1\input texinfo 2@setfilename gdbint.info 3@c $Id: gdbint.texinfo,v 1.2 1996/11/23 03:45:39 niklas Exp $ 4 5@ifinfo 6@format 7START-INFO-DIR-ENTRY 8* Gdb-Internals: (gdbint). The GNU debugger's internals. 9END-INFO-DIR-ENTRY 10@end format 11@end ifinfo 12 13@ifinfo 14This file documents the internals of the GNU debugger GDB. 15 16Copyright 1990, 1991, 1992, 1993, 1994 Free Software Foundation, Inc. 17Contributed by Cygnus Support. Written by John Gilmore. 18 19Permission is granted to make and distribute verbatim copies of 20this manual provided the copyright notice and this permission notice 21are preserved on all copies. 22 23@ignore 24Permission is granted to process this file through Tex and print the 25results, provided the printed document carries copying permission 26notice identical to this one except for the removal of this paragraph 27(this paragraph not being relevant to the printed manual). 28 29@end ignore 30Permission is granted to copy or distribute modified versions of this 31manual under the terms of the GPL (for which purpose this text may be 32regarded as a program in the language TeX). 33@end ifinfo 34 35@setchapternewpage off 36@settitle GDB Internals 37@titlepage 38@title{Working in GDB} 39@subtitle{A guide to the internals of the GNU debugger} 40@author John Gilmore 41@author Cygnus Support 42@page 43@tex 44\def\$#1${{#1}} % Kluge: collect RCS revision info without $...$ 45\xdef\manvers{\$Revision: 1.2 $} % For use in headers, footers too 46{\parskip=0pt 47\hfill Cygnus Support\par 48\hfill \manvers\par 49\hfill \TeX{}info \texinfoversion\par 50} 51@end tex 52 53@vskip 0pt plus 1filll 54Copyright @copyright{} 1990, 1991, 1992, 1993, 1994 Free Software Foundation, Inc. 55 56Permission is granted to make and distribute verbatim copies of 57this manual provided the copyright notice and this permission notice 58are preserved on all copies. 59 60@end titlepage 61 62@node Top 63@c Perhaps this should be the title of the document (but only for info, 64@c not for TeX). Existing GNU manuals seem inconsistent on this point. 65@top Scope of this Document 66 67This document documents the internals of the GNU debugger, GDB. It is 68intended to document aspects of GDB which apply across many different 69parts of GDB (for example, @pxref{Coding Style}), or which are global 70aspects of design (for example, what are the major modules and which 71files document them in detail?). Information which pertains to specific 72data structures, functions, variables, etc., should be put in comments 73in the source code, not here. It is more likely to get noticed and kept 74up to date there. Some of the information in this document should 75probably be moved into comments. 76 77@menu 78* README:: The README File 79* Getting Started:: Getting started working on GDB 80* Debugging GDB:: Debugging GDB with itself 81* New Architectures:: Defining a New Host or Target Architecture 82* Config:: Adding a New Configuration 83* Host:: Adding a New Host 84* Native:: Adding a New Native Configuration 85* Target:: Adding a New Target 86* Languages:: Defining New Source Languages 87* Releases:: Configuring GDB for Release 88* Partial Symbol Tables:: How GDB reads symbols quickly at startup 89* Types:: How GDB keeps track of types 90* BFD support for GDB:: How BFD and GDB interface 91* Symbol Reading:: Defining New Symbol Readers 92* Cleanups:: Cleanups 93* Wrapping:: Wrapping Output Lines 94* Frames:: Keeping track of function calls 95* Remote Stubs:: Code that runs in targets and talks to GDB 96* Longjmp Support:: Stepping through longjmp's in the target 97* Coding Style:: Strunk and White for GDB maintainers 98* Clean Design:: Frank Lloyd Wright for GDB maintainers 99* Submitting Patches:: How to get your changes into GDB releases 100* Host Conditionals:: What features exist in the host 101* Target Conditionals:: What features exist in the target 102* Native Conditionals:: Conditionals for when host and target are same 103* Obsolete Conditionals:: Conditionals that don't exist any more 104* XCOFF:: The Object file format used on IBM's RS/6000 105@end menu 106 107@node README 108@chapter The @file{README} File 109 110Check the @file{README} file, it often has useful information that does not 111appear anywhere else in the directory. 112 113@node Getting Started 114@chapter Getting Started Working on GDB 115 116GDB is a large and complicated program, and if you first starting to 117work on it, it can be hard to know where to start. Fortunately, if you 118know how to go about it, there are ways to figure out what is going on: 119 120@itemize @bullet 121@item 122This manual, the GDB Internals manual, has information which applies 123generally to many parts of GDB. 124 125@item 126Information about particular functions or data structures are located in 127comments with those functions or data structures. If you run across a 128function or a global variable which does not have a comment correctly 129explaining what is does, this can be thought of as a bug in GDB; feel 130free to submit a bug report, with a suggested comment if you can figure 131out what the comment should say (@pxref{Submitting Patches}). If you 132find a comment which is actually wrong, be especially sure to report that. 133 134Comments explaining the function of macros defined in host, target, or 135native dependent files can be in several places. Sometimes they are 136repeated every place the macro is defined. Sometimes they are where the 137macro is used. Sometimes there is a header file which supplies a 138default definition of the macro, and the comment is there. This manual 139also has a list of macros (@pxref{Host Conditionals}, @pxref{Target 140Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete 141Conditionals}) with some documentation. 142 143@item 144Start with the header files. Once you some idea of how GDB's internal 145symbol tables are stored (see @file{symtab.h}, @file{gdbtypes.h}), you 146will find it much easier to understand the code which uses and creates 147those symbol tables. 148 149@item 150You may wish to process the information you are getting somehow, to 151enhance your understanding of it. Summarize it, translate it to another 152language, add some (perhaps trivial or non-useful) feature to GDB, use 153the code to predict what a test case would do and write the test case 154and verify your prediction, etc. If you are reading code and your eyes 155are starting to glaze over, this is a sign you need to use a more active 156approach. 157 158@item 159Once you have a part of GDB to start with, you can find more 160specifically the part you are looking for by stepping through each 161function with the @code{next} command. Do not use @code{step} or you 162will quickly get distracted; when the function you are stepping through 163calls another function try only to get a big-picture understanding 164(perhaps using the comment at the beginning of the function being 165called) of what it does. This way you can identify which of the 166functions being called by the function you are stepping through is the 167one which you are interested in. You may need to examine the data 168structures generated at each stage, with reference to the comments in 169the header files explaining what the data structures are supposed to 170look like. 171 172Of course, this same technique can be used if you are just reading the 173code, rather than actually stepping through it. The same general 174principle applies---when the code you are looking at calls something 175else, just try to understand generally what the code being called does, 176rather than worrying about all its details. 177 178@item 179A good place to start when tracking down some particular area is with a 180command which invokes that feature. Suppose you want to know how 181single-stepping works. As a GDB user, you know that the @code{step} 182command invokes single-stepping. The command is invoked via command 183tables (see @file{command.h}); by convention the function which actually 184performs the command is formed by taking the name of the command and 185adding @samp{_command}, or in the case of an @code{info} subcommand, 186@samp{_info}. For example, the @code{step} command invokes the 187@code{step_command} function and the @code{info display} command invokes 188@code{display_info}. When this convention is not followed, you might 189have to use @code{grep} or @kbd{M-x tags-search} in emacs, or run GDB on 190itself and set a breakpoint in @code{execute_command}. 191 192@item 193If all of the above fail, it may be appropriate to ask for information 194on @code{bug-gdb}. But @emph{never} post a generic question like ``I was 195wondering if anyone could give me some tips about understanding 196GDB''---if we had some magic secret we would put it in this manual. 197Suggestions for improving the manual are always welcome, of course. 198@end itemize 199 200Good luck! 201 202@node Debugging GDB 203@chapter Debugging GDB with itself 204If GDB is limping on your machine, this is the preferred way to get it 205fully functional. Be warned that in some ancient Unix systems, like 206Ultrix 4.2, a program can't be running in one process while it is being 207debugged in another. Rather than typing the command @code{@w{./gdb 208./gdb}}, which works on Suns and such, you can copy @file{gdb} to 209@file{gdb2} and then type @code{@w{./gdb ./gdb2}}. 210 211When you run GDB in the GDB source directory, it will read a 212@file{.gdbinit} file that sets up some simple things to make debugging 213gdb easier. The @code{info} command, when executed without a subcommand 214in a GDB being debugged by gdb, will pop you back up to the top level 215gdb. See @file{.gdbinit} for details. 216 217If you use emacs, you will probably want to do a @code{make TAGS} after 218you configure your distribution; this will put the machine dependent 219routines for your local machine where they will be accessed first by 220@kbd{M-.} 221 222Also, make sure that you've either compiled GDB with your local cc, or 223have run @code{fixincludes} if you are compiling with gcc. 224 225@node New Architectures 226@chapter Defining a New Host or Target Architecture 227 228When building support for a new host and/or target, much of the work you 229need to do is handled by specifying configuration files; 230@pxref{Config,,Adding a New Configuration}. Further work can be 231divided into ``host-dependent'' (@pxref{Host,,Adding a New Host}) and 232``target-dependent'' (@pxref{Target,,Adding a New Target}). The 233following discussion is meant to explain the difference between hosts 234and targets. 235 236@heading What is considered ``host-dependent'' versus ``target-dependent''? 237 238@dfn{Host} refers to attributes of the system where GDB runs. 239@dfn{Target} refers to the system where the program being debugged 240executes. In most cases they are the same machine, in which case 241a third type of @dfn{Native} attributes come into play. 242 243Defines and include files needed to build on the host are host support. 244Examples are tty support, system defined types, host byte order, host 245float format. 246 247Defines and information needed to handle the target format are target 248dependent. Examples are the stack frame format, instruction set, 249breakpoint instruction, registers, and how to set up and tear down the stack 250to call a function. 251 252Information that is only needed when the host and target are the same, 253is native dependent. One example is Unix child process support; if the 254host and target are not the same, doing a fork to start the target 255process is a bad idea. The various macros needed for finding the 256registers in the @code{upage}, running @code{ptrace}, and such are all in the 257native-dependent files. 258 259Another example of native-dependent code is support for features 260that are really part of the target environment, but which require 261@code{#include} files that are only available on the host system. 262Core file handling and @code{setjmp} handling are two common cases. 263 264When you want to make GDB work ``native'' on a particular 265machine, you have to include all three kinds of information. 266 267The dependent information in GDB is organized into files by naming 268conventions. 269 270Host-Dependent Files 271@table @file 272@item config/*/*.mh 273Sets Makefile parameters 274@item config/*/xm-*.h 275Global #include's and #define's and definitions 276@item *-xdep.c 277Global variables and functions 278@end table 279 280Native-Dependent Files 281@table @file 282@item config/*/*.mh 283Sets Makefile parameters (for @emph{both} host and native) 284@item config/*/nm-*.h 285#include's and #define's and definitions. This file 286is only included by the small number of modules that need it, 287so beware of doing feature-test #define's from its macros. 288@item *-nat.c 289global variables and functions 290@end table 291 292Target-Dependent Files 293@table @file 294@item config/*/*.mt 295Sets Makefile parameters 296@item config/*/tm-*.h 297Global #include's and #define's and definitions 298@item *-tdep.c 299Global variables and functions 300@end table 301 302At this writing, most supported hosts have had their host and native 303dependencies sorted out properly. There are a few stragglers, which 304can be recognized by the absence of NATDEPFILES lines in their 305@file{config/*/*.mh}. 306 307@node Config 308@chapter Adding a New Configuration 309 310Most of the work in making GDB compile on a new machine is in specifying 311the configuration of the machine. This is done in a dizzying variety of 312header files and configuration scripts, which we hope to make more 313sensible soon. Let's say your new host is called an @var{xxx} (e.g. 314@samp{sun4}), and its full three-part configuration name is 315@code{@var{xarch}-@var{xvend}-@var{xos}} (e.g. @samp{sparc-sun-sunos4}). In 316particular: 317 318In the top level directory, edit @file{config.sub} and add @var{xarch}, 319@var{xvend}, and @var{xos} to the lists of supported architectures, 320vendors, and operating systems near the bottom of the file. Also, add 321@var{xxx} as an alias that maps to 322@code{@var{xarch}-@var{xvend}-@var{xos}}. You can test your changes by 323running 324 325@example 326./config.sub @var{xxx} 327@end example 328@noindent 329and 330@example 331./config.sub @code{@var{xarch}-@var{xvend}-@var{xos}} 332@end example 333@noindent 334which should both respond with @code{@var{xarch}-@var{xvend}-@var{xos}} 335and no error messages. 336 337Now, go to the @file{bfd} directory and 338create a new file @file{bfd/hosts/h-@var{xxx}.h}. Examine the 339other @file{h-*.h} files as templates, and create one that brings in the 340right include files for your system, and defines any host-specific 341macros needed by BFD, the Binutils, GNU LD, or the Opcodes directories. 342(They all share the bfd @file{hosts} directory and the @file{configure.host} 343file.) 344 345Then edit @file{bfd/configure.host}. Add a line to recognize your 346@code{@var{xarch}-@var{xvend}-@var{xos}} configuration, and set 347@code{my_host} to @var{xxx} when you recognize it. This will cause your 348file @file{h-@var{xxx}.h} to be linked to @file{sysdep.h} at configuration 349time. When creating the line that recognizes your configuration, 350only match the fields that you really need to match; e.g. don't 351match the architecture or manufacturer if the OS is sufficient 352to distinguish the configuration that your @file{h-@var{xxx}.h} file supports. 353Don't match the manufacturer name unless you really need to. 354This should make future ports easier. 355 356Also, if this host requires any changes to the Makefile, create a file 357@file{bfd/config/@var{xxx}.mh}, which includes the required lines. 358 359It's possible that the @file{libiberty} and @file{readline} directories 360won't need any changes for your configuration, but if they do, you can 361change the @file{configure.in} file there to recognize your system and 362map to an @file{mh-@var{xxx}} file. Then add @file{mh-@var{xxx}} 363to the @file{config/} subdirectory, to set any makefile variables you 364need. The only current options in there are things like @samp{-DSYSV}. 365(This @file{mh-@var{xxx}} naming convention differs from elsewhere 366in GDB, by historical accident. It should be cleaned up so that all 367such files are called @file{@var{xxx}.mh}.) 368 369Aha! Now to configure GDB itself! Edit 370@file{gdb/configure.in} to recognize your system and set @code{gdb_host} 371to @var{xxx}, and (unless your desired target is already available) also 372set @code{gdb_target} to something appropriate (for instance, 373@var{xxx}). To handle new hosts, modify the segment after the comment 374@samp{# per-host}; to handle new targets, modify after @samp{# 375per-target}. 376@c Would it be simpler to just use different per-host and per-target 377@c *scripts*, and call them from {configure} ? 378 379Finally, you'll need to specify and define GDB's host-, native-, and 380target-dependent @file{.h} and @file{.c} files used for your 381configuration; the next two chapters discuss those. 382 383 384@node Host 385@chapter Adding a New Host 386 387Once you have specified a new configuration for your host 388(@pxref{Config,,Adding a New Configuration}), there are three remaining 389pieces to making GDB work on a new machine. First, you have to make it 390host on the new machine (compile there, handle that machine's terminals 391properly, etc). If you will be cross-debugging to some other kind of 392system that's already supported, you are done. 393 394If you want to use GDB to debug programs that run on the new machine, 395you have to get it to understand the machine's object files, symbol 396files, and interfaces to processes; @pxref{Target,,Adding a New Target} 397and @pxref{Native,,Adding a New Native Configuration} 398 399Several files control GDB's configuration for host systems: 400 401@table @file 402@item gdb/config/@var{arch}/@var{xxx}.mh 403Specifies Makefile fragments needed when hosting on machine @var{xxx}. 404In particular, this lists the required machine-dependent object files, 405by defining @samp{XDEPFILES=@dots{}}. Also 406specifies the header file which describes host @var{xxx}, by defining 407@code{XM_FILE= xm-@var{xxx}.h}. You can also define @code{CC}, 408@code{REGEX} and @code{REGEX1}, @code{SYSV_DEFINE}, @code{XM_CFLAGS}, 409@code{XM_ADD_FILES}, @code{XM_CLIBS}, @code{XM_CDEPS}, 410etc.; see @file{Makefile.in}. 411 412@item gdb/config/@var{arch}/xm-@var{xxx}.h 413(@file{xm.h} is a link to this file, created by configure). 414Contains C macro definitions describing the host system environment, 415such as byte order, host C compiler and library, ptrace support, 416and core file structure. Crib from existing @file{xm-*.h} files 417to create a new one. 418 419@item gdb/@var{xxx}-xdep.c 420Contains any miscellaneous C code required for this machine 421as a host. On many machines it doesn't exist at all. If it does 422exist, put @file{@var{xxx}-xdep.o} into the @code{XDEPFILES} line 423in @file{gdb/config/mh-@var{xxx}}. 424@end table 425 426@subheading Generic Host Support Files 427 428There are some ``generic'' versions of routines that can be used by 429various systems. These can be customized in various ways by macros 430defined in your @file{xm-@var{xxx}.h} file. If these routines work for 431the @var{xxx} host, you can just include the generic file's name (with 432@samp{.o}, not @samp{.c}) in @code{XDEPFILES}. 433 434Otherwise, if your machine needs custom support routines, you will need 435to write routines that perform the same functions as the generic file. 436Put them into @code{@var{xxx}-xdep.c}, and put @code{@var{xxx}-xdep.o} 437into @code{XDEPFILES}. 438 439@table @file 440@item ser-bsd.c 441This contains serial line support for Berkeley-derived Unix systems. 442 443@item ser-go32.c 444This contains serial line support for 32-bit programs running under DOS 445using the GO32 execution environment. 446 447@item ser-termios.c 448This contains serial line support for System V-derived Unix systems. 449@end table 450 451Now, you are now ready to try configuring GDB to compile using your system 452as its host. From the top level (above @file{bfd}, @file{gdb}, etc), do: 453 454@example 455./configure @var{xxx} --target=vxworks960 456@end example 457 458This will configure your system to cross-compile for VxWorks on 459the Intel 960, which is probably not what you really want, but it's 460a test case that works at this stage. (You haven't set up to be 461able to debug programs that run @emph{on} @var{xxx} yet.) 462 463If this succeeds, you can try building it all with: 464 465@example 466make 467@end example 468 469Repeat until the program configures, compiles, links, and runs. 470When run, it won't be able to do much (unless you have a VxWorks/960 471board on your network) but you will know that the host support is 472pretty well done. 473 474Good luck! Comments and suggestions about this section are particularly 475welcome; send them to @samp{bug-gdb@@prep.ai.mit.edu}. 476 477@node Native 478@chapter Adding a New Native Configuration 479 480If you are making GDB run native on the @var{xxx} machine, you have 481plenty more work to do. Several files control GDB's configuration for 482native support: 483 484@table @file 485@item gdb/config/@var{xarch}/@var{xxx}.mh 486Specifies Makefile fragments needed when hosting @emph{or native} 487on machine @var{xxx}. 488In particular, this lists the required native-dependent object files, 489by defining @samp{NATDEPFILES=@dots{}}. Also 490specifies the header file which describes native support on @var{xxx}, 491by defining @samp{NAT_FILE= nm-@var{xxx}.h}. 492You can also define @samp{NAT_CFLAGS}, 493@samp{NAT_ADD_FILES}, @samp{NAT_CLIBS}, @samp{NAT_CDEPS}, 494etc.; see @file{Makefile.in}. 495 496@item gdb/config/@var{arch}/nm-@var{xxx}.h 497(@file{nm.h} is a link to this file, created by configure). 498Contains C macro definitions describing the native system environment, 499such as child process control and core file support. 500Crib from existing @file{nm-*.h} files to create a new one. 501 502@item gdb/@var{xxx}-nat.c 503Contains any miscellaneous C code required for this native support 504of this machine. On some machines it doesn't exist at all. 505@end table 506 507@subheading Generic Native Support Files 508 509There are some ``generic'' versions of routines that can be used by 510various systems. These can be customized in various ways by macros 511defined in your @file{nm-@var{xxx}.h} file. If these routines work for 512the @var{xxx} host, you can just include the generic file's name (with 513@samp{.o}, not @samp{.c}) in @code{NATDEPFILES}. 514 515Otherwise, if your machine needs custom support routines, you will need 516to write routines that perform the same functions as the generic file. 517Put them into @code{@var{xxx}-nat.c}, and put @code{@var{xxx}-nat.o} 518into @code{NATDEPFILES}. 519 520@table @file 521 522@item inftarg.c 523This contains the @emph{target_ops vector} that supports Unix child 524processes on systems which use ptrace and wait to control the child. 525 526@item procfs.c 527This contains the @emph{target_ops vector} that supports Unix child 528processes on systems which use /proc to control the child. 529 530@item fork-child.c 531This does the low-level grunge that uses Unix system calls 532to do a "fork and exec" to start up a child process. 533 534@item infptrace.c 535This is the low level interface to inferior processes for systems 536using the Unix @code{ptrace} call in a vanilla way. 537 538@item core-aout.c::fetch_core_registers() 539Support for reading registers out of a core file. This routine calls 540@code{register_addr()}, see below. 541Now that BFD is used to read core files, virtually all machines should 542use @code{core-aout.c}, and should just provide @code{fetch_core_registers} in 543@code{@var{xxx}-nat.c} (or @code{REGISTER_U_ADDR} in @code{nm-@var{xxx}.h}). 544 545@item core-aout.c::register_addr() 546If your @code{nm-@var{xxx}.h} file defines the macro 547@code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to 548set @code{addr} to the offset within the @samp{user} 549struct of GDB register number @code{regno}. @code{blockend} is the 550offset within the ``upage'' of @code{u.u_ar0}. 551If @code{REGISTER_U_ADDR} is defined, 552@file{core-aout.c} will define the @code{register_addr()} function and use 553the macro in it. If you do not define @code{REGISTER_U_ADDR}, but you 554are using the standard @code{fetch_core_registers()}, you will need to 555define your own version of @code{register_addr()}, put it into your 556@code{@var{xxx}-nat.c} file, and be sure @code{@var{xxx}-nat.o} is in 557the @code{NATDEPFILES} list. If you have your own 558@code{fetch_core_registers()}, you may not need a separate 559@code{register_addr()}. Many custom @code{fetch_core_registers()} 560implementations simply locate the registers themselves.@refill 561@end table 562 563When making GDB run native on a new operating system, 564to make it possible to debug 565core files, you will need to either write specific code for parsing your 566OS's core files, or customize @file{bfd/trad-core.c}. First, use 567whatever @code{#include} files your machine uses to define the struct of 568registers that is accessible (possibly in the u-area) in a core file 569(rather than @file{machine/reg.h}), and an include file that defines whatever 570header exists on a core file (e.g. the u-area or a @samp{struct core}). Then 571modify @code{trad_unix_core_file_p()} to use these values to set up the 572section information for the data segment, stack segment, any other 573segments in the core file (perhaps shared library contents or control 574information), ``registers'' segment, and if there are two discontiguous 575sets of registers (e.g. integer and float), the ``reg2'' segment. This 576section information basically delimits areas in the core file in a 577standard way, which the section-reading routines in BFD know how to seek 578around in. 579 580Then back in GDB, you need a matching routine called 581@code{fetch_core_registers()}. If you can use the generic one, it's in 582@file{core-aout.c}; if not, it's in your @file{@var{xxx}-nat.c} file. 583It will be passed a char pointer to the entire ``registers'' segment, 584its length, and a zero; or a char pointer to the entire ``regs2'' 585segment, its length, and a 2. The routine should suck out the supplied 586register values and install them into GDB's ``registers'' array. 587(@xref{New Architectures,,Defining a New Host or Target Architecture}, 588for more info about this.) 589 590If your system uses @file{/proc} to control processes, and uses ELF 591format core files, then you may be able to use the same routines 592for reading the registers out of processes and out of core files. 593 594@node Target 595@chapter Adding a New Target 596 597For a new target called @var{ttt}, first specify the configuration as 598described in @ref{Config,,Adding a New Configuration}. If your new 599target is the same as your new host, you've probably already done that. 600 601A variety of files specify attributes of the GDB target environment: 602 603@table @file 604@item gdb/config/@var{arch}/@var{ttt}.mt 605Contains a Makefile fragment specific to this target. 606Specifies what object files are needed for target @var{ttt}, by 607defining @samp{TDEPFILES=@dots{}}. 608Also specifies the header file which describes @var{ttt}, by defining 609@samp{TM_FILE= tm-@var{ttt}.h}. You can also define @samp{TM_CFLAGS}, 610@samp{TM_CLIBS}, @samp{TM_CDEPS}, 611and other Makefile variables here; see @file{Makefile.in}. 612 613@item gdb/config/@var{arch}/tm-@var{ttt}.h 614(@file{tm.h} is a link to this file, created by configure). 615Contains macro definitions about the target machine's 616registers, stack frame format and instructions. 617Crib from existing @file{tm-*.h} files when building a new one. 618 619@item gdb/@var{ttt}-tdep.c 620Contains any miscellaneous code required for this target machine. 621On some machines it doesn't exist at all. Sometimes the macros 622in @file{tm-@var{ttt}.h} become very complicated, so they are 623implemented as functions here instead, and the macro is simply 624defined to call the function. 625 626@item gdb/exec.c 627Defines functions for accessing files that are 628executable on the target system. These functions open and examine an 629exec file, extract data from one, write data to one, print information 630about one, etc. Now that executable files are handled with BFD, every 631target should be able to use the generic exec.c rather than its 632own custom code. 633 634@item gdb/@var{arch}-pinsn.c 635Prints (disassembles) the target machine's instructions. 636This file is usually shared with other target machines which use the 637same processor, which is why it is @file{@var{arch}-pinsn.c} rather 638than @file{@var{ttt}-pinsn.c}. 639 640@item gdb/@var{arch}-opcode.h 641Contains some large initialized 642data structures describing the target machine's instructions. 643This is a bit strange for a @file{.h} file, but it's OK since 644it is only included in one place. @file{@var{arch}-opcode.h} is shared 645between the debugger and the assembler, if the GNU assembler has been 646ported to the target machine. 647 648@item gdb/config/@var{arch}/tm-@var{arch}.h 649This often exists to describe the basic layout of the target machine's 650processor chip (registers, stack, etc). 651If used, it is included by @file{tm-@var{xxx}.h}. It can 652be shared among many targets that use the same processor. 653 654@item gdb/@var{arch}-tdep.c 655Similarly, there are often common subroutines that are shared by all 656target machines that use this particular architecture. 657@end table 658 659When adding support for a new target machine, there are various areas 660of support that might need change, or might be OK. 661 662If you are using an existing object file format (a.out or COFF), 663there is probably little to be done. See @file{bfd/doc/bfd.texinfo} 664for more information on writing new a.out or COFF versions. 665 666If you need to add a new object file format, you must first add it to 667BFD. This is beyond the scope of this document right now. Basically 668you must build a transfer vector (of type @code{bfd_target}), which will 669mean writing all the required routines, and add it to the list in 670@file{bfd/targets.c}. 671 672You must then arrange for the BFD code to provide access to the 673debugging symbols. Generally GDB will have to call swapping routines 674from BFD and a few other BFD internal routines to locate the debugging 675information. As much as possible, GDB should not depend on the BFD 676internal data structures. 677 678For some targets (e.g., COFF), there is a special transfer vector used 679to call swapping routines, since the external data structures on various 680platforms have different sizes and layouts. Specialized routines that 681will only ever be implemented by one object file format may be called 682directly. This interface should be described in a file 683@file{bfd/libxxx.h}, which is included by GDB. 684 685If you are adding a new operating system for an existing CPU chip, add a 686@file{tm-@var{xos}.h} file that describes the operating system 687facilities that are unusual (extra symbol table info; the breakpoint 688instruction needed; etc). Then write a 689@file{tm-@var{xarch}-@var{xos}.h} that just @code{#include}s 690@file{tm-@var{xarch}.h} and @file{tm-@var{xos}.h}. (Now that we have 691three-part configuration names, this will probably get revised to 692separate the @var{xos} configuration from the @var{xarch} 693configuration.) 694 695 696@node Languages 697@chapter Adding a Source Language to GDB 698 699To add other languages to GDB's expression parser, follow the following steps: 700 701@table @emph 702@item Create the expression parser. 703 704This should reside in a file @file{@var{lang}-exp.y}. Routines for building 705parsed expressions into a @samp{union exp_element} list are in @file{parse.c}. 706 707Since we can't depend upon everyone having Bison, and YACC produces 708parsers that define a bunch of global names, the following lines 709@emph{must} be included at the top of the YACC parser, to prevent 710the various parsers from defining the same global names: 711 712@example 713#define yyparse @var{lang}_parse 714#define yylex @var{lang}_lex 715#define yyerror @var{lang}_error 716#define yylval @var{lang}_lval 717#define yychar @var{lang}_char 718#define yydebug @var{lang}_debug 719#define yypact @var{lang}_pact 720#define yyr1 @var{lang}_r1 721#define yyr2 @var{lang}_r2 722#define yydef @var{lang}_def 723#define yychk @var{lang}_chk 724#define yypgo @var{lang}_pgo 725#define yyact @var{lang}_act 726#define yyexca @var{lang}_exca 727#define yyerrflag @var{lang}_errflag 728#define yynerrs @var{lang}_nerrs 729@end example 730 731At the bottom of your parser, define a @code{struct language_defn} and 732initialize it with the right values for your language. Define an 733@code{initialize_@var{lang}} routine and have it call 734@samp{add_language(@var{lang}_language_defn)} to tell the rest of GDB 735that your language exists. You'll need some other supporting variables 736and functions, which will be used via pointers from your 737@code{@var{lang}_language_defn}. See the declaration of @code{struct 738language_defn} in @file{language.h}, and the other @file{*-exp.y} files, 739for more information. 740 741@item Add any evaluation routines, if necessary 742 743If you need new opcodes (that represent the operations of the language), 744add them to the enumerated type in @file{expression.h}. Add support 745code for these operations in @code{eval.c:evaluate_subexp()}. Add cases 746for new opcodes in two functions from @file{parse.c}: 747@code{prefixify_subexp()} and @code{length_of_subexp()}. These compute 748the number of @code{exp_element}s that a given operation takes up. 749 750@item Update some existing code 751 752Add an enumerated identifier for your language to the enumerated type 753@code{enum language} in @file{defs.h}. 754 755Update the routines in @file{language.c} so your language is included. These 756routines include type predicates and such, which (in some cases) are 757language dependent. If your language does not appear in the switch 758statement, an error is reported. 759 760Also included in @file{language.c} is the code that updates the variable 761@code{current_language}, and the routines that translate the 762@code{language_@var{lang}} enumerated identifier into a printable 763string. 764 765Update the function @code{_initialize_language} to include your language. This 766function picks the default language upon startup, so is dependent upon 767which languages that GDB is built for. 768 769Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading 770code so that the language of each symtab (source file) is set properly. 771This is used to determine the language to use at each stack frame level. 772Currently, the language is set based upon the extension of the source 773file. If the language can be better inferred from the symbol 774information, please set the language of the symtab in the symbol-reading 775code. 776 777Add helper code to @code{expprint.c:print_subexp()} to handle any new 778expression opcodes you have added to @file{expression.h}. Also, add the 779printed representations of your operators to @code{op_print_tab}. 780 781@item Add a place of call 782 783Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in 784@code{parse.c:parse_exp_1()}. 785 786@item Use macros to trim code 787 788The user has the option of building GDB for some or all of the 789languages. If the user decides to build GDB for the language 790@var{lang}, then every file dependent on @file{language.h} will have the 791macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to 792leave out large routines that the user won't need if he or she is not 793using your language. 794 795Note that you do not need to do this in your YACC parser, since if GDB 796is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the 797compiled form of your parser) is not linked into GDB at all. 798 799See the file @file{configure.in} for how GDB is configured for different 800languages. 801 802@item Edit @file{Makefile.in} 803 804Add dependencies in @file{Makefile.in}. Make sure you update the macro 805variables such as @code{HFILES} and @code{OBJS}, otherwise your code may 806not get linked in, or, worse yet, it may not get @code{tar}red into the 807distribution! 808@end table 809 810 811@node Releases 812@chapter Configuring GDB for Release 813 814From the top level directory (containing @file{gdb}, @file{bfd}, 815@file{libiberty}, and so on): 816@example 817make -f Makefile.in gdb.tar.gz 818@end example 819 820This will properly configure, clean, rebuild any files that are 821distributed pre-built (e.g. @file{c-exp.tab.c} or @file{refcard.ps}), 822and will then make a tarfile. (If the top level directory has already 823been configured, you can just do @code{make gdb.tar.gz} instead.) 824 825This procedure requires: 826@itemize @bullet 827@item symbolic links 828@item @code{makeinfo} (texinfo2 level) 829@item @TeX{} 830@item @code{dvips} 831@item @code{yacc} or @code{bison} 832@end itemize 833@noindent 834@dots{} and the usual slew of utilities (@code{sed}, @code{tar}, etc.). 835 836@subheading TEMPORARY RELEASE PROCEDURE FOR DOCUMENTATION 837 838@file{gdb.texinfo} is currently marked up using the texinfo-2 macros, 839which are not yet a default for anything (but we have to start using 840them sometime). 841 842For making paper, the only thing this implies is the right generation of 843@file{texinfo.tex} needs to be included in the distribution. 844 845For making info files, however, rather than duplicating the texinfo2 846distribution, generate @file{gdb-all.texinfo} locally, and include the files 847@file{gdb.info*} in the distribution. Note the plural; @code{makeinfo} will 848split the document into one overall file and five or so included files. 849 850 851@node Partial Symbol Tables 852@chapter Partial Symbol Tables 853 854GDB has three types of symbol tables. 855 856@itemize @bullet 857@item full symbol tables (symtabs). These contain the main 858information about symbols and addresses. 859@item partial symbol tables (psymtabs). These contain enough 860information to know when to read the corresponding 861part of the full symbol table. 862@item minimal symbol tables (msymtabs). These contain information 863gleaned from non-debugging symbols. 864@end itemize 865 866This section describes partial symbol tables. 867 868A psymtab is constructed by doing a very quick pass over an executable 869file's debugging information. Small amounts of information are 870extracted -- enough to identify which parts of the symbol table will 871need to be re-read and fully digested later, when the user needs the 872information. The speed of this pass causes GDB to start up very 873quickly. Later, as the detailed rereading occurs, it occurs in small 874pieces, at various times, and the delay therefrom is mostly invisible to 875the user. (@xref{Symbol Reading}.) 876 877The symbols that show up in a file's psymtab should be, roughly, those 878visible to the debugger's user when the program is not running code from 879that file. These include external symbols and types, static 880symbols and types, and enum values declared at file scope. 881 882The psymtab also contains the range of instruction addresses that the 883full symbol table would represent. 884 885The idea is that there are only two ways for the user (or much of 886the code in the debugger) to reference a symbol: 887 888@itemize @bullet 889 890@item by its address 891(e.g. execution stops at some address which is inside a function 892in this file). The address will be noticed to be in the 893range of this psymtab, and the full symtab will be read in. 894@code{find_pc_function}, @code{find_pc_line}, and other @code{find_pc_@dots{}} 895functions handle this. 896 897@item by its name 898(e.g. the user asks to print a variable, or set a breakpoint on a 899function). Global names and file-scope names will be found in the 900psymtab, which will cause the symtab to be pulled in. Local names will 901have to be qualified by a global name, or a file-scope name, in which 902case we will have already read in the symtab as we evaluated the 903qualifier. Or, a local symbol can be referenced when 904we are "in" a local scope, in which case the first case applies. 905@code{lookup_symbol} does most of the work here. 906 907@end itemize 908 909The only reason that psymtabs exist is to cause a symtab to be read in 910at the right moment. Any symbol that can be elided from a psymtab, 911while still causing that to happen, should not appear in it. Since 912psymtabs don't have the idea of scope, you can't put local symbols in 913them anyway. Psymtabs don't have the idea of the type of a symbol, 914either, so types need not appear, unless they will be referenced by 915name. 916 917It is a bug for GDB to behave one way when only a psymtab has been read, 918and another way if the corresponding symtab has been read in. Such 919bugs are typically caused by a psymtab that does not contain all the 920visible symbols, or which has the wrong instruction address ranges. 921 922The psymtab for a particular section of a symbol-file (objfile) 923could be thrown away after the symtab has been read in. The symtab 924should always be searched before the psymtab, so the psymtab will 925never be used (in a bug-free environment). Currently, 926psymtabs are allocated on an obstack, and all the psymbols themselves 927are allocated in a pair of large arrays on an obstack, so there is 928little to be gained by trying to free them unless you want to do a lot 929more work. 930 931@node Types 932@chapter Types 933 934Fundamental Types (e.g., FT_VOID, FT_BOOLEAN). 935 936These are the fundamental types that GDB uses internally. Fundamental 937types from the various debugging formats (stabs, ELF, etc) are mapped into 938one of these. They are basically a union of all fundamental types that 939gdb knows about for all the languages that GDB knows about. 940 941Type Codes (e.g., TYPE_CODE_PTR, TYPE_CODE_ARRAY). 942 943Each time GDB builds an internal type, it marks it with one of these 944types. The type may be a fundamental type, such as TYPE_CODE_INT, or 945a derived type, such as TYPE_CODE_PTR which is a pointer to another 946type. Typically, several FT_* types map to one TYPE_CODE_* type, and 947are distinguished by other members of the type struct, such as whether 948the type is signed or unsigned, and how many bits it uses. 949 950Builtin Types (e.g., builtin_type_void, builtin_type_char). 951 952These are instances of type structs that roughly correspond to fundamental 953types and are created as global types for GDB to use for various ugly 954historical reasons. We eventually want to eliminate these. Note for 955example that builtin_type_int initialized in gdbtypes.c is basically the 956same as a TYPE_CODE_INT type that is initialized in c-lang.c for an 957FT_INTEGER fundamental type. The difference is that the builtin_type is 958not associated with any particular objfile, and only one instance exists, 959while c-lang.c builds as many TYPE_CODE_INT types as needed, with each 960one associated with some particular objfile. 961 962@node BFD support for GDB 963@chapter Binary File Descriptor Library Support for GDB 964 965BFD provides support for GDB in several ways: 966 967@table @emph 968@item identifying executable and core files 969BFD will identify a variety of file types, including a.out, coff, and 970several variants thereof, as well as several kinds of core files. 971 972@item access to sections of files 973BFD parses the file headers to determine the names, virtual addresses, 974sizes, and file locations of all the various named sections in files 975(such as the text section or the data section). GDB simply calls 976BFD to read or write section X at byte offset Y for length Z. 977 978@item specialized core file support 979BFD provides routines to determine the failing command name stored 980in a core file, the signal with which the program failed, and whether 981a core file matches (i.e. could be a core dump of) a particular executable 982file. 983 984@item locating the symbol information 985GDB uses an internal interface of BFD to determine where to find the 986symbol information in an executable file or symbol-file. GDB itself 987handles the reading of symbols, since BFD does not ``understand'' debug 988symbols, but GDB uses BFD's cached information to find the symbols, 989string table, etc. 990@end table 991 992@c The interface for symbol reading is described in @ref{Symbol 993@c Reading,,Symbol Reading}. 994 995 996@node Symbol Reading 997@chapter Symbol Reading 998 999GDB reads symbols from "symbol files". The usual symbol file is the 1000file containing the program which GDB is debugging. GDB can be directed 1001to use a different file for symbols (with the ``symbol-file'' 1002command), and it can also read more symbols via the ``add-file'' and ``load'' 1003commands, or while reading symbols from shared libraries. 1004 1005Symbol files are initially opened by @file{symfile.c} using the BFD 1006library. BFD identifies the type of the file by examining its header. 1007@code{symfile_init} then uses this identification to locate a 1008set of symbol-reading functions. 1009 1010Symbol reading modules identify themselves to GDB by calling 1011@code{add_symtab_fns} during their module initialization. The argument 1012to @code{add_symtab_fns} is a @code{struct sym_fns} which contains 1013the name (or name prefix) of the symbol format, the length of the prefix, 1014and pointers to four functions. These functions are called at various 1015times to process symbol-files whose identification matches the specified 1016prefix. 1017 1018The functions supplied by each module are: 1019 1020@table @code 1021@item @var{xxx}_symfile_init(struct sym_fns *sf) 1022 1023Called from @code{symbol_file_add} when we are about to read a new 1024symbol file. This function should clean up any internal state 1025(possibly resulting from half-read previous files, for example) 1026and prepare to read a new symbol file. Note that the symbol file 1027which we are reading might be a new "main" symbol file, or might 1028be a secondary symbol file whose symbols are being added to the 1029existing symbol table. 1030 1031The argument to @code{@var{xxx}_symfile_init} is a newly allocated 1032@code{struct sym_fns} whose @code{bfd} field contains the BFD 1033for the new symbol file being read. Its @code{private} field 1034has been zeroed, and can be modified as desired. Typically, 1035a struct of private information will be @code{malloc}'d, and 1036a pointer to it will be placed in the @code{private} field. 1037 1038There is no result from @code{@var{xxx}_symfile_init}, but it can call 1039@code{error} if it detects an unavoidable problem. 1040 1041@item @var{xxx}_new_init() 1042 1043Called from @code{symbol_file_add} when discarding existing symbols. 1044This function need only handle 1045the symbol-reading module's internal state; the symbol table data 1046structures visible to the rest of GDB will be discarded by 1047@code{symbol_file_add}. It has no arguments and no result. 1048It may be called after @code{@var{xxx}_symfile_init}, if a new symbol 1049table is being read, or may be called alone if all symbols are 1050simply being discarded. 1051 1052@item @var{xxx}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline) 1053 1054Called from @code{symbol_file_add} to actually read the symbols from a 1055symbol-file into a set of psymtabs or symtabs. 1056 1057@code{sf} points to the struct sym_fns originally passed to 1058@code{@var{xxx}_sym_init} for possible initialization. @code{addr} is the 1059offset between the file's specified start address and its true address 1060in memory. @code{mainline} is 1 if this is the main symbol table being 1061read, and 0 if a secondary symbol file (e.g. shared library or 1062dynamically loaded file) is being read.@refill 1063@end table 1064 1065In addition, if a symbol-reading module creates psymtabs when 1066@var{xxx}_symfile_read is called, these psymtabs will contain a pointer to 1067a function @code{@var{xxx}_psymtab_to_symtab}, which can be called from 1068any point in the GDB symbol-handling code. 1069 1070@table @code 1071@item @var{xxx}_psymtab_to_symtab (struct partial_symtab *pst) 1072 1073Called from @code{psymtab_to_symtab} (or the PSYMTAB_TO_SYMTAB 1074macro) if the psymtab has not already been read in and had its 1075@code{pst->symtab} pointer set. The argument is the psymtab 1076to be fleshed-out into a symtab. Upon return, pst->readin 1077should have been set to 1, and pst->symtab should contain a 1078pointer to the new corresponding symtab, or zero if there 1079were no symbols in that part of the symbol file. 1080@end table 1081 1082 1083@node Cleanups 1084@chapter Cleanups 1085 1086Cleanups are a structured way to deal with things that need to be done 1087later. When your code does something (like @code{malloc} some memory, or open 1088a file) that needs to be undone later (e.g. free the memory or close 1089the file), it can make a cleanup. The cleanup will be done at some 1090future point: when the command is finished, when an error occurs, or 1091when your code decides it's time to do cleanups. 1092 1093You can also discard cleanups, that is, throw them away without doing 1094what they say. This is only done if you ask that it be done. 1095 1096Syntax: 1097 1098@table @code 1099@item struct cleanup *@var{old_chain}; 1100Declare a variable which will hold a cleanup chain handle. 1101 1102@item @var{old_chain} = make_cleanup (@var{function}, @var{arg}); 1103Make a cleanup which will cause @var{function} to be called with @var{arg} 1104(a @code{char *}) later. The result, @var{old_chain}, is a handle that can be 1105passed to @code{do_cleanups} or @code{discard_cleanups} later. Unless you are 1106going to call @code{do_cleanups} or @code{discard_cleanups} yourself, 1107you can ignore the result from @code{make_cleanup}. 1108 1109 1110@item do_cleanups (@var{old_chain}); 1111Perform all cleanups done since @code{make_cleanup} returned @var{old_chain}. 1112E.g.: 1113@example 1114make_cleanup (a, 0); 1115old = make_cleanup (b, 0); 1116do_cleanups (old); 1117@end example 1118@noindent 1119will call @code{b()} but will not call @code{a()}. The cleanup that calls @code{a()} will remain 1120in the cleanup chain, and will be done later unless otherwise discarded.@refill 1121 1122@item discard_cleanups (@var{old_chain}); 1123Same as @code{do_cleanups} except that it just removes the cleanups from the 1124chain and does not call the specified functions. 1125 1126@end table 1127 1128Some functions, e.g. @code{fputs_filtered()} or @code{error()}, specify that they 1129``should not be called when cleanups are not in place''. This means 1130that any actions you need to reverse in the case of an error or 1131interruption must be on the cleanup chain before you call these functions, 1132since they might never return to your code (they @samp{longjmp} instead). 1133 1134 1135@node Wrapping 1136@chapter Wrapping Output Lines 1137 1138Output that goes through @code{printf_filtered} or @code{fputs_filtered} or 1139@code{fputs_demangled} needs only to have calls to @code{wrap_here} added 1140in places that would be good breaking points. The utility routines 1141will take care of actually wrapping if the line width is exceeded. 1142 1143The argument to @code{wrap_here} is an indentation string which is printed 1144@emph{only} if the line breaks there. This argument is saved away and used 1145later. It must remain valid until the next call to @code{wrap_here} or 1146until a newline has been printed through the @code{*_filtered} functions. 1147Don't pass in a local variable and then return! 1148 1149It is usually best to call @code{wrap_here()} after printing a comma or space. 1150If you call it before printing a space, make sure that your indentation 1151properly accounts for the leading space that will print if the line wraps 1152there. 1153 1154Any function or set of functions that produce filtered output must finish 1155by printing a newline, to flush the wrap buffer, before switching to 1156unfiltered (``@code{printf}'') output. Symbol reading routines that print 1157warnings are a good example. 1158 1159 1160@node Frames 1161@chapter Frames 1162 1163A frame is a construct that GDB uses to keep track of calling and called 1164functions. 1165 1166@table @code 1167@item FRAME_FP 1168in the machine description has no meaning to the machine-independent 1169part of GDB, except that it is used when setting up a new frame from 1170scratch, as follows: 1171 1172@example 1173 create_new_frame (read_register (FP_REGNUM), read_pc ())); 1174@end example 1175 1176Other than that, all the meaning imparted to @code{FP_REGNUM} is imparted by 1177the machine-dependent code. So, @code{FP_REGNUM} can have any value that 1178is convenient for the code that creates new frames. (@code{create_new_frame} 1179calls @code{INIT_EXTRA_FRAME_INFO} if it is defined; that is where you should 1180use the @code{FP_REGNUM} value, if your frames are nonstandard.) 1181 1182@item FRAME_CHAIN 1183Given a GDB frame, determine the address of the calling function's 1184frame. This will be used to create a new GDB frame struct, and then 1185@code{INIT_EXTRA_FRAME_INFO} and @code{INIT_FRAME_PC} will be called for 1186the new frame. 1187@end table 1188 1189@node Remote Stubs 1190@chapter Remote Stubs 1191 1192GDB's file @file{remote.c} talks a serial protocol to code that runs 1193in the target system. GDB provides several sample ``stubs'' that can 1194be integrated into target programs or operating systems for this purpose; 1195they are named @file{*-stub.c}. 1196 1197The GDB user's manual describes how to put such a stub into your target 1198code. What follows is a discussion of integrating the SPARC stub 1199into a complicated operating system (rather than a simple program), 1200by Stu Grossman, the author of this stub. 1201 1202The trap handling code in the stub assumes the following upon entry to 1203trap_low: 1204 1205@enumerate 1206@item %l1 and %l2 contain pc and npc respectively at the time of the trap 1207@item traps are disabled 1208@item you are in the correct trap window 1209@end enumerate 1210 1211As long as your trap handler can guarantee those conditions, then there is no 1212reason why you shouldn't be able to `share' traps with the stub. The stub has 1213no requirement that it be jumped to directly from the hardware trap vector. 1214That is why it calls @code{exceptionHandler()}, which is provided by the external 1215environment. For instance, this could setup the hardware traps to actually 1216execute code which calls the stub first, and then transfers to its own trap 1217handler. 1218 1219For the most point, there probably won't be much of an issue with `sharing' 1220traps, as the traps we use are usually not used by the kernel, and often 1221indicate unrecoverable error conditions. Anyway, this is all controlled by a 1222table, and is trivial to modify. 1223The most important trap for us is for @code{ta 1}. Without that, we 1224can't single step or do breakpoints. Everything else is unnecessary 1225for the proper operation of the debugger/stub. 1226 1227From reading the stub, it's probably not obvious how breakpoints work. They 1228are simply done by deposit/examine operations from GDB. 1229 1230@node Longjmp Support 1231@chapter Longjmp Support 1232 1233GDB has support for figuring out that the target is doing a 1234@code{longjmp} and for stopping at the target of the jump, if we are 1235stepping. This is done with a few specialized internal breakpoints, 1236which are visible in the @code{maint info breakpoint} command. 1237 1238To make this work, you need to define a macro called 1239@code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf} 1240structure and extract the longjmp target address. Since @code{jmp_buf} 1241is target specific, you will need to define it in the appropriate 1242@file{tm-xxx.h} file. Look in @file{tm-sun4os4.h} and 1243@file{sparc-tdep.c} for examples of how to do this. 1244 1245@node Coding Style 1246@chapter Coding Style 1247 1248GDB is generally written using the GNU coding standards, as described in 1249@file{standards.texi}, which is available for anonymous FTP from GNU 1250archive sites. There are some additional considerations for GDB 1251maintainers that reflect the unique environment and style of GDB 1252maintenance. If you follow these guidelines, GDB will be more 1253consistent and easier to maintain. 1254 1255GDB's policy on the use of prototypes is that prototypes are used 1256to @emph{declare} functions but never to @emph{define} them. Simple 1257macros are used in the declarations, so that a non-ANSI compiler can 1258compile GDB without trouble. The simple macro calls are used like 1259this: 1260 1261@example @code 1262extern int 1263memory_remove_breakpoint PARAMS ((CORE_ADDR, char *)); 1264@end example 1265 1266Note the double parentheses around the parameter types. This allows 1267an arbitrary number of parameters to be described, without freaking 1268out the C preprocessor. When the function has no parameters, it 1269should be described like: 1270 1271@example @code 1272void 1273noprocess PARAMS ((void)); 1274@end example 1275 1276The @code{PARAMS} macro expands to its argument in ANSI C, or to a simple 1277@code{()} in traditional C. 1278 1279All external functions should have a @code{PARAMS} declaration in a 1280header file that callers include. All static functions should have such 1281a declaration near the top of their source file. 1282 1283We don't have a gcc option that will properly check that these rules 1284have been followed, but it's GDB policy, and we periodically check it 1285using the tools available (plus manual labor), and clean up any remnants. 1286 1287@node Clean Design 1288@chapter Clean Design 1289 1290In addition to getting the syntax right, there's the little question of 1291semantics. Some things are done in certain ways in GDB because long 1292experience has shown that the more obvious ways caused various kinds of 1293trouble. In particular: 1294 1295@table @bullet 1296@item 1297You can't assume the byte order of anything that comes from a 1298target (including @var{value}s, object files, and instructions). Such 1299things must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in GDB, 1300or one of the swap routines defined in @file{bfd.h}, such as @code{bfd_get_32}. 1301 1302@item 1303You can't assume that you know what interface is being used to talk to 1304the target system. All references to the target must go through the 1305current @code{target_ops} vector. 1306 1307@item 1308You can't assume that the host and target machines are the same machine 1309(except in the ``native'' support modules). 1310In particular, you can't assume that the target machine's header files 1311will be available on the host machine. Target code must bring along its 1312own header files -- written from scratch or explicitly donated by their 1313owner, to avoid copyright problems. 1314 1315@item 1316Insertion of new @code{#ifdef}'s will be frowned upon. It's much better 1317to write the code portably than to conditionalize it for various systems. 1318 1319@item 1320New @code{#ifdef}'s which test for specific compilers or manufacturers 1321or operating systems are unacceptable. All @code{#ifdef}'s should test 1322for features. The information about which configurations contain which 1323features should be segregated into the configuration files. Experience 1324has proven far too often that a feature unique to one particular system 1325often creeps into other systems; and that a conditional based on 1326some predefined macro for your current system will become worthless 1327over time, as new versions of your system come out that behave differently 1328with regard to this feature. 1329 1330@item 1331Adding code that handles specific architectures, operating systems, target 1332interfaces, or hosts, is not acceptable in generic code. If a hook 1333is needed at that point, invent a generic hook and define it for your 1334configuration, with something like: 1335 1336@example 1337#ifdef WRANGLE_SIGNALS 1338 WRANGLE_SIGNALS (signo); 1339#endif 1340@end example 1341 1342In your host, target, or native configuration file, as appropriate, 1343define @code{WRANGLE_SIGNALS} to do the machine-dependent thing. Take 1344a bit of care in defining the hook, so that it can be used by other 1345ports in the future, if they need a hook in the same place. 1346 1347If the hook is not defined, the code should do whatever "most" machines 1348want. Using @code{#ifdef}, as above, is the preferred way to do this, 1349but sometimes that gets convoluted, in which case use 1350 1351@example 1352#ifndef SPECIAL_FOO_HANDLING 1353#define SPECIAL_FOO_HANDLING(pc, sp) (0) 1354#endif 1355@end example 1356 1357where the macro is used or in an appropriate header file. 1358 1359Whether to include a @dfn{small} hook, a hook around the exact pieces of 1360code which are system-dependent, or whether to replace a whole function 1361with a hook depends on the case. A good example of this dilemma can be 1362found in @code{get_saved_register}. All machines that GDB 2.8 ran on 1363just needed the @code{FRAME_FIND_SAVED_REGS} hook to find the saved 1364registers. Then the SPARC and Pyramid came along, and 1365@code{HAVE_REGISTER_WINDOWS} and @code{REGISTER_IN_WINDOW_P} were 1366introduced. Then the 29k and 88k required the @code{GET_SAVED_REGISTER} 1367hook. The first three are examples of small hooks; the latter replaces 1368a whole function. In this specific case, it is useful to have both 1369kinds; it would be a bad idea to replace all the uses of the small hooks 1370with @code{GET_SAVED_REGISTER}, since that would result in much 1371duplicated code. Other times, duplicating a few lines of code here or 1372there is much cleaner than introducing a large number of small hooks. 1373 1374Another way to generalize GDB along a particular interface is with an 1375attribute struct. For example, GDB has been generalized to handle 1376multiple kinds of remote interfaces -- not by #ifdef's everywhere, but 1377by defining the "target_ops" structure and having a current target (as 1378well as a stack of targets below it, for memory references). Whenever 1379something needs to be done that depends on which remote interface we are 1380using, a flag in the current target_ops structure is tested (e.g. 1381`target_has_stack'), or a function is called through a pointer in the 1382current target_ops structure. In this way, when a new remote interface 1383is added, only one module needs to be touched -- the one that actually 1384implements the new remote interface. Other examples of 1385attribute-structs are BFD access to multiple kinds of object file 1386formats, or GDB's access to multiple source languages. 1387 1388Please avoid duplicating code. For example, in GDB 3.x all the code 1389interfacing between @code{ptrace} and the rest of GDB was duplicated in 1390@file{*-dep.c}, and so changing something was very painful. In GDB 4.x, 1391these have all been consolidated into @file{infptrace.c}. 1392@file{infptrace.c} can deal with variations between systems the same way 1393any system-independent file would (hooks, #if defined, etc.), and 1394machines which are radically different don't need to use infptrace.c at 1395all. 1396 1397@item 1398@emph{Do} write code that doesn't depend on the sizes of C data types, 1399the format of the host's floating point numbers, the alignment of anything, 1400or the order of evaluation of expressions. In short, follow good 1401programming practices for writing portable C code. 1402 1403@end table 1404 1405@node Submitting Patches 1406@chapter Submitting Patches 1407 1408Thanks for thinking of offering your changes back to the community of 1409GDB users. In general we like to get well designed enhancements. 1410Thanks also for checking in advance about the best way to transfer the 1411changes. 1412 1413The two main problems with getting your patches in are, 1414 1415@table @bullet 1416@item 1417The GDB maintainers will only install ``cleanly designed'' patches. 1418You may not always agree on what is clean design. 1419@pxref{Coding Style}, @pxref{Clean Design}. 1420 1421@item 1422If the maintainers don't have time to put the patch in when it 1423arrives, or if there is any question about a patch, it 1424goes into a large queue with everyone else's patches and 1425bug reports. 1426@end table 1427 1428I don't know how to get past these problems except by continuing to try. 1429 1430There are two issues here -- technical and legal. 1431 1432The legal issue is that to incorporate substantial changes requires a 1433copyright assignment from you and/or your employer, granting ownership 1434of the changes to the Free Software Foundation. You can get the 1435standard document for doing this by sending mail to 1436@code{gnu@@prep.ai.mit.edu} and asking for it. I recommend that people 1437write in "All programs owned by the Free Software Foundation" as "NAME 1438OF PROGRAM", so that changes in many programs (not just GDB, but GAS, 1439Emacs, GCC, etc) can be contributed with only one piece of legalese 1440pushed through the bureacracy and filed with the FSF. I can't start 1441merging changes until this paperwork is received by the FSF (their 1442rules, which I follow since I maintain it for them). 1443 1444Technically, the easiest way to receive changes is to receive each 1445feature as a small context diff or unidiff, suitable for "patch". 1446Each message sent to me should include the changes to C code and 1447header files for a single feature, plus ChangeLog entries for each 1448directory where files were modified, and diffs for any changes needed 1449to the manuals (gdb/doc/gdb.texi or gdb/doc/gdbint.texi). If there 1450are a lot of changes for a single feature, they can be split down 1451into multiple messages. 1452 1453In this way, if I read and like the feature, I can add it to the 1454sources with a single patch command, do some testing, and check it in. 1455If you leave out the ChangeLog, I have to write one. If you leave 1456out the doc, I have to puzzle out what needs documenting. Etc. 1457 1458The reason to send each change in a separate message is that I will 1459not install some of the changes. They'll be returned to you with 1460questions or comments. If I'm doing my job, my message back to you 1461will say what you have to fix in order to make the change acceptable. 1462The reason to have separate messages for separate features is so 1463that other changes (which I @emph{am} willing to accept) can be installed 1464while one or more changes are being reworked. If multiple features 1465are sent in a single message, I tend to not put in the effort to sort 1466out the acceptable changes from the unacceptable, so none of the 1467features get installed until all are acceptable. 1468 1469If this sounds painful or authoritarian, well, it is. But I get a lot 1470of bug reports and a lot of patches, and most of them don't get 1471installed because I don't have the time to finish the job that the bug 1472reporter or the contributor could have done. Patches that arrive 1473complete, working, and well designed, tend to get installed on the day 1474they arrive. The others go into a queue and get installed if and when 1475I scan back over the queue -- which can literally take months 1476sometimes. It's in both our interests to make patch installation easy 1477-- you get your changes installed, and I make some forward progress on 1478GDB in a normal 12-hour day (instead of them having to wait until I 1479have a 14-hour or 16-hour day to spend cleaning up patches before I 1480can install them). 1481 1482Please send patches to @code{bug-gdb@@prep.ai.mit.edu}, if they are less 1483than about 25,000 characters. If longer than that, either make them 1484available somehow (e.g. anonymous FTP), and announce it on 1485@code{bug-gdb}, or send them directly to the GDB maintainers at 1486@code{gdb-patches@@cygnus.com}. 1487 1488@node Host Conditionals 1489@chapter Host Conditionals 1490 1491When GDB is configured and compiled, various macros are defined or left 1492undefined, to control compilation based on the attributes of the host 1493system. These macros and their meanings (or if the meaning is not 1494documented here, then one of the source files where they are used is 1495indicated) are: 1496 1497@emph{NOTE: For now, both host and target conditionals are here. 1498Eliminate target conditionals from this list as they are identified.} 1499 1500@table @code 1501 1502@item BLOCK_ADDRESS_FUNCTION_RELATIVE 1503dbxread.c 1504 1505@item GDBINIT_FILENAME 1506The default name of GDB's initialization file (normally @file{.gdbinit}). 1507 1508@item MEM_FNS_DECLARED 1509Your host config file defines this if it includes 1510declarations of @code{memcpy} and @code{memset}. Define this 1511to avoid conflicts between the native include 1512files and the declarations in @file{defs.h}. 1513 1514@item NO_SYS_FILE 1515Define this if your system does not have a @code{<sys/file.h>}. 1516 1517@item SIGWINCH_HANDLER 1518If your host defines @code{SIGWINCH}, you can define this to 1519be the name of a function to be called if @code{SIGWINCH} is received. 1520 1521@item SIGWINCH_HANDLER_BODY 1522Define this to expand into code that will define the function 1523named by the expansion of @code{SIGWINCH_HANDLER}. 1524 1525@item ADDITIONAL_OPTIONS 1526main.c 1527@item ADDITIONAL_OPTION_CASES 1528main.c 1529@item ADDITIONAL_OPTION_HANDLER 1530main.c 1531@item ADDITIONAL_OPTION_HELP 1532main.c 1533 1534@item AIX_BUGGY_PTRACE_CONTINUE 1535infptrace.c 1536 1537@item ALIGN_STACK_ON_STARTUP 1538Define this if your system is of a sort that will crash in @code{tgetent} 1539if the stack happens not to be longword-aligned when @code{main} is 1540called. This is a rare situation, but is known to occur on several 1541different types of systems. 1542 1543@item CFRONT_PRODUCER 1544dwarfread.c 1545@item DBX_PARM_SYMBOL_CLASS 1546stabsread.c 1547 1548@item DEFAULT_PROMPT 1549The default value of the prompt string (normally @code{"(gdb) "}). 1550 1551@item DEV_TTY 1552symmisc.c 1553@item DO_REGISTERS_INFO 1554infcmd.c 1555 1556@item FCLOSE_PROVIDED 1557Define this if the system declares @code{fclose} in the headers included in 1558@code{defs.h}. This isn't needed unless your compiler is unusually anal. 1559 1560@sc{ANSI} definition. 1561 1562@item FILES_INFO_HOOK 1563target.c 1564@item FLOAT_INFO 1565infcmd.c 1566 1567@item FOPEN_RB 1568Define this if binary files are opened the same way as text files. 1569 1570@item GCC2_COMPILED_FLAG_SYMBOL 1571dbxread.c 1572@item GCC_COMPILED_FLAG_SYMBOL 1573dbxread.c 1574@item GCC_MANGLE_BUG 1575symtab.c 1576@item GCC_PRODUCER 1577dwarfread.c 1578 1579@item GETENV_PROVIDED 1580Define this if the system declares @code{getenv} in its headers included in 1581@code{defs.h}. This isn't needed unless your compiler is unusually anal. 1582 1583@item GPLUS_PRODUCER 1584dwarfread.c 1585 1586@item HAVE_MMAP 1587In some cases, use the system call @code{mmap} for reading symbol 1588tables. For some machines this allows for sharing and quick updates. 1589 1590@item HAVE_SIGSETMASK 1591Define this if the host system has job control, but does not 1592define @code{sigsetmask()}. 1593Currently, this is only true of the RS/6000. 1594 1595@item HAVE_TERMIO 1596inflow.c 1597 1598@item HOST_BYTE_ORDER 1599The ordering of bytes in the host. 1600This must be defined to be either @code{BIG_ENDIAN} or @code{LITTLE_ENDIAN}. 1601 1602@item INT_MAX 1603@item INT_MIN 1604@item LONG_MAX 1605@item UINT_MAX 1606@item ULONG_MAX 1607Values for host-side constants. 1608 1609@item ISATTY 1610Substitute for isatty, if not available. 1611 1612@item KERNEL_DEBUGGING 1613tm-ultra3.h 1614 1615@item KERNEL_U_ADDR 1616Define this to the address of the @code{u} structure (the ``user struct'', 1617also known as the ``u-page'') in kernel virtual memory. GDB needs to know 1618this so that it can subtract this address from absolute addresses in 1619the upage, that are obtained via ptrace or from core files. On systems 1620that don't need this value, set it to zero. 1621 1622@item KERNEL_U_ADDR_BSD 1623Define this to cause GDB to determine the address of @code{u} at runtime, 1624by using Berkeley-style @code{nlist} on the kernel's image in the root 1625directory. 1626 1627@item KERNEL_U_ADDR_HPUX 1628Define this to cause GDB to determine the address of @code{u} at runtime, 1629by using HP-style @code{nlist} on the kernel's image in the root 1630directory. 1631 1632@item LCC_PRODUCER 1633dwarfread.c 1634 1635@item LONGEST 1636This is the longest integer type available on the host. 1637If not defined, it will default to @code{long long} or @code{long}, 1638depending on @code{CC_HAS_LONG_LONG}. 1639 1640@item CC_HAS_LONG_LONG 1641Define this if the host C compiler supports ``long long''. 1642This is set by the configure script. 1643 1644@item PRINTF_HAS_LONG_LONG 1645Define this if the host can handle printing of long long integers via a 1646format directive ``ll''. 1647 1648@item LSEEK_NOT_LINEAR 1649source.c 1650@item L_LNNO32 1651coffread.c 1652 1653@item L_SET 1654This macro is used as the argument to lseek (or, most commonly, bfd_seek). 1655FIXME, should be replaced by SEEK_SET instead, which is the POSIX equivalent. 1656 1657@item MAINTENANCE_CMDS 1658If the value of this is 1, then a number of optional maintenance commands 1659are compiled in. 1660 1661@item MALLOC_INCOMPATIBLE 1662Define this if the system's prototype for @code{malloc} differs from the 1663@sc{ANSI} definition. 1664 1665@item MMAP_BASE_ADDRESS 1666When using HAVE_MMAP, the first mapping should go at this address. 1667 1668@item MMAP_INCREMENT 1669when using HAVE_MMAP, this is the increment between mappings. 1670 1671@item NEED_POSIX_SETPGID 1672Define this to use the POSIX version of @code{setpgid} to determine 1673whether job control is available. 1674 1675@item NORETURN 1676If defined, this should be one or more tokens, such as @code{volatile}, 1677that can be used in both the declaration and definition of functions 1678to indicate that they never return. The default is already set 1679correctly if compiling with GCC. 1680This will almost never need to be defined. 1681 1682@item ATTR_NORETURN 1683If defined, this should be one or more tokens, such as 1684@code{__attribute__ ((noreturn))}, that can be used in the declarations 1685of functions to indicate that they never return. The default is already 1686set correctly if compiling with GCC. 1687This will almost never need to be defined. 1688 1689@item NOTICE_SIGNAL_HANDLING_CHANGE 1690infrun.c 1691@item NO_HIF_SUPPORT 1692remote-mm.c 1693@item NO_JOB_CONTROL 1694signals.h 1695 1696@item NO_MMALLOC 1697GDB will use the @code{mmalloc} library for memory allocation for symbol 1698reading, unless this symbol is defined. Define it on systems 1699on which @code{mmalloc} does not 1700work for some reason. One example is the DECstation, where its RPC 1701library can't cope with our redefinition of @code{malloc} to call 1702@code{mmalloc}. When defining @code{NO_MMALLOC}, you will also have 1703to override the setting of @code{MMALLOC_LIB} to empty, in the Makefile. 1704Therefore, this define is usually set on the command line by overriding 1705@code{MMALLOC_DISABLE} in @file{config/*/*.mh}, rather than by defining 1706it in @file{xm-*.h}. 1707 1708@item NO_MMCHECK 1709Define this if you are using @code{mmalloc}, but don't want the overhead 1710of checking the heap with @code{mmcheck}. Note that on some systems, 1711the C runtime makes calls to malloc prior to calling @code{main}, and if 1712@code{free} is ever called with these pointers after calling @code{mmcheck} 1713to enable checking, a memory corruption abort is certain to occur. These 1714systems can still use mmalloc, but must define NO_MMCHECK. 1715 1716@item MMCHECK_FORCE 1717Define this to 1 if the C runtime allocates memory prior to @code{mmcheck} 1718being called, but that memory is never freed so we don't have to worry 1719about it triggering a memory corruption abort. The default is 0, which 1720means that @code{mmcheck} will only install the heap checking functions 1721if there has not yet been any memory allocation calls, and if it fails 1722to install the functions, gdb will issue a warning. 1723 1724@item NO_SIGINTERRUPT 1725remote-adapt.c 1726@item NUMERIC_REG_NAMES 1727mips-tdep.c 1728@item N_SETV 1729dbxread.c 1730@item N_SET_MAGIC 1731hppabsd-tdep.c 1732@item ONE_PROCESS_WRITETEXT 1733breakpoint.c 1734@item O_BINARY 1735exec.c 1736@item O_RDONLY 1737xm-ultra3.h 1738@item PCC_SOL_BROKEN 1739dbxread.c 1740@item PC_LOAD_SEGMENT 1741stack.c 1742@item PRINT_RANDOM_SIGNAL 1743infcmd.c 1744@item PRINT_REGISTER_HOOK 1745infcmd.c 1746@item PROCESS_LINENUMBER_HOOK 1747buildsym.c 1748@item PROLOGUE_FIRSTLINE_OVERLAP 1749infrun.c 1750@item PUSH_ARGUMENTS 1751valops.c 1752@item PYRAMID_CONTROL_FRAME_DEBUGGING 1753pyr-xdep.c 1754@item PYRAMID_CORE 1755pyr-xdep.c 1756@item PYRAMID_PTRACE 1757pyr-xdep.c 1758@item REGISTER_BYTES 1759remote.c 1760@item REG_STACK_SEGMENT 1761exec.c 1762@item REG_STRUCT_HAS_ADDR 1763findvar.c 1764@item R_FP 1765dwarfread.c 1766@item R_OK 1767xm-altos.h 1768@item SEEK_END 1769state.c 1770@item SEEK_SET 1771state.c 1772@item SEM 1773coffread.c 1774 1775@item SHELL_COMMAND_CONCAT 1776infrun.c 1777@item SHELL_FILE 1778infrun.c 1779@item SHIFT_INST_REGS 1780breakpoint.c 1781@item SIGTRAP_STOP_AFTER_LOAD 1782infrun.c 1783@item STACK_ALIGN 1784valops.c 1785@item STOP_SIGNAL 1786main.c 1787@item SUN4_COMPILER_FEATURE 1788infrun.c 1789@item SUN_FIXED_LBRAC_BUG 1790dbxread.c 1791@item SVR4_SHARED_LIBS 1792solib.c 1793@item SYMBOL_RELOADING_DEFAULT 1794symfile.c 1795@item TIOCGETC 1796inflow.c 1797@item TIOCGLTC 1798inflow.c 1799@item TIOCGPGRP 1800inflow.c 1801@item TIOCLGET 1802inflow.c 1803@item TIOCLSET 1804inflow.c 1805@item TIOCNOTTY 1806inflow.c 1807@item UPAGES 1808altos-xdep.c 1809@item USE_O_NOCTTY 1810inflow.c 1811 1812@item USG 1813Means that System V (prior to SVR4) include files are in use. 1814(FIXME: This symbol is abused in @file{infrun.c}, @file{regex.c}, 1815@file{remote-nindy.c}, and @file{utils.c} for other things, at the moment.) 1816 1817@item WRS_ORIG 1818remote-vx.c 1819@item alloca 1820defs.h 1821@item const 1822defs.h 1823 1824@item lint 1825Define this to help lint in some stupid way. 1826 1827@item volatile 1828Define this to override the defaults of @code{__volatile__} or @code{/**/}. 1829 1830@end table 1831 1832Platform-specific host conditionals. 1833 1834@table @code 1835 1836@item ALTOS 1837altos-xdep.c 1838@item ALTOS_AS 1839xm-altos.h 1840@item MOTOROLA 1841xm-altos.h 1842@item NBPG 1843altos-xdep.c 1844 1845@item BCS 1846tm-delta88.h 1847 1848@item DELTA88 1849m88k-xdep.c 1850@item DGUX 1851m88k-xdep.c 1852 1853@item F_OK 1854xm-ultra3.h 1855 1856@end table 1857 1858Regex conditionals. 1859 1860@table @code 1861 1862@item C_ALLOCA 1863regex.c 1864@item NFAILURES 1865regex.c 1866@item RE_NREGS 1867regex.h 1868@item SIGN_EXTEND_CHAR 1869regex.c 1870@item SWITCH_ENUM_BUG 1871regex.c 1872@item SYNTAX_TABLE 1873regex.c 1874@item Sword 1875regex.c 1876@item sparc 1877regex.c 1878@item test 1879regex.c 1880 1881@end table 1882 1883@node Target Conditionals 1884@chapter Target Conditionals 1885 1886When GDB is configured and compiled, various macros are defined or left 1887undefined, to control compilation based on the attributes of the target 1888system. These macros and their meanings are: 1889 1890@emph{NOTE: For now, both host and target conditionals are here. 1891Eliminate host conditionals from this list as they are identified.} 1892 1893@table @code 1894 1895@item PUSH_DUMMY_FRAME 1896Used in @samp{call_function_by_hand} to create an artificial stack frame. 1897 1898@item POP_FRAME 1899Used in @samp{call_function_by_hand} to remove an artificial stack frame. 1900 1901@item BLOCK_ADDRESS_FUNCTION_RELATIVE 1902dbxread.c 1903@item PYRAMID_CONTROL_FRAME_DEBUGGING 1904pyr-xdep.c 1905@item ADDITIONAL_OPTIONS 1906main.c 1907@item ADDITIONAL_OPTION_CASES 1908main.c 1909@item ADDITIONAL_OPTION_HANDLER 1910main.c 1911@item ADDITIONAL_OPTION_HELP 1912main.c 1913 1914@item ADDR_BITS_REMOVE (addr) 1915If a raw machine address includes any bits that are not really part 1916of the address, then define this macro to expand into an expression 1917that zeros those bits in @var{addr}. For example, the two low-order 1918bits of a Motorola 88K address may be used by some kernels for their 1919own purposes, since addresses must always be 4-byte aligned, and so 1920are of no use for addressing. Those bits should be filtered out with 1921an expression such as @code{((addr) & ~3)}. 1922 1923@item ALIGN_STACK_ON_STARTUP 1924main.c 1925@item ALTOS 1926altos-xdep.c 1927@item ALTOS_AS 1928xm-altos.h 1929@item BCS 1930tm-delta88.h 1931 1932@item BEFORE_MAIN_LOOP_HOOK 1933Define this to expand into any code that you want to execute before 1934the main loop starts. Although this is not, strictly speaking, 1935a target conditional, that is how it is currently being used. 1936Note that if a configuration were to define it one way for a host 1937and a different way for the target, GDB will probably not compile, 1938let alone run correctly. 1939 1940@item BELIEVE_PCC_PROMOTION 1941coffread.c 1942@item BELIEVE_PCC_PROMOTION_TYPE 1943stabsread.c 1944 1945@item BITS_BIG_ENDIAN 1946Define this if the numbering of bits in the targets does *not* match 1947the endianness of the target byte order. 1948A value of 1 means that the bits are numbered in a big-endian order, 19490 means little-endian. 1950 1951@item BLOCK_ADDRESS_ABSOLUTE 1952dbxread.c 1953@item BREAKPOINT 1954tm-m68k.h 1955 1956@item CALL_DUMMY 1957valops.c 1958@item CALL_DUMMY_LOCATION 1959inferior.h 1960@item CALL_DUMMY_STACK_ADJUST 1961valops.c 1962 1963@item CANNOT_FETCH_REGISTER (regno) 1964A C expression that should be nonzero if @var{regno} cannot be 1965fetched from an inferior process. 1966This is only relevant if @code{FETCH_INFERIOR_REGISTERS} is not 1967defined. 1968 1969@item CANNOT_STORE_REGISTER (regno) 1970A C expression that should be nonzero if @var{regno} should not be 1971written to the target. This is often the case for program counters, 1972status words, and other special registers. If this is not defined, 1973GDB will assume that all registers may be written. 1974 1975@item CFRONT_PRODUCER 1976dwarfread.c 1977 1978@item DO_DEFERRED_STORES 1979@item CLEAR_DEFERRED_STORES 1980Define this to execute any deferred stores of registers into the inferior, 1981and to cancel any deferred stores. 1982 1983Currently only implemented correctly for native Sparc configurations? 1984 1985@item CPLUS_MARKER 1986Define this to expand into the character that G++ uses to 1987distinguish compiler-generated identifiers from programmer-specified 1988identifiers. By default, this expands into @code{'$'}. 1989Most System V targets should define this to @code{'.'}. 1990 1991@item DBX_PARM_SYMBOL_CLASS 1992stabsread.c 1993 1994@item DECR_PC_AFTER_BREAK 1995Define this to be the amount by which to decrement the PC after 1996the program encounters a breakpoint. 1997This is often the number of bytes in BREAKPOINT, though not always. 1998For most targets this value will be 0. 1999 2000@item DECR_PC_AFTER_HW_BREAK 2001Similarly, for hardware breakpoints. 2002 2003@item DELTA88 2004m88k-xdep.c 2005@item DEV_TTY 2006symmisc.c 2007@item DGUX 2008m88k-xdep.c 2009 2010@item DISABLE_UNSETTABLE_BREAK addr 2011If defined, this should evaluate to 1 if @var{addr} is in a shared 2012library in which breakpoints cannot be set and so should be disabled. 2013 2014@item DO_REGISTERS_INFO 2015infcmd.c 2016 2017@item END_OF_TEXT_DEFAULT 2018This is an expression that should designate the end of the text section 2019(? FIXME ?) 2020 2021@item EXTRACT_RETURN_VALUE 2022tm-m68k.h 2023@item EXTRACT_STRUCT_VALUE_ADDRESS 2024values.c 2025 2026@item EXTRA_FRAME_INFO 2027If defined, this must be a list of slots that may be inserted into 2028the @code{frame_info} structure defined in @code{frame.h}. 2029 2030@item EXTRA_SYMTAB_INFO 2031If defined, this must be a list of slots that may be inserted into 2032the @code{symtab} structure defined in @code{symtab.h}. 2033 2034@item FILES_INFO_HOOK 2035target.c 2036@item FLOAT_INFO 2037infcmd.c 2038@item FP0_REGNUM 2039a68v-xdep.c 2040@item FPC_REGNUM 2041mach386-xdep.c 2042@item FP_REGNUM 2043parse.c 2044@item FRAMELESS_FUNCTION_INVOCATION 2045blockframe.c 2046@item FRAME_ARGS_ADDRESS_CORRECT 2047stack.c 2048 2049@item FRAME_CHAIN 2050Given FRAME, return a pointer to the calling frame. 2051 2052@item FRAME_CHAIN_COMBINE 2053blockframe.c 2054@item FRAME_CHAIN_VALID 2055frame.h 2056@item FRAME_CHAIN_VALID_ALTERNATE 2057frame.h 2058@item FRAME_FIND_SAVED_REGS 2059stack.c 2060@item FRAME_GET_BASEREG_VALUE 2061frame.h 2062 2063@item FRAME_NUM_ARGS (val, fi) 2064For the frame described by fi, set val to the number of arguments 2065that are being passed. 2066 2067@item FRAME_SPECIFICATION_DYADIC 2068stack.c 2069 2070@item FRAME_SAVED_PC 2071Given FRAME, return the pc saved there. That is, the return address. 2072 2073@item FUNCTION_EPILOGUE_SIZE 2074For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the 2075function end symbol is 0. For such targets, you must define 2076@code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size 2077of a function's epilogue. 2078 2079@item GCC2_COMPILED_FLAG_SYMBOL 2080dbxread.c 2081@item GCC_COMPILED_FLAG_SYMBOL 2082dbxread.c 2083@item GCC_MANGLE_BUG 2084symtab.c 2085@item GCC_PRODUCER 2086dwarfread.c 2087 2088@item GDB_TARGET_IS_HPPA 2089This determines whether horrible kludge code in dbxread.c and partial-stab.h 2090is used to mangle multiple-symbol-table files from HPPA's. This should all 2091be ripped out, and a scheme like elfread.c used. 2092 2093@item GDB_TARGET_IS_MACH386 2094mach386-xdep.c 2095@item GDB_TARGET_IS_SUN3 2096a68v-xdep.c 2097@item GDB_TARGET_IS_SUN386 2098sun386-xdep.c 2099 2100@item GET_LONGJMP_TARGET 2101For most machines, this is a target-dependent parameter. On the DECstation 2102and the Iris, this is a native-dependent parameter, since <setjmp.h> is 2103needed to define it. 2104 2105This macro determines the target PC address that longjmp() will jump 2106to, assuming that we have just stopped at a longjmp breakpoint. It 2107takes a CORE_ADDR * as argument, and stores the target PC value through 2108this pointer. It examines the current state of the machine as needed. 2109 2110@item GET_SAVED_REGISTER 2111Define this if you need to supply your own definition for the 2112function @code{get_saved_register}. 2113Currently this is only done for the a29k. 2114 2115@item GPLUS_PRODUCER 2116dwarfread.c 2117 2118@item GR64_REGNUM 2119Very a29k-specific. 2120 2121@item HAVE_REGISTER_WINDOWS 2122Define this if the target has register windows. 2123@item REGISTER_IN_WINDOW_P regnum 2124Define this to be an expression that is 1 is the given register is 2125in the window. 2126 2127@item IBM6000_TARGET 2128Shows that we are configured for an IBM RS/6000 target. This conditional 2129should be eliminated (FIXME) and replaced by feature-specific macros. 2130It was introduced in haste and we are repenting at leisure. 2131 2132@item IEEE_FLOAT 2133Define this if the target system uses IEEE-format floating point numbers. 2134 2135@item IGNORE_SYMBOL type 2136This seems to be no longer used. 2137 2138@item INIT_EXTRA_FRAME_INFO (fromleaf, fci) 2139If defined, this should be a C expression or statement that fills 2140in the @code{EXTRA_FRAME_INFO} slots of the given frame @var{fci}. 2141 2142@item INIT_EXTRA_SYMTAB_INFO 2143symfile.c 2144 2145@item INIT_FRAME_PC (fromleaf, prev) 2146This is a C statement that sets the pc of the frame pointed 2147to by @var{prev}. [By default...] 2148 2149@item INNER_THAN 2150Define this to be either @code{<} if the target's stack grows 2151downward in memory, or @code{>} is the stack grows upwards. 2152 2153@item IN_SIGTRAMP (pc name) 2154Define this to return true if the given pc and/or name indicates 2155that the current function is a sigtramp. 2156 2157@item SIGTRAMP_START (pc) 2158@item SIGTRAMP_END (pc) 2159Define these to be the start and end address of the sigtramp for the given pc. 2160On machines where the address is just a compile time constant, the macro 2161expansion will typically just ignore the supplied pc. 2162 2163@item IN_SOLIB_TRAMPOLINE pc name 2164Define this to evaluate to nonzero if the program is stopped in 2165the trampoline that connects to a shared library. 2166 2167@item IS_TRAPPED_INTERNALVAR name 2168This is an ugly hook to allow the specification of special actions 2169that should occur as a side-effect of setting the value of a variable 2170internal to GDB. Currently only used by the h8500. 2171Note that this could be either a host or target conditional. 2172 2173@item KERNEL_DEBUGGING 2174tm-ultra3.h 2175@item LCC_PRODUCER 2176dwarfread.c 2177@item L_LNNO32 2178coffread.c 2179@item MIPSEL 2180mips-tdep.c 2181@item MOTOROLA 2182xm-altos.h 2183@item NBPG 2184altos-xdep.c 2185 2186@item NEED_TEXT_START_END 2187Define this if GDB should determine the start and end addresses 2188of the text section. (Seems dubious.) 2189 2190@item NOTICE_SIGNAL_HANDLING_CHANGE 2191infrun.c 2192@item NO_HIF_SUPPORT 2193remote-mm.c 2194@item NO_SIGINTERRUPT 2195remote-adapt.c 2196 2197@item NO_SINGLE_STEP 2198Define this if the target does not support single-stepping. 2199If this is defined, you must supply, in @code{*-tdep.c}, the function 2200@code{single_step}, which takes a target_signal as argument and returns nothing. 2201It must insert breakpoints at each possible destinations of the next 2202instruction. See @code{sparc-tdep.c} and @code{rs6000-tdep.c} 2203for examples. 2204 2205@item NUMERIC_REG_NAMES 2206mips-tdep.c 2207@item N_SETV 2208dbxread.c 2209@item N_SET_MAGIC 2210hppabsd-tdep.c 2211@item ONE_PROCESS_WRITETEXT 2212breakpoint.c 2213@item PCC_SOL_BROKEN 2214dbxread.c 2215@item PC_IN_CALL_DUMMY 2216inferior.h 2217@item PC_LOAD_SEGMENT 2218stack.c 2219 2220@item PC_REGNUM 2221If the program counter is kept in a register, then define this macro 2222to be the number of that register. 2223This need be defined only if @code{TARGET_WRITE_PC} is not defined. 2224 2225@item NPC_REGNUM 2226The number of the ``next program counter'' register, if defined. 2227 2228@item NNPC_REGNUM 2229The number of the ``next next program counter'' register, if defined. 2230Currently, this is only defined for the Motorola 88K. 2231 2232@item PRINT_RANDOM_SIGNAL 2233infcmd.c 2234@item PRINT_REGISTER_HOOK 2235infcmd.c 2236 2237@item PRINT_TYPELESS_INTEGER 2238This is an obscure substitute for @code{print_longest} that 2239seems to have been defined for the Convex target. 2240 2241@item PROCESS_LINENUMBER_HOOK 2242buildsym.c 2243@item PROLOGUE_FIRSTLINE_OVERLAP 2244infrun.c 2245@item PS_REGNUM 2246parse.c 2247@item PUSH_ARGUMENTS 2248valops.c 2249@item REGISTER_BYTES 2250remote.c 2251 2252@item REGISTER_NAMES 2253Define this to expand into an initializer of an array of strings. 2254Each string is the name of a register. 2255[more detail] 2256 2257@item REG_STACK_SEGMENT 2258exec.c 2259@item REG_STRUCT_HAS_ADDR 2260findvar.c 2261@item R_FP 2262dwarfread.c 2263@item R_OK 2264xm-altos.h 2265 2266@item SDB_REG_TO_REGNUM 2267Define this to convert sdb register numbers 2268into GDB regnums. If not defined, no conversion will be done. 2269 2270@item SEEK_END 2271state.c 2272@item SEEK_SET 2273state.c 2274@item SEM 2275coffread.c 2276@item SHELL_COMMAND_CONCAT 2277infrun.c 2278@item SHELL_FILE 2279infrun.c 2280@item SHIFT_INST_REGS 2281breakpoint.c 2282@item SIGTRAP_STOP_AFTER_LOAD 2283infrun.c 2284 2285@item SKIP_PROLOGUE 2286A C statement that advances the PC across any function entry 2287prologue instructions so as to reach ``real'' code. 2288 2289@item SKIP_PROLOGUE_FRAMELESS_P 2290A C statement that should behave similarly, but that can stop 2291as soon as the function is known to have a frame. 2292If not defined, @code{SKIP_PROLOGUE} will be used instead. 2293 2294@item SKIP_TRAMPOLINE_CODE (pc) 2295If the target machine has trampoline code that sits between callers 2296and the functions being called, then define this macro to return 2297a new PC that is at the start of the real function. 2298 2299@item SP_REGNUM 2300parse.c 2301 2302@item STAB_REG_TO_REGNUM 2303Define this to convert stab register numbers (as gotten from `r' declarations) 2304into GDB regnums. If not defined, no conversion will be done. 2305 2306@item STACK_ALIGN 2307valops.c 2308@item STOP_SIGNAL 2309main.c 2310 2311@item STORE_RETURN_VALUE (type, valbuf) 2312A C expression that stores a function return value of type @var{type}, 2313where @var{valbuf} is the address of the value to be stored. 2314 2315@item SUN4_COMPILER_FEATURE 2316infrun.c 2317@item SUN_FIXED_LBRAC_BUG 2318dbxread.c 2319@item SVR4_SHARED_LIBS 2320solib.c 2321@item SYMBOL_RELOADING_DEFAULT 2322symfile.c 2323 2324@item TARGET_BYTE_ORDER 2325The ordering of bytes in the target. 2326This must be defined to be either @code{BIG_ENDIAN} or @code{LITTLE_ENDIAN}. 2327 2328@item TARGET_CHAR_BIT 2329Number of bits in a char; defaults to 8. 2330 2331@item TARGET_COMPLEX_BIT 2332Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}. 2333 2334@item TARGET_DOUBLE_BIT 2335Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}. 2336 2337@item TARGET_DOUBLE_COMPLEX_BIT 2338Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}. 2339 2340@item TARGET_FLOAT_BIT 2341Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}. 2342 2343@item TARGET_INT_BIT 2344Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}. 2345 2346@item TARGET_LONG_BIT 2347Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}. 2348 2349@item TARGET_LONG_DOUBLE_BIT 2350Number of bits in a long double float; 2351defaults to @code{2 * TARGET_DOUBLE_BIT}. 2352 2353@item TARGET_LONG_LONG_BIT 2354Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}. 2355 2356@item TARGET_PTR_BIT 2357Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}. 2358 2359@item TARGET_SHORT_BIT 2360Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}. 2361 2362@item TARGET_READ_PC 2363@item TARGET_WRITE_PC (val, pid) 2364@item TARGET_READ_SP 2365@item TARGET_WRITE_SP 2366@item TARGET_READ_FP 2367@item TARGET_WRITE_FP 2368These change the behavior of @code{read_pc}, @code{write_pc}, 2369@code{read_sp}, @code{write_sp}, @code{read_fp} and @code{write_fp}. 2370For most targets, these may be left undefined. GDB will call the 2371read and write register functions with the relevant @code{_REGNUM} argument. 2372 2373These macros are useful when a target keeps one of these registers in a 2374hard to get at place; for example, part in a segment register and part 2375in an ordinary register. 2376 2377@item USE_STRUCT_CONVENTION (gcc_p, type) 2378If defined, this must be an expression that is nonzero if a value 2379of the given @var{type} being returned from a function must have 2380space allocated for it on the stack. @var{gcc_p} is true if the 2381function being considered is known to have been compiled by GCC; 2382this is helpful for systems where GCC is known to use different calling 2383convention than other compilers. 2384 2385@item VARIABLES_INSIDE_BLOCK (desc, gcc_p) 2386For dbx-style debugging information, if the compiler puts variable 2387declarations inside LBRAC/RBRAC blocks, this should be defined 2388to be nonzero. @var{desc} is the value of @code{n_desc} from the 2389@code{N_RBRAC} symbol, and @var{gcc_p} is true if GDB has noticed 2390the presence of either the @code{GCC_COMPILED_SYMBOL} or the 2391@code{GCC2_COMPILED_SYMBOL}. 2392By default, this is 0. 2393 2394@item OS9K_VARIABLES_INSIDE_BLOCK (desc, gcc_p) 2395Similarly, for OS/9000. Defaults to 1. 2396 2397@item WRS_ORIG 2398remote-vx.c 2399 2400@item test 2401(Define this to enable testing code in regex.c.) 2402 2403@end table 2404 2405Motorola M68K target conditionals. 2406 2407@table @code 2408 2409@item BPT_VECTOR 2410Define this to be the 4-bit location of the breakpoint trap vector. 2411If not defined, it will default to @code{0xf}. 2412 2413@item REMOTE_BPT_VECTOR 2414Defaults to @code{1}. 2415 2416@end table 2417 2418@node Native Conditionals 2419@chapter Native Conditionals 2420 2421When GDB is configured and compiled, various macros are defined or left 2422undefined, to control compilation when the host and target systems 2423are the same. These macros should be defined (or left undefined) 2424in @file{nm-@var{system}.h}. 2425 2426@table @code 2427 2428@item ATTACH_DETACH 2429If defined, then GDB will include support for the @code{attach} and 2430@code{detach} commands. 2431 2432@item CHILD_PREPARE_TO_STORE 2433If the machine stores all registers at once in the child process, 2434then define this to ensure that all values are correct. 2435This usually entails a read from the child. 2436 2437[Note that this is incorrectly defined in @file{xm-@var{system}.h} 2438files currently.] 2439 2440@item FETCH_INFERIOR_REGISTERS 2441Define this if the native-dependent code will provide its 2442own routines 2443@code{fetch_inferior_registers} and @code{store_inferior_registers} in 2444@file{@var{HOST}-nat.c}. 2445If this symbol is @emph{not} defined, and @file{infptrace.c} 2446is included in this configuration, the default routines in 2447@file{infptrace.c} are used for these functions. 2448 2449@item GET_LONGJMP_TARGET 2450For most machines, this is a target-dependent parameter. On the DECstation 2451and the Iris, this is a native-dependent parameter, since <setjmp.h> is 2452needed to define it. 2453 2454This macro determines the target PC address that longjmp() will jump 2455to, assuming that we have just stopped at a longjmp breakpoint. It 2456takes a CORE_ADDR * as argument, and stores the target PC value through 2457this pointer. It examines the current state of the machine as needed. 2458 2459@item PROC_NAME_FMT 2460Defines the format for the name of a @file{/proc} device. Should be 2461defined in @file{nm.h} @emph{only} in order to override the default 2462definition in @file{procfs.c}. 2463 2464@item PTRACE_FP_BUG 2465mach386-xdep.c 2466 2467@item PTRACE_ARG3_TYPE 2468The type of the third argument to the @code{ptrace} system call, if it exists 2469and is different from @code{int}. 2470 2471@item REGISTER_U_ADDR 2472Defines the offset of the registers in the ``u area''; @pxref{Host}. 2473 2474@item SOLIB_ADD (filename, from_tty, targ) 2475Define this to expand into an expression that will cause the symbols 2476in @var{filename} to be added to GDB's symbol table. 2477 2478@item SOLIB_CREATE_INFERIOR_HOOK 2479Define this to expand into any shared-library-relocation code 2480that you want to be run just after the child process has been forked. 2481 2482@item START_INFERIOR_TRAPS_EXPECTED 2483When starting an inferior, GDB normally expects to trap twice; 2484once when the shell execs, and once when the program itself execs. 2485If the actual number of traps is something other than 2, then 2486define this macro to expand into the number expected. 2487 2488@item USE_PROC_FS 2489This determines whether small routines in @file{*-tdep.c}, which 2490translate register values 2491between GDB's internal representation and the /proc representation, 2492are compiled. 2493 2494@item U_REGS_OFFSET 2495This is the offset of the registers in the upage. It need only be 2496defined if the generic ptrace register access routines in 2497@file{infptrace.c} are being used (that is, 2498@file{infptrace.c} is configured in, and 2499@code{FETCH_INFERIOR_REGISTERS} is not defined). If the default value 2500from @file{infptrace.c} is good enough, leave it undefined. 2501 2502The default value means that u.u_ar0 @emph{points to} the location of the 2503registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means that 2504u.u_ar0 @emph{is} the location of the registers. 2505 2506@item CLEAR_SOLIB 2507objfiles.c 2508 2509@item DEBUG_PTRACE 2510Define this to debug ptrace calls. 2511 2512@end table 2513 2514@node Obsolete Conditionals 2515@chapter Obsolete Conditionals 2516 2517Fragments of old code in GDB sometimes reference or set the following 2518configuration macros. They should not be used by new code, and 2519old uses should be removed as those parts of the debugger are 2520otherwise touched. 2521 2522@table @code 2523 2524@item STACK_END_ADDR 2525This macro used to define where the end of the stack appeared, for use 2526in interpreting core file formats that don't record this address in the 2527core file itself. This information is now configured in BFD, and GDB 2528gets the info portably from there. The values in GDB's configuration 2529files should be moved into BFD configuration files (if needed there), 2530and deleted from all of GDB's config files. 2531 2532Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR 2533is so old that it has never been converted to use BFD. Now that's old! 2534 2535@end table 2536 2537@node XCOFF 2538@chapter The XCOFF Object File Format 2539 2540The IBM RS/6000 running AIX uses an object file format called xcoff. 2541The COFF sections, symbols, and line numbers are used, but debugging 2542symbols are dbx-style stabs whose strings are located in the 2543@samp{.debug} section (rather than the string table). For more 2544information, @xref{Top,,,stabs,The Stabs Debugging Format}, and search 2545for XCOFF. 2546 2547The shared library scheme has a nice clean interface for figuring out 2548what shared libraries are in use, but the catch is that everything which 2549refers to addresses (symbol tables and breakpoints at least) needs to be 2550relocated for both shared libraries and the main executable. At least 2551using the standard mechanism this can only be done once the program has 2552been run (or the core file has been read). 2553 2554@contents 2555@bye 2556