2 @setfilename stabs.info
7 * Stabs:: The "stabs" debugging information format.
13 This document describes the stabs debugging symbol tables.
15 Copyright 1992 Free Software Foundation, Inc.
16 Contributed by Cygnus Support. Written by Julia Menapace.
18 Permission is granted to make and distribute verbatim copies of
19 this manual provided the copyright notice and this permission notice
20 are preserved on all copies.
23 Permission is granted to process this file through Tex and print the
24 results, provided the printed document carries copying permission
25 notice identical to this one except for the removal of this paragraph
26 (this paragraph not being relevant to the printed manual).
29 Permission is granted to copy or distribute modified versions of this
30 manual under the terms of the GPL (for which purpose this text may be
31 regarded as a program in the language TeX).
34 @setchapternewpage odd
37 @title The ``stabs'' debug format
38 @author Julia Menapace
39 @author Cygnus Support
42 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
43 \xdef\manvers{\$Revision$} % For use in headers, footers too
45 \hfill Cygnus Support\par
47 \hfill \TeX{}info \texinfoversion\par
51 @vskip 0pt plus 1filll
52 Copyright @copyright{} 1992 Free Software Foundation, Inc.
53 Contributed by Cygnus Support.
55 Permission is granted to make and distribute verbatim copies of
56 this manual provided the copyright notice and this permission notice
57 are preserved on all copies.
63 @top The "stabs" representation of debugging information
65 This document describes the stabs debugging format.
68 * Overview:: Overview of stabs
69 * Program structure:: Encoding of the structure of the program
70 * Constants:: Constants
71 * Example:: A comprehensive example in C
73 * Types:: Type definitions
74 * Symbol Tables:: Symbol information in symbol tables
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
94 @chapter Overview of stabs
96 @dfn{Stabs} refers to a format for information that describes a program
97 to a debugger. This format was apparently invented by
98 @c FIXME! <<name of inventor>> at
99 the University of California at Berkeley, for the @code{pdx} Pascal
100 debugger; the format has spread widely since then.
102 This document is one of the few published sources of documentation on
103 stabs. It is believed to be completely comprehensive for stabs used by
104 C. The lists of symbol descriptors (@pxref{Symbol Descriptors}) and
105 type descriptors (@pxref{Type Descriptors}) are believed to be completely
106 comprehensive. There are known to be stabs for C++ and COBOL which are
107 poorly documented here. Stabs specific to other languages (e.g. Pascal,
108 Modula-2) are probably not as well documented as they should be.
110 Other sources of information on stabs are @cite{dbx and dbxtool
111 interfaces}, 2nd edition, by Sun, circa 1988, and @cite{AIX Version 3.2
112 Files Reference}, Fourth Edition, September 1992, "dbx Stabstring
113 Grammar" in the a.out section, page 2-31. This document is believed to
114 incorporate the information from those two sources except where it
115 explictly directs you to them for more information.
118 * Flow:: Overview of debugging information flow
119 * Stabs Format:: Overview of stab format
120 * C example:: A simple example in C source
121 * Assembly code:: The simple example at the assembly level
125 @section Overview of debugging information flow
127 The GNU C compiler compiles C source in a @file{.c} file into assembly
128 language in a @file{.s} file, which is translated by the assembler into
129 a @file{.o} file, and then linked with other @file{.o} files and
130 libraries to produce an executable file.
132 With the @samp{-g} option, GCC puts additional debugging information in
133 the @file{.s} file, which is slightly transformed by the assembler and
134 linker, and carried through into the final executable. This debugging
135 information describes features of the source file like line numbers,
136 the types and scopes of variables, and functions, their parameters and
139 For some object file formats, the debugging information is
140 encapsulated in assembler directives known collectively as `stab' (symbol
141 table) directives, interspersed with the generated code. Stabs are
142 the native format for debugging information in the a.out and xcoff
143 object file formats. The GNU tools can also emit stabs in the coff
144 and ecoff object file formats.
146 The assembler adds the information from stabs to the symbol information
147 it places by default in the symbol table and the string table of the
148 @file{.o} file it is building. The linker consolidates the @file{.o}
149 files into one executable file, with one symbol table and one string
150 table. Debuggers use the symbol and string tables in the executable as
151 a source of debugging information about the program.
154 @section Overview of stab format
156 There are three overall formats for stab assembler directives
157 differentiated by the first word of the stab. The name of the directive
158 describes what combination of four possible data fields will follow. It
159 is either @code{.stabs} (string), @code{.stabn} (number), or
160 @code{.stabd} (dot). IBM's xcoff uses @code{.stabx} (and some other
161 directives such as @code{.file} and @code{.bi}) instead of
162 @code{.stabs}, @code{.stabn} or @code{.stabd}.
164 The overall format of each class of stab is:
167 .stabs "@var{string}",@var{type},0,@var{desc},@var{value}
168 .stabx "@var{string}",@var{value},@var{type},@var{sdb-type}
169 .stabn @var{type},0,@var{desc},@var{value}
170 .stabd @var{type},0,@var{desc}
173 @c what is the correct term for "current file location"? My AIX
174 @c assembler manual calls it "the value of the current location counter".
175 For @code{.stabn} and @code{.stabd}, there is no string (the
176 @code{n_strx} field is zero, @pxref{Symbol Tables}). For @code{.stabd}
177 the value field is implicit and has the value of the current file
178 location. The @var{sdb-type} field to @code{.stabx} is unused for stabs
179 and can always be set to 0.
181 The number in the type field gives some basic information about what
182 type of stab this is (or whether it @emph{is} a stab, as opposed to an
183 ordinary symbol). Each possible type number defines a different stab
184 type. The stab type further defines the exact interpretation of, and
185 possible values for, any remaining @code{"@var{string}"}, @var{desc}, or
186 @var{value} fields present in the stab. @xref{Stab Types}, for a list
187 in numeric order of the possible type field values for stab directives.
189 For @code{.stabs} the @code{"@var{string}"} field holds the meat of the
190 debugging information. The generally unstructured nature of this field
191 is what makes stabs extensible. For some stab types the string field
192 contains only a name. For other stab types the contents can be a great
195 The overall format is of the @code{"@var{string}"} field is:
198 "@var{name}:@var{symbol-descriptor} @var{type-information}"
201 @var{name} is the name of the symbol represented by the stab.
202 @var{name} can be omitted, which means the stab represents an unnamed
203 object. For example, @samp{:t10=*2} defines type 10 as a pointer to
204 type 2, but does not give the type a name. Omitting the @var{name}
205 field is supported by AIX dbx and GDB after about version 4.8, but not
206 other debuggers. GCC sometimes uses a single space as the name instead
207 of omitting the name altogether; apparently that is supported by most
210 The @var{symbol_descriptor} following the @samp{:} is an alphabetic
211 character that tells more specifically what kind of symbol the stab
212 represents. If the @var{symbol_descriptor} is omitted, but type
213 information follows, then the stab represents a local variable. For a
214 list of symbol descriptors, see @ref{Symbol Descriptors,,Table C: Symbol
217 The @samp{c} symbol descriptor is an exception in that it is not
218 followed by type information. @xref{Constants}.
220 Type information is either a @var{type_number}, or a
221 @samp{@var{type_number}=}. The @var{type_number} alone is a type
222 reference, referring directly to a type that has already been defined.
224 The @samp{@var{type_number}=} is a type definition, where the number
225 represents a new type which is about to be defined. The type definition
226 may refer to other types by number, and those type numbers may be
227 followed by @samp{=} and nested definitions.
229 In a type definition, if the character that follows the equals sign is
230 non-numeric then it is a @var{type_descriptor}, and tells what kind of
231 type is about to be defined. Any other values following the
232 @var{type_descriptor} vary, depending on the @var{type_descriptor}. If
233 a number follows the @samp{=} then the number is a @var{type_reference}.
234 This is described more thoroughly in the section on types. @xref{Type
235 Descriptors,,Table D: Type Descriptors}, for a list of
236 @var{type_descriptor} values.
238 There is an AIX extension for type attributes. Following the @samp{=}
239 is any number of type attributes. Each one starts with @samp{@@} and
240 ends with @samp{;}. Debuggers, including AIX's dbx, skip any type
241 attributes they do not recognize. GDB 4.9 does not do this---it will
242 ignore the entire symbol containing a type attribute. Hopefully this
243 will be fixed in the next GDB release. Because of a conflict with C++
244 (@pxref{Cplusplus}), new attributes should not be defined which begin
245 with a digit, @samp{(}, or @samp{-}; GDB may be unable to distinguish
246 those from the C++ type descriptor @samp{@@}. The attributes are:
249 @item a@var{boundary}
250 @var{boundary} is an integer specifying the alignment. I assume it
251 applies to all variables of this type.
254 Size in bits of a variable of this type.
257 Pointer class (for checking). Not sure what this means, or how
258 @var{integer} is interpreted.
261 Indicate this is a packed type, meaning that structure fields or array
262 elements are placed more closely in memory, to save memory at the
266 All this can make the @code{"@var{string}"} field quite long. All
267 versions of GDB, and some versions of DBX, can handle arbitrarily long
268 strings. But many versions of DBX cretinously limit the strings to
269 about 80 characters, so compilers which must work with such DBX's need
270 to split the @code{.stabs} directive into several @code{.stabs}
271 directives. Each stab duplicates exactly all but the
272 @code{"@var{string}"} field. The @code{"@var{string}"} field of
273 every stab except the last is marked as continued with a
274 double-backslash at the end. Removing the backslashes and concatenating
275 the @code{"@var{string}"} fields of each stab produces the original,
279 @section A simple example in C source
281 To get the flavor of how stabs describe source information for a C
282 program, let's look at the simple program:
287 printf("Hello world");
291 When compiled with @samp{-g}, the program above yields the following
292 @file{.s} file. Line numbers have been added to make it easier to refer
293 to parts of the @file{.s} file in the description of the stabs that
297 @section The simple example at the assembly level
301 2 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0
302 3 .stabs "hello.c",100,0,0,Ltext0
305 6 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
306 7 .stabs "char:t2=r2;0;127;",128,0,0,0
307 8 .stabs "long int:t3=r1;-2147483648;2147483647;",128,0,0,0
308 9 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
309 10 .stabs "long unsigned int:t5=r1;0;-1;",128,0,0,0
310 11 .stabs "short int:t6=r1;-32768;32767;",128,0,0,0
311 12 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
312 13 .stabs "short unsigned int:t8=r1;0;65535;",128,0,0,0
313 14 .stabs "long long unsigned int:t9=r1;0;-1;",128,0,0,0
314 15 .stabs "signed char:t10=r1;-128;127;",128,0,0,0
315 16 .stabs "unsigned char:t11=r1;0;255;",128,0,0,0
316 17 .stabs "float:t12=r1;4;0;",128,0,0,0
317 18 .stabs "double:t13=r1;8;0;",128,0,0,0
318 19 .stabs "long double:t14=r1;8;0;",128,0,0,0
319 20 .stabs "void:t15=15",128,0,0,0
322 23 .ascii "Hello, world!\12\0"
337 38 sethi %hi(LC0),%o1
338 39 or %o1,%lo(LC0),%o0
349 50 .stabs "main:F1",36,0,0,_main
350 51 .stabn 192,0,0,LBB2
351 52 .stabn 224,0,0,LBE2
354 This simple ``hello world'' example demonstrates several of the stab
355 types used to describe C language source files.
357 @node Program structure
358 @chapter Encoding for the structure of the program
361 * Source Files:: The path and name of the source file
368 @section The path and name of the source files
370 Before any other stabs occur, there must be a stab specifying the source
371 file. This information is contained in a symbol of stab type
372 @code{N_SO}; the string contains the name of the file. The value of the
373 symbol is the start address of portion of the text section corresponding
376 With the Sun Solaris2 compiler, the @code{desc} field contains a
377 source-language code.
379 Some compilers (for example, gcc2 and SunOS4 @file{/bin/cc}) also
380 include the directory in which the source was compiled, in a second
381 @code{N_SO} symbol preceding the one containing the file name. This
382 symbol can be distinguished by the fact that it ends in a slash. Code
383 from the cfront C++ compiler can have additional @code{N_SO} symbols for
384 nonexistent source files after the @code{N_SO} for the real source file;
385 these are believed to contain no useful information.
390 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0 ; 100 is N_SO
391 .stabs "hello.c",100,0,0,Ltext0
396 Instead of @code{N_SO} symbols, XCOFF uses a @code{.file} assembler
397 directive which assembles to a standard COFF @code{.file} symbol;
398 explaining this in detail is outside the scope of this document.
400 There are several different schemes for dealing with include files: the
401 traditional @code{N_SOL} approach, Sun's @code{N_BINCL} scheme, and the
402 XCOFF @code{C_BINCL} (which despite the similar name has little in
403 common with @code{N_BINCL}).
405 An @code{N_SOL} symbol specifies which include file subsequent symbols
406 refer to. The string field is the name of the file and the value is the
407 text address corresponding to the start of the previous include file and
408 the start of this one. To specify the main source file again, use an
409 @code{N_SOL} symbol with the name of the main source file.
411 A @code{N_BINCL} symbol specifies the start of an include file. In an
412 object file, only the name is significant. The Sun linker puts data
413 into some of the other fields. The end of the include file is marked by
414 a @code{N_EINCL} symbol (which has no name field). In an ojbect file,
415 there is no significant data in the @code{N_EINCL} symbol; the Sun
416 linker puts data into some of the fields. @code{N_BINCL} and
417 @code{N_EINCL} can be nested. If the linker detects that two source
418 files have identical stabs with a @code{N_BINCL} and @code{N_EINCL} pair
419 (as will generally be the case for a header file), then it only puts out
420 the stabs once. Each additional occurance is replaced by an
421 @code{N_EXCL} symbol. I believe the Sun (SunOS4, not sure about
422 Solaris) linker is the only one which supports this feature.
424 For the start of an include file in XCOFF, use the @file{.bi} assembler
425 directive which generates a @code{C_BINCL} symbol. A @file{.ei}
426 directive, which generates a @code{C_EINCL} symbol, denotes the end of
427 the include file. Both directives are followed by the name of the
428 source file in quotes, which becomes the string for the symbol. The
429 value of each symbol, produced automatically by the assembler and
430 linker, is an offset into the executable which points to the beginning
431 (inclusive, as you'd expect) and end (inclusive, as you would not
432 expect) of the portion of the COFF linetable which corresponds to this
433 include file. @code{C_BINCL} and @code{C_EINCL} do not nest.
436 @section Line Numbers
438 A @code{N_SLINE} symbol represents the start of a source line. The
439 @var{desc} field contains the line number and the @var{value} field
440 contains the code address for the start of that source line. On most
441 machines the address is absolute; for Sun's stabs-in-elf, it is relative
442 to the function in which the @code{N_SLINE} symbol occurs.
444 GNU documents @code{N_DSLINE} and @code{N_BSLINE} symbols for line
445 numbers in the data or bss segments, respectively. They are identical
446 to @code{N_SLINE} but are relocated differently by the linker. They
447 were intended to be used to describe the source location of a variable
448 declaration, but I believe that gcc2 actually puts the line number in
449 the desc field of the stab for the variable itself. GDB has been
450 ignoring these symbols (unless they contain a string field) at least
453 XCOFF uses COFF line numbers instead, which are outside the scope of
454 this document, ammeliorated by adequate marking of include files
455 (@pxref{Source Files}).
457 For single source lines that generate discontiguous code, such as flow
458 of control statements, there may be more than one line number entry for
459 the same source line. In this case there is a line number entry at the
460 start of each code range, each with the same line number.
465 All of the following stabs use the @samp{N_FUN} symbol type.
467 A function is represented by a @samp{F} symbol descriptor for a global
468 (extern) function, and @samp{f} for a static (local) function. The next
469 @samp{N_SLINE} symbol can be used to find the line number of the start
470 of the function. The value field is the address of the start of the
471 function. The type information of the stab represents the return type
472 of the function; thus @samp{foo:f5} means that foo is a function
475 The type information of the stab is optionally followed by type
476 information for each argument, with each argument preceded by @samp{;}.
477 An argument type of 0 means that additional arguments are being passed,
478 whose types and number may vary (@samp{...} in ANSI C). This extension
479 is used by Sun's Solaris compiler. GDB has tolerated it (i.e. at least
480 parsed the syntax, if not necessarily used the information) at least
481 since version 4.8; I don't know whether all versions of dbx will
482 tolerate it. The argument types given here are not merely redundant
483 with the symbols for the arguments themselves (@pxref{Parameters}), they
484 are the types of the arguments as they are passed, before any
485 conversions might take place. For example, if a C function which is
486 declared without a prototype takes a @code{float} argument, the value is
487 passed as a @code{double} but then converted to a @code{float}.
488 Debuggers need to use the types given in the arguments when printing
489 values, but if calling the function they need to use the types given in
490 the symbol defining the function.
492 If the return type and types of arguments of a function which is defined
493 in another source file are specified (i.e. a function prototype in ANSI
494 C), traditionally compilers emit no stab; the only way for the debugger
495 to find the information is if the source file where the function is
496 defined was also compiled with debugging symbols. As an extension the
497 Solaris compiler uses symbol descriptor @samp{P} followed by the return
498 type of the function, followed by the arguments, each preceded by
499 @samp{;}, as in a stab with symbol descriptor @samp{f} or @samp{F}.
500 This use of symbol descriptor @samp{P} can be distinguished from its use
501 for register parameters (@pxref{Parameters}) by the fact that it has
502 symbol type @code{N_FUN}.
504 The AIX documentation also defines symbol descriptor @samp{J} as an
505 internal function. I assume this means a function nested within another
506 function. It also says Symbol descriptor @samp{m} is a module in
507 Modula-2 or extended Pascal.
509 Procedures (functions which do not return values) are represented as
510 functions returning the void type in C. I don't see why this couldn't
511 be used for all languages (inventing a void type for this purpose if
512 necessary), but the AIX documentation defines @samp{I}, @samp{P}, and
513 @samp{Q} for internal, global, and static procedures, respectively.
514 These symbol descriptors are unusual in that they are not followed by
517 For any of the above symbol descriptors, after the symbol descriptor and
518 the type information, there is optionally a comma, followed by the name
519 of the procedure, followed by a comma, followed by a name specifying the
520 scope. The first name is local to the scope specified. I assume then
521 that the name of the symbol (before the @samp{:}), if specified, is some
522 sort of global name. I assume the name specifying the scope is the name
523 of a function specifying that scope. This feature is an AIX extension,
524 and this information is based on the manual; I haven't actually tried
527 The stab representing a procedure is located immediately following the
528 code of the procedure. This stab is in turn directly followed by a
529 group of other stabs describing elements of the procedure. These other
530 stabs describe the procedure's parameters, its block local variables and
538 The @code{.stabs} entry after this code fragment shows the @var{name} of
539 the procedure (@code{main}); the type descriptor @var{desc} (@code{F},
540 for a global procedure); a reference to the predefined type @code{int}
541 for the return type; and the starting @var{address} of the procedure.
543 Here is an exploded summary (with whitespace introduced for clarity),
544 followed by line 50 of our sample assembly output, which has this form:
548 @var{desc} @r{(global proc @samp{F})}
549 @var{return_type_ref} @r{(int)}
555 50 .stabs "main:F1",36,0,0,_main
558 @node Block Structure
559 @section Block Structure
561 The program's block structure is represented by the @code{N_LBRAC} (left
562 brace) and the @code{N_RBRAC} (right brace) stab types. The variables
563 defined inside a block preceded the @code{N_LBRAC} symbol for most
564 compilers, including GCC. Other compilers, such as the Convex, Acorn
565 RISC machine, and Sun acc compilers, put the variables after the
566 @code{N_LBRAC} symbol. The values of the @code{N_LBRAC} and
567 @code{N_RBRAC} symbols are the start and end addresses of the code of
568 the block, respectively. For most machines, they are relative to the
569 starting address of this source file. For the Gould NP1, they are
570 absolute. For Sun's stabs-in-elf, they are relative to the function in
573 The @code{N_LBRAC} and @code{N_RBRAC} stabs that describe the block
574 scope of a procedure are located after the @code{N_FUN} stab that
575 represents the procedure itself.
577 Sun documents the @code{desc} field of @code{N_LBRAC} and
578 @code{N_RBRAC} symbols as containing the nesting level of the block.
579 However, dbx seems not to care, and GCC just always set @code{desc} to
585 The @samp{c} symbol descriptor indicates that this stab represents a
586 constant. This symbol descriptor is an exception to the general rule
587 that symbol descriptors are followed by type information. Instead, it
588 is followed by @samp{=} and one of the following:
592 Boolean constant. @var{value} is a numeric value; I assume it is 0 for
596 Character constant. @var{value} is the numeric value of the constant.
598 @item e @var{type-information} , @var{value}
599 Constant whose value can be represented as integral.
600 @var{type-information} is the type of the constant, as it would appear
601 after a symbol descriptor (@pxref{Stabs Format}). @var{value} is the
602 numeric value of the constant. GDB 4.9 does not actually get the right
603 value if @var{value} does not fit in a host @code{int}, but it does not
604 do anything violent, and future debuggers could be extended to accept
605 integers of any size (whether unsigned or not). This constant type is
606 usually documented as being only for enumeration constants, but GDB has
607 never imposed that restriction; I don't know about other debuggers.
610 Integer constant. @var{value} is the numeric value. The type is some
611 sort of generic integer type (for GDB, a host @code{int}); to specify
612 the type explicitly, use @samp{e} instead.
615 Real constant. @var{value} is the real value, which can be @samp{INF}
616 (optionally preceded by a sign) for infinity, @samp{QNAN} for a quiet
617 NaN (not-a-number), or @samp{SNAN} for a signalling NaN. If it is a
618 normal number the format is that accepted by the C library function
622 String constant. @var{string} is a string enclosed in either @samp{'}
623 (in which case @samp{'} characters within the string are represented as
624 @samp{\'} or @samp{"} (in which case @samp{"} characters within the
625 string are represented as @samp{\"}).
627 @item S @var{type-information} , @var{elements} , @var{bits} , @var{pattern}
628 Set constant. @var{type-information} is the type of the constant, as it
629 would appear after a symbol descriptor (@pxref{Stabs Format}).
630 @var{elements} is the number of elements in the set (Does this means
631 how many bits of @var{pattern} are actually used, which would be
632 redundant with the type, or perhaps the number of bits set in
633 @var{pattern}? I don't get it), @var{bits} is the number of bits in the
634 constant (meaning it specifies the length of @var{pattern}, I think),
635 and @var{pattern} is a hexadecimal representation of the set. AIX
636 documentation refers to a limit of 32 bytes, but I see no reason why
637 this limit should exist. This form could probably be used for arbitrary
638 constants, not just sets; the only catch is that @var{pattern} should be
639 understood to be target, not host, byte order and format.
642 The boolean, character, string, and set constants are not supported by
643 GDB 4.9, but it will ignore them. GDB 4.8 and earlier gave an error
644 message and refused to read symbols from the file containing the
647 This information is followed by @samp{;}.
650 @chapter A Comprehensive Example in C
652 Now we'll examine a second program, @code{example2}, which builds on the
653 first example to introduce the rest of the stab types, symbol
654 descriptors, and type descriptors used in C.
655 @xref{Example2.c} for the complete @file{.c} source,
656 and @pxref{Example2.s} for the @file{.s} assembly code.
657 This description includes parts of those files.
659 @section Flow of control and nested scopes
665 @code{N_SLINE}, @code{N_LBRAC}, @code{N_RBRAC} (cont.)
668 Consider the body of @code{main}, from @file{example2.c}. It shows more
669 about how @code{N_SLINE}, @code{N_RBRAC}, and @code{N_LBRAC} stabs are used.
673 21 static float s_flap;
675 23 for (times=0; times < s_g_repeat; times++)@{
677 25 printf ("Hello world\n");
682 Here we have a single source line, the @samp{for} line, that generates
683 non-linear flow of control, and non-contiguous code. In this case, an
684 @code{N_SLINE} stab with the same line number proceeds each block of
685 non-contiguous code generated from the same source line.
687 The example also shows nested scopes. The @code{N_LBRAC} and
688 @code{N_LBRAC} stabs that describe block structure are nested in the
689 same order as the corresponding code blocks, those of the for loop
690 inside those for the body of main.
693 This is the label for the @code{N_LBRAC} (left brace) stab marking the
694 start of @code{main}.
701 In the first code range for C source line 23, the @code{for} loop
702 initialize and test, @code{N_SLINE} (68) records the line number:
709 58 .stabn 68,0,23,LM2
713 62 sethi %hi(_s_g_repeat),%o0
715 64 ld [%o0+%lo(_s_g_repeat)],%o0
720 @exdent label for the @code{N_LBRAC} (start block) marking the start of @code{for} loop
723 69 .stabn 68,0,25,LM3
725 71 sethi %hi(LC0),%o1
726 72 or %o1,%lo(LC0),%o0
729 75 .stabn 68,0,26,LM4
732 @exdent label for the @code{N_RBRAC} (end block) stab marking the end of the @code{for} loop
738 Now we come to the second code range for source line 23, the @code{for}
739 loop increment and return. Once again, @code{N_SLINE} (68) records the
743 .stabn, N_SLINE, NIL,
747 78 .stabn 68,0,23,LM5
755 86 .stabn 68,0,27,LM6
758 @exdent label for the @code{N_RBRAC} (end block) stab marking the end of the @code{for} loop
761 89 .stabn 68,0,27,LM7
766 94 .stabs "main:F1",36,0,0,_main
767 95 .stabs "argc:p1",160,0,0,68
768 96 .stabs "argv:p20=*21=*2",160,0,0,72
769 97 .stabs "s_flap:V12",40,0,0,_s_flap.0
770 98 .stabs "times:1",128,0,0,-20
774 Here is an illustration of stabs describing nested scopes. The scope
775 nesting is reflected in the nested bracketing stabs (@code{N_LBRAC},
779 .stabn N_LBRAC,NIL,NIL,
780 @var{block-start-address}
782 99 .stabn 192,0,0,LBB2 ## begin proc label
783 100 .stabs "inner:1",128,0,0,-24
784 101 .stabn 192,0,0,LBB3 ## begin for label
788 @code{N_RBRAC} (224), ``right brace'' ends a lexical block (scope).
791 .stabn N_RBRAC,NIL,NIL,
792 @var{block-end-address}
794 102 .stabn 224,0,0,LBE3 ## end for label
795 103 .stabn 224,0,0,LBE2 ## end proc label
802 * Automatic variables:: locally scoped
804 * Register variables::
805 * Initialized statics::
806 * Un-initialized statics::
810 @node Automatic variables
811 @section Locally scoped automatic variables
818 @item Symbol Descriptor:
822 In addition to describing types, the @code{N_LSYM} stab type also
823 describes locally scoped automatic variables. Refer again to the body
824 of @code{main} in @file{example2.c}. It allocates two automatic
825 variables: @samp{times} is scoped to the body of @code{main}, and
826 @samp{inner} is scoped to the body of the @code{for} loop.
827 @samp{s_flap} is locally scoped but not automatic, and will be discussed
832 21 static float s_flap;
834 23 for (times=0; times < s_g_repeat; times++)@{
836 25 printf ("Hello world\n");
841 The @code{N_LSYM} stab for an automatic variable is located just before the
842 @code{N_LBRAC} stab describing the open brace of the block to which it is
846 @exdent @code{N_LSYM} (128): automatic variable, scoped locally to @code{main}
849 @var{type information}",
851 @var{frame-pointer-offset}
853 98 .stabs "times:1",128,0,0,-20
854 99 .stabn 192,0,0,LBB2 ## begin `main' N_LBRAC
856 @exdent @code{N_LSYM} (128): automatic variable, scoped locally to the @code{for} loop
859 @var{type information}",
861 @var{frame-pointer-offset}
863 100 .stabs "inner:1",128,0,0,-24
864 101 .stabn 192,0,0,LBB3 ## begin `for' loop N_LBRAC
867 The symbol descriptor is omitted for automatic variables. Since type
868 information should being with a digit, @samp{-}, or @samp{(}, only
869 digits, @samp{-}, and @samp{(} are precluded from being used for symbol
870 descriptors by this fact. However, the Acorn RISC machine (ARM) is said
871 to get this wrong: it puts out a mere type definition here, without the
872 preceding @code{@var{typenumber}=}. This is a bad idea; there is no
873 guarantee that type descriptors are distinct from symbol descriptors.
875 @node Global Variables
876 @section Global Variables
883 @item Symbol Descriptor:
887 Global variables are represented by the @code{N_GSYM} stab type. The symbol
888 descriptor, following the colon in the string field, is @samp{G}. Following
889 the @samp{G} is a type reference or type definition. In this example it is a
890 type reference to the basic C type, @code{char}. The first source line in
898 yields the following stab. The stab immediately precedes the code that
899 allocates storage for the variable it describes.
902 @exdent @code{N_GSYM} (32): global symbol
907 N_GSYM, NIL, NIL, NIL
909 21 .stabs "g_foo:G2",32,0,0,0
916 The address of the variable represented by the @code{N_GSYM} is not contained
917 in the @code{N_GSYM} stab. The debugger gets this information from the
918 external symbol for the global variable.
920 @node Register variables
921 @section Register variables
923 @c According to an old version of this manual, AIX uses C_RPSYM instead
924 @c of C_RSYM. I am skeptical; this should be verified.
925 Register variables have their own stab type, @code{N_RSYM}, and their
926 own symbol descriptor, @code{r}. The stab's value field contains the
927 number of the register where the variable data will be stored.
929 The value is the register number.
931 AIX defines a separate symbol descriptor @samp{d} for floating point
932 registers. This seems incredibly stupid---why not just just give
933 floating point registers different register numbers? I have not
934 verified whether the compiler actually uses @samp{d}.
936 If the register is explicitly allocated to a global variable, but not
940 register int g_bar asm ("%g5");
943 the stab may be emitted at the end of the object file, with
944 the other bss symbols.
946 @node Initialized statics
947 @section Initialized static variables
954 @item Symbol Descriptors:
955 @code{S} (file scope), @code{V} (procedure scope)
958 Initialized static variables are represented by the @code{N_STSYM} stab
959 type. The symbol descriptor part of the string field shows if the
960 variable is file scope static (@samp{S}) or procedure scope static
961 (@samp{V}). The source line
964 3 static int s_g_repeat = 2;
968 yields the following code. The stab is located immediately preceding
969 the storage for the variable it represents. Since the variable in
970 this example is file scope static the symbol descriptor is @samp{S}.
973 @exdent @code{N_STSYM} (38): initialized static variable (data seg w/internal linkage)
981 26 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
988 @node Un-initialized statics
989 @section Un-initialized static variables
996 @item Symbol Descriptors:
997 @code{S} (file scope), @code{V} (procedure scope)
1000 Un-initialized static variables are represented by the @code{N_LCSYM}
1001 stab type. The symbol descriptor part of the string shows if the
1002 variable is file scope static (@samp{S}) or procedure scope static
1003 (@samp{V}). In this example it is procedure scope static. The source
1004 line allocating @code{s_flap} immediately follows the open brace for the
1005 procedure @code{main}.
1009 21 static float s_flap;
1012 The code that reserves storage for the variable @code{s_flap} precedes the
1013 body of body of @code{main}.
1016 39 .reserve _s_flap.0,4,"bss",4
1019 But since @code{s_flap} is scoped locally to @code{main}, its stab is
1020 located with the other stabs representing symbols local to @code{main}.
1021 The stab for @code{s_flap} is located just before the @code{N_LBRAC} for
1025 @exdent @code{N_LCSYM} (40): uninitialized static var (BSS seg w/internal linkage)
1033 97 .stabs "s_flap:V12",40,0,0,_s_flap.0
1034 98 .stabs "times:1",128,0,0,-20
1035 99 .stabn 192,0,0,LBB2 # N_LBRAC for main.
1038 @c ............................................................
1043 The symbol descriptor @samp{p} is used to refer to parameters which are
1044 in the arglist. Symbols have symbol type @samp{N_PSYM}. The value of
1045 the symbol is the offset relative to the argument list.
1047 If the parameter is passed in a register, then the traditional way to do
1048 this is to provide two symbols for each argument:
1051 .stabs "arg:p1" . . . ; N_PSYM
1052 .stabs "arg:r1" . . . ; N_RSYM
1055 Debuggers are expected to use the second one to find the value, and the
1056 first one to know that it is an argument.
1058 Because this is kind of ugly, some compilers use symbol descriptor
1059 @samp{P} or @samp{R} to indicate an argument which is in a register.
1060 The symbol value is the register number. @samp{P} and @samp{R} mean the
1061 same thing, the difference is that @samp{P} is a GNU invention and
1062 @samp{R} is an IBM (xcoff) invention. As of version 4.9, GDB should
1063 handle either one. Symbol type @samp{C_RPSYM} is used with @samp{R} and
1064 @samp{N_RSYM} is used with @samp{P}.
1066 AIX, according to the documentation, uses @samp{D} for a parameter
1067 passed in a floating point register. This strikes me as incredibly
1068 bogus---why doesn't it just use @samp{R} with a register number which
1069 indicates that it's a floating point register? I haven't verified
1070 whether the system actually does what the documentation indicates.
1072 There is at least one case where GCC uses a @samp{p}/@samp{r} pair
1073 rather than @samp{P}; this is where the argument is passed in the
1074 argument list and then loaded into a register.
1076 On the sparc and hppa, for a @samp{P} symbol whose type is a structure
1077 or union, the register contains the address of the structure. On the
1078 sparc, this is also true of a @samp{p}/@samp{r} pair (using Sun cc) or a
1079 @samp{p} symbol. However, if a (small) structure is really in a
1080 register, @samp{r} is used. And, to top it all off, on the hppa it
1081 might be a structure which was passed on the stack and loaded into a
1082 register and for which there is a @samp{p}/@samp{r} pair! I believe
1083 that symbol descriptor @samp{i} is supposed to deal with this case, (it
1084 is said to mean "value parameter by reference, indirect access", I don't
1085 know the source for this information) but I don't know details or what
1086 compilers or debuggers use it, if any (not GDB or GCC). It is not clear
1087 to me whether this case needs to be dealt with differently than
1088 parameters passed by reference (see below).
1090 There is another case similar to an argument in a register, which is an
1091 argument which is actually stored as a local variable. Sometimes this
1092 happens when the argument was passed in a register and then the compiler
1093 stores it as a local variable. If possible, the compiler should claim
1094 that it's in a register, but this isn't always done. Some compilers use
1095 the pair of symbols approach described above ("arg:p" followed by
1096 "arg:"); this includes gcc1 (not gcc2) on the sparc when passing a small
1097 structure and gcc2 (sometimes) when the argument type is float and it is
1098 passed as a double and converted to float by the prologue (in the latter
1099 case the type of the "arg:p" symbol is double and the type of the "arg:"
1100 symbol is float). GCC, at least on the 960, uses a single @samp{p}
1101 symbol descriptor for an argument which is stored as a local variable
1102 but uses @samp{N_LSYM} instead of @samp{N_PSYM}. In this case the value
1103 of the symbol is an offset relative to the local variables for that
1104 function, not relative to the arguments (on some machines those are the
1105 same thing, but not on all).
1107 If the parameter is passed by reference (e.g. Pascal VAR parameters),
1108 then type symbol descriptor is @samp{v} if it is in the argument list,
1109 or @samp{a} if it in a register. Other than the fact that these contain
1110 the address of the parameter other than the parameter itself, they are
1111 identical to @samp{p} and @samp{R}, respectively. I believe @samp{a} is
1112 an AIX invention; @samp{v} is supported by all stabs-using systems as
1115 @c Is this paragraph correct? It is based on piecing together patchy
1116 @c information and some guesswork
1117 Conformant arrays refer to a feature of Modula-2, and perhaps other
1118 languages, in which the size of an array parameter is not known to the
1119 called function until run-time. Such parameters have two stabs, a
1120 @samp{x} for the array itself, and a @samp{C}, which represents the size
1121 of the array. The value of the @samp{x} stab is the offset in the
1122 argument list where the address of the array is stored (it this right?
1123 it is a guess); the value of the @samp{C} stab is the offset in the
1124 argument list where the size of the array (in elements? in bytes?) is
1127 The following are also said to go with @samp{N_PSYM}:
1130 "name" -> "param_name:#type"
1132 -> pF FORTRAN function parameter
1133 -> X (function result variable)
1134 -> b (based variable)
1136 value -> offset from the argument pointer (positive).
1139 As a simple example, the code
1151 .stabs "main:F1",36,0,0,_main ; 36 is N_FUN
1152 .stabs "argc:p1",160,0,0,68 ; 160 is N_PSYM
1153 .stabs "argv:p20=*21=*2",160,0,0,72
1156 The type definition of argv is interesting because it contains several
1157 type definitions. Type 21 is pointer to type 2 (char) and argv (type 20) is
1161 @chapter Type Definitions
1163 Now let's look at some variable definitions involving complex types.
1164 This involves understanding better how types are described. In the
1165 examples so far types have been described as references to previously
1166 defined types or defined in terms of subranges of or pointers to
1167 previously defined types. The section that follows will talk about
1168 the various other type descriptors that may follow the = sign in a
1172 * Builtin types:: Integers, floating point, void, etc.
1173 * Miscellaneous Types:: Pointers, sets, files, etc.
1174 * Cross-references:: Referring to a type not yet defined.
1175 * Subranges:: A type with a specific range.
1176 * Arrays:: An aggregate type of same-typed elements.
1177 * Strings:: Like an array but also has a length.
1178 * Enumerations:: Like an integer but the values have names.
1179 * Structures:: An aggregate type of different-typed elements.
1180 * Typedefs:: Giving a type a name.
1181 * Unions:: Different types sharing storage.
1186 @section Builtin types
1188 Certain types are built in (@code{int}, @code{short}, @code{void},
1189 @code{float}, etc.); the debugger recognizes these types and knows how
1190 to handle them. Thus don't be surprised if some of the following ways
1191 of specifying builtin types do not specify everything that a debugger
1192 would need to know about the type---in some cases they merely specify
1193 enough information to distinguish the type from other types.
1195 The traditional way to define builtin types is convolunted, so new ways
1196 have been invented to describe them. Sun's ACC uses the @samp{b} and
1197 @samp{R} type descriptors, and IBM uses negative type numbers. GDB can
1198 accept all three, as of version 4.8; dbx just accepts the traditional
1199 builtin types and perhaps one of the other two formats.
1202 * Traditional Builtin Types:: Put on your seatbelts and prepare for kludgery
1203 * Builtin Type Descriptors:: Builtin types with special type descriptors
1204 * Negative Type Numbers:: Builtin types using negative type numbers
1207 @node Traditional Builtin Types
1208 @subsection Traditional Builtin types
1210 Often types are defined as subranges of themselves. If the array bounds
1211 can fit within an @code{int}, then they are given normally. For example:
1214 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0 ; 128 is N_LSYM
1215 .stabs "char:t2=r2;0;127;",128,0,0,0
1218 Builtin types can also be described as subranges of @code{int}:
1221 .stabs "unsigned short:t6=r1;0;65535;",128,0,0,0
1224 If the lower bound of a subrange is 0 and the upper bound is -1, it
1225 means that the type is an unsigned integral type whose bounds are too
1226 big to describe in an int. Traditionally this is only used for
1227 @code{unsigned int} and @code{unsigned long}; GCC also sometimes uses it
1228 for @code{long long} and @code{unsigned long long}, and the only way to
1229 tell those types apart is to look at their names. On other machines GCC
1230 puts out bounds in octal, with a leading 0. In this case a negative
1231 bound consists of a number which is a 1 bit followed by a bunch of 0
1232 bits, and a positive bound is one in which a bunch of bits are 1.
1235 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
1236 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
1239 If the lower bound of a subrange is 0 and the upper bound is negative,
1240 it means that it is an unsigned integral type whose size in bytes is the
1241 absolute value of the upper bound. I believe this is a Convex
1242 convention for @code{unsigned long long}.
1244 If the lower bound of a subrange is negative and the upper bound is 0,
1245 it means that the type is a signed integral type whose size in bytes is
1246 the absolute value of the lower bound. I believe this is a Convex
1247 convention for @code{long long}. To distinguish this from a legitimate
1248 subrange, the type should be a subrange of itself. I'm not sure whether
1249 this is the case for Convex.
1251 If the upper bound of a subrange is 0, it means that this is a floating
1252 point type, and the lower bound of the subrange indicates the number of
1256 .stabs "float:t12=r1;4;0;",128,0,0,0
1257 .stabs "double:t13=r1;8;0;",128,0,0,0
1260 However, GCC writes @code{long double} the same way it writes
1261 @code{double}; the only way to distinguish them is by the name:
1264 .stabs "long double:t14=r1;8;0;",128,0,0,0
1267 Complex types are defined the same way as floating-point types; the only
1268 way to distinguish a single-precision complex from a double-precision
1269 floating-point type is by the name.
1271 The C @code{void} type is defined as itself:
1274 .stabs "void:t15=15",128,0,0,0
1277 I'm not sure how a boolean type is represented.
1279 @node Builtin Type Descriptors
1280 @subsection Defining Builtin Types using Builtin Type Descriptors
1282 There are various type descriptors to define builtin types:
1285 @c FIXME: clean up description of width and offset, once we figure out
1287 @item b @var{signed} @var{char-flag} @var{width} ; @var{offset} ; @var{nbits} ;
1288 Define an integral type. @var{signed} is @samp{u} for unsigned or
1289 @samp{s} for signed. @var{char-flag} is @samp{c} which indicates this
1290 is a character type, or is omitted. I assume this is to distinguish an
1291 integral type from a character type of the same size, for example it
1292 might make sense to set it for the C type @code{wchar_t} so the debugger
1293 can print such variables differently (Solaris does not do this). Sun
1294 sets it on the C types @code{signed char} and @code{unsigned char} which
1295 arguably is wrong. @var{width} and @var{offset} appear to be for small
1296 objects stored in larger ones, for example a @code{short} in an
1297 @code{int} register. @var{width} is normally the number of bytes in the
1298 type. @var{offset} seems to always be zero. @var{nbits} is the number
1299 of bits in the type.
1301 Note that type descriptor @samp{b} used for builtin types conflicts with
1302 its use for Pascal space types (@pxref{Miscellaneous Types}); they can
1303 be distinguished because the character following the type descriptor
1304 will be a digit, @samp{(}, or @samp{-} for a Pascal space type, or
1305 @samp{u} or @samp{s} for a builtin type.
1308 Documented by AIX to define a wide character type, but their compiler
1309 actually uses negative type numbers (@pxref{Negative Type Numbers}).
1311 @item R @var{fp_type} ; @var{bytes} ;
1312 Define a floating point type. @var{fp_type} has one of the following values:
1316 IEEE 32-bit (single precision) floating point format.
1319 IEEE 64-bit (double precision) floating point format.
1321 @item 3 (NF_COMPLEX)
1322 @item 4 (NF_COMPLEX16)
1323 @item 5 (NF_COMPLEX32)
1324 @c "GDB source" really means @file{include/aout/stab_gnu.h}, but trying
1325 @c to put that here got an overfull hbox.
1326 These are for complex numbers. A comment in the GDB source describes
1327 them as Fortran complex, double complex, and complex*16, respectively,
1328 but what does that mean? (i.e. Single precision? Double precison?).
1330 @item 6 (NF_LDOUBLE)
1331 Long double. This should probably only be used for Sun format long
1332 double, and new codes should be used for other floating point formats
1333 (NF_DOUBLE can be used if a long double is really just an IEEE double,
1337 @var{bytes} is the number of bytes occupied by the type. This allows a
1338 debugger to perform some operations with the type even if it doesn't
1339 understand @var{fp_code}.
1341 @item g @var{type-information} ; @var{nbits}
1342 Documented by AIX to define a floating type, but their compiler actually
1343 uses negative type numbers (@pxref{Negative Type Numbers}).
1345 @item c @var{type-information} ; @var{nbits}
1346 Documented by AIX to define a complex type, but their compiler actually
1347 uses negative type numbers (@pxref{Negative Type Numbers}).
1350 The C @code{void} type is defined as a signed integral type 0 bits long:
1352 .stabs "void:t19=bs0;0;0",128,0,0,0
1354 The Solaris compiler seems to omit the trailing semicolon in this case.
1355 Getting sloppy in this way is not a swift move because if a type is
1356 embedded in a more complex expression it is necessary to be able to tell
1359 I'm not sure how a boolean type is represented.
1361 @node Negative Type Numbers
1362 @subsection Negative Type numbers
1364 Since the debugger knows about the builtin types anyway, the idea of
1365 negative type numbers is simply to give a special type number which
1366 indicates the built in type. There is no stab defining these types.
1368 I'm not sure whether anyone has tried to define what this means if
1369 @code{int} can be other than 32 bits (or other types can be other than
1370 their customary size). If @code{int} has exactly one size for each
1371 architecture, then it can be handled easily enough, but if the size of
1372 @code{int} can vary according the compiler options, then it gets hairy.
1373 I guess the consistent way to do this would be to define separate
1374 negative type numbers for 16-bit @code{int} and 32-bit @code{int};
1375 therefore I have indicated below the customary size (and other format
1376 information) for each type. The information below is currently correct
1377 because AIX on the RS6000 is the only system which uses these type
1378 numbers. If these type numbers start to get used on other systems, I
1379 suspect the correct thing to do is to define a new number in cases where
1380 a type does not have the size and format indicated below.
1382 Also note that part of the definition of the negative type number is
1383 the name of the type. Types with identical size and format but
1384 different names have different negative type numbers.
1388 @code{int}, 32 bit signed integral type.
1391 @code{char}, 8 bit type holding a character. Both GDB and dbx on AIX
1392 treat this as signed. GCC uses this type whether @code{char} is signed
1393 or not, which seems like a bad idea. The AIX compiler (xlc) seems to
1394 avoid this type; it uses -5 instead for @code{char}.
1397 @code{short}, 16 bit signed integral type.
1400 @code{long}, 32 bit signed integral type.
1403 @code{unsigned char}, 8 bit unsigned integral type.
1406 @code{signed char}, 8 bit signed integral type.
1409 @code{unsigned short}, 16 bit unsigned integral type.
1412 @code{unsigned int}, 32 bit unsigned integral type.
1415 @code{unsigned}, 32 bit unsigned integral type.
1418 @code{unsigned long}, 32 bit unsigned integral type.
1421 @code{void}, type indicating the lack of a value.
1424 @code{float}, IEEE single precision.
1427 @code{double}, IEEE double precision.
1430 @code{long double}, IEEE double precision. The compiler claims the size
1431 will increase in a future release, and for binary compatibility you have
1432 to avoid using @code{long double}. I hope when they increase it they
1433 use a new negative type number.
1436 @code{integer}. 32 bit signed integral type.
1439 @code{boolean}. Only one bit is used, not sure about the actual size of the
1443 @code{short real}. IEEE single precision.
1446 @code{real}. IEEE double precision.
1449 @code{stringptr}. @xref{Strings}.
1452 @code{character}, 8 bit unsigned type.
1455 @code{logical*1}, 8 bit unsigned integral type.
1458 @code{logical*2}, 16 bit unsigned integral type.
1461 @code{logical*4}, 32 bit unsigned integral type.
1464 @code{logical}, 32 bit unsigned integral type.
1467 @code{complex}. A complex type consisting of two IEEE single-precision
1468 floating point values.
1471 @code{complex}. A complex type consisting of two IEEE double-precision
1472 floating point values.
1475 @code{integer*1}, 8 bit signed integral type.
1478 @code{integer*2}, 16 bit signed integral type.
1481 @code{integer*4}, 32 bit signed integral type.
1484 @code{wchar}. Wide character, 16 bits wide (Unicode format?). This is
1485 not used for the C type @code{wchar_t}.
1488 @node Miscellaneous Types
1489 @section Miscellaneous Types
1492 @item b @var{type-information} ; @var{bytes}
1493 Pascal space type. This is documented by IBM; what does it mean?
1495 Note that this use of the @samp{b} type descriptor can be distinguished
1496 from its use for builtin integral types (@pxref{Builtin Type
1497 Descriptors}) because the character following the type descriptor is
1498 always a digit, @samp{(}, or @samp{-}.
1500 @item B @var{type-information}
1501 A volatile-qualified version of @var{type-information}. This is a Sun
1502 extension. A volatile-qualified type means that references and stores
1503 to a variable of that type must not be optimized or cached; they must
1504 occur as the user specifies them.
1506 @item d @var{type-information}
1507 File of type @var{type-information}. As far as I know this is only used
1510 @item k @var{type-information}
1511 A const-qualified version of @var{type-information}. This is a Sun
1512 extension. A const-qualified type means that a variable of this type
1515 @item M @var{type-information} ; @var{length}
1516 Multiple instance type. The type seems to composed of @var{length}
1517 repetitions of @var{type-information}, for example @code{character*3} is
1518 represented by @samp{M-2;3}, where @samp{-2} is a reference to a
1519 character type (@pxref{Negative Type Numbers}). I'm not sure how this
1520 differs from an array. This appears to be a FORTRAN feature.
1521 @var{length} is a bound, like those in range types, @xref{Subranges}.
1523 @item S @var{type-information}
1524 Pascal set type. @var{type-information} must be a small type such as an
1525 enumeration or a subrange, and the type is a bitmask whose length is
1526 specified by the number of elements in @var{type-information}.
1528 @item * @var{type-information}
1529 Pointer to @var{type-information}.
1532 @node Cross-references
1533 @section Cross-references to other types
1535 If a type is used before it is defined, one common way to deal with this
1536 is just to use a type reference to a type which has not yet been
1537 defined. The debugger is expected to be able to deal with this.
1539 Another way is with the @samp{x} type descriptor, which is followed by
1540 @samp{s} for a structure tag, @samp{u} for a union tag, or @samp{e} for
1541 a enumerator tag, followed by the name of the tag, followed by @samp{:}.
1542 for example the following C declarations:
1552 .stabs "bar:G16=*17=xsfoo:",32,0,0,0
1555 Not all debuggers support the @samp{x} type descriptor, so on some
1556 machines GCC does not use it. I believe that for the above example it
1557 would just emit a reference to type 17 and never define it, but I
1558 haven't verified that.
1560 Modula-2 imported types, at least on AIX, use the @samp{i} type
1561 descriptor, which is followed by the name of the module from which the
1562 type is imported, followed by @samp{:}, followed by the name of the
1563 type. There is then optionally a comma followed by type information for
1564 the type (This differs from merely naming the type (@pxref{Typedefs}) in
1565 that it identifies the module; I don't understand whether the name of
1566 the type given here is always just the same as the name we are giving
1567 it, or whether this type descriptor is used with a nameless stab
1568 (@pxref{Stabs Format}), or what). The symbol ends with @samp{;}.
1571 @section Subrange types
1573 The @samp{r} type descriptor defines a type as a subrange of another
1574 type. It is followed by type information for the type which it is a
1575 subrange of, a semicolon, an integral lower bound, a semicolon, an
1576 integral upper bound, and a semicolon. The AIX documentation does not
1577 specify the trailing semicolon, in an effort to specify array indexes
1578 more cleanly, but a subrange which is not an array index has always
1579 included a trailing semicolon (@pxref{Arrays}).
1581 Instead of an integer, either bound can be one of the following:
1584 @item A @var{offset}
1585 The bound is passed by reference on the stack at offset @var{offset}
1586 from the argument list. @xref{Parameters}, for more information on such
1589 @item T @var{offset}
1590 The bound is passed by value on the stack at offset @var{offset} from
1593 @item a @var{register-number}
1594 The bound is pased by reference in register number
1595 @var{register-number}.
1597 @item t @var{register-number}
1598 The bound is passed by value in register number @var{register-number}.
1604 Subranges are also used for builtin types, @xref{Traditional Builtin Types}.
1607 @section Array types
1609 Arrays use the @samp{a} type descriptor. Following the type descriptor
1610 is the type of the index and the type of the array elements. If the
1611 index type is a range type, it will end in a semicolon; if it is not a
1612 range type (for example, if it is a type reference), there does not
1613 appear to be any way to tell where the types are separated. In an
1614 effort to clean up this mess, IBM documents the two types as being
1615 separated by a semicolon, and a range type as not ending in a semicolon
1616 (but this is not right for range types which are not array indexes,
1617 @pxref{Subranges}). I think probably the best solution is to specify
1618 that a semicolon ends a range type, and that the index type and element
1619 type of an array are separated by a semicolon, but that if the index
1620 type is a range type, the extra semicolon can be omitted. GDB (at least
1621 through version 4.9) doesn't support any kind of index type other than a
1622 range anyway; I'm not sure about dbx.
1624 It is well established, and widely used, that the type of the index,
1625 unlike most types found in the stabs, is merely a type definition, not
1626 type information (@pxref{Stabs Format}) (that is, it need not start with
1627 @var{type-number}@code{=} if it is defining a new type). According to a
1628 comment in GDB, this is also true of the type of the array elements; it
1629 gives @samp{ar1;1;10;ar1;1;10;4} as a legitimate way to express a two
1630 dimensional array. According to AIX documentation, the element type
1631 must be type information. GDB accepts either.
1633 The type of the index is often a range type, expressed as the letter r
1634 and some parameters. It defines the size of the array. In the example
1635 below, the range @code{r1;0;2;} defines an index type which is a
1636 subrange of type 1 (integer), with a lower bound of 0 and an upper bound
1637 of 2. This defines the valid range of subscripts of a three-element C
1640 For example, the definition
1643 char char_vec[3] = @{'a','b','c'@};
1650 .stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
1659 If an array is @dfn{packed}, it means that the elements are spaced more
1660 closely than normal, saving memory at the expense of speed. For
1661 example, an array of 3-byte objects might, if unpacked, have each
1662 element aligned on a 4-byte boundary, but if packed, have no padding.
1663 One way to specify that something is packed is with type attributes
1664 (@pxref{Stabs Format}), in the case of arrays another is to use the
1665 @samp{P} type descriptor instead of @samp{a}. Other than specifying a
1666 packed array, @samp{P} is identical to @samp{a}.
1668 @c FIXME-what is it? A pointer?
1669 An open array is represented by the @samp{A} type descriptor followed by
1670 type information specifying the type of the array elements.
1672 @c FIXME: what is the format of this type? A pointer to a vector of pointers?
1673 An N-dimensional dynamic array is represented by
1676 D @var{dimensions} ; @var{type-information}
1679 @c Does dimensions really have this meaning? The AIX documentation
1681 @var{dimensions} is the number of dimensions; @var{type-information}
1682 specifies the type of the array elements.
1684 @c FIXME: what is the format of this type? A pointer to some offsets in
1686 A subarray of an N-dimensional array is represented by
1689 E @var{dimensions} ; @var{type-information}
1692 @c Does dimensions really have this meaning? The AIX documentation
1694 @var{dimensions} is the number of dimensions; @var{type-information}
1695 specifies the type of the array elements.
1700 Some languages, like C or the original Pascal, do not have string types,
1701 they just have related things like arrays of characters. But most
1702 Pascals and various other languages have string types, which are
1703 indicated as follows:
1706 @item n @var{type-information} ; @var{bytes}
1707 @var{bytes} is the maximum length. I'm not sure what
1708 @var{type-information} is; I suspect that it means that this is a string
1709 of @var{type-information} (thus allowing a string of integers, a string
1710 of wide characters, etc., as well as a string of characters). Not sure
1711 what the format of this type is. This is an AIX feature.
1713 @item z @var{type-information} ; @var{bytes}
1714 Just like @samp{n} except that this is a gstring, not an ordinary
1715 string. I don't know the difference.
1718 Pascal Stringptr. What is this? This is an AIX feature.
1722 @section Enumerations
1724 Enumerations are defined with the @samp{e} type descriptor.
1726 @c FIXME: Where does this information properly go? Perhaps it is
1727 @c redundant with something we already explain.
1728 The source line below declares an enumeration type. It is defined at
1729 file scope between the bodies of main and s_proc in example2.c.
1730 The type definition is located after the N_RBRAC that marks the end of
1731 the previous procedure's block scope, and before the N_FUN that marks
1732 the beginning of the next procedure's block scope. Therefore it does not
1733 describe a block local symbol, but a file local one.
1738 enum e_places @{first,second=3,last@};
1742 generates the following stab
1745 .stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0
1748 The symbol descriptor (T) says that the stab describes a structure,
1749 enumeration, or type tag. The type descriptor e, following the 22= of
1750 the type definition narrows it down to an enumeration type. Following
1751 the e is a list of the elements of the enumeration. The format is
1752 name:value,. The list of elements ends with a ;.
1754 There is no standard way to specify the size of an enumeration type; it
1755 is determined by the architecture (normally all enumerations types are
1756 32 bits). There should be a way to specify an enumeration type of
1757 another size; type attributes would be one way to do this @xref{Stabs
1767 @code{N_LSYM} or @code{C_DECL}
1768 @item Symbol Descriptor:
1770 @item Type Descriptor:
1774 The following source code declares a structure tag and defines an
1775 instance of the structure in global scope. Then a typedef equates the
1776 structure tag with a new type. A seperate stab is generated for the
1777 structure tag, the structure typedef, and the structure instance. The
1778 stabs for the tag and the typedef are emited when the definitions are
1779 encountered. Since the structure elements are not initialized, the
1780 stab and code for the structure variable itself is located at the end
1781 of the program in .common.
1787 9 char s_char_vec[8];
1788 10 struct s_tag* s_next;
1791 13 typedef struct s_tag s_typedef;
1794 The structure tag is an N_LSYM stab type because, like the enum, the
1795 symbol is file scope. Like the enum, the symbol descriptor is T, for
1796 enumeration, struct or tag type. The symbol descriptor s following
1797 the 16= of the type definition narrows the symbol type to struct.
1799 Following the struct symbol descriptor is the number of bytes the
1800 struct occupies, followed by a description of each structure element.
1801 The structure element descriptions are of the form name:type, bit
1802 offset from the start of the struct, and number of bits in the
1807 <128> N_LSYM - type definition
1808 .stabs "name:sym_desc(struct tag) Type_def(16)=type_desc(struct type)
1810 elem_name:type_ref(int),bit_offset,field_bits;
1811 elem_name:type_ref(float),bit_offset,field_bits;
1812 elem_name:type_def(17)=type_desc(array)
1813 index_type(range of int from 0 to 7);
1814 element_type(char),bit_offset,field_bits;;",
1817 30 .stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;
1818 s_char_vec:17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0
1821 In this example, two of the structure elements are previously defined
1822 types. For these, the type following the name: part of the element
1823 description is a simple type reference. The other two structure
1824 elements are new types. In this case there is a type definition
1825 embedded after the name:. The type definition for the array element
1826 looks just like a type definition for a standalone array. The s_next
1827 field is a pointer to the same kind of structure that the field is an
1828 element of. So the definition of structure type 16 contains an type
1829 definition for an element which is a pointer to type 16.
1832 @section Giving a type a name
1834 To give a type a name, use the @samp{t} symbol descriptor. For example,
1837 .stabs "s_typedef:t16",128,0,0,0
1840 specifies that @code{s_typedef} refers to type number 16. Such stabs
1841 have symbol type @code{N_LSYM} or @code{C_DECL}.
1843 If instead, you are specifying the tag name for a structure, union, or
1844 enumeration, use the @samp{T} symbol descriptor instead. I believe C is
1845 the only language with this feature.
1847 If the type is an opaque type (I believe this is a Modula-2 feature),
1848 AIX provides a type descriptor to specify it. The type descriptor is
1849 @samp{o} and is followed by a name. I don't know what the name
1850 means---is it always the same as the name of the type, or is this type
1851 descriptor used with a nameless stab (@pxref{Stabs Format})? There
1852 optionally follows a comma followed by type information which defines
1853 the type of this type. If omitted, a semicolon is used in place of the
1854 comma and the type information, and, the type is much like a generic
1855 pointer type---it has a known size but little else about it is
1861 Next let's look at unions. In example2 this union type is declared
1862 locally to a procedure and an instance of the union is defined.
1872 This code generates a stab for the union tag and a stab for the union
1873 variable. Both use the N_LSYM stab type. Since the union variable is
1874 scoped locally to the procedure in which it is defined, its stab is
1875 located immediately preceding the N_LBRAC for the procedure's block
1878 The stab for the union tag, however is located preceding the code for
1879 the procedure in which it is defined. The stab type is N_LSYM. This
1880 would seem to imply that the union type is file scope, like the struct
1881 type s_tag. This is not true. The contents and position of the stab
1882 for u_type do not convey any infomation about its procedure local
1887 .stabs "name:sym_desc(union tag)type_def(22)=type_desc(union)
1889 elem_name:type_ref(int),bit_offset(0),bit_size(32);
1890 elem_name:type_ref(float),bit_offset(0),bit_size(32);
1891 elem_name:type_ref(ptr to char),bit_offset(0),bit_size(32);;"
1892 N_LSYM, NIL, NIL, NIL
1896 105 .stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
1900 The symbol descriptor, T, following the name: means that the stab
1901 describes an enumeration, struct or type tag. The type descriptor u,
1902 following the 23= of the type definition, narrows it down to a union
1903 type definition. Following the u is the number of bytes in the union.
1904 After that is a list of union element descriptions. Their format is
1905 name:type, bit offset into the union, and number of bytes for the
1908 The stab for the union variable follows. Notice that the frame
1909 pointer offset for local variables is negative.
1912 <128> N_LSYM - local variable (with no symbol descriptor)
1913 .stabs "name:type_ref(u_tag)", N_LSYM, NIL, NIL, frame_ptr_offset
1917 130 .stabs "an_u:23",128,0,0,-20
1920 @node Function Types
1921 @section Function types
1923 There are various types for function variables. These types are not
1924 used in defining functions; see symbol descriptor @samp{f}; they are
1925 used for things like pointers to functions.
1927 The simple, traditional, type is type descriptor @samp{f} is followed by
1928 type information for the return type of the function, followed by a
1931 This does not deal with functions the number and type of whose
1932 parameters are part of their type, as found in Modula-2 or ANSI C. AIX
1933 provides extensions to specify these, using the @samp{f}, @samp{F},
1934 @samp{p}, and @samp{R} type descriptors.
1936 First comes the type descriptor. Then, if it is @samp{f} or @samp{F},
1937 this is a function, and the type information for the return type of the
1938 function follows, followed by a comma. Then comes the number of
1939 parameters to the function and a semicolon. Then, for each parameter,
1940 there is the name of the parameter followed by a colon (this is only
1941 present for type descriptors @samp{R} and @samp{F} which represent
1942 Pascal function or procedure parameters), type information for the
1943 parameter, a comma, @samp{0} if passed by reference or @samp{1} if
1944 passed by value, and a semicolon. The type definition ends with a
1954 generates the following code:
1957 .stabs "g_pf:G24=*25=f1",32,0,0,0
1958 .common _g_pf,4,"bss"
1961 The variable defines a new type, 24, which is a pointer to another new
1962 type, 25, which is defined as a function returning int.
1965 @chapter Symbol information in symbol tables
1967 This section examines more closely the format of symbol table entries
1968 and how stab assembler directives map to them. It also describes what
1969 transformations the assembler and linker make on data from stabs.
1971 Each time the assembler encounters a stab in its input file it puts
1972 each field of the stab into corresponding fields in a symbol table
1973 entry of its output file. If the stab contains a string field, the
1974 symbol table entry for that stab points to a string table entry
1975 containing the string data from the stab. Assembler labels become
1976 relocatable addresses. Symbol table entries in a.out have the format:
1979 struct internal_nlist @{
1980 unsigned long n_strx; /* index into string table of name */
1981 unsigned char n_type; /* type of symbol */
1982 unsigned char n_other; /* misc info (usually empty) */
1983 unsigned short n_desc; /* description field */
1984 bfd_vma n_value; /* value of symbol */
1988 For .stabs directives, the n_strx field holds the character offset
1989 from the start of the string table to the string table entry
1990 containing the "string" field. For other classes of stabs (.stabn and
1991 .stabd) this field is null.
1993 Symbol table entries with n_type fields containing a value greater or
1994 equal to 0x20 originated as stabs generated by the compiler (with one
1995 random exception). Those with n_type values less than 0x20 were
1996 placed in the symbol table of the executable by the assembler or the
1999 The linker concatenates object files and does fixups of externally
2000 defined symbols. You can see the transformations made on stab data by
2001 the assembler and linker by examining the symbol table after each pass
2002 of the build, first the assemble and then the link.
2004 To do this use nm with the -ap options. This dumps the symbol table,
2005 including debugging information, unsorted. For stab entries the
2006 columns are: value, other, desc, type, string. For assembler and
2007 linker symbols, the columns are: value, type, string.
2009 There are a few important things to notice about symbol tables. Where
2010 the value field of a stab contains a frame pointer offset, or a
2011 register number, that value is unchanged by the rest of the build.
2013 Where the value field of a stab contains an assembly language label,
2014 it is transformed by each build step. The assembler turns it into a
2015 relocatable address and the linker turns it into an absolute address.
2016 This source line defines a static variable at file scope:
2019 3 static int s_g_repeat
2023 The following stab describes the symbol.
2026 26 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
2030 The assembler transforms the stab into this symbol table entry in the
2031 @file{.o} file. The location is expressed as a data segment offset.
2034 21 00000084 - 00 0000 STSYM s_g_repeat:S1
2038 in the symbol table entry from the executable, the linker has made the
2039 relocatable address absolute.
2042 22 0000e00c - 00 0000 STSYM s_g_repeat:S1
2045 Stabs for global variables do not contain location information. In
2046 this case the debugger finds location information in the assembler or
2047 linker symbol table entry describing the variable. The source line:
2057 21 .stabs "g_foo:G2",32,0,0,0
2060 The variable is represented by the following two symbol table entries
2061 in the object file. The first one originated as a stab. The second
2062 one is an external symbol. The upper case D signifies that the n_type
2063 field of the symbol table contains 7, N_DATA with local linkage (see
2064 Table B). The value field following the file's line number is empty
2065 for the stab entry. For the linker symbol it contains the
2066 rellocatable address corresponding to the variable.
2069 19 00000000 - 00 0000 GSYM g_foo:G2
2070 20 00000080 D _g_foo
2074 These entries as transformed by the linker. The linker symbol table
2075 entry now holds an absolute address.
2078 21 00000000 - 00 0000 GSYM g_foo:G2
2080 215 0000e008 D _g_foo
2084 @chapter GNU C++ stabs
2087 * Basic Cplusplus types::
2090 * Methods:: Method definition
2092 * Method Modifiers:: (const, volatile, const volatile)
2095 * Virtual Base Classes::
2099 @subsection type descriptors added for C++ descriptions
2103 method type (two ## if minimal debug)
2106 Member (class and variable) type. It is followed by type information
2107 for the offset basetype, a comma, and type information for the type of
2108 the field being pointed to. (FIXME: this is acknowledged to be
2109 gibberish. Can anyone say what really goes here?).
2111 Note that there is a conflict between this and type attributes
2112 (@pxref{Stabs Format}); both use type descriptor @samp{@@}.
2113 Fortunately, the @samp{@@} type descriptor used in this C++ sense always
2114 will be followed by a digit, @samp{(}, or @samp{-}, and type attributes
2115 never start with those things.
2118 @node Basic Cplusplus types
2119 @section Basic types for C++
2121 << the examples that follow are based on a01.C >>
2124 C++ adds two more builtin types to the set defined for C. These are
2125 the unknown type and the vtable record type. The unknown type, type
2126 16, is defined in terms of itself like the void type.
2128 The vtable record type, type 17, is defined as a structure type and
2129 then as a structure tag. The structure has four fields, delta, index,
2130 pfn, and delta2. pfn is the function pointer.
2132 << In boilerplate $vtbl_ptr_type, what are the fields delta,
2133 index, and delta2 used for? >>
2135 This basic type is present in all C++ programs even if there are no
2136 virtual methods defined.
2139 .stabs "struct_name:sym_desc(type)type_def(17)=type_desc(struct)struct_bytes(8)
2140 elem_name(delta):type_ref(short int),bit_offset(0),field_bits(16);
2141 elem_name(index):type_ref(short int),bit_offset(16),field_bits(16);
2142 elem_name(pfn):type_def(18)=type_desc(ptr to)type_ref(void),
2143 bit_offset(32),field_bits(32);
2144 elem_name(delta2):type_def(short int);bit_offset(32),field_bits(16);;"
2149 .stabs "$vtbl_ptr_type:t17=s8
2150 delta:6,0,16;index:6,16,16;pfn:18=*15,32,32;delta2:6,32,16;;"
2155 .stabs "name:sym_dec(struct tag)type_ref($vtbl_ptr_type)",N_LSYM,NIL,NIL,NIL
2159 .stabs "$vtbl_ptr_type:T17",128,0,0,0
2162 @node Simple classes
2163 @section Simple class definition
2165 The stabs describing C++ language features are an extension of the
2166 stabs describing C. Stabs representing C++ class types elaborate
2167 extensively on the stab format used to describe structure types in C.
2168 Stabs representing class type variables look just like stabs
2169 representing C language variables.
2171 Consider the following very simple class definition.
2177 int Ameth(int in, char other);
2181 The class baseA is represented by two stabs. The first stab describes
2182 the class as a structure type. The second stab describes a structure
2183 tag of the class type. Both stabs are of stab type N_LSYM. Since the
2184 stab is not located between an N_FUN and a N_LBRAC stab this indicates
2185 that the class is defined at file scope. If it were, then the N_LSYM
2186 would signify a local variable.
2188 A stab describing a C++ class type is similar in format to a stab
2189 describing a C struct, with each class member shown as a field in the
2190 structure. The part of the struct format describing fields is
2191 expanded to include extra information relevent to C++ class members.
2192 In addition, if the class has multiple base classes or virtual
2193 functions the struct format outside of the field parts is also
2196 In this simple example the field part of the C++ class stab
2197 representing member data looks just like the field part of a C struct
2198 stab. The section on protections describes how its format is
2199 sometimes extended for member data.
2201 The field part of a C++ class stab representing a member function
2202 differs substantially from the field part of a C struct stab. It
2203 still begins with `name:' but then goes on to define a new type number
2204 for the member function, describe its return type, its argument types,
2205 its protection level, any qualifiers applied to the method definition,
2206 and whether the method is virtual or not. If the method is virtual
2207 then the method description goes on to give the vtable index of the
2208 method, and the type number of the first base class defining the
2211 When the field name is a method name it is followed by two colons
2212 rather than one. This is followed by a new type definition for the
2213 method. This is a number followed by an equal sign and then the
2214 symbol descriptor `##', indicating a method type. This is followed by
2215 a type reference showing the return type of the method and a
2218 The format of an overloaded operator method name differs from that
2219 of other methods. It is "op$::XXXX." where XXXX is the operator name
2220 such as + or +=. The name ends with a period, and any characters except
2221 the period can occur in the XXXX string.
2223 The next part of the method description represents the arguments to
2224 the method, preceeded by a colon and ending with a semi-colon. The
2225 types of the arguments are expressed in the same way argument types
2226 are expressed in C++ name mangling. In this example an int and a char
2229 This is followed by a number, a letter, and an asterisk or period,
2230 followed by another semicolon. The number indicates the protections
2231 that apply to the member function. Here the 2 means public. The
2232 letter encodes any qualifier applied to the method definition. In
2233 this case A means that it is a normal function definition. The dot
2234 shows that the method is not virtual. The sections that follow
2235 elaborate further on these fields and describe the additional
2236 information present for virtual methods.
2240 .stabs "class_name:sym_desc(type)type_def(20)=type_desc(struct)struct_bytes(4)
2241 field_name(Adat):type(int),bit_offset(0),field_bits(32);
2243 method_name(Ameth)::type_def(21)=type_desc(method)return_type(int);
2244 :arg_types(int char);
2245 protection(public)qualifier(normal)virtual(no);;"
2250 .stabs "baseA:t20=s4Adat:1,0,32;Ameth::21=##1;:ic;2A.;;",128,0,0,0
2252 .stabs "class_name:sym_desc(struct tag)",N_LSYM,NIL,NIL,NIL
2254 .stabs "baseA:T20",128,0,0,0
2257 @node Class instance
2258 @section Class instance
2260 As shown above, describing even a simple C++ class definition is
2261 accomplished by massively extending the stab format used in C to
2262 describe structure types. However, once the class is defined, C stabs
2263 with no modifications can be used to describe class instances. The
2273 yields the following stab describing the class instance. It looks no
2274 different from a standard C stab describing a local variable.
2277 .stabs "name:type_ref(baseA)", N_LSYM, NIL, NIL, frame_ptr_offset
2281 .stabs "AbaseA:20",128,0,0,-20
2285 @section Method defintion
2287 The class definition shown above declares Ameth. The C++ source below
2292 baseA::Ameth(int in, char other)
2299 This method definition yields three stabs following the code of the
2300 method. One stab describes the method itself and following two
2301 describe its parameters. Although there is only one formal argument
2302 all methods have an implicit argument which is the `this' pointer.
2303 The `this' pointer is a pointer to the object on which the method was
2304 called. Note that the method name is mangled to encode the class name
2305 and argument types. << Name mangling is not described by this
2306 document - Is there already such a doc? >>
2309 .stabs "name:symbol_desriptor(global function)return_type(int)",
2310 N_FUN, NIL, NIL, code_addr_of_method_start
2312 .stabs "Ameth__5baseAic:F1",36,0,0,_Ameth__5baseAic
2315 Here is the stab for the `this' pointer implicit argument. The name
2316 of the `this' pointer is always `this.' Type 19, the `this' pointer is
2317 defined as a pointer to type 20, baseA, but a stab defining baseA has
2318 not yet been emited. Since the compiler knows it will be emited
2319 shortly, here it just outputs a cross reference to the undefined
2320 symbol, by prefixing the symbol name with xs.
2323 .stabs "name:sym_desc(register param)type_def(19)=
2324 type_desc(ptr to)type_ref(baseA)=
2325 type_desc(cross-reference to)baseA:",N_RSYM,NIL,NIL,register_number
2327 .stabs "this:P19=*20=xsbaseA:",64,0,0,8
2330 The stab for the explicit integer argument looks just like a parameter
2331 to a C function. The last field of the stab is the offset from the
2332 argument pointer, which in most systems is the same as the frame
2336 .stabs "name:sym_desc(value parameter)type_ref(int)",
2337 N_PSYM,NIL,NIL,offset_from_arg_ptr
2339 .stabs "in:p1",160,0,0,72
2342 << The examples that follow are based on A1.C >>
2345 @section Protections
2348 In the simple class definition shown above all member data and
2349 functions were publicly accessable. The example that follows
2350 contrasts public, protected and privately accessable fields and shows
2351 how these protections are encoded in C++ stabs.
2353 Protections for class member data are signified by two characters
2354 embeded in the stab defining the class type. These characters are
2355 located after the name: part of the string. /0 means private, /1
2356 means protected, and /2 means public. If these characters are omited
2357 this means that the member is public. The following C++ source:
2371 generates the following stab to describe the class type all_data.
2374 .stabs "class_name:sym_desc(type)type_def(19)=type_desc(struct)struct_bytes
2375 data_name:/protection(private)type_ref(int),bit_offset,num_bits;
2376 data_name:/protection(protected)type_ref(char),bit_offset,num_bits;
2377 data_name:(/num omited, private)type_ref(float),bit_offset,num_bits;;"
2382 .stabs "all_data:t19=s12
2383 priv_dat:/01,0,32;prot_dat:/12,32,8;pub_dat:12,64,32;;",128,0,0,0
2386 Protections for member functions are signified by one digit embeded in
2387 the field part of the stab describing the method. The digit is 0 if
2388 private, 1 if protected and 2 if public. Consider the C++ class
2392 class all_methods @{
2394 int priv_meth(int in)@{return in;@};
2396 char protMeth(char in)@{return in;@};
2398 float pubMeth(float in)@{return in;@};
2402 It generates the following stab. The digit in question is to the left
2403 of an `A' in each case. Notice also that in this case two symbol
2404 descriptors apply to the class name struct tag and struct type.
2407 .stabs "class_name:sym_desc(struct tag&type)type_def(21)=
2408 sym_desc(struct)struct_bytes(1)
2409 meth_name::type_def(22)=sym_desc(method)returning(int);
2410 :args(int);protection(private)modifier(normal)virtual(no);
2411 meth_name::type_def(23)=sym_desc(method)returning(char);
2412 :args(char);protection(protected)modifier(normal)virual(no);
2413 meth_name::type_def(24)=sym_desc(method)returning(float);
2414 :args(float);protection(public)modifier(normal)virtual(no);;",
2419 .stabs "all_methods:Tt21=s1priv_meth::22=##1;:i;0A.;protMeth::23=##2;:c;1A.;
2420 pubMeth::24=##12;:f;2A.;;",128,0,0,0
2423 @node Method Modifiers
2424 @section Method Modifiers (const, volatile, const volatile)
2428 In the class example described above all the methods have the normal
2429 modifier. This method modifier information is located just after the
2430 protection information for the method. This field has four possible
2431 character values. Normal methods use A, const methods use B, volatile
2432 methods use C, and const volatile methods use D. Consider the class
2438 int ConstMeth (int arg) const @{ return arg; @};
2439 char VolatileMeth (char arg) volatile @{ return arg; @};
2440 float ConstVolMeth (float arg) const volatile @{return arg; @};
2444 This class is described by the following stab:
2447 .stabs "class(A):sym_desc(struct)type_def(20)=type_desc(struct)struct_bytes(1)
2448 meth_name(ConstMeth)::type_def(21)sym_desc(method)
2449 returning(int);:arg(int);protection(public)modifier(const)virtual(no);
2450 meth_name(VolatileMeth)::type_def(22)=sym_desc(method)
2451 returning(char);:arg(char);protection(public)modifier(volatile)virt(no)
2452 meth_name(ConstVolMeth)::type_def(23)=sym_desc(method)
2453 returning(float);:arg(float);protection(public)modifer(const volatile)
2454 virtual(no);;", @dots{}
2458 .stabs "A:T20=s1ConstMeth::21=##1;:i;2B.;VolatileMeth::22=##2;:c;2C.;
2459 ConstVolMeth::23=##12;:f;2D.;;",128,0,0,0
2462 @node Virtual Methods
2463 @section Virtual Methods
2465 << The following examples are based on a4.C >>
2467 The presence of virtual methods in a class definition adds additional
2468 data to the class description. The extra data is appended to the
2469 description of the virtual method and to the end of the class
2470 description. Consider the class definition below:
2476 virtual int A_virt (int arg) @{ return arg; @};
2480 This results in the stab below describing class A. It defines a new
2481 type (20) which is an 8 byte structure. The first field of the class
2482 struct is Adat, an integer, starting at structure offset 0 and
2485 The second field in the class struct is not explicitly defined by the
2486 C++ class definition but is implied by the fact that the class
2487 contains a virtual method. This field is the vtable pointer. The
2488 name of the vtable pointer field starts with $vf and continues with a
2489 type reference to the class it is part of. In this example the type
2490 reference for class A is 20 so the name of its vtable pointer field is
2491 $vf20, followed by the usual colon.
2493 Next there is a type definition for the vtable pointer type (21).
2494 This is in turn defined as a pointer to another new type (22).
2496 Type 22 is the vtable itself, which is defined as an array, indexed by
2497 a range of integers between 0 and 1, and whose elements are of type
2498 17. Type 17 was the vtable record type defined by the boilerplate C++
2499 type definitions, as shown earlier.
2501 The bit offset of the vtable pointer field is 32. The number of bits
2502 in the field are not specified when the field is a vtable pointer.
2504 Next is the method definition for the virtual member function A_virt.
2505 Its description starts out using the same format as the non-virtual
2506 member functions described above, except instead of a dot after the
2507 `A' there is an asterisk, indicating that the function is virtual.
2508 Since is is virtual some addition information is appended to the end
2509 of the method description.
2511 The first number represents the vtable index of the method. This is a
2512 32 bit unsigned number with the high bit set, followed by a
2515 The second number is a type reference to the first base class in the
2516 inheritence hierarchy defining the virtual member function. In this
2517 case the class stab describes a base class so the virtual function is
2518 not overriding any other definition of the method. Therefore the
2519 reference is to the type number of the class that the stab is
2522 This is followed by three semi-colons. One marks the end of the
2523 current sub-section, one marks the end of the method field, and the
2524 third marks the end of the struct definition.
2526 For classes containing virtual functions the very last section of the
2527 string part of the stab holds a type reference to the first base
2528 class. This is preceeded by `~%' and followed by a final semi-colon.
2531 .stabs "class_name(A):type_def(20)=sym_desc(struct)struct_bytes(8)
2532 field_name(Adat):type_ref(int),bit_offset(0),field_bits(32);
2533 field_name(A virt func ptr):type_def(21)=type_desc(ptr to)type_def(22)=
2534 sym_desc(array)index_type_ref(range of int from 0 to 1);
2535 elem_type_ref(vtbl elem type),
2537 meth_name(A_virt)::typedef(23)=sym_desc(method)returning(int);
2538 :arg_type(int),protection(public)normal(yes)virtual(yes)
2539 vtable_index(1);class_first_defining(A);;;~%first_base(A);",
2543 @c FIXME: bogus line break.
2545 .stabs "A:t20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2546 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2550 @section Inheritence
2552 Stabs describing C++ derived classes include additional sections that
2553 describe the inheritence hierarchy of the class. A derived class stab
2554 also encodes the number of base classes. For each base class it tells
2555 if the base class is virtual or not, and if the inheritence is private
2556 or public. It also gives the offset into the object of the portion of
2557 the object corresponding to each base class.
2559 This additional information is embeded in the class stab following the
2560 number of bytes in the struct. First the number of base classes
2561 appears bracketed by an exclamation point and a comma.
2563 Then for each base type there repeats a series: two digits, a number,
2564 a comma, another number, and a semi-colon.
2566 The first of the two digits is 1 if the base class is virtual and 0 if
2567 not. The second digit is 2 if the derivation is public and 0 if not.
2569 The number following the first two digits is the offset from the start
2570 of the object to the part of the object pertaining to the base class.
2572 After the comma, the second number is a type_descriptor for the base
2573 type. Finally a semi-colon ends the series, which repeats for each
2576 The source below defines three base classes A, B, and C and the
2584 virtual int A_virt (int arg) @{ return arg; @};
2590 virtual int B_virt (int arg) @{return arg; @};
2596 virtual int C_virt (int arg) @{return arg; @};
2599 class D : A, virtual B, public C @{
2602 virtual int A_virt (int arg ) @{ return arg+1; @};
2603 virtual int B_virt (int arg) @{ return arg+2; @};
2604 virtual int C_virt (int arg) @{ return arg+3; @};
2605 virtual int D_virt (int arg) @{ return arg; @};
2609 Class stabs similar to the ones described earlier are generated for
2612 @c FIXME!!! the linebreaks in the following example probably make the
2613 @c examples literally unusable, but I don't know any other way to get
2614 @c them on the page.
2615 @c One solution would be to put some of the type definitions into
2616 @c separate stabs, even if that's not exactly what the compiler actually
2619 .stabs "A:T20=s8Adat:1,0,32;$vf20:21=*22=ar1;0;1;17,32;
2620 A_virt::23=##1;:i;2A*-2147483647;20;;;~%20;",128,0,0,0
2622 .stabs "B:Tt25=s8Bdat:1,0,32;$vf25:21,32;B_virt::26=##1;
2623 :i;2A*-2147483647;25;;;~%25;",128,0,0,0
2625 .stabs "C:Tt28=s8Cdat:1,0,32;$vf28:21,32;C_virt::29=##1;
2626 :i;2A*-2147483647;28;;;~%28;",128,0,0,0
2629 In the stab describing derived class D below, the information about
2630 the derivation of this class is encoded as follows.
2633 .stabs "derived_class_name:symbol_descriptors(struct tag&type)=
2634 type_descriptor(struct)struct_bytes(32)!num_bases(3),
2635 base_virtual(no)inheritence_public(no)base_offset(0),
2636 base_class_type_ref(A);
2637 base_virtual(yes)inheritence_public(no)base_offset(NIL),
2638 base_class_type_ref(B);
2639 base_virtual(no)inheritence_public(yes)base_offset(64),
2640 base_class_type_ref(C); @dots{}
2643 @c FIXME! fake linebreaks.
2645 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:
2646 1,160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt:
2647 :32:i;2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;
2648 28;;D_virt::32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2651 @node Virtual Base Classes
2652 @section Virtual Base Classes
2654 A derived class object consists of a concatination in memory of the
2655 data areas defined by each base class, starting with the leftmost and
2656 ending with the rightmost in the list of base classes. The exception
2657 to this rule is for virtual inheritence. In the example above, class
2658 D inherits virtually from base class B. This means that an instance
2659 of a D object will not contain it's own B part but merely a pointer to
2660 a B part, known as a virtual base pointer.
2662 In a derived class stab, the base offset part of the derivation
2663 information, described above, shows how the base class parts are
2664 ordered. The base offset for a virtual base class is always given as
2665 0. Notice that the base offset for B is given as 0 even though B is
2666 not the first base class. The first base class A starts at offset 0.
2668 The field information part of the stab for class D describes the field
2669 which is the pointer to the virtual base class B. The vbase pointer
2670 name is $vb followed by a type reference to the virtual base class.
2671 Since the type id for B in this example is 25, the vbase pointer name
2674 @c FIXME!! fake linebreaks below
2676 .stabs "D:Tt31=s32!3,000,20;100,25;0264,28;$vb25:24,128;Ddat:1,
2677 160,32;A_virt::32=##1;:i;2A*-2147483647;20;;B_virt::32:i;
2678 2A*-2147483647;25;;C_virt::32:i;2A*-2147483647;28;;D_virt:
2679 :32:i;2A*-2147483646;31;;;~%20;",128,0,0,0
2682 Following the name and a semicolon is a type reference describing the
2683 type of the virtual base class pointer, in this case 24. Type 24 was
2684 defined earlier as the type of the B class `this` pointer. The
2685 `this' pointer for a class is a pointer to the class type.
2688 .stabs "this:P24=*25=xsB:",64,0,0,8
2691 Finally the field offset part of the vbase pointer field description
2692 shows that the vbase pointer is the first field in the D object,
2693 before any data fields defined by the class. The layout of a D class
2694 object is a follows, Adat at 0, the vtable pointer for A at 32, Cdat
2695 at 64, the vtable pointer for C at 96, the virtual ase pointer for B
2696 at 128, and Ddat at 160.
2699 @node Static Members
2700 @section Static Members
2702 The data area for a class is a concatenation of the space used by the
2703 data members of the class. If the class has virtual methods, a vtable
2704 pointer follows the class data. The field offset part of each field
2705 description in the class stab shows this ordering.
2707 << How is this reflected in stabs? See Cygnus bug #677 for some info. >>
2710 @appendix Example2.c - source code for extended example
2714 2 register int g_bar asm ("%g5");
2715 3 static int s_g_repeat = 2;
2721 9 char s_char_vec[8];
2722 10 struct s_tag* s_next;
2725 13 typedef struct s_tag s_typedef;
2727 15 char char_vec[3] = @{'a','b','c'@};
2729 17 main (argc, argv)
2733 21 static float s_flap;
2735 23 for (times=0; times < s_g_repeat; times++)@{
2737 25 printf ("Hello world\n");
2741 29 enum e_places @{first,second=3,last@};
2743 31 static s_proc (s_arg, s_ptr_arg, char_vec)
2745 33 s_typedef* s_ptr_arg;
2759 @appendix Example2.s - assembly code for extended example
2763 2 .stabs "/cygint/s1/users/jcm/play/",100,0,0,Ltext0
2764 3 .stabs "example2.c",100,0,0,Ltext0
2767 6 .stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
2768 7 .stabs "char:t2=r2;0;127;",128,0,0,0
2769 8 .stabs "long int:t3=r1;-2147483648;2147483647;",128,0,0,0
2770 9 .stabs "unsigned int:t4=r1;0;-1;",128,0,0,0
2771 10 .stabs "long unsigned int:t5=r1;0;-1;",128,0,0,0
2772 11 .stabs "short int:t6=r1;-32768;32767;",128,0,0,0
2773 12 .stabs "long long int:t7=r1;0;-1;",128,0,0,0
2774 13 .stabs "short unsigned int:t8=r1;0;65535;",128,0,0,0
2775 14 .stabs "long long unsigned int:t9=r1;0;-1;",128,0,0,0
2776 15 .stabs "signed char:t10=r1;-128;127;",128,0,0,0
2777 16 .stabs "unsigned char:t11=r1;0;255;",128,0,0,0
2778 17 .stabs "float:t12=r1;4;0;",128,0,0,0
2779 18 .stabs "double:t13=r1;8;0;",128,0,0,0
2780 19 .stabs "long double:t14=r1;8;0;",128,0,0,0
2781 20 .stabs "void:t15=15",128,0,0,0
2782 21 .stabs "g_foo:G2",32,0,0,0
2787 26 .stabs "s_g_repeat:S1",38,0,0,_s_g_repeat
2791 @c FIXME! fake linebreak in line 30
2792 30 .stabs "s_tag:T16=s20s_int:1,0,32;s_float:12,32,32;s_char_vec:
2793 17=ar1;0;7;2,64,64;s_next:18=*16,128,32;;",128,0,0,0
2794 31 .stabs "s_typedef:t16",128,0,0,0
2795 32 .stabs "char_vec:G19=ar1;0;2;2",32,0,0,0
2796 33 .global _char_vec
2802 39 .reserve _s_flap.0,4,"bss",4
2806 43 .ascii "Hello world\12\0"
2811 48 .stabn 68,0,20,LM1
2814 51 save %sp,-144,%sp
2821 58 .stabn 68,0,23,LM2
2825 62 sethi %hi(_s_g_repeat),%o0
2827 64 ld [%o0+%lo(_s_g_repeat)],%o0
2832 69 .stabn 68,0,25,LM3
2834 71 sethi %hi(LC0),%o1
2835 72 or %o1,%lo(LC0),%o0
2838 75 .stabn 68,0,26,LM4
2841 78 .stabn 68,0,23,LM5
2849 86 .stabn 68,0,27,LM6
2852 89 .stabn 68,0,27,LM7
2857 94 .stabs "main:F1",36,0,0,_main
2858 95 .stabs "argc:p1",160,0,0,68
2859 96 .stabs "argv:p20=*21=*2",160,0,0,72
2860 97 .stabs "s_flap:V12",40,0,0,_s_flap.0
2861 98 .stabs "times:1",128,0,0,-20
2862 99 .stabn 192,0,0,LBB2
2863 100 .stabs "inner:1",128,0,0,-24
2864 101 .stabn 192,0,0,LBB3
2865 102 .stabn 224,0,0,LBE3
2866 103 .stabn 224,0,0,LBE2
2867 104 .stabs "e_places:T22=efirst:0,second:3,last:4,;",128,0,0,0
2868 @c FIXME: fake linebreak in line 105
2869 105 .stabs "u_tag:T23=u4u_int:1,0,32;u_float:12,0,32;u_char:21,0,32;;",
2874 109 .stabn 68,0,35,LM8
2877 112 save %sp,-120,%sp
2883 118 .stabn 68,0,41,LM9
2886 121 .stabn 68,0,41,LM10
2891 126 .stabs "s_proc:f1",36,0,0,_s_proc
2892 127 .stabs "s_arg:p16",160,0,0,0
2893 128 .stabs "s_ptr_arg:p18",160,0,0,72
2894 129 .stabs "char_vec:p21",160,0,0,76
2895 130 .stabs "an_u:23",128,0,0,-20
2896 131 .stabn 192,0,0,LBB4
2897 132 .stabn 224,0,0,LBE4
2898 133 .stabs "g_bar:r1",64,0,0,5
2899 134 .stabs "g_pf:G24=*25=f1",32,0,0,0
2900 135 .common _g_pf,4,"bss"
2901 136 .stabs "g_an_s:G16",32,0,0,0
2902 137 .common _g_an_s,20,"bss"
2906 @appendix Values for the Stab Type Field
2908 These are all the possible values for the stab type field, for
2909 @code{a.out} files. This does not apply to XCOFF.
2911 The following types are used by the linker and assembler; there is
2912 nothing stabs-specific about them. Since this document does not attempt
2913 to describe aspects of object file format other than the debugging
2914 format, no details are given.
2916 @c Try to get most of these to fit on a single line.
2926 File scope absolute symbol
2928 @item 0x3 N_ABS | N_EXT
2929 External absolute symbol
2932 File scope text symbol
2934 @item 0x5 N_TEXT | N_EXT
2935 External text symbol
2938 File scope data symbol
2940 @item 0x7 N_DATA | N_EXT
2941 External data symbol
2944 File scope BSS symbol
2946 @item 0x9 N_BSS | N_EXT
2950 Same as N_FN, for Sequent compilers
2953 Symbol is indirected to another symbol
2956 Common sym -- visable after shared lib dynamic link
2959 Absolute set element
2962 Text segment set element
2965 Data segment set element
2968 BSS segment set element
2971 Pointer to set vector
2973 @item 0x1e N_WARNING
2974 Print a warning message during linking
2977 File name of a .o file
2980 The following symbol types indicate that this is a stab. This is the
2981 full list of stab numbers, including stab types that are used in
2982 languages other than C.
2986 Global symbol, @xref{N_GSYM}.
2989 Function name (for BSD Fortran), @xref{N_FNAME}.
2992 Function name or text segment variable for C, @xref{N_FUN}.
2995 Static symbol (data segment variable with internal linkage), @xref{N_STSYM}.
2998 .lcomm symbol (BSS segment variable with internal linkage), @xref{N_LCSYM}.
3001 Name of main routine (not used in C), @xref{N_MAIN}.
3003 @c FIXME: discuss this in the main body of the text where we talk about
3004 @c using N_FUN for variables.
3006 Read-only data symbol (Solaris2). Most systems use N_FUN for this.
3009 Global symbol (for Pascal), @xref{N_PC}.
3012 Number of symbols (according to Ultrix V4.0), @xref{N_NSYMS}.
3015 No DST map for sym (according to Ultrix V4.0), @xref{N_NOMAP}.
3017 @c FIXME: describe this solaris feature in the body of the text (see
3018 @c comments in include/aout/stab.def).
3020 Object file (Solaris2).
3022 @c See include/aout/stab.def for (a little) more info.
3024 Debugger options (Solaris2).
3027 Register variable, @xref{N_RSYM}.
3030 Modula-2 compilation unit, @xref{N_M2C}.
3033 Line number in text segment, @xref{Line Numbers}.
3036 Line number in data segment, @xref{Line Numbers}.
3039 Line number in bss segment, @xref{Line Numbers}.
3042 Sun source code browser, path to .cb file, @xref{N_BROWS}.
3045 Gnu Modula2 definition module dependency, @xref{N_DEFD}.
3048 Function start/body/end line numbers (Solaris2).
3051 Gnu C++ exception variable, @xref{N_EHDECL}.
3054 Modula2 info "for imc" (according to Ultrix V4.0), @xref{N_MOD2}.
3057 Gnu C++ "catch" clause, @xref{N_CATCH}.
3060 Structure of union element, @xref{N_SSYM}.
3063 Last stab for module (Solaris2).
3066 Path and name of source file , @xref{Source Files}.
3069 Automatic var in the stack or type definition, @xref{N_LSYM}, @xref{Typedefs}.
3072 Beginning of an include file (Sun only), @xref{Source Files}.
3075 Name of include file, @xref{Source Files}.
3078 Parameter variable, @xref{Parameters}.
3081 End of an include file, @xref{Source Files}.
3084 Alternate entry point, @xref{N_ENTRY}.
3087 Beginning of a lexical block, @xref{Block Structure}.
3090 Place holder for a deleted include file, @xref{Source Files}.
3093 Modula2 scope information (Sun linker), @xref{N_SCOPE}.
3096 End of a lexical block, @xref{Block Structure}.
3099 Begin named common block, @xref{N_BCOMM}.
3102 End named common block, @xref{N_ECOMM}.
3105 End common (local name), @xref{N_ECOML}.
3107 @c FIXME: How does this really work? Move it to main body of document.
3109 Pascal @code{with} statement: type,,0,0,offset (Solaris2).
3112 Gould non-base registers, @xref{Gould}.
3115 Gould non-base registers, @xref{Gould}.
3118 Gould non-base registers, @xref{Gould}.
3121 Gould non-base registers, @xref{Gould}.
3124 Gould non-base registers, @xref{Gould}.
3127 @c Restore the default table indent
3132 @node Symbol Descriptors
3133 @appendix Table of Symbol Descriptors
3135 @c Please keep this alphabetical
3137 @c In TeX, this looks great, digit is in italics. But makeinfo insists
3138 @c on putting it in `', not realizing that @var should override @code.
3139 @c I don't know of any way to make makeinfo do the right thing. Seems
3140 @c like a makeinfo bug to me.
3144 Local variable, @xref{Automatic variables}.
3147 Parameter passed by reference in register, @xref{Parameters}.
3150 Constant, @xref{Constants}.
3153 Conformant array bound (Pascal, maybe other languages),
3154 @xref{Parameters}. Name of a caught exception (GNU C++). These can be
3155 distinguished because the latter uses N_CATCH and the former uses
3156 another symbol type.
3159 Floating point register variable, @xref{Register variables}.
3162 Parameter in floating point register, @xref{Parameters}.
3165 Static function, @xref{Procedures}.
3168 Global function, @xref{Procedures}.
3171 Global variable, @xref{Global Variables}.
3177 Internal (nested) procedure, @xref{Procedures}.
3180 Internal (nested) function, @xref{Procedures}.
3183 Label name (documented by AIX, no further information known).
3186 Module, @xref{Procedures}.
3189 Argument list parameter, @xref{Parameters}.
3195 FORTRAN Function parameter, @xref{Parameters}.
3198 Unfortunately, three separate meanings have been independently invented
3199 for this symbol descriptor. At least the GNU and Sun uses can be
3200 distinguished by the symbol type. Global Procedure (AIX) (symbol type
3201 used unknown), @xref{Procedures}. Register parameter (GNU) (symbol type
3202 N_PSYM), @xref{Parameters}. Prototype of function referenced by this
3203 file (Sun acc) (symbol type N_FUN).
3206 Static Procedure, @xref{Procedures}.
3209 Register parameter @xref{Parameters}.
3212 Register variable, @xref{Register variables}.
3215 Static file scope variable @xref{Initialized statics},
3216 @xref{Un-initialized statics}.
3219 Type name, @xref{Typedefs}.
3222 enumeration, struct or union tag, @xref{Typedefs}.
3225 Parameter passed by reference, @xref{Parameters}.
3228 Static procedure scope variable @xref{Initialized statics},
3229 @xref{Un-initialized statics}.
3232 Conformant array, @xref{Parameters}.
3235 Function return variable, @xref{Parameters}.
3238 @node Type Descriptors
3239 @appendix Table of Type Descriptors
3244 Type reference, @xref{Stabs Format}.
3247 Reference to builtin type, @xref{Negative Type Numbers}.
3250 Method (C++), @xref{Cplusplus}.
3253 Pointer, @xref{Miscellaneous Types}.
3259 Type Attributes (AIX), @xref{Stabs Format}. Member (class and variable)
3260 type (GNU C++), @xref{Cplusplus}.
3263 Array, @xref{Arrays}.
3266 Open array, @xref{Arrays}.
3269 Pascal space type (AIX), @xref{Miscellaneous Types}. Builtin integer
3270 type (Sun), @xref{Builtin Type Descriptors}.
3273 Volatile-qualified type, @xref{Miscellaneous Types}.
3276 Complex builtin type, @xref{Builtin Type Descriptors}.
3279 COBOL Picture type. See AIX documentation for details.
3282 File type, @xref{Miscellaneous Types}.
3285 N-dimensional dynamic array, @xref{Arrays}.
3288 Enumeration type, @xref{Enumerations}.
3291 N-dimensional subarray, @xref{Arrays}.
3294 Function type, @xref{Function Types}.
3297 Pascal function parameter, @xref{Function Types}
3300 Builtin floating point type, @xref{Builtin Type Descriptors}.
3303 COBOL Group. See AIX documentation for details.
3306 Imported type, @xref{Cross-references}.
3309 Const-qualified type, @xref{Miscellaneous Types}.
3312 COBOL File Descriptor. See AIX documentation for details.
3315 Multiple instance type, @xref{Miscellaneous Types}.
3318 String type, @xref{Strings}.
3321 Stringptr, @xref{Strings}.
3324 Opaque type, @xref{Typedefs}.
3327 Procedure, @xref{Function Types}.
3330 Packed array, @xref{Arrays}.
3333 Range type, @xref{Subranges}.
3336 Builtin floating type, @xref{Builtin Type Descriptors} (Sun). Pascal
3337 subroutine parameter, @xref{Function Types} (AIX). Detecting this
3338 conflict is possible with careful parsing (hint: a Pascal subroutine
3339 parameter type will always contain a comma, and a builtin type
3340 descriptor never will).
3343 Structure type, @xref{Structures}.
3346 Set type, @xref{Miscellaneous Types}.
3349 Union, @xref{Unions}.
3352 Variant record. This is a Pascal and Modula-2 feature which is like a
3353 union within a struct in C. See AIX documentation for details.
3356 Wide character, @xref{Builtin Type Descriptors}.
3359 Cross-reference, @xref{Cross-references}.
3362 gstring, @xref{Strings}.
3365 @node Expanded reference
3366 @appendix Expanded reference by stab type.
3368 @c FIXME: This appendix should go away, see N_PSYM or N_SO for an example.
3370 For a full list of stab types, and cross-references to where they are
3371 described, @xref{Stab Types}. This appendix just duplicates certain
3372 information from the main body of this document; eventually the
3373 information will all be in one place.
3377 The first line is the symbol type expressed in decimal, hexadecimal,
3378 and as a #define (see devo/include/aout/stab.def).
3380 The second line describes the language constructs the symbol type
3383 The third line is the stab format with the significant stab fields
3384 named and the rest NIL.
3386 Subsequent lines expand upon the meaning and possible values for each
3387 significant stab field. # stands in for the type descriptor.
3389 Finally, any further information.
3392 * N_GSYM:: Global variable
3393 * N_FNAME:: Function name (BSD Fortran)
3394 * N_FUN:: C Function name or text segment variable
3395 * N_STSYM:: Initialized static symbol
3396 * N_LCSYM:: Uninitialized static symbol
3397 * N_MAIN:: Name of main routine (not for C)
3398 * N_PC:: Pascal global symbol
3399 * N_NSYMS:: Number of symbols
3400 * N_NOMAP:: No DST map
3401 * N_RSYM:: Register variable
3402 * N_M2C:: Modula-2 compilation unit
3403 * N_BROWS:: Path to .cb file for Sun source code browser
3404 * N_DEFD:: GNU Modula2 definition module dependency
3405 * N_EHDECL:: GNU C++ exception variable
3406 * N_MOD2:: Modula2 information "for imc"
3407 * N_CATCH:: GNU C++ "catch" clause
3408 * N_SSYM:: Structure or union element
3409 * N_LSYM:: Automatic variable
3410 * N_ENTRY:: Alternate entry point
3411 * N_SCOPE:: Modula2 scope information (Sun only)
3412 * N_BCOMM:: Begin named common block
3413 * N_ECOMM:: End named common block
3414 * N_ECOML:: End common
3415 * Gould:: non-base register symbols used on Gould systems
3416 * N_LENG:: Length of preceding entry
3420 @section 32 - 0x20 - N_GYSM
3425 .stabs "name", N_GSYM, NIL, NIL, NIL
3429 "name" -> "symbol_name:#type"
3433 Only the "name" field is significant. The location of the variable is
3434 obtained from the corresponding external symbol.
3437 @section 34 - 0x22 - N_FNAME
3438 Function name (for BSD Fortran)
3441 .stabs "name", N_FNAME, NIL, NIL, NIL
3445 "name" -> "function_name"
3448 Only the "name" field is significant. The location of the symbol is
3449 obtained from the corresponding extern symbol.
3452 @section 36 - 0x24 - N_FUN
3454 Function name (@pxref{Procedures}) or text segment variable
3455 (@pxref{Variables}).
3457 @exdent @emph{For functions:}
3458 "name" -> "proc_name:#return_type"
3459 # -> F (global function)
3461 desc -> line num for proc start. (GCC doesn't set and DBX doesn't miss it.)
3462 value -> Code address of proc start.
3464 @exdent @emph{For text segment variables:}
3465 <<How to create one?>>
3469 @section 38 - 0x26 - N_STSYM
3470 Initialized static symbol (data segment w/internal linkage).
3473 .stabs "name", N_STSYM, NIL, NIL, value
3477 "name" -> "symbol_name#type"
3478 # -> S (scope global to compilation unit)
3479 -> V (scope local to a procedure)
3480 value -> Data Address
3484 @section 40 - 0x28 - N_LCSYM
3485 Unitialized static (.lcomm) symbol(BSS segment w/internal linkage).
3488 .stabs "name", N_LCLSYM, NIL, NIL, value
3492 "name" -> "symbol_name#type"
3493 # -> S (scope global to compilation unit)
3494 -> V (scope local to procedure)
3495 value -> BSS Address
3499 @section 42 - 0x2a - N_MAIN
3500 Name of main routine (not used in C)
3503 .stabs "name", N_MAIN, NIL, NIL, NIL
3507 "name" -> "name_of_main_routine"
3511 @section 48 - 0x30 - N_PC
3512 Global symbol (for Pascal)
3515 .stabs "name", N_PC, NIL, NIL, value
3519 "name" -> "symbol_name" <<?>>
3520 value -> supposedly the line number (stab.def is skeptical)
3526 global pascal symbol: name,,0,subtype,line
3531 @section 50 - 0x32 - N_NSYMS
3532 Number of symbols (according to Ultrix V4.0)
3535 0, files,,funcs,lines (stab.def)
3539 @section 52 - 0x34 - N_NOMAP
3540 no DST map for sym (according to Ultrix V4.0)
3543 name, ,0,type,ignored (stab.def)
3547 @section 64 - 0x40 - N_RSYM
3551 .stabs "name:type",N_RSYM,0,RegSize,RegNumber (Sun doc)
3555 @section 66 - 0x42 - N_M2C
3556 Modula-2 compilation unit
3559 .stabs "name", N_M2C, 0, desc, value
3563 "name" -> "unit_name,unit_time_stamp[,code_time_stamp]
3565 value -> 0 (main unit)
3570 @section 72 - 0x48 - N_BROWS
3571 Sun source code browser, path to .cb file
3574 "path to associated .cb file"
3576 Note: type field value overlaps with N_BSLINE
3579 @section 74 - 0x4a - N_DEFD
3580 GNU Modula2 definition module dependency
3582 GNU Modula-2 definition module dependency. Value is the modification
3583 time of the definition file. Other is non-zero if it is imported with
3584 the GNU M2 keyword %INITIALIZE. Perhaps N_M2C can be used if there
3585 are enough empty fields?
3588 @section 80 - 0x50 - N_EHDECL
3589 GNU C++ exception variable <<?>>
3591 "name is variable name"
3593 Note: conflicts with N_MOD2.
3596 @section 80 - 0x50 - N_MOD2
3597 Modula2 info "for imc" (according to Ultrix V4.0)
3599 Note: conflicts with N_EHDECL <<?>>
3602 @section 84 - 0x54 - N_CATCH
3603 GNU C++ "catch" clause
3605 GNU C++ `catch' clause. Value is its address. Desc is nonzero if
3606 this entry is immediately followed by a CAUGHT stab saying what
3607 exception was caught. Multiple CAUGHT stabs means that multiple
3608 exceptions can be caught here. If Desc is 0, it means all exceptions
3612 @section 96 - 0x60 - N_SSYM
3613 Structure or union element
3615 Value is offset in the structure.
3617 <<?looking at structs and unions in C I didn't see these>>
3620 @section 128 - 0x80 - N_LSYM
3621 Automatic var in the stack (also used for type descriptors.)
3624 .stabs "name" N_LSYM, NIL, NIL, value
3628 @exdent @emph{For stack based local variables:}
3630 "name" -> name of the variable
3631 value -> offset from frame pointer (negative)
3633 @exdent @emph{For type descriptors:}
3635 "name" -> "name_of_the_type:#type"
3638 type -> type_ref (or) type_def
3640 type_ref -> type_number
3641 type_def -> type_number=type_desc etc.
3644 Type may be either a type reference or a type definition. A type
3645 reference is a number that refers to a previously defined type. A
3646 type definition is the number that will refer to this type, followed
3647 by an equals sign, a type descriptor and the additional data that
3648 defines the type. See the Table D for type descriptors and the
3649 section on types for what data follows each type descriptor.
3652 @section 164 - 0xa4 - N_ENTRY
3654 Alternate entry point.
3655 Value is its address.
3659 @section 196 - 0xc4 - N_SCOPE
3661 Modula2 scope information (Sun linker)
3665 @section 226 - 0xe2 - N_BCOMM
3667 Begin named common block.
3669 Only the name is significant.
3673 @section 228 - 0xe4 - N_ECOMM
3675 End named common block.
3677 Only the name is significant and it should match the N_BCOMM
3681 @section 232 - 0xe8 - N_ECOML
3683 End common (local name)
3689 @section Non-base registers on Gould systems
3691 These are used on Gould systems for non-base registers syms.
3693 However, the following values are not the values used by Gould; they are
3694 the values which GNU has been documenting for these values for a long
3695 time, without actually checking what Gould uses. I include these values
3696 only because perhaps some someone actually did something with the GNU
3697 information (I hope not, why GNU knowingly assigned wrong values to
3698 these in the header file is a complete mystery to me).
3701 240 0xf0 N_NBTEXT ??
3702 242 0xf2 N_NBDATA ??
3709 @section - 0xfe - N_LENG
3711 Second symbol entry containing a length-value for the preceding entry.
3712 The value is the length.
3715 @appendix Questions and anomalies
3719 For GNU C stabs defining local and global variables (N_LSYM and
3720 N_GSYM), the desc field is supposed to contain the source line number
3721 on which the variable is defined. In reality the desc field is always
3722 0. (This behavour is defined in dbxout.c and putting a line number in
3723 desc is controlled by #ifdef WINNING_GDB which defaults to false). Gdb
3724 supposedly uses this information if you say 'list var'. In reality
3725 var can be a variable defined in the program and gdb says `function
3729 In GNU C stabs there seems to be no way to differentiate tag types:
3730 structures, unions, and enums (symbol descriptor T) and typedefs
3731 (symbol descriptor t) defined at file scope from types defined locally
3732 to a procedure or other more local scope. They all use the N_LSYM
3733 stab type. Types defined at procedure scope are emited after the
3734 N_RBRAC of the preceding function and before the code of the
3735 procedure in which they are defined. This is exactly the same as
3736 types defined in the source file between the two procedure bodies.
3737 GDB overcompensates by placing all types in block #1, the block for
3738 symbols of file scope. This is true for default, -ansi and
3739 -traditional compiler options. (Bugs gcc/1063, gdb/1066.)
3742 What ends the procedure scope? Is it the proc block's N_RBRAC or the
3743 next N_FUN? (I believe its the first.)
3746 The comment in xcoff.h says DBX_STATIC_CONST_VAR_CODE is used for
3747 static const variables. DBX_STATIC_CONST_VAR_CODE is set to N_FUN by
3748 default, in dbxout.c. If included, xcoff.h redefines it to N_STSYM.
3749 But testing the default behaviour, my Sun4 native example shows
3750 N_STSYM not N_FUN is used to describe file static initialized
3751 variables. (the code tests for TREE_READONLY(decl) &&
3752 !TREE_THIS_VOLATILE(decl) and if true uses DBX_STATIC_CONST_VAR_CODE).
3755 Global variable stabs don't have location information. This comes
3756 from the external symbol for the same variable. The external symbol
3757 has a leading underbar on the _name of the variable and the stab does
3758 not. How do we know these two symbol table entries are talking about
3759 the same symbol when their names are different?
3762 Can gcc be configured to output stabs the way the Sun compiler
3763 does, so that their native debugging tools work? <NO?> It doesn't by
3764 default. GDB reads either format of stab. (gcc or SunC). How about
3768 @node xcoff-differences
3769 @appendix Differences between GNU stabs in a.out and GNU stabs in xcoff
3771 @c FIXME: Merge *all* these into the main body of the document.
3772 (The AIX/RS6000 native object file format is xcoff with stabs). This
3773 appendix only covers those differences which are not covered in the main
3774 body of this document.
3778 BSD a.out stab types correspond to AIX xcoff storage classes. In general the
3779 mapping is N_STABTYPE becomes C_STABTYPE. Some stab types in a.out
3780 are not supported in xcoff. See Table E. for full mappings.
3783 initialised static N_STSYM and un-initialized static N_LCSYM both map
3784 to the C_STSYM storage class. But the destinction is preserved
3785 because in xcoff N_STSYM and N_LCSYM must be emited in a named static
3786 block. Begin the block with .bs s[RW] data_section_name for N_STSYM
3787 or .bs s bss_section_name for N_LCSYM. End the block with .es
3790 If the xcoff stab is a N_FUN (C_FUN) then follow the string field with
3791 ,. instead of just ,
3795 (I think that's it for .s file differences. They could stand to be
3796 better presented. This is just a list of what I have noticed so far.
3797 There are a *lot* of differences in the information in the symbol
3798 tables of the executable and object files.)
3800 Table E: mapping a.out stab types to xcoff storage classes
3803 stab type storage class
3804 -------------------------------
3813 N_RPSYM (0x8e) C_RPSYM
3823 N_DECL (0x8c) C_DECL
3840 @node Sun-differences
3841 @appendix Differences between GNU stabs and Sun native stabs.
3843 @c FIXME: Merge all this stuff into the main body of the document.
3847 GNU C stabs define *all* types, file or procedure scope, as
3848 N_LSYM. Sun doc talks about using N_GSYM too.
3851 Sun C stabs use type number pairs in the format (a,b) where a is a
3852 number starting with 1 and incremented for each sub-source file in the
3853 compilation. b is a number starting with 1 and incremented for each
3854 new type defined in the compilation. GNU C stabs use the type number
3855 alone, with no source file number.
3859 @appendix Using stabs with the ELF object file format.
3861 The ELF object file format allows tools to create object files with custom
3862 sections containing any arbitrary data. To use stabs in ELF object files,
3863 the tools create two custom sections, a ".stab" section which contains
3864 an array of fixed length structures, one struct per stab, and a ".stabstr"
3865 section containing all the variable length strings that are referenced by
3866 stabs in the ".stab" section.
3868 The first stab in the ".stab" section for each object file is a "synthetic
3869 stab", generated entirely by the assembler, with no corresponding ".stab"
3870 directive as input to the assembler. This stab contains the following
3875 Offset in the ".stabstr" section to the source filename.
3881 Unused field, always zero.
3884 Count of upcoming symbols. I.E. the number of remaining stabs for this
3888 Size of the string table fragment associated with this object module, in
3893 The ".stabstr" section always starts with a null byte (so that string
3894 offsets of zero reference a null string), followed by random length strings,
3895 each of which is null byte terminated.
3897 The ELF section header for the ".stab" section has it's sh_link member set
3898 to the section number of the ".stabstr" section, and the ".stabstr" section
3899 has it's ELF section header sh_type member set to SHT_STRTAB to mark it as