1 \chapter{Program Scope Entries}
2 \label{chap:programscopeentries}
3 This section describes debugging information entries that
4 relate to different levels of program scope: compilation,
5 module, subprogram, and so on. Except for separate type
6 entries (see Section \refersec{chap:typeunitentries}),
7 these entries may be thought of as
8 ranges of text addresses within the program.
10 \section{Unit Entries}
11 \label{chap:unitentries}
12 A DWARF object file is an object file that contains one or more
13 DWARF compilation units, of which there are these kinds:
14 \addtoindexx{unit|see {compilation unit}}
15 \addtoindexx{compilation unit}
17 \item A \definition{full compilation unit} describes
18 a complete compilation, possibly in combination with
19 related partial compilation units and/or type units.
21 \item A \definition{partial compilation unit} describes
22 a part of a compilation (generally corresponding to an
23 imported module) which is imported into one or more
24 related full compilation units.
26 \item A \definition{type unit} is a specialized unit
27 (similar to a compilation unit) that represents a type
28 whose description may be usefully shared by multiple
32 \index{conventional compilation unit|see{
33 full compilation unit, partial compilation unit, type unit}}
35 \textit{These first three kinds of compilation unit are
36 sometimes called \doublequote{conventional} compilation
37 units--they are kinds of compilation units that were
38 defined prior to \DWARFVersionV. Conventional compilation units
39 are part of the same object file as the compiled code and
40 data (whether relocatable, executable, shared and so on).
41 The word \doublequote{conventional} is usually
42 omitted in these names, unless needed to distinguish them
43 from the similar split compilation units below.}
47 \item A \definition{skeleton compilation unit} represents
48 the DWARF debugging information for a compilation using a
49 minimal description that identifies a separate split
50 compilation unit that provides the remainder (and most)
54 \textit{A skeleton compilation acts as a minimal conventional full
55 compilation (see above) that identifies and is paired with a
56 corresponding split full compilation (as described below). Like
57 the conventional compilation units, a skeleton compilation unit
58 is part of the same object file as the compiled code and data.}
62 \definition{split compilation unit} describes
63 a complete compilation, possibly in combination with
64 related type compilation units. It corresponds
65 to a specific skeleton compilation unit.
67 \item A \definition{split type unit} is a specialized
68 compilation unit that represents a type whose description may
69 be usefully shared by multiple other units.
73 \textit{Split compilation units and split type units may be
74 contained in object files separate from those containing the
75 program code and data.
76 These object files are not processed by a linker; thus,
77 split units do not depend on underlying object file relocations.}
79 \textit{Either a full compilation unit or a partial compilation
80 unit may be logically incorporated into another compilation unit
81 using an \addtoindex{imported unit entry}
82 (see Section \refersec{chap:importedunitentries}).}
85 combined split and partial
86 compilation unit kind is not defined.}
88 \textit{In the remainder of this document, the word
89 \doublequote{compilation} in the phrase \doublequote{compilation unit}
90 is generally omitted, unless it is deemed needed for clarity
93 \subsection{Full and Partial Compilation Unit Entries}
94 \label{chap:fullandpartialcompilationunitentries}
95 A \addtoindex{full compilation unit}\addtoindexx{compilation unit!full}
96 is represented by a debugging information entry with the tag
97 \DWTAGcompileunitTARG.
98 A \addtoindex{partial compilation unit}\addtoindexx{compilation unit!partial}
99 is represented by a debugging information entry with the tag
100 \DWTAGpartialunitTARG.
103 In a simple compilation, a single compilation unit with
105 \DWTAGcompileunit{} represents a complete object file
107 \DWTAGpartialunit{} (as well as tag \DWTAGtypeunit) is not used.
109 employing the DWARF space compression and duplicate elimination
111 Appendix \refersec{app:usingcompilationunits},
112 multiple compilation units using
115 \DWTAGpartialunit{} and/or
117 are used to represent portions of an object file.
120 \textit{A full compilation unit typically represents the text and
121 data contributed to an executable by a single relocatable
122 object file. It may be derived from several source files,
123 including pre-processed header files.
124 A \addtoindex{partial compilation unit} typically represents a part
125 of the text and data of a relocatable object file, in a manner that
126 can potentially be shared with the results of other compilations
127 to save space. It may be derived from an \doublequote{include file,}
128 template instantiation, or other implementation-dependent
129 portion of a compilation. A full compilation unit can also
130 function in a manner similar to a partial compilation unit
132 See Appendix \refersec{app:dwarfcompressionandduplicateeliminationinformative}
133 for discussion of related compression techniques.}
135 A compilation unit entry owns debugging information
136 entries that represent all or part of the declarations
137 made in the corresponding compilation. In the case of a
138 partial compilation unit, the containing scope of its owned
139 declarations is indicated by imported unit entries in one
140 or more other compilation unit entries that refer to that
141 partial compilation unit (see
142 Section \refersec{chap:importedunitentries}).
144 Compilation unit entries may have the following
146 \begin{enumerate}[1. ]
147 \item Either a \DWATlowpc{} and
148 \DWAThighpc{} pair of
149 \addtoindexx{high PC attribute}
151 \addtoindexx{low PC attribute}
153 \addtoindexx{ranges attribute}
155 \DWATranges{} attribute
156 \addtoindexx{ranges attribute}
158 \addtoindexx{discontiguous address ranges|see{non-contiguous address ranges}}
160 non-contiguous address ranges, respectively,
161 of the machine instructions generated for the compilation
162 unit (see Section \refersec{chap:codeaddressesandranges}).
164 A \DWATlowpc{} attribute
165 may also be specified in combination
166 \addtoindexx{ranges attribute}
168 \DWATranges{} to specify the
169 \addtoindexx{ranges attribute}
170 default base address for use in
171 \addtoindexx{location list}
172 location lists (see Section
173 \refersec{chap:locationlists}) and range lists
174 \addtoindexx{range list}
175 (see Section \refersec{chap:noncontiguousaddressranges}).
177 \item \hypertarget{chap:DWATnamepathnameofcompilationsource}{}
178 A \DWATnameDEFN{} attribute \addtoindexx{name attribute}
179 whose value is a null-terminated string
180 containing the full or relative path name
181 (relative to the value of the \DWATcompdir{} attribute,
182 see below) of the primary
183 source file from which the compilation unit was derived.
185 \item \hypertarget{chap:DWATlanguageprogramminglanguage}{}
186 A \DWATlanguageDEFN{} attribute\addtoindexx{language attribute}
187 whose constant value is an integer code
188 indicating the source language of the compilation
189 unit. The set of language names and their meanings are given
190 in Table \refersec{tab:languagenames}.
195 \setlength{\extrarowheight}{0.1cm}
196 \begin{longtable}{l|l}
197 \caption{Language names} \label{tab:languagenames} \\
198 \hline \bfseries Language name & \bfseries Meaning \\ \hline
200 \bfseries Language name & \bfseries Meaning \\ \hline
202 \hline \emph{Continued on next page}
205 \addtoindexx{ISO-defined language names}
206 \DWLANGAdaeightythreeTARG{} \dag & ISO Ada:1983 \addtoindexx{Ada:1983 (ISO)} \\
207 \DWLANGAdaninetyfiveTARG{} \dag & ISO Ada:1995 \addtoindexx{Ada:1995 (ISO)} \\
208 \DWLANGCTARG & Non-standardized C, such as K\&R \addtoindexx{C!non-standard} \\*
209 \DWLANGCeightynineTARG & ISO C:1989 \addtoindexx{C:1989 (ISO)} \\*
210 \DWLANGCninetynineTARG & ISO C:1999 \addtoindexx{C:1999 (ISO)} \\*
211 \DWLANGCelevenTARG & ISO C:2011 \addtoindexx{C:2011 (ISO)} \\*
212 \DWLANGCplusplusTARG & ISO C++98 \addtoindexx{C++98 (ISO)} \\
213 \DWLANGCpluspluszerothreeTARG & ISO C++03 \addtoindexx{C++03 (ISO)} \\
214 \DWLANGCpluspluselevenTARG & ISO C++11 \addtoindexx{C++11 (ISO)} \\
215 \DWLANGCplusplusfourteenTARG & ISO C++14 \addtoindexx{C++14 (ISO)}
217 \DWLANGCobolseventyfourTARG & ISO COBOL:1974 \addtoindexx{COBOL:1974 (ISO)} \\
218 \DWLANGCoboleightyfiveTARG & ISO COBOL:1985 \addtoindexx{COBOL:1985 (ISO)} \\
219 \DWLANGDTARG{}~\dag & D \addtoindexx{D language} \\
220 \DWLANGDylanTARG~\dag & Dylan \addtoindexx{Dylan} \\
221 \DWLANGFortranseventysevenTARG & ISO FORTRAN:1977 \addtoindexx{FORTRAN:1977 (ISO)} \\
222 \DWLANGFortranninetyTARG & ISO Fortran:1990 \addtoindexx{Fortran:1990 (ISO)} \\
223 \DWLANGFortranninetyfiveTARG & ISO Fortran:1995 \addtoindexx{Fortran:1995 (ISO)} \\
224 \DWLANGFortranzerothreeTARG & ISO Fortran:2004 \addtoindexx{Fortran:2004 (ISO)} \\
225 \DWLANGFortranzeroeightTARG & ISO Fortran:2010 \addtoindexx{Fortran:2010 (ISO)} \\
226 \DWLANGGoTARG{}~\dag & \addtoindex{Go} \\
227 \DWLANGHaskellTARG{} \dag & \addtoindex{Haskell} \\
228 \DWLANGJavaTARG{} & \addtoindex{Java}\\
229 \DWLANGJuliaTARG{}~\dag & \addtoindex{Julia} \\
230 \DWLANGModulatwoTARG & ISO Modula\dash 2:1996 \addtoindexx{Modula-2:1996 (ISO)} \\
231 \DWLANGModulathreeTARG & \addtoindex{Modula-3} \\
232 \DWLANGObjCTARG{} & \addtoindex{Objective C} \\
233 \DWLANGObjCplusplusTARG{} & \addtoindex{Objective C++} \\
234 \DWLANGOCamlTARG{}~\dag & \addtoindex{OCaml}\index{Objective Caml|see{OCaml}} \\
235 \DWLANGOpenCLTARG{}~\dag & \addtoindex{OpenCL} \\
236 \DWLANGPascaleightythreeTARG & ISO Pascal:1983 \addtoindexx{Pascal:1983 (ISO)} \\
237 \DWLANGPLITARG{}~\dag & ANSI PL/I:1976 \addtoindexx{PL/I:1976 (ANSI)} \\
238 \DWLANGPythonTARG{}~\dag & \addtoindex{Python} \\
239 \DWLANGRenderScriptTARG~\dag &
241 \addtoindex{RenderScript Kernel Language}
244 \DWLANGRustTARG{}~\dag & \addtoindex{Rust} \\
246 & \addtoindex{Swift} \\
247 \DWLANGUPCTARG{} & UPC (Unified Parallel C) \addtoindexx{UPC}
248 \index{Unified Parallel C|see{UPC}} \\
250 \dag \ \ \textit{Support for these languages is limited}& \\
255 \item \hypertarget{chap:DWATstmtlistlinenumberinformationforunit}{}
256 A \DWATstmtlistDEFN{}\addtoindexx{statement list attribute}
257 attribute whose value is a
258 \addtoindexx{section offset!in statement list attribute}
259 section offset to the line number information for this compilation
262 This information is placed in a separate object file
263 section from the debugging information entries themselves. The
264 value of the statement list attribute is the offset in the
265 \dotdebugline{} section of the first byte of the line number
266 information for this compilation unit
267 (see Section \refersec{chap:linenumberinformation}).
269 \item A \DWATmacrosDEFN{}\hypertarget{chap:DWATmacrosmacroinformation}{}
271 \addtoindexx{macro information attribute}
273 \addtoindexx{section offset!in macro information attribute}
274 section offset to the macro information for this compilation unit.
276 This information is placed in a separate object file section
277 from the debugging information entries themselves. The
278 value of the macro information attribute is the offset in
279 the \dotdebugmacro{} section of the first byte of the macro
280 information for this compilation unit
281 (see Section \refersec{chap:macroinformation}).
283 \textit{The \DWATmacrosNAME{} attribute is new in \DWARFVersionV,
285 \DWATmacroinfoDEFN{} attribute of earlier DWARF versions.
286 \livetarg{chap:DWATmacroinfomacroinformation}{}
287 While \DWATmacrosNAME{} and \DWATmacroinfoNAME{} attributes cannot both occur in the same
288 compilation unit, both may be found in the set of units that make up an executable
289 or shared object file. The two attributes have distinct encodings to facilitate such
293 \item \hypertarget{chap:DWATcompdircompilationdirectory}{}
294 A \DWATcompdirDEFN{} attribute\addtoindexx{compilation directory attribute}
296 null-terminated string containing the current working directory
297 of the compilation command that produced this compilation
298 unit in whatever form makes sense for the host system.
300 \item \hypertarget{chap:DWATproducercompileridentification}{}
301 A \DWATproducerDEFN{} attribute\addtoindexx{producer attribute}
302 whose value is a null-terminated string containing
303 information about the compiler that produced the compilation unit.
305 \textit{The actual contents of
306 the string will be specific to each producer, but should
307 begin with the name of the compiler vendor or some other
308 identifying character sequence that will avoid confusion
309 with other producer values.}
312 \item \hypertarget{chap:DWATidentifiercaseidentifiercaserule}{}
313 A \DWATidentifiercaseDEFN{} attribute
314 \addtoindexx{identifier case attribute} whose integer
315 constant value is a code describing the treatment
316 of identifiers within this compilation unit. The
317 set of identifier case codes is given in
318 Table \refersec{tab:identifiercasecodes}.
320 \begin{simplenametable}{Identifier case codes}{tab:identifiercasecodes}
321 \DWIDcasesensitive{} \\
324 \DWIDcaseinsensitive{} \\
325 \end{simplenametable}
327 \DWIDcasesensitiveTARG{} is the default for all compilation units
328 that do not have this attribute. It indicates that names given
329 as the values of \DWATname{} attributes
330 \addtoindexx{name attribute}
331 in debugging information
332 entries for the compilation unit reflect the names as they
333 appear in the source program.
335 \textit{A debugger should be sensitive
336 to the case of \addtoindex{identifier names} when doing identifier
340 \DWIDupcaseTARG{} means that the
341 producer of the debugging
342 information for this compilation unit converted all source
343 names to upper case. The values of the name attributes may not
344 reflect the names as they appear in the source program.
346 \textit{A debugger should convert all names to upper case
349 \DWIDdowncaseTARG{} means that the producer of the debugging
350 information for this compilation unit converted all source
351 names to lower case. The values of the name attributes may not
352 reflect the names as they appear in the source program.
354 \textit{A debugger should convert all names to lower case
358 \DWIDcaseinsensitiveTARG{} means that the values of the name
359 attributes reflect the names as they appear in the source
360 program but that case is not significant.
362 \textit{A debugger should ignore case when doing lookups.}
365 \item \hypertarget{chap:DWATbasetypesprimitivedatatypesofcompilationunit}{}
366 A \DWATbasetypesDEFN{} attribute\addtoindexx{base types attribute}
367 whose value is a \livelink{chap:classreference}{reference}.
368 This attribute points to a debugging information entry
369 representing another compilation unit. It may be used
370 to specify the compilation unit containing the base type
371 entries used by entries in the current compilation unit
372 (see Section \refersec{chap:basetypeentries}).
375 \textit{This attribute provides a consumer a way to find the definition
376 of base types for a compilation unit that does not itself
377 contain such definitions. This allows a consumer, for example,
378 to interpret a type conversion to a base type correctly.}
380 \item \hypertarget{chap:DWATuseUTF8compilationunitusesutf8strings}{}
381 A \DWATuseUTFeightDEFN{} attribute,
382 \addtoindexx{use UTF8 attribute}\addtoindexx{UTF-8}
383 which is a \livelink{chap:classflag}{flag} whose
384 presence indicates that all strings (such as the names of
385 declared entities in the source program, or filenames in the line number table)
386 are represented using the UTF-8 representation.
390 \hypertarget{chap:DWATmainsubprogramunitcontainingmainorstartingsubprogram}{}
391 A \DWATmainsubprogramDEFN{} attribute,\addtoindexx{main subprogram attribute}
392 which is a \livelink{chap:classflag}{flag},
393 whose presence indicates that the compilation unit contains a
394 subprogram that has been identified as the starting
395 subprogram of the program. If more than one compilation unit contains
396 this \nolink{flag}, any one of them may contain the starting
399 \textit{\addtoindex{Fortran} has a \addtoindex{PROGRAM statement}
401 to specify and provide a user-specified name for the main
402 subroutine of a program.
403 \addtoindex{C} uses the name \doublequote{main} to identify
404 the main subprogram of a program. Some other languages provide
405 similar or other means to identify the main subprogram of
406 a program. The \DWATmainsubprogram{} attribute may also be used to
407 identify such subprograms (see
408 Section \refersec{chap:generalsubroutineandentrypointinformation}).}
411 \hypertarget{chap:DWATentrypcofcompileunit}{}
412 \hypertarget{chap:DWATentrypcofpartialunit}{}
413 A \DWATentrypc{} attribute whose value is the address of the first
414 \addtoindexx{entry pc attribute}
415 executable instruction of the unit (see
416 Section \refersec{chap:entryaddress}).
419 \item \hypertarget{chap:DWATstroffsetbaseforindirectstringtable}{}
420 A \DWATstroffsetsbaseDEFN\addtoindexx{string offset base attribute}
421 attribute, whose value is of class \CLASSstroffsetsptr.
422 This attribute points to the first string
423 offset of the compilation unit's contribution to the
424 \dotdebugstroffsets{} (or \dotdebugstroffsetsdwo{}) section.
425 Indirect string references
426 (using \DWFORMstrx) within the compilation unit are
427 interpreted as indices relative to this base.
430 \item \hypertarget{chap:DWATaddrbaseforaddresstable}{}
431 A \DWATaddrbaseDEFN\addtoindexx{address table base attribute}
432 attribute, whose value is of class \CLASSaddrptr.
433 This attribute points to the beginning of the compilation
434 unit's contribution to the \dotdebugaddr{} section.
435 Indirect references (using \DWFORMaddrx, \DWOPaddrx,
438 \DWLLEbaseaddressx{}, \DWLLEstartxendx{}, \DWLLEstartxlength{},
439 \DWRLEbaseaddressx{}, \DWRLEstartxendx{} or \DWRLEstartxlength)
441 within the compilation unit are interpreted as indices
442 relative to this base.
445 \item \hypertarget{chap:DWATrnglistsbaseforrnglists}{}
446 A \DWATrnglistsbaseDEFN\addtoindexx{ranges table base attribute}
447 attribute, whose value is of class \CLASSrnglistsptr.
448 This attribute points to the
450 beginning of the offsets table (immediately following the header)
453 unit's contribution to the \dotdebugrnglists{} section.
454 References to range lists (using \DWFORMrnglistx)
455 within the compilation unit are
456 interpreted relative to this base.
458 \item \hypertarget{chap:DWATloclistsbaseinlocationlist}{}
459 A \DWATloclistsbaseDEFN{}\addtoindexx{location table base attribute}
460 attribute, whose value is of class \CLASSloclistsptr.
461 This attribute points to the
463 beginning of the offsets table (immediately following the header)
466 unit's contribution to the \dotdebugloclists{} section. References
467 to location lists (using \DWFORMloclistx) within the compilation
468 unit are interpreted relative to this base.
471 The base address of a compilation unit is defined as the
472 value of the \DWATlowpc{} attribute, if present; otherwise,
473 it is undefined. If the base address is undefined, then any
474 DWARF entry or structure defined in terms of the base address
475 of that compilation unit is not valid.
478 \subsection{Skeleton Compilation Unit Entries}
479 \label{chap:skeletoncompilationunitentries}
480 \addtoindexx{compilation unit!skeleton}
481 \addtoindexx{skeleton compilation unit}
482 When generating a \splitDWARFobjectfile{} (see
483 Section \refersec{datarep:splitdwarfobjectfiles}), the
484 compilation unit in the \dotdebuginfo{} section is a "skeleton"
485 compilation unit with the tag
486 \DWTAGskeletonunitTARG, which contains a
487 \DWATdwoname{} attribute as well as a subset of the
488 attributes of a full or partial compilation unit. In general,
489 it contains those attributes that are necessary for the consumer
490 to locate the object file where the split full compilation unit
491 can be found, and for the consumer to interpret references to
492 addresses in the program.
495 A skeleton compilation unit has no children.
497 A skeleton compilation unit has a \DWATdwoname{} attribute:
499 \begin{enumerate}[1. ]
501 \item \livetarg{chap:DWATdwonameforunit}{}
502 A \DWATdwonameDEFN{} attribute
503 \addtoindexx{split DWARF object file name attribute}
505 null-terminated string containing the full or relative
506 path name (relative to the value of the \DWATcompdir{} attribute,
507 see below) of the object file that contains the full
510 The value in the \HFNdwoid{} field of the unit header for
511 this unit is the same as the value in the \HFNdwoid{} field
512 of the unit header of the corresponding full compilation
513 unit (see Section \refersec{datarep:unitheaders}).
515 \textit{The means of determining a compilation unit ID does
516 not need to be similar or related to the means of
517 determining a \TUsignature. However, it should
518 be suitable for detecting file version skew or other
519 kinds of mismatched files and for looking up a full
520 split unit in a DWARF package file
521 (see Section \refersec{datarep:dwarfpackagefiles}).}
525 A skeleton compilation unit may have additional attributes,
526 which are the same as for conventional compilation unit entries
527 except as noted, from among the following:
528 \begin{enumerate}[1. ]
529 \addtocounter{enumi}{1}
530 \item Either a \DWATlowpc{} and \DWAThighpc{} pair of attributes
531 or a \DWATranges{} attribute.
532 \item A \DWATstmtlist{} attribute.
533 \item A \DWATcompdir{} attribute.
536 \item A \DWATuseUTFeight{} attribute.
538 \textit{This attribute applies to strings referred to by the skeleton
539 compilation unit entry itself, and strings in the associated line
541 The representation for strings in the object file referenced
542 by the \DWATdwoname{} attribute is determined by the presence
543 of a \DWATuseUTFeight{} attribute in the full compilation unit
544 (see Section \refersec{chap:splitfullcompilationunitentries}).}
546 \item A \DWATstroffsetsbase{} attribute, for indirect strings references
547 from the skeleton compilation unit.
548 \item A \DWATaddrbase{} attribute.
552 All other attributes of a compilation unit entry (described
553 in Section \refersec{chap:fullandpartialcompilationunitentries})
554 are placed in the split full compilation unit
555 (see \refersec{chap:splitfullcompilationunitentries}).
556 The attributes provided by the skeleton compilation
557 unit entry do not need to be repeated in the full compilation
560 \textit{The \DWATaddrbase{}
561 and \DWATstroffsetsbase{} attributes provide context that may be
562 necessary to interpret the contents
563 of the corresponding \splitDWARFobjectfile.}
565 \textit{The \DWATbasetypes{} attribute is not defined for a
566 skeleton compilation unit.}
569 \subsection{Split Full Compilation Unit Entries}
570 \label{chap:splitfullcompilationunitentries}
571 A \definition{split full compilation unit} is represented by a
572 debugging information entry with tag \DWTAGcompileunit.
573 It is very similar to a conventional full compilation unit but
574 is logically paired with a specific skeleton compilation unit while
575 being physically separate.
577 A split full compilation unit
578 may have the following attributes,
579 which are the same as for conventional compilation unit entries
581 \begin{enumerate}[1. ]
582 \item A \DWATname{} attribute.
583 \item A \DWATlanguage{} attribute.
584 \item A \DWATmacros{} attribute.
585 The value of this attribute is of class \CLASSmacptr{}, which is
586 an offset relative to the \dotdebugmacrodwo{} section.
588 \item A \DWATproducer{} attribute.
589 \item A \DWATidentifiercase{} attribute.
590 \item A \DWATmainsubprogram{} attribute.
591 \item A \DWATentrypc{} attribute.
592 \item A \DWATuseUTFeight{} attribute.
596 \textit{The following attributes are not part of a
597 split full compilation unit entry but instead are inherited
598 (if present) from the corresponding skeleton compilation unit:
599 \DWATlowpc, \DWAThighpc, \DWATranges, \DWATstmtlist, \DWATcompdir,
600 \DWATstroffsetsbase, \DWATaddrbase{} and
603 \textit{The \DWATbasetypes{} attribute is not defined for a
604 split full compilation unit.}
608 \subsection{Type Unit Entries}
609 \label{chap:typeunitentries}
610 \addtoindexx{type unit}
611 \addtoindexx{type unit|see{\textit{also} compilation unit}}
612 \addtoindexx{compilation unit!\textit{see also} type unit}
613 An object file may contain any number of separate type
614 unit entries, each representing a single complete type
616 Each \addtoindex{type unit} must be uniquely identified by
617 an 8-byte signature, stored as part of the type unit, which
618 can be used to reference the type definition from debugging
619 information entries in other compilation units and type units.
621 Conventional and split type units are identical except for
622 the sections in which they are represented
623 (see \refersec{datarep:splitdwarfobjectfiles} for details).
624 \addtoindexx{conventional type unit}
625 \addtoindexx{split type unit}
626 Moreover, the \DWATstroffsetsbase{} attribute (see below) is not
627 used in a split type unit.
629 A type unit is represented by a debugging information entry
630 with the tag \DWTAGtypeunitTARG.
631 A \addtoindex{type unit entry} owns debugging
632 information entries that represent the definition of a single
633 type, plus additional debugging information entries that may
634 be necessary to include as part of the definition of the type.
637 A type unit entry may have the following attributes:
638 \begin{enumerate}[1. ]
641 \DWATlanguage{} attribute,
643 \addtoindexx{language attribute}
644 constant value is an integer code indicating the source
645 language used to define the type. The set of language names
646 and their meanings are given in Table \refersec{tab:languagenames}.
649 \item A \DWATstmtlist{} attribute\addtoindexx{statement list attribute}
650 whose value of class \CLASSlineptr{} points to the line number
651 information for this type unit.
653 \textit{Because type units do not describe any code, they
654 do not actually need a line number table, but the line number
655 headers contain a list of directories and file names that
656 may be referenced by the \DWATdeclfile{} attribute of the
657 type or part of its description.}
659 \textit{In an object file with a conventional compilation
660 unit entry, the type unit entries may refer to (share) the
661 line number table used by the compilation unit. In a type
662 unit located in a split compilation unit, the
663 \DWATstmtlistNAME{} attribute refers to a \doublequote{specialized}
664 line number table in the \dotdebuglinedwo{} section, which
665 contains only the list of directories and file names.}
667 \textit{All type unit entries in a \splitDWARFobjectfile{} may
668 (but are not required to) refer to the same
669 \addtoindex{specialized line number table}.}
671 \item A \DWATuseUTFeight{} attribute, which is a flag
672 whose presence indicates that all strings referred to by this type
673 unit entry, its children, and its associated
674 \addtoindex{specialized line number table},
675 are represented using the UTF-8 representation.
678 \item A \DWATstroffsetsbase\addtoindexx{string offsets base attribute}
679 attribute, whose value is of class \CLASSstroffsetsptr.
680 This attribute points
681 to the first string offset of the type unit's contribution to
682 the \dotdebugstroffsets{} section. Indirect string references
683 (using \DWFORMstrx) within the type unit are interpreted
684 as indices relative to this base.
688 A \addtoindex{type unit} entry for a given type T owns a debugging
689 information entry that represents a defining declaration
690 of type T. If the type is nested within enclosing types or
691 namespaces, the debugging information entry for T is nested
692 within debugging information entries describing its containers;
693 otherwise, T is a direct child of the type unit entry.
695 A type unit entry may also own additional debugging information
696 entries that represent declarations of additional types that
697 are referenced by type T and have not themselves been placed in
698 separate type units. Like T, if an additional type U is nested
699 within enclosing types or namespaces, the debugging information
700 entry for U is nested within entries describing its containers;
701 otherwise, U is a direct child of the type unit entry.
703 The containing entries for types T and U are declarations,
704 and the outermost containing entry for any given type T or
705 U is a direct child of the type unit entry. The containing
706 entries may be shared among the additional types and between
707 T and the additional types.
709 \textit{Examples of these kinds of relationships are found in
710 Section \refersec{app:signaturecomputationexample} and
711 Section \refersec{app:declarationscompletingnondefiningdeclarations}.}
714 \textit{Types are not required to be placed in type units. In general,
715 only large types such as structure, class, enumeration, and
716 union types included from header files should be considered
717 for separate type units. Base types and other small types
718 are not usually worth the overhead of placement in separate
719 type units. Types that are unlikely to be replicated, such
720 as those defined in the main source file, are also better
721 left in the main compilation unit.}
723 \section{Module, Namespace and Importing Entries}
724 \textit{Modules and namespaces provide a means to collect related
725 entities into a single entity and to manage the names of
729 \subsection{Module Entries}
730 \label{chap:moduleentries}
731 \textit{Several languages have the concept of a \doublequote{module.}
732 \addtoindexx{Modula-2}
733 A Modula\dash 2 definition module
734 \addtoindexx{Modula-2!definition module}
735 may be represented by a module
737 \addtoindex{declaration attribute}
738 (\DWATdeclaration). A
739 \addtoindex{Fortran 90} module
740 \addtoindexx{Fortran!module (Fortran 90)}
741 may also be represented by a module entry
742 (but no declaration attribute is warranted because \addtoindex{Fortran}
743 has no concept of a corresponding module body).}
745 A module is represented by a debugging information entry
747 tag \DWTAGmoduleTARG.
748 Module entries may own other
749 debugging information entries describing program entities
750 whose declaration scopes end at the end of the module itself.
752 If the module has a name, the module entry has a
753 \DWATname{} attribute
754 \addtoindexx{name attribute}
755 whose value is a null\dash terminated string containing
758 The \addtoindex{module entry} may have either a
762 \addtoindexx{high PC attribute}
764 \addtoindexx{low PC attribute}
766 \DWATranges{} attribute
767 \addtoindexx{ranges attribute}
768 whose values encode the contiguous or non-contiguous address
769 ranges, respectively, of the machine instructions generated for
770 the module initialization
771 code\hypertarget{chap:DWATentrypcentryaddressofmoduleinitialization}{}
772 (see Section \refersec{chap:codeaddressesandranges}).
774 \addtoindexx{entry PC attribute!for module initialization}
775 \DWATentrypc{} attribute whose value is the address of
776 the first executable instruction of that initialization code
777 (see Section \refersec{chap:entryaddress}).
780 If\hypertarget{chap:DWATprioritymodulepriority}{}
781 the module has been assigned a priority, it may have a
782 \addtoindexx{priority attribute}
783 \DWATpriorityDEFN{} attribute.
784 The value of this attribute is a
785 reference to another debugging information entry describing
786 a variable with a constant value. The value of this variable
787 is the actual constant value of the module\textquoteright s priority,
788 represented as it would be on the target architecture.
790 \subsection{Namespace Entries}
791 \label{chap:namespaceentries}
792 \textit{\addtoindex{C++} has the notion of a namespace, which provides a way to
793 \addtoindexx{namespace (C++)}
794 implement name hiding, so that names of unrelated things
795 do not accidentally clash in the
796 \addtoindex{global namespace} when an
797 application is linked together.}
799 A namespace is represented by a debugging information entry
800 with the tag \DWTAGnamespaceTARG. A namespace extension
801 is\hypertarget{chap:DWATextensionpreviousnamespaceextensionororiginalnamespace}{}
802 represented by a \DWTAGnamespaceNAME{} entry with a
803 \DWATextensionDEFN{}\addtoindexx{extension attribute}
804 attribute referring to the previous extension, or if there
805 is no previous extension, to the original
806 \DWTAGnamespaceNAME{}
807 entry. A namespace extension entry does not need to duplicate
808 information in a previous extension entry of the namespace
809 nor need it duplicate information in the original namespace
810 entry. (Thus, for a namespace with a name,
811 a \DWATname{} attribute
812 \addtoindexx{name attribute}
813 need only be attached directly to the original
814 \DWTAGnamespaceNAME{} entry.)
817 Namespace and namespace extension entries may own
818 \addtoindexx{namespace extension entry}
820 \addtoindexx{namespace declaration entry}
821 debugging information entries describing program entities
822 whose declarations occur in the namespace.
824 A namespace may have a
825 \DWATexportsymbolsDEFN{}\livetarg{chap:DWATexportsymbolsofnamespace}{}
826 attribute\addtoindexx{export symbols attribute}
827 \addtoindexx{inline namespace|see{\textit{also} export symbols attribute}}
828 which is a \CLASSflag{} which
829 indicates that all member names defined within the
830 namespace may be referenced as if they were defined within
831 the containing namespace.
833 \textit{This may be used to describe an \addtoindex{inline namespace} in \addtoindex{C++}}.
835 If a type, variable, or function declared in a namespace is
836 defined outside of the body of the namespace declaration,
837 that type, variable, or function definition entry has a
838 \DWATspecification{} attribute
839 \addtoindexx{specification attribute}
840 whose value is a \livelink{chap:classreference}{reference} to the
841 debugging information entry representing the declaration of
842 the type, variable or function. Type, variable, or function
844 \DWATspecification{} attribute
845 \addtoindexx{specification attribute}
847 to duplicate information provided by the declaration entry
848 referenced by the specification attribute.
850 \textit{The \addtoindex{C++} \addtoindex{global namespace}
852 \addtoindexx{global namespace|see{namespace (C++), global}}
854 \addtoindexx{namespace (C++)!global}
856 \texttt{::f}, for example) is not explicitly represented in
857 DWARF with a namespace entry (thus mirroring the situation
858 in \addtoindex{C++} source).
859 Global items may be simply declared with no
860 reference to a namespace.}
862 \textit{The \addtoindex{C++}
863 compilation unit specific \doublequote{unnamed namespace} may
864 \addtoindexx{namespace (C++)!unnamed}
865 \addtoindexx{unnamed namespace|see {namespace (C++), unnamed}}
866 be represented by a namespace entry with no name attribute in
867 the original namespace declaration entry (and therefore no name
868 attribute in any namespace extension entry of this namespace).
869 C++ states that declarations in the unnamed namespace are
870 implicitly available in the containing scope; a producer
871 should make this effect explicit with the \DWATexportsymbols{}
872 attribute, or by using a \DWTAGimportedmodule{} that is a
873 sibling of the namespace entry and references it.}
875 \textit{A compiler emitting namespace information may choose to
876 explicitly represent namespace extensions, or to represent the
877 final namespace declaration of a compilation unit; this is a
878 quality-of-implementation issue and no specific requirements
879 are given here. If only the final namespace is represented,
880 \addtoindexx{namespace (C++)!using declaration}
881 it is impossible for a debugger to interpret using declaration
882 references in exactly the manner defined by the
883 \addtoindex{C++} language.}
885 \textit{For \addtoindex{C++} namespace examples,
886 see Appendix \refersec{app:namespaceexamples}.}
890 \subsection{Imported (or Renamed) Declaration Entries}
891 \label{chap:importedorrenameddeclarationentries}
893 \textit{Some languages support the concept of importing into or
894 making accessible in a given unit certain declarations that occur
895 in a different module or scope. An imported declaration may
896 sometimes be given another name.}
899 An imported declaration is represented by one or
900 \addtoindexx{imported declaration entry}
901 more debugging information entries with the
902 tag \DWTAGimporteddeclarationTARG.
903 When\hypertarget{chap:DWATimportimporteddeclaration}{}
904 an overloaded entity is imported, there is one imported
905 declaration entry for each overloading.
906 Each imported declaration entry has a
907 \DWATimportDEFN{} attribute,\addtoindexx{import attribute}
908 whose value is a \livelink{chap:classreference}{reference} to the
909 debugging information entry representing the declaration that
912 An imported declaration may also have a \DWATname{}
913 attribute\addtoindexx{name attribute}
914 whose value is a null-terminated string containing the
916 imported entity is to be known in the context of the imported
917 declaration entry (which may be different than the name of
918 the entity being imported). If no name is present, then the
919 name by which the entity is to be known is the same as the
920 name of the entity being imported.
922 An imported declaration entry with a name attribute may be
923 used as a general means to rename or provide an alias for
924 \addtoindexx{alias declaration|see{imported declaration entry}}
925 an entity, regardless of the context in which the importing
926 declaration or the imported entity occurs.
928 \textit{A \addtoindex{C++}
929 namespace alias\hypertarget{chap:DWATimportnamespacealias}{}
930 may be represented by an imported declaration entry
931 \addtoindexx{namespace (C++)!alias}
932 with a name attribute whose value is
933 a null-terminated string containing the alias name
934 and a \DWATimportDEFN{} attribute
935 whose value is a \livelink{chap:classreference}{reference} to the
936 applicable original namespace or namespace extension entry.}
938 \textit{A \addtoindex{C++} using declaration may be represented
940 imported\hypertarget{chap:DWATimportnamespaceusingdeclaration}{}
941 \addtoindexx{namespace (C++)!using declaration}
942 declaration entries. When the using declaration
943 refers to an overloaded function, there is one imported
944 declaration entry corresponding to each overloading. Each
945 imported declaration entry has no name attribute but it does
946 have a \DWATimportDEFN{} attribute that refers to the entry for the
947 entity being imported. (\addtoindex{C++}
948 provides no means to \doublequote{rename}
949 an imported entity, other than a namespace).}
952 \textit{A \addtoindex{Fortran} use statement
953 \addtoindexx{Fortran!use statement}
954 \addtoindexx{use statement|see {Fortran, use statement}}
955 with an \doublequote{only list} may be
956 represented by a series of imported declaration entries,
957 one (or more) for each entity that is imported. An entity
958 \addtoindexx{renamed declaration|see{imported declaration entry}}
959 that is renamed in the importing context may be represented
960 by an imported declaration entry with a name attribute that
961 specifies the new local name.
964 \subsection{Imported Module Entries}
965 \label{chap:importedmoduleentries}
967 \textit{Some languages support the concept of importing into or making
968 accessible in a given unit all of the declarations contained
969 within a separate module or namespace.
972 An imported module declaration is represented by a debugging
973 information entry with
974 \addtoindexx{imported module attribute}
976 \addtoindexx{imported module entry}
977 tag \DWTAGimportedmoduleTARG.
979 imported module entry contains a
980 \DWATimport{} attribute
981 \addtoindexx{import attribute}
982 whose value is a \livelink{chap:classreference}{reference}
983 to the module or namespace entry
984 containing the definition and/or declaration entries for
985 the entities that are to be imported into the context of the
986 imported module entry.
988 An imported module declaration may own a set of imported
989 declaration entries, each of which refers to an entry in the
990 module whose corresponding entity is to be known in the context
991 of the imported module declaration by a name other than its
992 name in that module. Any entity in the module that is not
993 renamed in this way is known in the context of the imported
994 module entry by the same name as it is declared in the module.
996 \textit{A \addtoindex{C++} using directive
997 \addtoindexx{namespace (C++)!using directive}
998 \addtoindexx{using directive|see {namespace (C++), using directive}}
999 may be represented by an imported
1000 module\hypertarget{chap:DWATimportnamespaceusingdirective}{}
1001 entry, with a \DWATimportDEFN{} attribute referring to the namespace
1002 entry of the appropriate extension of the namespace (which
1003 might be the original namespace entry) and no owned entries.
1006 \textit{A \addtoindex{Fortran} use statement
1007 \addtoindexx{Fortran!use statement}
1008 with a \doublequote{rename list} may be
1009 represented by an imported module entry with an import
1010 attribute referring to the module and owned entries
1011 corresponding to those entities that are renamed as part of
1015 \textit{A \addtoindex{Fortran} use statement
1016 \addtoindexx{Fortran!use statement}
1017 with neither a \doublequote{rename list} nor
1018 an \doublequote{only list} may be represented by an imported module
1019 entry with an import attribute referring to the module and
1020 no owned child entries.
1023 \textit{A use statement with an \doublequote{only list} is represented by a
1024 series of individual imported declaration entries as described
1025 in Section \refersec{chap:importedorrenameddeclarationentries}.
1029 \textit{A \addtoindex{Fortran} use statement for an entity in a module that is
1030 \addtoindexx{Fortran!use statement}
1031 itself imported by a use statement without an explicit mention
1032 may be represented by an imported declaration entry that refers
1033 to the original debugging information entry. For example, given}
1034 \par % Needed to end paragraph before listing so that it gets a line number
1050 \textit{the imported declaration entry for Q within module C refers
1051 directly to the variable declaration entry for X in module A
1052 because there is no explicit representation for X in module B.
1055 \textit{A similar situation arises for a \addtoindex{C++} using declaration
1056 \addtoindexx{namespace (C++)!using declaration}
1057 \addtoindexx{using declaration|see {namespace (C++), using declaration}}
1058 that imports an entity in terms of a namespace alias. See
1059 Appendix \refersec{app:namespaceexamples}
1063 \subsection{Imported Unit Entries}
1064 \label{chap:importedunitentries}
1065 \hypertarget{chap:DWATimportimportedunit}{}
1066 The place where a normal or partial compilation unit is imported is
1067 represented by a debugging information entry with the
1068 \addtoindexx{imported unit entry}
1069 tag \DWTAGimportedunitTARG.
1070 An imported unit entry contains a
1071 \DWATimportDEFN{} attribute\addtoindexx{import attribute}
1072 whose value is a \livelink{chap:classreference}{reference} to the
1073 normal or partial compilation unit whose declarations logically
1074 belong at the place of the imported unit entry.
1076 \textit{An imported unit entry does not necessarily correspond to
1077 any entity or construct in the source program. It is merely
1078 \doublequote{glue} used to relate a partial unit, or a compilation
1079 unit used as a partial unit, to a place in some other
1082 \section{Subroutine and Entry Point Entries}
1083 \label{chap:subroutineandentrypointentries}
1085 The following tags exist to describe
1086 debugging information entries
1087 \addtoindexx{function entry|see{subroutine entry}}
1089 \addtoindexx{subroutine entry}
1091 \addtoindexx{subprogram entry}
1093 % FIXME: is entry point entry the right index 'entry'?
1094 \addtoindexx{entry point entry}
1098 \begin{tabular}{lp{9.0cm}}
1099 \DWTAGsubprogramTARG{} & A subroutine or function \\
1100 \DWTAGinlinedsubroutine{} & A particular inlined
1101 \addtoindexx{inlined subprogram entry}
1102 instance of a subroutine or function \\
1103 \DWTAGentrypointTARG{} & An alternate entry point \\
1105 \par\condlinenumbers
1108 \subsection{General Subroutine and Entry Point Information}
1109 \label{chap:generalsubroutineandentrypointinformation}
1110 The subroutine or entry point entry has a \DWATname{}
1111 attribute whose value is a null-terminated string containing the
1112 subroutine or entry point name.
1113 It may also have a \DWATlinkagename{} attribute as
1114 described in Section \refersec{chap:linkagenames}.
1116 If the name of the subroutine described by an entry with the
1117 \addtoindexx{subprogram entry}
1118 tag \DWTAGsubprogram{}\hypertarget{chap:DWATexternalexternalsubroutine}{}
1119 is visible outside of its containing
1120 compilation unit, that entry has a
1121 \DWATexternalDEFN{} attribute,\addtoindexx{external attribute}
1122 which is a \livelink{chap:classflag}{flag}.
1124 \textit{Additional attributes for functions that are members of a
1125 class or structure are described in
1126 Section \refersec{chap:memberfunctionentries}.
1129 A\hypertarget{chap:DWATmainsubprogrammainorstartingsubprogram}{}
1130 subroutine entry may contain a
1131 \DWATmainsubprogramDEFN{} attribute
1132 \addtoindexx{main subprogram attribute}
1134 a \CLASSflag{} whose presence indicates that the
1135 subroutine has been identified as the starting function of
1136 the program. If more than one subprogram contains this
1138 any one of them may be the starting subroutine of the program.
1140 \textit{See also Section \refersec{chap:unitentries}) regarding the
1141 related use of this attribute to indicate that a compilation
1142 unit contains the main subroutine of a program.}
1144 \subsubsection{Calling Convention Information}
1145 \hypertarget{chap:DWATcallingconventionforsubprograms}{}
1146 A subroutine entry may contain a
1147 \DWATcallingconventionDEFN{}
1148 \addtoindexx{calling convention attribute!for subprogram}
1149 attribute, whose value is an
1150 \livelink{chap:classconstant}{integer constant}. The set of
1151 \addtoindexi{calling convention codes for subroutines}{calling convention codes!for subroutines}
1152 is given in Table \refersec{tab:callingconventioncodesforsubroutines}.
1154 \begin{simplenametable}[1.4in]{Calling convention codes for subroutines}{tab:callingconventioncodesforsubroutines}
1158 \end{simplenametable}
1160 If this attribute is not present, or its value is the constant
1161 \DWCCnormalTARG, then the subroutine may be safely called by
1162 obeying the \doublequote{standard} calling conventions of the target
1163 architecture. If the value of the calling convention attribute
1164 is the constant \DWCCnocallTARG, the subroutine does not obey
1165 standard calling conventions, and it may not be safe for the
1166 debugger to call this subroutine.
1168 \textit{Note that \DWCCnormal{} is also used as a calling convention
1169 code for certain types
1170 (see Table \refersec{tab:callingconventioncodesfortypes}).}
1172 If the semantics of the language of the compilation unit
1173 containing the subroutine entry distinguishes between ordinary
1174 subroutines and subroutines that can serve as the \doublequote{main
1175 program,} that is, subroutines that cannot be called
1176 directly according to the ordinary calling conventions,
1177 then the debugging information entry for such a subroutine
1178 may have a calling convention attribute whose value is the
1179 constant \DWCCprogramTARG.
1181 \textit{A common debugger feature is to allow the debugger user to call
1182 a subroutine within the subject program. In certain cases,
1183 however, the generated code for a subroutine will not obey
1184 the standard calling conventions for the target architecture
1185 and will therefore not be safe to call from within a debugger.}
1187 \textit{The \DWCCprogram{}
1188 value is intended to support \addtoindex{Fortran} main
1189 \addtoindexx{Fortran!main program}
1190 programs which in some implementations may not be callable
1191 or which must be invoked in a special way. It is not intended
1192 as a way of finding the entry address for the program.}
1195 \subsubsection{Miscellaneous Subprogram Properties}
1196 \textit{In \addtoindex{C}
1197 there is a difference between the types of functions
1198 declared using function prototype style declarations and
1199 those declared using non-prototype declarations.}
1201 A subroutine entry declared with a function prototype style
1202 declaration may have a
1203 \addtoindexx{prototyped attribute}
1204 \DWATprototypedDEFN{} attribute, which is
1206 The attribute indicates whether a subroutine entry point corresponds
1207 to a function declaration that includes parameter prototype information.
1209 A subprogram entry may have
1210 a\hypertarget{chap:DWATelementalelementalpropertyofasubroutine}{}
1211 \DWATelementalDEFN{} attribute,\addtoindexx{elemental attribute}
1212 which is a \livelink{chap:classflag}{flag}.
1213 The attribute indicates whether the subroutine
1214 or entry point was declared with the \doublequote{elemental} keyword
1217 A\hypertarget{chap:DWATpurepurepropertyofasubroutine}{}
1218 subprogram entry may have a
1219 \addtoindexx{pure attribute}
1220 \DWATpureDEFN{} attribute, which is
1221 a \livelink{chap:classflag}{flag}.
1222 The attribute indicates whether the subroutine was
1223 declared with the \doublequote{pure} keyword or property.
1225 A\hypertarget{chap:DWATrecursiverecursivepropertyofasubroutine}{}
1226 subprogram entry may have a
1227 \addtoindexx{recursive attribute}
1228 \DWATrecursiveDEFN{} attribute, which
1229 is a \livelink{chap:classflag}{flag}.
1230 The attribute indicates whether the subroutine
1231 or entry point was declared with the \doublequote{recursive} keyword
1234 A subprogram entry may have a
1236 \livetargi{chap:DWATnoreturnofsubprogram}{attribute}{noreturn attribute},
1237 which is a \CLASSflag. The attribute
1238 indicates whether the subprogram was declared with the \doublequote{noreturn} keyword or property
1239 indicating that the subprogram can be called, but will never return to its caller.
1242 \textit{The \addtoindex{Fortran}
1243 language allows the keywords \texttt{elemental}, \texttt{pure}
1244 and \texttt{recursive} to be included as part of the declaration of
1245 a subroutine; these attributes reflect that usage. These
1246 attributes are not relevant for languages that do not support
1247 similar keywords or syntax. In particular, the \DWATrecursiveNAME{}
1248 attribute is neither needed nor appropriate in languages such
1249 as \addtoindex{C} where functions support recursion by default.}
1252 \subsubsection{Call Site-Related Attributes}
1253 \textit{While subprogram attributes in the previous section provide
1254 information about the subprogram and its entry point(s) as a whole,
1255 the following attributes provide summary information about the calls
1256 that occur within a subprogram.}
1258 A subroutine entry may have \DWATcallalltailcalls,
1259 \DWATcallallcalls{} and/or \DWATcallallsourcecalls{}
1260 attributes, each of which is a \CLASSflag.
1261 \addtoindexx{call site summary information}
1262 \addtoindexx{subroutine call site summary attributes}
1263 These flags indicate the completeness of the call site
1264 information provided by call site entries (see
1265 Section \refersec{chap:callsiteentries}) within the subprogram.
1267 The \DWATcallalltailcallsDEFN{}
1268 \livetargi{chap:DWATcallalltailcallsofasubprogram}{attribute}{all tail calls summary attribute}
1269 indicates that every tail call
1270 that occurs in the code for the subprogram is described by a
1271 \DWTAGcallsite{} entry.
1272 (There may or may not be other non-tail calls to some of the same
1273 target subprograms.)
1275 The \DWATcallallcallsDEFN{}
1276 \livetargi{chap:DWATcallallcallsofasubprogram}{attribute}{all calls summary attribute}
1277 indicates that every non-inlined call
1278 (either a tail call or a normal call) that occurs in the code for the subprogram
1279 is described by a \DWTAGcallsite{} entry.
1281 The \DWATcallallsourcecallsDEFN{}
1282 \livetargi{chap:DWATcallallsourcecallsofasubprogram}{attribute}{all source calls summary attribute}
1283 indicates that every call that occurs in the
1284 code for the subprogram, including every call inlined into it, is described by either a
1285 \DWTAGcallsite{} entry or a \DWTAGinlinedsubroutine{} entry; further, any call
1286 that is optimized out is nonetheless also described using a \DWTAGcallsite{} entry
1287 that has neither a \DWATcallpc{} nor \DWATcallreturnpc{} attribute.
1289 \textit{The \DWATcallallsourcecallsNAME{} attribute is intended for debugging
1290 information format consumers that analyze call graphs.}
1292 If the the \DWATcallallsourcecalls{} attribute is present then the
1293 \DWATcallallcalls{} and \DWATcallalltailcalls{} attributes are
1294 also implicitly present. Similarly, if the
1295 \DWATcallallcalls{} attribute is present then the \DWATcallalltailcalls{}
1296 attribute is implicitly present.
1299 \subsection{Subroutine and Entry Point Return Types}
1300 \label{chap:subroutineandentrypointreturntypes}
1302 If\hypertarget{chap:DWATtypetypeofsubroutinereturn}{}
1303 the subroutine or entry point
1304 \addtoindexx{return type of subroutine}
1305 is a function that returns a
1306 value, then its debugging information entry has
1307 \addtoindexx{type attribute}
1308 a \DWATtypeDEFN{} attribute
1309 to denote the type returned by that function.
1311 \textit{Debugging information entries for
1312 \addtoindex{C} void functions should
1313 not have an attribute for the return type. }
1315 \textit{Debugging information entries for declarations of \addtoindex{C++}
1316 member functions with an
1317 \autoreturntype{} specifier should use an unspecified type entry (see
1318 Section \refersec{chap:unspecifiedtypeentries}).
1319 The debugging information entry for the corresponding definition
1320 should provide the deduced return type. This practice causes the description of
1321 the containing class to be consistent across compilation units, allowing the class
1322 declaration to be placed into a separate type unit if desired.}
1325 \subsection{Subroutine and Entry Point Locations}
1326 \label{chap:subroutineandentrypointlocations}
1328 A subroutine entry may have either a \DWATlowpc{} and
1329 \DWAThighpc{} pair of attributes or a \DWATranges{} attribute
1330 \addtoindexx{ranges attribute}
1332 \addtoindexx{high PC attribute}
1334 \addtoindexx{low PC attribute}
1335 encode the contiguous or non-contiguous address
1336 ranges, respectively, of the machine instructions generated
1337 for the subroutine (see
1338 Section \refersec{chap:codeaddressesandranges}).
1340 A\hypertarget{chap:DWATentrypcentryaddressofsubprogram}{}
1341 subroutine entry may also have a
1342 \addtoindexx{entry PC attribute!for subroutine}
1343 \DWATentrypc{} attribute
1344 whose value is the address of the first executable instruction
1345 of the subroutine (see
1346 Section \refersec{chap:entryaddress}).
1348 An entry point has a \DWATlowpc{} attribute whose value is the
1349 relocated address of the first machine instruction generated
1350 for the entry point.
1353 %\textit{While the \DWATentrypc{} attribute
1354 %\addtoindexx{entry pc attribute!for subroutine}
1355 %might also seem appropriate for this purpose, historically the
1356 %\DWATlowpc{} attribute was used before the
1357 %\DWATentrypc{} was introduced (in
1358 %\addtoindex{DWARF Version 3}).
1359 %There is insufficient reason to change this.}
1361 Subroutines and entry points may also have
1362 \DWATsegment{}\hypertarget{chap:DWATaddressclasssubroutineorsubroutinetype}{}
1363 \addtoindexx{segment attribute} and
1364 \DWATaddressclassDEFN{}\addtoindexx{address class attribute}
1365 attributes, as appropriate, to specify
1366 which segments the code for the subroutine resides in and
1367 the addressing mode to be used in calling that subroutine.
1369 A subroutine entry representing a subroutine declaration
1370 that is not also a definition does not have code address or
1374 \subsection{Declarations Owned by Subroutines and Entry Points}
1375 \label{chap:declarationsownedbysubroutinesandentrypoints}
1376 \addtoindexx{subroutine formal parameters}
1377 The declarations enclosed by a subroutine or entry point are
1378 represented by debugging information entries that are owned
1379 by the subroutine or entry point entry. Entries representing
1380 \addtoindexx{formal parameter}
1381 the formal parameters of the subroutine or entry point appear
1382 in the same order as the corresponding declarations in the
1386 \textit{There is no ordering requirement for entries for declarations
1387 other than formal parameters. The formal parameter
1388 entries may be interspersed with other entries used by formal
1389 parameter entries, such as type entries.}
1391 The unspecified (sometimes called \doublequote{varying})
1392 parameters of a subroutine parameter list are
1393 represented by a debugging information
1394 entry\addtoindexx{unspecified parameters entry}
1395 with the tag \DWTAGunspecifiedparametersTARG.
1398 The entry for a subroutine that includes a
1399 \addtoindex{Fortran}
1400 \addtoindexx{Fortran!common block}
1401 \livelink{chap:fortrancommonblock}{common}
1402 \livelink{chap:commonblockentry}{block}
1403 \addtoindexx{common block|see{Fortran common block}}
1404 has a child entry with the
1405 tag \DWTAGcommoninclusionTARG.
1406 The\hypertarget{chap:commonreferencecommonblockusage}{}
1407 common inclusion entry has a
1408 \DWATcommonreferenceDEFN{} attribute
1409 \addtoindexx{common block reference attribute}
1410 whose value is a \livelink{chap:classreference}{reference}
1411 to the debugging information entry
1412 for the common \nolink{block} being included
1413 (see Section \refersec{chap:commonblockentries}).
1415 \subsection{Low-Level Information}
1416 \label{chap:lowlevelinformation}
1418 A\hypertarget{chap:DWATreturnaddrsubroutinereturnaddresssavelocation}{}
1419 subroutine or entry point entry may have a
1420 \addtoindexx{return address attribute}
1421 \DWATreturnaddrDEFN{}
1422 attribute, whose value is a location description. The location
1423 specified is the place where the return address for the
1424 subroutine or entry point is stored.
1426 A\hypertarget{chap:DWATframebasesubroutineframebaseaddress}{}
1427 subroutine or entry point entry may also have a
1428 \addtoindexx{frame base attribute}
1429 \DWATframebaseDEFN{} attribute, whose value is a location
1430 description that describes the \doublequote{frame base} for the
1431 subroutine or entry point. If the location description is
1432 a simple register location description, the given register
1433 contains the frame base address. If the location description is
1434 a DWARF expression, the result of evaluating that expression
1435 is the frame base address. Finally, for a
1436 \addtoindex{location list},
1437 this interpretation applies to each location description
1438 contained in the list of \addtoindex{location list} entries.
1440 \textit{The use of one of the \DWOPregn{}
1441 operations in this context is equivalent to using
1442 \DWOPbregn(0) but more
1443 compact. However, these are not equivalent in general.}
1446 \textit{The frame base for a subprogram is typically an address
1447 relative to the first unit of storage allocated for the
1448 subprogram\textquoteright s stack frame. The \DWATframebase{} attribute
1449 can be used in several ways:}
1450 \begin{enumerate}[1. ]
1451 \item \textit{In subprograms that need
1452 \addtoindexx{location list}
1453 location lists to locate local
1454 variables, the \DWATframebase{} can hold the needed location
1455 list, while all variables\textquoteright\ location descriptions can be
1456 simpler ones involving the frame base.}
1458 \item \textit{It can be used in resolving \doublequote{up\dash level} addressing
1459 within nested routines.
1460 (See also \DWATstaticlink, below)}
1464 \textit{Some languages support nested subroutines. In such languages,
1465 it is possible to reference the local variables of an
1466 outer subroutine from within an inner subroutine. The
1467 \DWATstaticlink{} and \DWATframebase{} attributes allow
1468 debuggers to support this same kind of referencing.}
1470 If\hypertarget{chap:DWATstaticlinklocationofuplevelframe}{}
1471 a subroutine or entry point is nested, it may have a
1472 \addtoindexx{address!uplevel|see {static link attribute}}
1473 \addtoindexx{uplevel address|see {static link attribute}}
1474 \DWATstaticlinkDEFN{} attribute, whose value is a location
1475 description that computes the frame base of the relevant
1476 instance of the subroutine that immediately encloses the
1477 subroutine or entry point.
1479 In the context of supporting nested subroutines, the
1480 \DWATframebase{} attribute value obeys the following constraints:
1482 \begin{enumerate}[1. ]
1484 It computes a value that does not change during the
1485 life of the subprogram, and
1487 \item The computed value is unique among instances of
1488 the same subroutine.
1490 \textit{For typical \DWATframebase{} use, this
1491 means that a recursive subroutine\textquoteright s stack frame must have
1496 \textit{If a debugger is attempting to resolve an up\dash level reference
1497 to a variable, it uses the nesting structure of DWARF to
1498 determine which subroutine is the lexical parent and the
1499 \DWATstaticlink{} value to identify the appropriate active
1500 frame of the parent. It can then attempt to find the reference
1501 within the context of the parent.}
1505 \subsection{Types Thrown by Exceptions}
1506 \label{chap:typesthrownbyexceptions}
1508 \textit{In \addtoindex{C++} a subroutine may declare a set of types which
1509 it may validly throw.}
1511 If a subroutine explicitly declares that it may throw
1512 \addtoindexx{exception thrown|see{thrown type entry}}
1514 \addtoindexx{thrown exception|see{thrown type entry}}
1515 exception of one or more types, each such type is
1516 represented by a debugging information entry with
1517 \addtoindexx{thrown type entry}
1519 \DWTAGthrowntypeTARG.
1520 Each such entry is a child of the entry
1521 representing the subroutine that may throw this type. Each
1522 thrown type entry contains
1523 \addtoindexx{type attribute}
1524 a \DWATtype{} attribute, whose
1525 value is a \livelink{chap:classreference}{reference}
1526 to an entry describing the type of the
1527 exception that may be thrown.
1529 \subsection{Function Template Instantiations}
1530 \label{chap:functiontemplateinstantiations}
1532 \textit{In \addtoindex{C++}, a function template is a generic definition of
1533 a function that is instantiated differently for calls with
1534 values of different types. DWARF does not represent the generic
1535 template definition, but does represent each instantiation.}
1538 A \addtoindex{function template instantiation}\addtoindexx{template instantiation!function}
1539 is represented by a debugging information entry with the
1540 \addtoindexx{subprogram entry!use for template instantiation}
1541 tag \DWTAGsubprogram.
1543 exceptions, such an entry will contain the same attributes and
1544 will have the same types of child entries as would an entry
1545 for a subroutine defined explicitly using the instantiation
1546 types and values. The exceptions are:
1548 \begin{enumerate}[1. ]
1549 \item Template parameters are described and referenced as specified in
1550 Section \refersec{chap:templateparameters}.
1553 \item If the compiler has generated a separate compilation unit
1554 to hold the template instantiation and that compilation unit
1555 has a different name from the compilation unit containing
1556 the template definition, the name attribute for the debugging
1557 information entry representing that compilation unit is empty
1560 \item If the subprogram entry representing the template
1561 instantiation or any of its child entries contain declaration
1562 coordinate attributes, those attributes refer to the source
1563 for the template definition, not to any source generated
1564 artificially by the compiler for this instantiation.
1569 \subsection{Inlinable and Inlined Subroutines}
1570 \label{chap:inlinedsubroutines}
1571 A declaration or a definition of an inlinable subroutine
1572 is represented by a debugging information entry with the
1573 tag \DWTAGsubprogram.
1574 The entry for a subroutine
1575 \addtoindexx{subprogram entry!use in inlined subprogram}
1576 that is\hypertarget{chap:DWATinlineinlinedsubroutine}{}
1577 explicitly declared to be available for inline expansion or
1578 that was expanded inline implicitly by the compiler has a
1579 \addtoindexx{inline attribute}
1580 \DWATinlineDEFN{} attribute whose value is an
1581 \livelink{chap:classconstant}{integer constant}. The
1582 set of values for the \DWATinline{} attribute is given in
1583 Table \refersec{tab:inlinecodes}.
1587 \caption{Inline codes}
1588 \label{tab:inlinecodes}
1589 \begin{tabular}{l|P{8cm}}
1591 Name&Meaning\\ \hline
1592 \DWINLnotinlinedTARG{} & Not declared inline nor inlined by the
1593 \mbox{compiler} (equivalent to the absence of the
1594 containing \DWATinline{} attribute) \\
1595 \DWINLinlinedTARG{} & Not declared inline but inlined by the \mbox{compiler} \\
1596 \DWINLdeclarednotinlinedTARG{} & Declared inline but
1597 not inlined by the \mbox{compiler} \\
1598 \DWINLdeclaredinlinedTARG{} & Declared inline and inlined by the
1604 \textit{In \addtoindex{C++}, a function or a constructor declared with
1605 \addttindex{constexpr} is implicitly declared inline. The abstract
1606 instance (see Section \refersec{chap:abstractinstances})
1607 is represented by a debugging information
1608 entry with the tag \DWTAGsubprogram. Such an entry has a
1609 \DWATinline{} attribute whose value is \DWINLinlined.}
1612 \subsubsection{Abstract Instances}
1613 \label{chap:abstractinstances}
1614 Any subroutine entry that contains a
1615 \DWATinlineDEFN{} attribute\addtoindexx{inline attribute}
1616 whose value is other than
1618 is known as an \definition{abstract instance root}.
1619 \addtoindexx{abstract instance!root}
1620 \hypertarget{chap:DWATinlineabstracttinstance}{}
1621 Any debugging information entry that is owned (either
1622 directly or indirectly) by an abstract instance root
1624 \definition{abstract instance entry.}\addtoindexx{abstract instance!entry}
1625 Any set of abstract instance entries that are all
1626 children (either directly or indirectly) of some abstract
1627 instance root, together with the root itself, is known as an
1628 \definition{abstract instance tree.}\addtoindexx{abstract instance!tree}
1629 However, in the case where an abstract instance tree is
1630 nested within another abstract instance tree, the entries in the
1631 \addtoindex{nested abstract instance}
1632 tree are not considered to be entries in the outer abstract
1636 Each abstract instance root is either part of a larger
1637 \addtoindexx{abstract instance!root}
1638 tree (which gives a context for the root) or
1639 \addtoindexx{specification attribute}
1641 \DWATspecification{}
1642 to refer to the declaration in context.
1644 \textit{For example, in \addtoindex{C++} the context might be a namespace
1645 declaration or a class declaration.}
1647 \textit{Abstract instance trees are defined so that no entry is part
1648 of more than one abstract instance tree.}
1650 Attributes and children in an abstract instance are shared
1651 by all concrete instances (see Section \refersec{chap:concreteinstances}).
1653 A debugging information entry that is a member of an abstract
1654 instance tree may not contain any attributes which describe
1655 aspects of the subroutine which vary between distinct inlined
1656 expansions or distinct out-of-line expansions.
1658 \textit{For example,
1659 \addtoindexx{entry pc attribute!and abstract instance}
1660 the \DWATlowpc,\addtoindexx{low PC attribute!and abstract instance}
1661 \DWAThighpc,\addtoindexx{high PC attribute!and abstract instance}
1662 \DWATranges,\addtoindexx{ranges attribute!and abstract instance}
1663 \DWATentrypc,\addtoindexx{entry PC attribute!and abstract instance}
1664 \DWATlocation,\addtoindexx{location attribute!and abstract instance}
1665 \DWATreturnaddr,\addtoindexx{return address attribute!and abstract instance}
1666 \DWATstartscope,\addtoindexx{start scope attribute!and abstract instance}
1668 \DWATsegment{}\addtoindexx{segment attribute!and abstract instance}
1669 attributes typically should be omitted; however, this list is not
1673 \textit{It would not make sense normally to put these attributes into
1674 abstract instance entries since such entries do not represent
1675 actual (concrete) instances and thus do not actually exist at
1676 run\dash time. However,
1677 see Appendix \refersec{app:inlineouteronenormalinner}
1678 for a contrary example.}
1680 The rules for the relative location of entries belonging to
1681 abstract instance trees are exactly the same as for other
1682 similar types of entries that are not abstract. Specifically,
1683 the rule that requires that an entry representing a declaration
1684 be a direct child of the entry representing the scope of the
1685 declaration applies equally to both abstract and non-abstract
1686 entries. Also, the ordering rules for formal parameter entries,
1687 member entries, and so on, all apply regardless of whether
1688 or not a given entry is abstract.
1691 \subsubsection{Concrete Instances}
1692 \label{chap:concreteinstances}
1694 Each inline expansion of a subroutine is represented
1695 by a debugging information entry with the
1696 tag \DWTAGinlinedsubroutineTARG. Each such entry is a direct
1697 child of the entry that represents the scope within which
1698 the inlining occurs.
1701 Each inlined subroutine entry may have either a
1703 and \DWAThighpc{} pair of attributes
1704 \addtoindexx{high PC attribute}
1705 \addtoindexx{low PC attribute}
1706 or a \DWATranges{}\addtoindexx{ranges attribute}
1707 attribute whose values encode the contiguous or non-contiguous
1708 address ranges, respectively, of the machine instructions
1709 generated for the inlined subroutine (see
1710 Section \referfol{chap:codeaddressesandranges}).
1711 An\hypertarget{chap:DWATentrypcentryaddressofinlinedsubprogram}{}
1712 inlined subroutine entry may
1713 \addtoindexx{inlined subprogram entry!in concrete instance}
1715 \addtoindexx{inlined subprogram entry}
1717 \addtoindexx{entry PC attribute!for inlined subprogram}
1720 attribute, representing the first executable instruction of
1721 the inline expansion (see
1722 Section \refersec{chap:entryaddress}).
1724 % Positions of the 3 targets here is a bit arbitrary.
1725 An inlined\hypertarget{chap:DWATcalllinelinenumberofinlinedsubroutinecall}{}
1726 subroutine\hypertarget{chap:DWATcallcolumncolumnpositionofinlinedsubroutinecall}{}
1727 entry\hypertarget{chap:DWATcallfilefilecontaininginlinedsubroutinecall}{}
1728 may also have \DWATcallfileDEFN,
1729 \DWATcalllineDEFN{} and \DWATcallcolumnDEFN{} attributes,
1730 \addtoindexx{inlined call location attributes}
1732 value is an \livelink{chap:classconstant}{integer constant}.
1733 These attributes represent the
1734 source file, source line number, and source column number,
1735 respectively, of the first character of the statement or
1736 expression that caused the inline expansion. The call file,
1737 call line, and call column attributes are interpreted in
1738 the same way as the declaration file, declaration line, and
1739 declaration column attributes, respectively (see
1740 Section \refersec{chap:declarationcoordinates}).
1742 \textit{The call file, call line and call column coordinates do not
1743 describe the coordinates of the subroutine declaration that
1744 was inlined, rather they describe the coordinates of the call.
1747 An inlined subroutine entry may have
1748 a\hypertarget{chap:DWATconstexprcompiletimeconstantfunction}{}
1749 \DWATconstexprDEFN{} attribute,\addtoindexx{constant expression attribute}
1750 which is a \livelink{chap:classflag}{flag}
1751 whose presence indicates that the
1752 subroutine has been evaluated as a compile\dash time constant. Such
1753 an entry may also have a \DWATconstvalue{} attribute,
1754 whose value may be of any form that is appropriate for the
1755 representation of the subroutine's return value. The value of
1756 this attribute is the actual return value of the subroutine,
1757 represented as it would be on the target architecture.
1759 \textit{In \addtoindex{C++}, if a function or a constructor declared with
1760 \addttindex{constexpr}
1761 is called with constant expressions, then the corresponding
1762 concrete inlined instance has a
1763 \DWATconstexpr{} attribute,
1764 as well as a \DWATconstvalue{} attribute whose value represents
1765 the actual return value of the concrete inlined instance.}
1768 Any debugging information entry that is owned (either
1769 directly or indirectly) by a debugging information entry
1770 with the tag \DWTAGinlinedsubroutine{} is referred to as a
1771 \doublequote{concrete inlined instance entry.} Any entry that has
1773 \DWTAGinlinedsubroutine{}
1774 is known as a \doublequote{concrete inlined instance root.}
1775 Any set of concrete inlined instance
1776 entries that are all children (either directly or indirectly)
1777 of some concrete inlined instance root, together with the root
1778 itself, is known as a \doublequote{concrete inlined instance tree.}
1779 However, in the case where a concrete inlined instance tree
1780 is nested within another concrete instance tree, the entries
1781 in the \addtoindex{nested concrete inline instance} tree
1782 are not considered to
1783 be entries in the outer concrete instance tree.
1786 \textit{Concrete inlined instance trees are defined so that no entry
1787 is part of more than one concrete inlined instance tree. This
1788 simplifies later descriptions.}
1790 Each concrete inlined instance tree is uniquely associated
1791 with one (and only one) abstract instance tree.
1793 \textit{Note, however, that the reverse is not true. Any given abstract
1794 instance tree may be associated with several different concrete
1795 inlined instance trees, or may even be associated with zero
1796 concrete inlined instance trees.}
1798 Concrete inlined instance entries may omit attributes that
1799 are not specific to the concrete instance (but present in
1800 the abstract instance) and need include only attributes that
1801 are specific to the concrete instance (but omitted in the
1802 abstract instance). In place of these omitted attributes,
1803 each\hypertarget{chap:DWATabstractorigininlineinstance}{}
1804 concrete inlined instance entry has a
1805 \addtoindexx{abstract origin attribute}
1806 \DWATabstractoriginDEFN{}
1807 attribute that may be used to obtain the missing information
1808 (indirectly) from the associated abstract instance entry. The
1809 value of the abstract origin attribute is a reference to the
1810 associated abstract instance entry.
1812 If an entry within a concrete inlined instance tree contains
1813 attributes describing the
1814 \addtoindexx{declaration coordinates!in concrete instance}
1815 \livelink{chap:declarationcoordinates}{declaration coordinates}
1816 of that entry, then those attributes refer to the file, line
1817 and column of the original declaration of the subroutine,
1818 not to the point at which it was inlined. As a consequence,
1819 they may usually be omitted from any entry that has an abstract
1823 For each pair of entries that are associated via a
1824 \addtoindexx{abstract origin attribute}
1825 \DWATabstractorigin{} attribute, both members of the pair
1826 have the same tag. So, for example, an entry with the tag
1827 \DWTAGvariable{} can only be associated with another entry
1828 that also has the tag \DWTAGvariable. The only exception
1829 to this rule is that the root of a concrete instance tree
1830 (which must always have the tag \DWTAGinlinedsubroutine)
1831 can only be associated with the root of its associated abstract
1832 instance tree (which must have the tag \DWTAGsubprogram).
1835 In general, the structure and content of any given concrete
1836 inlined instance tree will be closely analogous to the
1837 structure and content of its associated abstract instance
1838 tree. There are a few exceptions:
1840 \begin{enumerate}[1. ]
1841 \item An entry in the concrete instance tree may be omitted if
1843 \addtoindexx{abstract origin attribute}
1844 \DWATabstractorigin{} attribute and either
1845 has no children, or its children are omitted. Such entries
1846 would provide no useful information. In C\dash like languages,
1847 such entries frequently include types, including structure,
1848 union, class, and interface types; and members of types. If any
1849 entry within a concrete inlined instance tree needs to refer
1850 to an entity declared within the scope of the relevant inlined
1851 subroutine and for which no concrete instance entry exists,
1852 the reference refers to the abstract instance entry.
1855 \item Entries in the concrete instance tree which are associated
1856 with entries in the abstract instance tree such that neither
1857 has a \DWATname{} attribute,
1858 \addtoindexx{name attribute}
1859 and neither is referenced by
1860 any other debugging information entry, may be omitted. This
1861 may happen for debugging information entries in the abstract
1862 instance trees that became unnecessary in the concrete instance
1863 tree because of additional information available there. For
1864 example, an anonymous variable might have been created and
1865 described in the abstract instance tree, but because of
1866 the actual parameters for a particular inlined expansion,
1867 it could be described as a constant value without the need
1868 for that separate debugging information entry.
1871 \item A concrete instance tree may contain entries which do
1872 not correspond to entries in the abstract instance tree
1873 to describe new entities that are specific to a particular
1874 inlined expansion. In that case, they will not have associated
1875 entries in the abstract instance tree, do not contain
1876 \addtoindexx{abstract origin attribute}
1877 \DWATabstractorigin{} attributes, and must contain all their
1878 own attributes directly. This allows an abstract instance tree
1879 to omit debugging information entries for anonymous entities
1880 that are unlikely to be needed in most inlined expansions. In
1881 any expansion which deviates from that expectation, the
1882 entries can be described in its concrete inlined instance tree.
1886 \subsubsection{Out-of-Line Instances of Inlined Subroutines}
1887 \label{chap:outoflineinstancesofinlinedsubroutines}
1888 Under some conditions, compilers may need to generate concrete
1889 executable instances of inlined subroutines other than at
1890 points where those subroutines are actually called. Such
1891 concrete instances of inlined subroutines are referred to as
1892 \doublequote{concrete out\dash of\dash line instances.}
1894 \textit{In \addtoindex{C++}, for example,
1895 taking the address of a function declared
1896 to be inline can necessitate the generation of a concrete
1897 out\dash of\dash line instance of the given function.}
1899 The DWARF representation of a concrete out-of-line instance
1900 of an inlined subroutine is essentially the same as for a
1901 concrete inlined instance of that subroutine (as described in
1902 the preceding section). The representation of such a concrete
1903 % It is critical that the hypertarget and livelink be
1904 % separated to avoid problems with latex.
1906 \addtoindexx{abstract origin attribute}
1908 \hypertarget{chap:DWATabstractoriginoutoflineinstance}{}
1910 \DWATabstractoriginDEFN{}
1911 attributes in exactly the same way as they are used for
1912 a concrete inlined instance (that is, as references to
1913 corresponding entries within the associated abstract instance
1916 The differences between the DWARF representation of a
1917 concrete out\dash of\dash line instance of a given subroutine and the
1918 representation of a concrete inlined instance of that same
1919 subroutine are as follows:
1920 \begin{enumerate}[1. ]
1921 \item The root entry for a concrete out\dash of\dash line instance
1922 of a given inlined subroutine has the same tag as does its
1923 associated (abstract) inlined subroutine entry (that is, tag
1924 \DWTAGsubprogram{} rather than \DWTAGinlinedsubroutine).
1926 \item The root entry for a concrete out\dash of\dash line instance tree
1927 is normally owned by the same parent entry that also owns
1928 the root entry of the associated abstract instance. However,
1929 it is not required that the abstract and out\dash of\dash line instance
1930 trees be owned by the same parent entry.
1934 \subsubsection{Nested Inlined Subroutines}
1935 \label{nestedinlinedsubroutines}
1936 Some languages and compilers may permit the logical nesting of
1937 a subroutine within another subroutine, and may permit either
1938 the outer or the nested subroutine, or both, to be inlined.
1940 For a non-inlined subroutine nested within an inlined
1941 subroutine, the nested subroutine is described normally in
1942 both the abstract and concrete inlined instance trees for
1943 the outer subroutine. All rules pertaining to the abstract
1944 and concrete instance trees for the outer subroutine apply
1945 also to the abstract and concrete instance entries for the
1949 For an inlined subroutine nested within another inlined
1950 subroutine, the following rules apply to their abstract and
1951 \addtoindexx{abstract instance!nested}
1952 \addtoindexx{concrete instance!nested}
1953 concrete instance trees:
1955 \begin{enumerate}[1. ]
1956 \item The abstract instance tree for the nested subroutine is
1957 described within the abstract instance tree for the outer
1958 subroutine according to the rules in
1959 Section \refersec{chap:abstractinstances}, and
1960 without regard to the fact that it is within an outer abstract
1963 \item Any abstract instance tree for a nested subroutine is
1964 always omitted within the concrete instance tree for an
1967 \item A concrete instance tree for a nested subroutine is
1968 always omitted within the abstract instance tree for an
1971 \item The concrete instance tree for any inlined or
1972 \addtoindexx{out-of-line instance}
1974 \addtoindexx{out-of-line instance|see{\textit{also} concrete out-of-line instance}}
1975 expansion of the nested subroutine is described within a
1976 concrete instance tree for the outer subroutine according
1978 Sections \refersec{chap:concreteinstances} or
1979 \referfol{chap:outoflineinstancesofinlinedsubroutines}
1981 and without regard to the fact that it is within an outer
1982 concrete instance tree.
1985 \textit{See Appendix \refersec{app:inliningexamples}
1986 for discussion and examples.}
1988 \subsection{Trampolines}
1989 \label{chap:trampolines}
1991 \textit{A trampoline is a compiler\dash generated subroutine that serves
1992 as\hypertarget{chap:DWATtrampolinetargetsubroutine}{}
1993 an intermediary in making a call to another subroutine. It may
1994 adjust parameters and/or the result (if any) as appropriate
1995 to the combined calling and called execution contexts.}
1997 A trampoline is represented by a debugging information entry
1998 \addtoindexx{trampoline (subprogram) entry}
1999 with the tag \DWTAGsubprogram{} or \DWTAGinlinedsubroutine{}
2001 \addtoindexx{trampoline attribute}
2002 a \DWATtrampolineDEFN{} attribute.
2004 attribute indicates the target subroutine of the trampoline,
2005 that is, the subroutine to which the trampoline passes
2006 control. (A trampoline entry may but need not also have a
2007 \DWATartificial{} attribute.)
2010 The value of the trampoline attribute may be represented
2011 using any of the following forms:
2014 \item If the value is of class \CLASSreference{}, then the value
2015 specifies the debugging information entry of the target
2018 \item If the value is of class \CLASSaddress{}, then the value is
2019 the relocated address of the target subprogram.
2022 \item If the value is of class \CLASSstring{}, then the value is the
2023 (possibly mangled) \addtoindexx{mangled names}
2024 name of the target subprogram.
2026 \item If the value is of class \CLASSflag, then the value true
2027 indicates that the containing subroutine is a trampoline but
2028 that the target subroutine is not known.
2032 The target subprogram may itself be a trampoline. (A sequence
2033 of trampolines necessarily ends with a non-trampoline
2036 \textit{In \addtoindex{C++}, trampolines may be used to implement
2037 derived virtual member functions; such trampolines typically
2039 \texttt{this} parameter\index{this parameter@\texttt{this} parameter}
2040 in the course of passing control.
2041 Other languages and environments may use trampolines in a manner
2042 sometimes known as transfer functions or transfer vectors.}
2044 \textit{Trampolines may sometimes pass control to the target
2045 subprogram using a branch or jump instruction instead of a
2046 call instruction, thereby leaving no trace of their existence
2047 in the subsequent execution context. }
2049 \textit{This attribute helps make it feasible for a debugger to arrange
2050 that stepping into a trampoline or setting a breakpoint in
2051 a trampoline will result in stepping into or setting the
2052 breakpoint in the target subroutine instead. This helps to
2053 hide the compiler generated subprogram from the user. }
2055 \section{Call Site Entries and Parameters}
2056 \label{chap:callsiteentriesandparameters}
2058 A call site entry describes a call from one subprogram to another in the
2059 source program. It provides information about the actual parameters of
2060 the call so that they may be more easily accessed by a debugger. When
2061 used together with call frame information
2062 (see Section \refersec{chap:callframeinformation}),
2063 call site entries can be useful for computing the value of an actual parameter
2064 passed by a caller, even when the location description for the callee's
2065 corresponding formal parameter does not provide a current location for
2066 the formal parameter.}
2068 \textit{The DWARF expression for computing the value of an actual parameter at
2069 a call site may refer to registers or memory locations. The expression
2070 assumes these contain the values they would have at the point where the
2071 call is executed. After the called subprogram has been entered, these
2072 registers and memory locations might have been modified. In order to
2073 recover the values that existed at the point of the call (to allow
2074 evaluation of the DWARF expression for the actual parameter), a debugger
2075 may virtually unwind the subprogram activation
2076 (see Section \refersec{chap:callframeinformation}). Any
2077 register or memory location that cannot be recovered is referred to as
2078 "clobbered by the call."}
2080 A source call can be compiled into different types of machine code:
2083 A \textit{normal call} uses a call-like instruction which transfers
2084 control to the start of some subprogram and preserves the call site
2085 location for use by the callee.
2088 A \textit{tail call} uses a jump-like instruction which
2089 transfers control to the start of some subprogram, but
2090 there is no call site location address to preserve
2091 (and thus none is available using the
2092 virtual unwind information).
2095 A \textit{tail recursion call} is a call
2096 to the current subroutine which is compiled as a jump
2097 to the current subroutine.
2101 An \textit{inline (or inlined) call} is a call to an inlined subprogram,
2102 where at least one instruction has the location of the inlined subprogram
2103 or any of its blocks or inlined subprograms.
2107 There are also different types of \doublequote{optimized out} calls:
2110 An \textit{optimized out (normal) call} is a call that is in unreachable code that
2111 has not been emitted (such as, for example, the call to \texttt{foo} in
2112 \texttt{if (0) foo();}).
2114 An \textit{optimized out inline call}
2115 is a call to an inlined subprogram which either did not expand to any instructions
2116 or only parts of instructions belong to it and for debug information purposes those
2117 instructions are given a location in the caller.
2120 \DWTAGcallsite{} entries describe normal and tail calls but not tail recursion calls,
2121 while \DWTAGinlinedsubroutine{} entries describe inlined calls
2122 (see Section \refersec{chap:inlinedsubroutines}).
2123 Call site entries cannot describe tail recursion or optimized out calls.
2125 \subsection{Call Site Entries}
2126 \label{chap:callsiteentries}
2127 A call site is represented by a debugging information entry with the tag
2128 \DWTAGcallsiteTARG{}\addtoindexx{call site entry}.
2129 The entry for a call site is owned by the innermost
2130 debugging information entry representing the scope within which the
2131 call is present in the source program.
2134 \textit{A scope entry (for example, a lexical block) that would not
2135 otherwise be present in the debugging information of a subroutine
2136 need not be introduced solely to represent the immediately containing scope
2139 The call site entry may have a
2140 \DWATcallreturnpcDEFN{}\addtoindexx{call site return pc attribute}
2141 \livetargi{chap:DWATcallreturnpcofcallsite}{attribute}{call return pc attribute}
2142 which is the return address after the call.
2143 The value of this attribute corresponds to the return address
2144 computed by call frame information in the called subprogram
2145 (see Section \refersec{datarep:callframeinformation}).
2147 \textit{On many architectures the return address is the
2148 address immediately following the call instruction, but
2149 on architectures with delay slots it might
2150 be an address after the delay slot of the call.}
2152 The call site entry may have a
2153 \DWATcallpcDEFN{}\addtoindexx{call pc attribute}
2154 \livetargi{chap:DWATcallpcofcallsite}{attribute}{call pc attribute}
2155 which is the address of the
2156 call-like instruction for a normal call or the jump-like
2157 instruction for a tail call.
2159 If the call site entry corresponds to a tail call, it has the
2160 \DWATcalltailcallDEFN{}\addtoindexx{call tail call attribute}
2161 \livetargi{chap:DWATcalltailcallofcallsite}{attribute}{call tail call attribute},
2162 which is a \CLASSflag.
2164 The call site entry may have a
2165 \DWATcalloriginDEFN{}\addtoindexx{call origin attribute}
2166 \livetargi{chap:DWATcalloriginofcallsite}{attribute}{call origin attribute}
2167 which is a \CLASSreference. For direct calls or jumps where the called
2168 subprogram is known it is a reference to the called subprogram's debugging
2169 information entry. For indirect calls it may be a reference to a
2170 \DWTAGvariable{}, \DWTAGformalparameter{} or \DWTAGmember{} entry representing
2171 the subroutine pointer that is called.
2174 The call site may have a
2175 \DWATcalltargetDEFN{}\addtoindexx{call target attribute}
2176 \livetargi{chap:DWATcalltargetofcallsite}{attribute}{call target attribute} which is
2177 a DWARF expression. For indirect calls or jumps where it is unknown at
2178 compile time which subprogram will be called the expression computes the
2179 address of the subprogram that will be called.
2181 \textit{The DWARF expression should
2182 not use register or memory locations that might be clobbered by the call.}
2185 The call site entry may have a
2186 \DWATcalltargetclobberedDEFN{}\addtoindexx{call target clobbered attribute}
2187 \livetargi{chap:DWATcalltargetclobberedofcallsite}{attribute}{call target clobbered attribute}
2188 which is a DWARF expression. For indirect calls or jumps where the
2189 address is not computable without use of registers or memory locations that
2190 might be clobbered by the call the \DWATcalltargetclobberedNAME{}
2191 attribute is used instead of the \DWATcalltarget{} attribute.
2193 \textit{The expression of a call target clobbered attribute may only be
2194 valid at the time the call or call-like transfer of control is executed.}
2196 The call site entry may have a \DWATtypeDEFN{}\addtoindexx{call type attribute}
2197 \livetargi{chap:DWATtypeofcallsite}{attribute}{type attribute!of call site entry}
2198 referencing a debugging information entry for the type of the called function.
2200 \textit{When \DWATcallorigin{} is present, \DWATtypeNAME{} is usually omitted.}
2202 The call site entry may have
2203 \DWATcallfileDEFN{}\addtoindexx{call file attribute},
2204 \DWATcalllineDEFN{}\addtoindexx{call line attribute} and
2205 \DWATcallcolumnDEFN{}\addtoindexx{call column attribute}
2206 \livetargi{chap:DWATcallfileofcallsite}{attributes,}{call file attribute!of call site entry}
2207 \livetargi{chap:DWATcalllineofcallsite}{}{call line attribute!of call site entry}
2208 \livetargi{chap:DWATcallcolumnofcallsite}{}{call column attribute!of call site entry}
2209 each of whose value is an integer constant.
2210 These attributes represent the source file, source line number, and source
2211 column number, respectively, of the first character of the call statement or
2212 expression. The call file, call line, and call column attributes are
2213 interpreted in the same way as the declaration file, declaration
2214 line, and declaration column attributes, respectively
2215 (see Section \refersec{chap:declarationcoordinates}).
2217 \textit{The call file, call line and call column coordinates do
2218 not describe the coordinates of the subroutine declaration that
2219 was called, rather they describe the coordinates of the call.}
2222 \subsection{Call Site Parameters}
2223 \label{chap:callsiteparameters}
2224 The call site entry may own
2225 \DWTAGcallsiteparameterTARG{}\index{call site parameter entry}
2226 debugging information entries representing the parameters passed
2228 Call site parameter entries occur in the same order as the
2229 corresponding parameters in the source.
2230 Each such entry has a \DWATlocation{} attribute which is a location
2231 description. This location description
2232 describes where the parameter is passed
2233 (usually either some register, or a memory location expressible as
2234 the contents of the stack register plus some offset).
2237 Each \DWTAGcallsiteparameter{} entry may have a
2238 \DWATcallvalueDEFN{}\addtoindexx{call value attribute}
2239 \livetargi{chap:DWATcallvalueofcallparameter}{attribute}{call value attribute}
2240 which is a DWARF expression
2241 which when evaluated yields the value of the parameter at the time of the call.
2243 \textit{If it is not
2244 possible to avoid registers or memory locations that might be clobbered by
2245 the call in the expression, then the \DWATcallvalueNAME{} attribute should
2246 not be provided. The reason for the restriction is that the value of the parameter may be
2247 needed in the midst of the callee, where the call clobbered registers or
2248 memory might be already clobbered, and if the consumer is not assured by
2249 the producer it can safely use those values, the consumer can not safely
2250 use the values at all.}
2252 For parameters passed by reference, where the code passes a pointer to
2253 a location which contains the parameter, or for reference type parameters,
2254 the \DWTAGcallsiteparameter{} entry may also have a
2255 \DWATcalldatalocationDEFN{}\addtoindexx{call data location attribute}
2256 \livetargi{chap:DWATcalldatalocationofcallparameter}{attribute}{call data location attribute}
2257 whose value is a location description and a
2258 \DWATcalldatavalueDEFN{}\addtoindexx{call data value attribute}
2259 \livetargi{chap:DWATcalldatavalueofcallparameter}{attribute}{call data value attribute}
2260 whose value is a DWARF expression. The \DWATcalldatalocationNAME{} attribute
2261 \addtoindexx{call data location attribute}
2262 describes where the referenced value lives during the call. If it is just
2263 \DWOPpushobjectaddress{}, it may be left out. The
2264 \DWATcalldatavalueNAME{} attribute describes the value in that location.
2265 The expression should not use registers or memory
2266 locations that might be clobbered by the call, as it might be evaluated after
2267 virtually unwinding from the called function back to the caller.
2270 Each call site parameter entry may also have a
2271 \DWATcallparameterDEFN{}\addtoindexx{call parameter attribute}
2272 \livetargi{chap:DWATcallparameterofcallparameter}{attribute}{call parameter attribute}
2273 which contains a reference to a \DWTAGformalparameter{} entry,
2274 \DWATtype{} attribute referencing the type of the parameter or
2275 \DWATname{} attribute describing the parameter's name.
2277 \textit{Examples using call site entries and related attributes are
2278 found in Appendix \refersec{app:callsiteexamples}.}
2281 \section{Lexical Block Entries}
2282 \label{chap:lexicalblockentries}
2285 lexical \livetargi{chap:lexicalblock}{block}{lexical block}
2287 \addtoindexx{lexical block}
2288 a bracketed sequence of source statements
2289 that may contain any number of declarations. In some languages
2290 (including \addtoindex{C} and \addtoindex{C++}),
2291 \nolink{blocks} can be nested within other
2292 \nolink{blocks} to any depth.}
2294 % We do not need to link to the preceding paragraph.
2295 A lexical \nolink{block} is represented by a debugging information
2297 tag \DWTAGlexicalblockTARG.
2299 The lexical \livetargi{chap:lexicalblockentry}{block}{lexical block entry}
2301 either a \DWATlowpc{} and
2302 \DWAThighpc{} pair of
2304 \addtoindexx{high PC attribute}
2306 \addtoindexx{low PC attribute}
2308 \DWATranges{} attribute
2309 \addtoindexx{ranges attribute}
2310 whose values encode the contiguous or non-contiguous address
2311 ranges, respectively, of the machine instructions generated
2312 for the lexical \nolink{block}
2313 (see Section \refersec{chap:codeaddressesandranges}).
2315 A\hypertarget{chap:DWATentrypcoflexicalblock}{}
2316 lexical block entry may also have a
2317 \addtoindexx{entry PC attribute!for lexical block}
2318 \DWATentrypc{} attribute
2319 whose value is the address of the first executable instruction
2320 of the lexical block (see
2321 Section \refersec{chap:entryaddress}).
2323 If a name has been given to the lexical \nolink{block}
2324 in the source program, then the corresponding
2325 lexical \nolink{block} entry has a
2326 \DWATname{} attribute whose
2327 \addtoindexx{name attribute}
2328 value is a null-terminated string
2329 containing the name of the lexical \nolink{block}.
2331 \textit{This is not the same as a \addtoindex{C} or
2332 \addtoindex{C++} label (see Section \refersec{chap:labelentries}).}
2334 The lexical \nolink{block} entry owns debugging
2335 information entries that describe the declarations
2336 within that lexical \nolink{block}. There is
2337 one such debugging information entry for each local declaration
2338 of an identifier or inner lexical \nolink{block}.
2341 \section{Label Entries}
2342 \label{chap:labelentries}
2343 \textit{A label is a way of identifying a source location.
2344 A labeled statement is usually the target of one or more
2345 \doublequote{go to} statements.}
2348 A label is represented by a debugging information entry with
2349 \addtoindexx{label entry} the tag \DWTAGlabelTARG.
2350 The entry for a label is owned by
2351 the debugging information entry representing the scope within
2352 which the name of the label could be legally referenced within
2355 The label entry has a \DWATlowpc{} attribute whose value
2356 is the address of the first executable instruction for the
2357 location identified by the label in
2358 the source program. The label entry also has a
2359 \DWATname{} attribute
2360 \addtoindexx{name attribute}
2361 whose value is a null-terminated string containing
2362 the name of the label.
2365 \section{With Statement Entries}
2366 \label{chap:withstatemententries}
2368 \textit{Both \addtoindex{Pascal} and
2369 \addtoindexx{Modula-2}
2370 Modula-2 support the concept of a \doublequote{with}
2371 statement. The with statement specifies a sequence of
2372 executable statements within which the fields of a record
2373 variable may be referenced, unqualified by the name of the
2376 A with statement is represented by a
2377 \addtoindexi{debugging information entry}{with statement entry}
2378 with the tag \DWTAGwithstmtTARG.
2380 A with statement entry may have either a
2382 \DWAThighpc{} pair of attributes
2383 \addtoindexx{low PC attribute}
2384 \addtoindexx{high PC attribute}
2386 \DWATranges{} attribute
2387 \addtoindexx{ranges attribute}
2388 whose values encode the contiguous or non-contiguous address
2389 ranges, respectively, of the machine instructions generated
2390 for the with statement
2391 (see Section \refersec{chap:codeaddressesandranges}).
2393 A\hypertarget{chap:DWATentrypcofwithstmt}{}
2394 with statement entry may also have a
2395 \addtoindexx{entry PC attribute!for with statement}
2396 \DWATentrypc{} attribute
2397 whose value is the address of the first executable instruction
2398 of the with statement (see
2399 Section \refersec{chap:entryaddress}).
2402 The with statement entry has a
2403 \addtoindexx{type attribute}
2404 \DWATtype{} attribute, denoting
2405 the type of record whose fields may be referenced without full
2406 qualification within the body of the statement. It also has
2407 \addtoindexx{location attribute}
2408 a \DWATlocation{} attribute, describing how to find the base
2409 address of the record object referenced within the body of
2413 \section{Try and Catch Block Entries}
2414 \label{chap:tryandcatchblockentries}
2415 \livetarg{chap:tryandcatchblockentries}{}
2416 \textit{In \addtoindex{C++}, a \livelink{chap:lexicalblock}{lexical block} may be
2417 designated as a \doublequote{catch \nolink{block}.}
2418 A catch \nolink{block} is an exception handler that
2419 handles exceptions thrown by an immediately preceding
2420 \doublequote{try \nolink{block}.}
2421 A catch \nolink{block}
2422 designates the type of the exception that it can handle.}
2424 A \livetarg{chap:tryblock}{try block} is represented
2425 by a debugging information entry
2426 \addtoindexx{try block entry}
2427 with the tag \DWTAGtryblockTARG.
2428 A \livetarg{chap:catchblock}{catch block} is represented by
2429 a debugging information entry
2430 \addtoindexx{catch block entry}
2431 with the tag \DWTAGcatchblockTARG.
2433 Both try and catch \nolink{block} entries may have either a
2435 \DWAThighpc{} pair of attributes
2436 \addtoindexx{low PC attribute}
2437 \addtoindexx{high PC attribute}
2439 \DWATranges{} attribute
2440 \addtoindexx{ranges attribute}
2441 whose values encode the contiguous
2442 or non-contiguous address ranges, respectively, of the
2443 machine instructions generated for the \nolink{block}
2444 (see Section \refersec{chap:codeaddressesandranges}).
2446 A\hypertarget{chap:DWATentrypcoftryblock}{}
2447 try or catch\hypertarget{chap:DWATentrypcofcatchblock}{}
2448 block entry may also have a
2449 \addtoindexx{entry PC attribute!for try block}
2450 \addtoindexx{entry PC attribute!for catch block}
2451 \DWATentrypc{} attribute
2452 whose value is the address of the first executable instruction
2453 of the try or catch block
2454 (see Section \refersec{chap:entryaddress}).
2457 Catch \nolink{block} entries have at least one child entry,
2458 an entry representing the type of exception accepted by
2459 that catch \nolink{block}.
2460 This child entry has one of the tags
2461 \DWTAGformalparameter{}\addtoindexx{formal parameter entry!in catch block}
2463 \DWTAGunspecifiedparameters{},
2464 \addtoindexx{unspecified parameters entry!in catch block}
2465 and will have the same form as other parameter entries.
2467 The siblings immediately following a try \nolink{block}
2468 entry are its corresponding catch \nolink{block} entries.
2471 \section{Declarations with Reduced Scope}
2472 \label{declarationswithreducedscope}
2473 \hypertarget{chap:DWATstartscopeofdeclaration}{}
2474 Any debugging information entry for a declaration
2475 (including objects, subprograms, types and modules) whose scope
2476 has an address range that is a subset of the address range for
2477 the lexical scope most closely enclosing the declared entity
2479 \DWATstartscopeDEFN{}\addtoindexx{start scope attribute}
2480 attribute to specify that reduced range of addresses.
2482 There are two cases:
2483 \begin{enumerate}[1. ]
2484 \item If the address range for the scope of the entry
2485 includes all of addresses for the containing scope except
2486 for a contiguous sequence of bytes at the beginning of the
2487 address range for the containing scope, then the address is
2488 specified using a value of class \CLASSconstant.
2490 \begin{enumerate}[a) ]
2491 \item If the address
2492 range of the containing scope is contiguous, the value of
2493 this attribute is the offset in bytes of the beginning of
2494 the address range for the scope of the object from the low
2495 PC value of the debugging information entry that defines
2496 that containing scope.
2497 \item If the address range of the containing
2498 scope is non-contiguous
2499 (see \refersec{chap:noncontiguousaddressranges})
2500 the value of this attribute is the offset in bytes of the
2501 beginning of the address range for the scope of the entity
2502 from the beginning of the first \addtoindex{range list} entry
2503 for the containing scope that is not a base
2504 address entry, a default location
2505 entry or an end-of-list entry.
2509 \item Otherwise, the set of addresses for the scope of the
2510 entity is specified using a value of class \CLASSrnglistsptr{}.
2511 This value indicates the beginning of a \addtoindex{range list}
2512 (see Section \refersec{chap:noncontiguousaddressranges}).
2515 \textit{For example, the scope of a variable may begin somewhere
2516 in the midst of a lexical \livelink{chap:lexicalblock}{block} in a
2517 language that allows executable code in a
2518 \nolink{block} before a variable declaration, or where one declaration
2519 containing initialization code may change the scope of a
2520 subsequent declaration.}
2523 \textit{Consider the following example \addtoindex{C} code:}
2524 \par % Needed to end paragraph before listing so that it gets a line number
2535 \textit{\addtoindex{C} scoping rules require that the value of the
2536 variable \texttt{x} assigned to the variable \texttt{f} in the
2537 initialization sequence is the value of the global variable \texttt{x},
2538 rather than the local \texttt{x}, because the scope of the local variable
2539 \texttt{x} only starts after the full declarator for the local \texttt{x}.}
2541 \textit{Due to optimization, the scope of an object may be
2542 non-contiguous and require use of a \addtoindex{range list} even when
2543 the containing scope is contiguous. Conversely, the scope of
2544 an object may not require its own \addtoindex{range list} even when the
2545 containing scope is non-contiguous.}