info. >>
2736 @appendix Source code for extended example
2740 2 register int g_bar asm ("%g5");
2741 3 static int s_g_repeat = 2;
2747 9 char s_char_vec[8];
2748 10 struct s_tag* s_next;
2751 13 typedef struct s_tag s_typedef;
2753 15 char char_vec[3] = @{'a','b','c'@};
2755 17 main (argc, argv)
2759 21 static float s_flap;
2761 23 for (times=0; times < s_g_repeat; times++)@{
2763 25 printf ("Hello world\n");
2767 29 enum e_places @{first,second=3,last@};
2769 31 static s_proc (s_arg, s_ptr_arg, char_vec)
2771 33 s_typedef* s_ptr_arg;
2785 @appendix Assembly code for extended example
2789 2 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0
2790 3 .stabs "example2.c",100,0,0,Ltext0
2793 6 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
2794 7 .stabs "char:t2=r2;0;127;",128,0,0,0
2795 8 .stabs "long int:t3=r1;-2147483648;2147483647;",128,0,0,0
2796 9 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
2797 10 .stabs "long unsigned int:t5=r1;0;-1;",128,0,0,0
2798 11 .stabs "short int:t6=r1;-32768;32767;",128,0,0,0
2799 12 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
2800 13 .stabs "short unsigned int:t8=r1;0;65535;",128,0,0,0
2801 14 .stabs "long long unsigned int:t9=r1;0;-1;",128,0,0,0
2802 15 .stabs "signed char:t10=r1;-128;127;",128,0,0,0
2803 16 .stabs "unsigned char:t11=r1;0;255;",128,0,0,0
2804 17 .stabs "float:t12=r1;4;0;",128,0,0,0
2805 18 .stabs "double:t13=r1;8;0;",128,0,0,0
2806 19 .stabs "long double:t14=r1;8;0;",128,0,0,0
2807 20 .stabs "void:t15=15",128,0,0,0
2808 21 .stabs "g_foo:G2",32,0,0,0
2813 26 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
2817 @c FIXME! fake linebreak in line 30
2818 30 .stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;s_char_vec:
2819 17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0
2820 31 .stabs "s_typedef:t16",128,0,0,0
2821 32 .stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
2822 33 .global _char_vec
2828 39 .reserve _s_flap.0,4,"bss",4
2832 43 .ascii "Hello world\12\0"
2837 48 .stabn 68,0,20,LM1
2840 51 save %sp,-144,%sp
2847 58 .stabn 68,0,23,LM2
2851 62 sethi %hi(_s_g_repeat),%o0
2853 64 ld [%o0+%lo(_s_g_repeat)],%o0
2858 69 .stabn 68,0,25,LM3
2860 71 sethi %hi(LC0),%o1
2861 72 or %o1,%lo(LC0),%o0
2864 75 .stabn 68,0,26,LM4
2867 78 .stabn 68,0,23,LM5
2875 86 .stabn 68,0,27,LM6
2878 89 .stabn 68,0,27,LM7
2883 94 .stabs "main:F1",36,0,0,_main
2884 95 .stabs "argc:p1",160,0,0,68
2885 96 .stabs "argv:p20=*21=*2",160,0,0,72
2886 97 .stabs "s_flap:V12",40,0,0,_s_flap.0
2887 98 .stabs "times:1",128,0,0,-20
2888 99 .stabn 192,0,0,LBB2
2889 100 .stabs "inner:1",128,0,0,-24
2890 101 .stabn 192,0,0,LBB3
2891 102 .stabn 224,0,0,LBE3
2892 103 .stabn 224,0,0,LBE2
2893 104 .stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0
2894 @c FIXME: fake linebreak in line 105
2895 105 .stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
2900 109 .stabn 68,0,35,LM8
2903 112 save %sp,-120,%sp
2909 118 .stabn 68,0,41,LM9
2912 121 .stabn 68,0,41,LM10
2917 126 .stabs "s_proc:f1",36,0,0,_s_proc
2918 127 .stabs "s_arg:p16",160,0,0,0
2919 128 .stabs "s_ptr_arg:p18",160,0,0,72
2920 129 .stabs "char_vec:p21",160,0,0,76
2921 130 .stabs "an_u:23",128,0,0,-20
2922 131 .stabn 192,0,0,LBB4
2923 132 .stabn 224,0,0,LBE4
2924 133 .stabs "g_bar:r1",64,0,0,5
2925 134 .stabs "g_pf:G24=*25=f1",32,0,0,0
2926 135 .common _g_pf,4,"bss"
2927 136 .stabs "g_an_s:G16",32,0,0,0
2928 137 .common _g_an_s,20,"bss"
2932 @appendix Table of stab types
2934 The following are all the possible values for the stab type field, for
2935 @code{a.out} files, in numeric order. This does not apply to XCOFF.
2937 The symbolic names are defined in the file @file{include/aout/stabs.def}.
2940 * Non-stab symbol types::
2941 * Stab symbol types::
2944 @node Non-stab symbol types
2945 @appendixsec Non-stab symbol types
2947 The following types are used by the linker and assembler, not by stab
2948 directives. Since this document does not attempt to describe aspects of
2949 object file format other than the debugging format, no details are
2952 @c Try to get most of these to fit on a single line.
2962 File scope absolute symbol
2964 @item 0x3 N_ABS | N_EXT
2965 External absolute symbol
2968 File scope text symbol
2970 @item 0x5 N_TEXT | N_EXT
2971 External text symbol
2974 File scope data symbol
2976 @item 0x7 N_DATA | N_EXT
2977 External data symbol
2980 File scope BSS symbol
2982 @item 0x9 N_BSS | N_EXT
2986 Same as @code{N_FN}, for Sequent compilers
2989 Symbol is indirected to another symbol
2992 Common sym -- visable after shared lib dynamic link
2995 Absolute set element
2998 Text segment set element
3001 Data segment set element
3004 BSS segment set element
3007 Pointer to set vector
3009 @item 0x1e N_WARNING
3010 Print a warning message during linking
3013 File name of a @file{.o} file
3016 @node Stab symbol types
3017 @appendixsec Stab symbol types
3019 The following symbol types indicate that this is a stab. This is the
3020 full list of stab numbers, including stab types that are used in
3021 languages other than C.
3022 @xref{Expanded reference}, for more information about the stab types.
3026 Global symbol; see @ref{N_GSYM}.
3029 Function name (for BSD Fortran); see @ref{N_FNAME}.
3032 Function name (@pxref{Procedures}) or text segment variable
3036 Data segment file-scope variable; see @ref{Statics}.
3039 BSS segment file-scope variable; see @ref{Statics}.
3042 Name of main routine; see @ref{Main Program}.
3044 @c FIXME: discuss this in the Statics node where we talk about
3045 @c the fact that the n_type indicates the section.
3047 Variable in @code{.rodata} section; see @ref{Statics}.
3050 Global symbol (for Pascal); see @ref{N_PC}.
3053 Number of symbols (according to Ultrix V4.0); see @ref{N_NSYMS}.
3056 No DST map; see @ref{N_NOMAP}.
3058 @c FIXME: describe this solaris feature in the body of the text (see
3059 @c comments in include/aout/stab.def).
3061 Object file (Solaris2).
3063 @c See include/aout/stab.def for (a little) more info.
3065 Debugger options (Solaris2).
3068 Register variable; see @ref{N_RSYM}.
3071 Modula-2 compilation unit; see @ref{N_M2C}.
3074 Line number in text segment; see @ref{Line Numbers}.
3077 Line number in data segment; see @ref{Line Numbers}.
3080 Line number in bss segment; see @ref{Line Numbers}.
3083 Sun source code browser, path to @file{.cb} file; see @ref{N_BROWS}.
3086 GNU Modula2 definition module dependency; see @ref{N_DEFD}.
3089 Function start/body/end line numbers (Solaris2).
3092 GNU C++ exception variable; see @ref{N_EHDECL}.
3095 Modula2 info "for imc" (according to Ultrix V4.0); see @ref{N_MOD2}.
3098 GNU C++ @code{catch} clause; see @ref{N_CATCH}.
3101 Structure of union element; see @ref{N_SSYM}.
3104 Last stab for module (Solaris2).
3107 Path and name of source file; see @ref{Source Files}.
3110 Stack variable (@pxref{Stack Variables}) or type (@pxref{Typedefs}).
3113 Beginning of an include file (Sun only); see @ref{Source Files}.
3116 Name of include file; see @ref{Source Files}.
3119 Parameter variable; see @ref{Parameters}.
3122 End of an include file; see @ref{Source Files}.
3125 Alternate entry point; see @ref{N_ENTRY}.
3128 Beginning of a lexical block; see @ref{Block Structure}.
3131 Place holder for a deleted include file; see @ref{Source Files}.
3134 Modula2 scope information (Sun linker); see @ref{N_SCOPE}.
3137 End of a lexical block; see @ref{Block Structure}.
3140 Begin named common block; see @ref{Common Blocks}.
3143 End named common block; see @ref{Common Blocks}.
3146 Member of a common block; see @ref{Common Blocks}.
3148 @c FIXME: How does this really work? Move it to main body of document.
3150 Pascal @code{with} statement: type,,0,0,offset (Solaris2).
3153 Gould non-base registers; see @ref{Gould}.
3156 Gould non-base registers; see @ref{Gould}.
3159 Gould non-base registers; see @ref{Gould}.
3162 Gould non-base registers; see @ref{Gould}.
3165 Gould non-base registers; see @ref{Gould}.
3168 @c Restore the default table indent
3173 @node Symbol Descriptors
3174 @appendix Table of Symbol Descriptors
3176 These tell in the .stabs @var{string} field what kind of symbol the stab
3177 represents. They follow the symbol name and a colon. @xref{String
3178 Field}, for more information about their use.
3180 @c Please keep this alphabetical
3182 @c In TeX, this looks great, digit is in italics. But makeinfo insists
3183 @c on putting it in `', not realizing that @var should override @code.
3184 @c I don't know of any way to make makeinfo do the right thing. Seems
3185 @c like a makeinfo bug to me.
3189 Variable on the stack; see @ref{Stack Variables}.
3192 Parameter passed by reference in register; see @ref{Parameters}.
3195 Constant; see @ref{Constants}.
3198 Conformant array bound (Pascal, maybe other languages),
3199 @xref{Parameters}. Name of a caught exception (GNU C++). These can be
3200 distinguished because the latter uses N_CATCH and the former uses
3201 another symbol type.
3204 Floating point register variable; see @ref{Register variables}.
3207 Parameter in floating point register; see @ref{Parameters}.
3210 File scope function; see @ref{Procedures}.
3213 Global function; see @ref{Procedures}.
3216 Global variable; see @ref{Global Variables}.
3222 Internal (nested) procedure; see @ref{Procedures}.
3225 Internal (nested) function; see @ref{Procedures}.
3228 Label name (documented by AIX, no further information known).
3231 Module; see @ref{Procedures}.
3234 Argument list parameter; see @ref{Parameters}.
3240 Fortran Function parameter; see @ref{Parameters}.
3243 Unfortunately, three separate meanings have been independently invented
3244 for this symbol descriptor. At least the GNU and Sun uses can be
3245 distinguished by the symbol type. Global Procedure (AIX) (symbol type
3246 used unknown); see @ref{Procedures}. Register parameter (GNU) (symbol type
3247 N_PSYM); see @ref{Parameters}. Prototype of function referenced by this
3248 file (Sun acc) (symbol type N_FUN).
3251 Static Procedure; see @ref{Procedures}.
3254 Register parameter @xref{Parameters}.
3257 Register variable; see @ref{Register variables}.
3260 File scope variable; see @ref{Statics}.
3263 Type name; see @ref{Typedefs}.
3266 enumeration, struct or union tag; see @ref{Typedefs}.
3269 Parameter passed by reference; see @ref{Parameters}.
3272 Procedure scope static variable; see @ref{Statics}.
3275 Conformant array; see @ref{Parameters}.
3278 Function return variable; see @ref{Parameters}.
3281 @node Type Descriptors
3282 @appendix Table of Type Descriptors
3284 These tell in the .stabs @var{string} field what kind of type is being
3285 defined. They follow the type number and an equals sign.
3286 @xref{Overview}, for more information about their use.
3291 Type reference; see @ref{Stabs Format}.
3294 Reference to builtin type; see @ref{Negative Type Numbers}.
3297 Method (C++); see @ref{Cplusplus}.
3300 Pointer; see @ref{Miscellaneous Types}.
3306 Type Attributes (AIX); see @ref{Stabs Format}. Member (class and variable)
3307 type (GNU C++); see @ref{Cplusplus}.
3310 Array; see @ref{Arrays}.
3313 Open array; see @ref{Arrays}.
3316 Pascal space type (AIX); see @ref{Miscellaneous Types}. Builtin integer
3317 type (Sun); see @ref{Builtin Type Descriptors}.
3320 Volatile-qualified type; see @ref{Miscellaneous Types}.
3323 Complex builtin type; see @ref{Builtin Type Descriptors}.
3326 COBOL Picture type. See AIX documentation for details.
3329 File type; see @ref{Miscellaneous Types}.
3332 N-dimensional dynamic array; see @ref{Arrays}.
3335 Enumeration type; see @ref{Enumerations}.
3338 N-dimensional subarray; see @ref{Arrays}.
3341 Function type; see @ref{Function Types}.
3344 Pascal function parameter; see @ref{Function Types}
3347 Builtin floating point type; see @ref{Builtin Type Descriptors}.
3350 COBOL Group. See AIX documentation for details.
3353 Imported type; see @ref{Cross-references}.
3356 Const-qualified type; see @ref{Miscellaneous Types}.
3359 COBOL File Descriptor. See AIX documentation for details.
3362 Multiple instance type; see @ref{Miscellaneous Types}.
3365 String type; see @ref{Strings}.
3368 Stringptr; see @ref{Strings}.
3371 Opaque type; see @ref{Typedefs}.
3374 Procedure; see @ref{Function Types}.
3377 Packed array; see @ref{Arrays}.
3380 Range type; see @ref{Subranges}.
3383 Builtin floating type; see @ref{Builtin Type Descriptors} (Sun). Pascal
3384 subroutine parameter; see @ref{Function Types} (AIX). Detecting this
3385 conflict is possible with careful parsing (hint: a Pascal subroutine
3386 parameter type will always contain a comma, and a builtin type
3387 descriptor never will).
3390 Structure type; see @ref{Structures}.
3393 Set type; see @ref{Miscellaneous Types}.
3396 Union; see @ref{Unions}.
3399 Variant record. This is a Pascal and Modula-2 feature which is like a
3400 union within a struct in C. See AIX documentation for details.
3403 Wide character; see @ref{Builtin Type Descriptors}.
3406 Cross-reference; see @ref{Cross-references}.
3409 gstring; see @ref{Strings}.
3412 @node Expanded reference
3413 @appendix Expanded reference by stab type
3415 @c FIXME: This appendix should go away, see N_PSYM or N_SO for an example.
3417 For a full list of stab types, and cross-references to where they are
3418 described, see @ref{Stab Types}. This appendix just duplicates certain
3419 information from the main body of this document; eventually the
3420 information will all be in one place.
3424 The first line is the symbol type expressed in decimal, hexadecimal,
3425 and as a #define (see devo/include/aout/stab.def).
3427 The second line describes the language constructs the symbol type
3430 The third line is the stab format with the significant stab fields
3431 named and the rest NIL.
3433 Subsequent lines expand upon the meaning and possible values for each
3434 significant stab field. # stands in for the type descriptor.
3436 Finally, any further information.
3439 * N_GSYM:: Global variable
3440 * N_FNAME:: Function name (BSD Fortran)
3441 * N_PC:: Pascal global symbol
3442 * N_NSYMS:: Number of symbols
3443 * N_NOMAP:: No DST map
3444 * N_RSYM:: Register variable
3445 * N_M2C:: Modula-2 compilation unit
3446 * N_BROWS:: Path to .cb file for Sun source code browser
3447 * N_DEFD:: GNU Modula2 definition module dependency
3448 * N_EHDECL:: GNU C++ exception variable
3449 * N_MOD2:: Modula2 information "for imc"
3450 * N_CATCH:: GNU C++ "catch" clause
3451 * N_SSYM:: Structure or union element
3452 * N_ENTRY:: Alternate entry point
3453 * N_SCOPE:: Modula2 scope information (Sun only)
3454 * Gould:: non-base register symbols used on Gould systems
3455 * N_LENG:: Length of preceding entry
3459 @section 32 - 0x20 - N_GYSM
3464 .stabs "name", N_GSYM, NIL, NIL, NIL
3468 "name" -> "symbol_name:#type"
3472 Only the @var{name} field is significant. The location of the variable is
3473 obtained from the corresponding external symbol.
3476 @section 34 - 0x22 - N_FNAME
3477 Function name (for BSD Fortran)
3480 .stabs "name", N_FNAME, NIL, NIL, NIL
3484 "name" -> "function_name"
3487 Only the "name" field is significant. The location of the symbol is
3488 obtained from the corresponding extern symbol.
3491 @section 48 - 0x30 - N_PC
3492 Global symbol (for Pascal)
3495 .stabs "name", N_PC, NIL, NIL, value
3499 "name" -> "symbol_name" <<?>>
3500 value -> supposedly the line number (stab.def is skeptical)
3506 global pascal symbol: name,,0,subtype,line
3511 @section 50 - 0x32 - N_NSYMS
3512 Number of symbols (according to Ultrix V4.0)
3515 0, files,,funcs,lines (stab.def)
3519 @section 52 - 0x34 - N_NOMAP
3520 No DST map for symbol (according to Ultrix V4.0). I think this means a
3521 variable has been optimized out.
3524 name, ,0,type,ignored (stab.def)
3528 @section 64 - 0x40 - N_RSYM
3532 .stabs "name:type",N_RSYM,0,RegSize,RegNumber (Sun doc)
3536 @section 66 - 0x42 - N_M2C
3537 Modula-2 compilation unit
3540 .stabs "name", N_M2C, 0, desc, value
3544 "name" -> "unit_name,unit_time_stamp[,code_time_stamp]
3546 value -> 0 (main unit)
3551 @section 72 - 0x48 - N_BROWS
3552 Sun source code browser, path to @file{.cb} file
3555 "path to associated .cb file"
3557 Note: type field value overlaps with N_BSLINE
3560 @section 74 - 0x4a - N_DEFD
3561 GNU Modula2 definition module dependency
3563 GNU Modula-2 definition module dependency. Value is the modification
3564 time of the definition file. Other is non-zero if it is imported with
3565 the GNU M2 keyword %INITIALIZE. Perhaps N_M2C can be used if there
3566 are enough empty fields?
3569 @section 80 - 0x50 - N_EHDECL
3570 GNU C++ exception variable <<?>>
3572 "name is variable name"
3574 Note: conflicts with N_MOD2.
3577 @section 80 - 0x50 - N_MOD2
3578 Modula2 info "for imc" (according to Ultrix V4.0)
3580 Note: conflicts with N_EHDECL <<?>>
3583 @section 84 - 0x54 - N_CATCH
3584 GNU C++ @code{catch} clause
3586 GNU C++ @code{catch} clause. Value is its address. Desc is nonzero if
3587 this entry is immediately followed by a CAUGHT stab saying what
3588 exception was caught. Multiple CAUGHT stabs means that multiple
3589 exceptions can be caught here. If Desc is 0, it means all exceptions
3593 @section 96 - 0x60 - N_SSYM
3594 Structure or union element
3596 Value is offset in the structure.
3598 <<?looking at structs and unions in C I didn't see these>>
3601 @section 164 - 0xa4 - N_ENTRY
3603 Alternate entry point.
3604 Value is its address.
3608 @section 196 - 0xc4 - N_SCOPE
3610 Modula2 scope information (Sun linker)
3614 @section Non-base registers on Gould systems
3616 These are used on Gould systems for non-base registers syms.
3618 However, the following values are not the values used by Gould; they are
3619 the values which GNU has been documenting for these values for a long
3620 time, without actually checking what Gould uses. I include these values
3621 only because perhaps some someone actually did something with the GNU
3622 information (I hope not, why GNU knowingly assigned wrong values to
3623 these in the header file is a complete mystery to me).
3626 240 0xf0 N_NBTEXT ??
3627 242 0xf2 N_NBDATA ??
3634 @section - 0xfe - N_LENG
3636 Second symbol entry containing a length-value for the preceding entry.
3637 The value is the length.
3640 @appendix Questions and anomalies
3644 For GNU C stabs defining local and global variables (@code{N_LSYM} and
3645 @code{N_GSYM}), the @var{desc} field is supposed to contain the source line number
3646 on which the variable is defined. In reality the @var{desc} field is always
3647 0. (This behavior is defined in @file{dbxout.c} and putting a line number in
3648 @var{desc} is controlled by @samp{#ifdef WINNING_GDB}, which defaults to false). GDB
3649 supposedly uses this information if you say @samp{list @var{var}}. In reality,
3650 @var{var} can be a variable defined in the program and GDB says @samp{function
3651 @var{var} not defined}.
3654 In GNU C stabs, there seems to be no way to differentiate tag types:
3655 structures, unions, and enums (symbol descriptor @samp{T}) and typedefs
3656 (symbol descriptor @samp{t}) defined at file scope from types defined locally
3657 to a procedure or other more local scope. They all use the @code{N_LSYM}
3658 stab type. Types defined at procedure scope are emited after the
3659 @code{N_RBRAC} of the preceding function and before the code of the
3660 procedure in which they are defined. This is exactly the same as
3661 types defined in the source file between the two procedure bodies.
3662 GDB overcompensates by placing all types in block #1, the block for
3663 symbols of file scope. This is true for default, @samp{-ansi} and
3664 @samp{-traditional} compiler options. (Bugs gcc/1063, gdb/1066.)
3667 What ends the procedure scope? Is it the proc block's @code{N_RBRAC} or the
3668 next @code{N_FUN}? (I believe its the first.)
3671 @c FIXME: This should go with the other stuff about global variables.
3672 Global variable stabs don't have location information. This comes
3673 from the external symbol for the same variable. The external symbol
3674 has a leading underbar on the _name of the variable and the stab does
3675 not. How do we know these two symbol table entries are talking about
3676 the same symbol when their names are different? (Answer: the debugger
3677 knows that external symbols have leading underbars).
3679 @c FIXME: This is absurdly vague; there all kinds of differences, some
3680 @c of which are the same between gnu & sun, and some of which aren't.
3682 Can GCC be configured to output stabs the way the Sun compiler
3683 does, so that their native debugging tools work? <NO?> It doesn't by
3684 default. GDB reads either format of stab. (GCC or SunC). How about
3688 @node xcoff-differences
3689 @appendix Differences between GNU stabs in a.out and GNU stabs in xcoff
3691 @c FIXME: Merge *all* these into the main body of the document.
3692 The AIX/RS6000 native object file format is xcoff with stabs. This
3693 appendix only covers those differences which are not covered in the main
3694 body of this document.
3698 BSD a.out stab types correspond to AIX xcoff storage classes. In general the
3699 mapping is @code{N_STABTYPE} becomes @code{C_STABTYPE}. Some stab types in a.out
3700 are not supported in xcoff.
3702 @c FIXME: Get C_* types for the block, figure out whether it is always
3703 @c used (I suspect not), explain clearly, and move to node Statics.
3705 initialised static @code{N_STSYM} and un-initialized static @code{N_LCSYM} both map
3706 to the @code{C_STSYM} storage class. But the destinction is preserved
3707 because in xcoff @code{N_STSYM} and @code{N_LCSYM} must be emited in a named static
3708 block. Begin the block with @samp{.bs s[RW] data_section_name} for @code{N_STSYM}
3709 or @samp{.bs s bss_section_name} for @code{N_LCSYM}. End the block with @samp{.es}.
3711 @c FIXME: I think they are trying to say something about whether the
3712 @c assembler defaults the value to the location counter.
3714 If the xcoff stab is a @code{N_FUN} (@code{C_FUN}) then follow the string field with
3715 @samp{,.} instead of just @samp{,}.
3718 I think that's it for @file{.s} file differences. They could stand to be
3719 better presented. This is just a list of what I have noticed so far.
3720 There are a @emph{lot} of differences in the information in the symbol
3721 tables of the executable and object files.
3723 Mapping of a.out stab types to xcoff storage classes:
3726 stab type storage class
3727 -------------------------------
3736 N_RPSYM (0x8e) C_RPSYM
3746 N_DECL (0x8c) C_DECL
3763 @node Sun-differences
3764 @appendix Differences between GNU stabs and Sun native stabs
3766 @c FIXME: Merge all this stuff into the main body of the document.
3770 GNU C stabs define @emph{all} types, file or procedure scope, as
3771 @code{N_LSYM}. Sun doc talks about using @code{N_GSYM} too.
3774 Sun C stabs use type number pairs in the format (@var{a},@var{b}) where
3775 @var{a} is a number starting with 1 and incremented for each sub-source
3776 file in the compilation. @var{b} is a number starting with 1 and
3777 incremented for each new type defined in the compilation. GNU C stabs
3778 use the type number alone, with no source file number.
3782 @appendix Using stabs with the ELF object file format
3784 The ELF object file format allows tools to create object files with
3785 custom sections containing any arbitrary data. To use stabs in ELF
3786 object files, the tools create two custom sections, a section named
3787 @code{.stab} which contains an array of fixed length structures, one
3788 struct per stab, and a section named @code{.stabstr} containing all the
3789 variable length strings that are referenced by stabs in the @code{.stab}
3790 section. The byte order of the stabs binary data matches the byte order
3791 of the ELF file itself, as determined from the @code{EI_DATA} field in
3792 the @code{e_ident} member of the ELF header.
3794 @c Is "source file" the right term for this concept? We don't mean that
3795 @c there is a separate one for include files (but "object file" or
3796 @c "object module" isn't quite right either; the output from ld -r is a
3797 @c single object file but contains many source files).
3798 The first stab in the @code{.stab} section for each source file is
3799 synthetic, generated entirely by the assembler, with no corresponding
3800 @code{.stab} directive as input to the assembler. This stab contains
3801 the following fields:
3805 Offset in the @code{.stabstr} section to the source filename.
3811 Unused field, always zero.
3814 Count of upcoming symbols, i.e., the number of remaining stabs for this
3818 Size of the string table fragment associated with this source file, in
3822 The @code{.stabstr} section always starts with a null byte (so that string
3823 offsets of zero reference a null string), followed by random length strings,
3824 each of which is null byte terminated.
3826 The ELF section header for the @code{.stab} section has its
3827 @code{sh_link} member set to the section number of the @code{.stabstr}
3828 section, and the @code{.stabstr} section has its ELF section
3829 header @code{sh_type} member set to @code{SHT_STRTAB} to mark it as a
3832 Because the linker does not process the @code{.stab} section in any
3833 special way, none of the addresses in the @code{n_value} field of the
3834 stabs are relocated by the linker. Instead they are relative to the
3835 source file (or some entity smaller than a source file, like a
3836 function). To find the address of each section corresponding to a given
3837 source file, the (compiler? assembler?) puts out symbols giving the
3838 address of each section for a given source file. Since these are normal
3839 ELF symbols, the linker can relocate them correctly. They are
3840 named @code{Bbss.bss} for the bss section, @code{Ddata.data} for
3841 the data section, and @code{Drodata.rodata} for the rodata section. I
3842 haven't yet figured out how the debugger gets the address for the text