1 \input texinfo @c -*- texinfo -*-
2 @setfilename gdbint.info
4 @dircategory Programming & development tools.
6 * Gdb-Internals: (gdbint). The GNU debugger's internals.
10 This file documents the internals of the GNU debugger @value{GDBN}.
11 Copyright 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001,2002
12 Free Software Foundation, Inc.
13 Contributed by Cygnus Solutions. Written by John Gilmore.
14 Second Edition by Stan Shebs.
16 Permission is granted to copy, distribute and/or modify this document
17 under the terms of the GNU Free Documentation License, Version 1.1 or
18 any later version published by the Free Software Foundation; with no
19 Invariant Sections, with the Front-Cover Texts being ``A GNU Manual,''
20 and with the Back-Cover Texts as in (a) below.
22 (a) The FSF's Back-Cover Text is: ``You have freedom to copy and modify
23 this GNU Manual, like GNU software. Copies published by the Free
24 Software Foundation raise funds for GNU development.''
27 @setchapternewpage off
28 @settitle @value{GDBN} Internals
34 @title @value{GDBN} Internals
35 @subtitle{A guide to the internals of the GNU debugger}
37 @author Cygnus Solutions
38 @author Second Edition:
40 @author Cygnus Solutions
43 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
44 \xdef\manvers{\$Revision$} % For use in headers, footers too
46 \hfill Cygnus Solutions\par
48 \hfill \TeX{}info \texinfoversion\par
52 @vskip 0pt plus 1filll
53 Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001
54 Free Software Foundation, Inc.
56 Permission is granted to copy, distribute and/or modify this document
57 under the terms of the GNU Free Documentation License, Version 1.1 or
58 any later version published by the Free Software Foundation; with no
59 Invariant Sections, with the Front-Cover Texts being ``A GNU Manual,''
60 and with the Back-Cover Texts as in (a) below.
62 (a) The FSF's Back-Cover Text is: ``You have freedom to copy and modify
63 this GNU Manual, like GNU software. Copies published by the Free
64 Software Foundation raise funds for GNU development.''
70 @c Perhaps this should be the title of the document (but only for info,
71 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
72 @top Scope of this Document
74 This document documents the internals of the GNU debugger, @value{GDBN}. It
75 includes description of @value{GDBN}'s key algorithms and operations, as well
76 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
87 * Target Architecture Definition::
88 * Target Vector Definition::
97 * GNU Free Documentation License:: The license for this documentation
103 @chapter Requirements
104 @cindex requirements for @value{GDBN}
106 Before diving into the internals, you should understand the formal
107 requirements and other expectations for @value{GDBN}. Although some
108 of these may seem obvious, there have been proposals for @value{GDBN}
109 that have run counter to these requirements.
111 First of all, @value{GDBN} is a debugger. It's not designed to be a
112 front panel for embedded systems. It's not a text editor. It's not a
113 shell. It's not a programming environment.
115 @value{GDBN} is an interactive tool. Although a batch mode is
116 available, @value{GDBN}'s primary role is to interact with a human
119 @value{GDBN} should be responsive to the user. A programmer hot on
120 the trail of a nasty bug, and operating under a looming deadline, is
121 going to be very impatient of everything, including the response time
122 to debugger commands.
124 @value{GDBN} should be relatively permissive, such as for expressions.
125 While the compiler should be picky (or have the option to be made
126 picky), since source code lives for a long time usually, the
127 programmer doing debugging shouldn't be spending time figuring out to
128 mollify the debugger.
130 @value{GDBN} will be called upon to deal with really large programs.
131 Executable sizes of 50 to 100 megabytes occur regularly, and we've
132 heard reports of programs approaching 1 gigabyte in size.
134 @value{GDBN} should be able to run everywhere. No other debugger is
135 available for even half as many configurations as @value{GDBN}
139 @node Overall Structure
141 @chapter Overall Structure
143 @value{GDBN} consists of three major subsystems: user interface,
144 symbol handling (the @dfn{symbol side}), and target system handling (the
147 The user interface consists of several actual interfaces, plus
150 The symbol side consists of object file readers, debugging info
151 interpreters, symbol table management, source language expression
152 parsing, type and value printing.
154 The target side consists of execution control, stack frame analysis, and
155 physical target manipulation.
157 The target side/symbol side division is not formal, and there are a
158 number of exceptions. For instance, core file support involves symbolic
159 elements (the basic core file reader is in BFD) and target elements (it
160 supplies the contents of memory and the values of registers). Instead,
161 this division is useful for understanding how the minor subsystems
164 @section The Symbol Side
166 The symbolic side of @value{GDBN} can be thought of as ``everything
167 you can do in @value{GDBN} without having a live program running''.
168 For instance, you can look at the types of variables, and evaluate
169 many kinds of expressions.
171 @section The Target Side
173 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
174 Although it may make reference to symbolic info here and there, most
175 of the target side will run with only a stripped executable
176 available---or even no executable at all, in remote debugging cases.
178 Operations such as disassembly, stack frame crawls, and register
179 display, are able to work with no symbolic info at all. In some cases,
180 such as disassembly, @value{GDBN} will use symbolic info to present addresses
181 relative to symbols rather than as raw numbers, but it will work either
184 @section Configurations
188 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
189 @dfn{Target} refers to the system where the program being debugged
190 executes. In most cases they are the same machine, in which case a
191 third type of @dfn{Native} attributes come into play.
193 Defines and include files needed to build on the host are host support.
194 Examples are tty support, system defined types, host byte order, host
197 Defines and information needed to handle the target format are target
198 dependent. Examples are the stack frame format, instruction set,
199 breakpoint instruction, registers, and how to set up and tear down the stack
202 Information that is only needed when the host and target are the same,
203 is native dependent. One example is Unix child process support; if the
204 host and target are not the same, doing a fork to start the target
205 process is a bad idea. The various macros needed for finding the
206 registers in the @code{upage}, running @code{ptrace}, and such are all
207 in the native-dependent files.
209 Another example of native-dependent code is support for features that
210 are really part of the target environment, but which require
211 @code{#include} files that are only available on the host system. Core
212 file handling and @code{setjmp} handling are two common cases.
214 When you want to make @value{GDBN} work ``native'' on a particular machine, you
215 have to include all three kinds of information.
223 @value{GDBN} uses a number of debugging-specific algorithms. They are
224 often not very complicated, but get lost in the thicket of special
225 cases and real-world issues. This chapter describes the basic
226 algorithms and mentions some of the specific target definitions that
232 @cindex call stack frame
233 A frame is a construct that @value{GDBN} uses to keep track of calling
234 and called functions.
236 @findex create_new_frame
238 @code{FRAME_FP} in the machine description has no meaning to the
239 machine-independent part of @value{GDBN}, except that it is used when
240 setting up a new frame from scratch, as follows:
243 create_new_frame (read_register (FP_REGNUM), read_pc ()));
246 @cindex frame pointer register
247 Other than that, all the meaning imparted to @code{FP_REGNUM} is
248 imparted by the machine-dependent code. So, @code{FP_REGNUM} can have
249 any value that is convenient for the code that creates new frames.
250 (@code{create_new_frame} calls @code{INIT_EXTRA_FRAME_INFO} if it is
251 defined; that is where you should use the @code{FP_REGNUM} value, if
252 your frames are nonstandard.)
255 Given a @value{GDBN} frame, define @code{FRAME_CHAIN} to determine the
256 address of the calling function's frame. This will be used to create
257 a new @value{GDBN} frame struct, and then @code{INIT_EXTRA_FRAME_INFO}
258 and @code{INIT_FRAME_PC} will be called for the new frame.
260 @section Breakpoint Handling
263 In general, a breakpoint is a user-designated location in the program
264 where the user wants to regain control if program execution ever reaches
267 There are two main ways to implement breakpoints; either as ``hardware''
268 breakpoints or as ``software'' breakpoints.
270 @cindex hardware breakpoints
271 @cindex program counter
272 Hardware breakpoints are sometimes available as a builtin debugging
273 features with some chips. Typically these work by having dedicated
274 register into which the breakpoint address may be stored. If the PC
275 (shorthand for @dfn{program counter})
276 ever matches a value in a breakpoint registers, the CPU raises an
277 exception and reports it to @value{GDBN}.
279 Another possibility is when an emulator is in use; many emulators
280 include circuitry that watches the address lines coming out from the
281 processor, and force it to stop if the address matches a breakpoint's
284 A third possibility is that the target already has the ability to do
285 breakpoints somehow; for instance, a ROM monitor may do its own
286 software breakpoints. So although these are not literally ``hardware
287 breakpoints'', from @value{GDBN}'s point of view they work the same;
288 @value{GDBN} need not do nothing more than set the breakpoint and wait
289 for something to happen.
291 Since they depend on hardware resources, hardware breakpoints may be
292 limited in number; when the user asks for more, @value{GDBN} will
293 start trying to set software breakpoints. (On some architectures,
294 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
295 whether there's enough hardware resources to insert all the hardware
296 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
297 an error message only when the program being debugged is continued.)
299 @cindex software breakpoints
300 Software breakpoints require @value{GDBN} to do somewhat more work.
301 The basic theory is that @value{GDBN} will replace a program
302 instruction with a trap, illegal divide, or some other instruction
303 that will cause an exception, and then when it's encountered,
304 @value{GDBN} will take the exception and stop the program. When the
305 user says to continue, @value{GDBN} will restore the original
306 instruction, single-step, re-insert the trap, and continue on.
308 Since it literally overwrites the program being tested, the program area
309 must be writable, so this technique won't work on programs in ROM. It
310 can also distort the behavior of programs that examine themselves,
311 although such a situation would be highly unusual.
313 Also, the software breakpoint instruction should be the smallest size of
314 instruction, so it doesn't overwrite an instruction that might be a jump
315 target, and cause disaster when the program jumps into the middle of the
316 breakpoint instruction. (Strictly speaking, the breakpoint must be no
317 larger than the smallest interval between instructions that may be jump
318 targets; perhaps there is an architecture where only even-numbered
319 instructions may jumped to.) Note that it's possible for an instruction
320 set not to have any instructions usable for a software breakpoint,
321 although in practice only the ARC has failed to define such an
325 The basic definition of the software breakpoint is the macro
328 Basic breakpoint object handling is in @file{breakpoint.c}. However,
329 much of the interesting breakpoint action is in @file{infrun.c}.
331 @section Single Stepping
333 @section Signal Handling
335 @section Thread Handling
337 @section Inferior Function Calls
339 @section Longjmp Support
341 @cindex @code{longjmp} debugging
342 @value{GDBN} has support for figuring out that the target is doing a
343 @code{longjmp} and for stopping at the target of the jump, if we are
344 stepping. This is done with a few specialized internal breakpoints,
345 which are visible in the output of the @samp{maint info breakpoint}
348 @findex GET_LONGJMP_TARGET
349 To make this work, you need to define a macro called
350 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
351 structure and extract the longjmp target address. Since @code{jmp_buf}
352 is target specific, you will need to define it in the appropriate
353 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
354 @file{sparc-tdep.c} for examples of how to do this.
359 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
360 breakpoints}) which break when data is accessed rather than when some
361 instruction is executed. When you have data which changes without
362 your knowing what code does that, watchpoints are the silver bullet to
363 hunt down and kill such bugs.
365 @cindex hardware watchpoints
366 @cindex software watchpoints
367 Watchpoints can be either hardware-assisted or not; the latter type is
368 known as ``software watchpoints.'' @value{GDBN} always uses
369 hardware-assisted watchpoints if they are available, and falls back on
370 software watchpoints otherwise. Typical situations where @value{GDBN}
371 will use software watchpoints are:
375 The watched memory region is too large for the underlying hardware
376 watchpoint support. For example, each x86 debug register can watch up
377 to 4 bytes of memory, so trying to watch data structures whose size is
378 more than 16 bytes will cause @value{GDBN} to use software
382 The value of the expression to be watched depends on data held in
383 registers (as opposed to memory).
386 Too many different watchpoints requested. (On some architectures,
387 this situation is impossible to detect until the debugged program is
388 resumed.) Note that x86 debug registers are used both for hardware
389 breakpoints and for watchpoints, so setting too many hardware
390 breakpoints might cause watchpoint insertion to fail.
393 No hardware-assisted watchpoints provided by the target
397 Software watchpoints are very slow, since @value{GDBN} needs to
398 single-step the program being debugged and test the value of the
399 watched expression(s) after each instruction. The rest of this
400 section is mostly irrelevant for software watchpoints.
402 @value{GDBN} uses several macros and primitives to support hardware
406 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
407 @item TARGET_HAS_HARDWARE_WATCHPOINTS
408 If defined, the target supports hardware watchpoints.
410 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
411 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
412 Return the number of hardware watchpoints of type @var{type} that are
413 possible to be set. The value is positive if @var{count} watchpoints
414 of this type can be set, zero if setting watchpoints of this type is
415 not supported, and negative if @var{count} is more than the maximum
416 number of watchpoints of type @var{type} that can be set. @var{other}
417 is non-zero if other types of watchpoints are currently enabled (there
418 are architectures which cannot set watchpoints of different types at
421 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
422 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
423 Return non-zero if hardware watchpoints can be used to watch a region
424 whose address is @var{addr} and whose length in bytes is @var{len}.
426 @findex TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT
427 @item TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT (@var{size})
428 Return non-zero if hardware watchpoints can be used to watch a region
429 whose size is @var{size}. @value{GDBN} only uses this macro as a
430 fall-back, in case @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is not
433 @findex TARGET_DISABLE_HW_WATCHPOINTS
434 @item TARGET_DISABLE_HW_WATCHPOINTS (@var{pid})
435 Disables watchpoints in the process identified by @var{pid}. This is
436 used, e.g., on HP-UX which provides operations to disable and enable
437 the page-level memory protection that implements hardware watchpoints
440 @findex TARGET_ENABLE_HW_WATCHPOINTS
441 @item TARGET_ENABLE_HW_WATCHPOINTS (@var{pid})
442 Enables watchpoints in the process identified by @var{pid}. This is
443 used, e.g., on HP-UX which provides operations to disable and enable
444 the page-level memory protection that implements hardware watchpoints
447 @findex target_insert_watchpoint
448 @findex target_remove_watchpoint
449 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
450 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
451 Insert or remove a hardware watchpoint starting at @var{addr}, for
452 @var{len} bytes. @var{type} is the watchpoint type, one of the
453 possible values of the enumerated data type @code{target_hw_bp_type},
454 defined by @file{breakpoint.h} as follows:
457 enum target_hw_bp_type
459 hw_write = 0, /* Common (write) HW watchpoint */
460 hw_read = 1, /* Read HW watchpoint */
461 hw_access = 2, /* Access (read or write) HW watchpoint */
462 hw_execute = 3 /* Execute HW breakpoint */
467 These two macros should return 0 for success, non-zero for failure.
469 @cindex insert or remove hardware breakpoint
470 @findex target_remove_hw_breakpoint
471 @findex target_insert_hw_breakpoint
472 @item target_remove_hw_breakpoint (@var{addr}, @var{shadow})
473 @itemx target_insert_hw_breakpoint (@var{addr}, @var{shadow})
474 Insert or remove a hardware-assisted breakpoint at address @var{addr}.
475 Returns zero for success, non-zero for failure. @var{shadow} is the
476 real contents of the byte where the breakpoint has been inserted; it
477 is generally not valid when hardware breakpoints are used, but since
478 no other code touches these values, the implementations of the above
479 two macros can use them for their internal purposes.
481 @findex target_stopped_data_address
482 @item target_stopped_data_address ()
483 If the inferior has some watchpoint that triggered, return the address
484 associated with that watchpoint. Otherwise, return zero.
486 @findex DECR_PC_AFTER_HW_BREAK
487 @item DECR_PC_AFTER_HW_BREAK
488 If defined, @value{GDBN} decrements the program counter by the value
489 of @code{DECR_PC_AFTER_HW_BREAK} after a hardware break-point. This
490 overrides the value of @code{DECR_PC_AFTER_BREAK} when a breakpoint
491 that breaks is a hardware-assisted breakpoint.
493 @findex HAVE_STEPPABLE_WATCHPOINT
494 @item HAVE_STEPPABLE_WATCHPOINT
495 If defined to a non-zero value, it is not necessary to disable a
496 watchpoint to step over it.
498 @findex HAVE_NONSTEPPABLE_WATCHPOINT
499 @item HAVE_NONSTEPPABLE_WATCHPOINT
500 If defined to a non-zero value, @value{GDBN} should disable a
501 watchpoint to step the inferior over it.
503 @findex HAVE_CONTINUABLE_WATCHPOINT
504 @item HAVE_CONTINUABLE_WATCHPOINT
505 If defined to a non-zero value, it is possible to continue the
506 inferior after a watchpoint has been hit.
508 @findex CANNOT_STEP_HW_WATCHPOINTS
509 @item CANNOT_STEP_HW_WATCHPOINTS
510 If this is defined to a non-zero value, @value{GDBN} will remove all
511 watchpoints before stepping the inferior.
513 @findex STOPPED_BY_WATCHPOINT
514 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
515 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
516 the type @code{struct target_waitstatus}, defined by @file{target.h}.
519 @subsection x86 Watchpoints
520 @cindex x86 debug registers
521 @cindex watchpoints, on x86
523 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
524 registers designed to facilitate debugging. @value{GDBN} provides a
525 generic library of functions that x86-based ports can use to implement
526 support for watchpoints and hardware-assisted breakpoints. This
527 subsection documents the x86 watchpoint facilities in @value{GDBN}.
529 To use the generic x86 watchpoint support, a port should do the
533 @findex I386_USE_GENERIC_WATCHPOINTS
535 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
536 target-dependent headers.
539 Include the @file{config/i386/nm-i386.h} header file @emph{after}
540 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
543 Add @file{i386-nat.o} to the value of the Make variable
544 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
545 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
548 Provide implementations for the @code{I386_DR_LOW_*} macros described
549 below. Typically, each macro should call a target-specific function
550 which does the real work.
553 The x86 watchpoint support works by maintaining mirror images of the
554 debug registers. Values are copied between the mirror images and the
555 real debug registers via a set of macros which each target needs to
559 @findex I386_DR_LOW_SET_CONTROL
560 @item I386_DR_LOW_SET_CONTROL (@var{val})
561 Set the Debug Control (DR7) register to the value @var{val}.
563 @findex I386_DR_LOW_SET_ADDR
564 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
565 Put the address @var{addr} into the debug register number @var{idx}.
567 @findex I386_DR_LOW_RESET_ADDR
568 @item I386_DR_LOW_RESET_ADDR (@var{idx})
569 Reset (i.e.@: zero out) the address stored in the debug register
572 @findex I386_DR_LOW_GET_STATUS
573 @item I386_DR_LOW_GET_STATUS
574 Return the value of the Debug Status (DR6) register. This value is
575 used immediately after it is returned by
576 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
580 For each one of the 4 debug registers (whose indices are from 0 to 3)
581 that store addresses, a reference count is maintained by @value{GDBN},
582 to allow sharing of debug registers by several watchpoints. This
583 allows users to define several watchpoints that watch the same
584 expression, but with different conditions and/or commands, without
585 wasting debug registers which are in short supply. @value{GDBN}
586 maintains the reference counts internally, targets don't have to do
587 anything to use this feature.
589 The x86 debug registers can each watch a region that is 1, 2, or 4
590 bytes long. The ia32 architecture requires that each watched region
591 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
592 region on 4-byte boundary. However, the x86 watchpoint support in
593 @value{GDBN} can watch unaligned regions and regions larger than 4
594 bytes (up to 16 bytes) by allocating several debug registers to watch
595 a single region. This allocation of several registers per a watched
596 region is also done automatically without target code intervention.
598 The generic x86 watchpoint support provides the following API for the
599 @value{GDBN}'s application code:
602 @findex i386_region_ok_for_watchpoint
603 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
604 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
605 this function. It counts the number of debug registers required to
606 watch a given region, and returns a non-zero value if that number is
607 less than 4, the number of debug registers available to x86
610 @findex i386_stopped_data_address
611 @item i386_stopped_data_address (void)
612 The macros @code{STOPPED_BY_WATCHPOINT} and
613 @code{target_stopped_data_address} are set to call this function. The
614 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
615 function examines the breakpoint condition bits in the DR6 Debug
616 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
617 macro, and returns the address associated with the first bit that is
620 @findex i386_insert_watchpoint
621 @findex i386_remove_watchpoint
622 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
623 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
624 Insert or remove a watchpoint. The macros
625 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
626 are set to call these functions. @code{i386_insert_watchpoint} first
627 looks for a debug register which is already set to watch the same
628 region for the same access types; if found, it just increments the
629 reference count of that debug register, thus implementing debug
630 register sharing between watchpoints. If no such register is found,
631 the function looks for a vacant debug register, sets its mirrored
632 value to @var{addr}, sets the mirrored value of DR7 Debug Control
633 register as appropriate for the @var{len} and @var{type} parameters,
634 and then passes the new values of the debug register and DR7 to the
635 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
636 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
637 required to cover the given region, the above process is repeated for
640 @code{i386_remove_watchpoint} does the opposite: it resets the address
641 in the mirrored value of the debug register and its read/write and
642 length bits in the mirrored value of DR7, then passes these new
643 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
644 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
645 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
646 decrements the reference count, and only calls
647 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
648 the count goes to zero.
650 @findex i386_insert_hw_breakpoint
651 @findex i386_remove_hw_breakpoint
652 @item i386_insert_hw_breakpoint (@var{addr}, @var{shadow}
653 @itemx i386_remove_hw_breakpoint (@var{addr}, @var{shadow})
654 These functions insert and remove hardware-assisted breakpoints. The
655 macros @code{target_insert_hw_breakpoint} and
656 @code{target_remove_hw_breakpoint} are set to call these functions.
657 These functions work like @code{i386_insert_watchpoint} and
658 @code{i386_remove_watchpoint}, respectively, except that they set up
659 the debug registers to watch instruction execution, and each
660 hardware-assisted breakpoint always requires exactly one debug
663 @findex i386_stopped_by_hwbp
664 @item i386_stopped_by_hwbp (void)
665 This function returns non-zero if the inferior has some watchpoint or
666 hardware breakpoint that triggered. It works like
667 @code{i386_stopped_data_address}, except that it doesn't return the
668 address whose watchpoint triggered.
670 @findex i386_cleanup_dregs
671 @item i386_cleanup_dregs (void)
672 This function clears all the reference counts, addresses, and control
673 bits in the mirror images of the debug registers. It doesn't affect
674 the actual debug registers in the inferior process.
681 x86 processors support setting watchpoints on I/O reads or writes.
682 However, since no target supports this (as of March 2001), and since
683 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
684 watchpoints, this feature is not yet available to @value{GDBN} running
688 x86 processors can enable watchpoints locally, for the current task
689 only, or globally, for all the tasks. For each debug register,
690 there's a bit in the DR7 Debug Control register that determines
691 whether the associated address is watched locally or globally. The
692 current implementation of x86 watchpoint support in @value{GDBN}
693 always sets watchpoints to be locally enabled, since global
694 watchpoints might interfere with the underlying OS and are probably
695 unavailable in many platforms.
700 @chapter User Interface
702 @value{GDBN} has several user interfaces. Although the command-line interface
703 is the most common and most familiar, there are others.
705 @section Command Interpreter
707 @cindex command interpreter
709 The command interpreter in @value{GDBN} is fairly simple. It is designed to
710 allow for the set of commands to be augmented dynamically, and also
711 has a recursive subcommand capability, where the first argument to
712 a command may itself direct a lookup on a different command list.
714 For instance, the @samp{set} command just starts a lookup on the
715 @code{setlist} command list, while @samp{set thread} recurses
716 to the @code{set_thread_cmd_list}.
720 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
721 the main command list, and should be used for those commands. The usual
722 place to add commands is in the @code{_initialize_@var{xyz}} routines at
723 the ends of most source files.
725 @cindex deprecating commands
726 @findex deprecate_cmd
727 Before removing commands from the command set it is a good idea to
728 deprecate them for some time. Use @code{deprecate_cmd} on commands or
729 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
730 @code{struct cmd_list_element} as it's first argument. You can use the
731 return value from @code{add_com} or @code{add_cmd} to deprecate the
732 command immediately after it is created.
734 The first time a command is used the user will be warned and offered a
735 replacement (if one exists). Note that the replacement string passed to
736 @code{deprecate_cmd} should be the full name of the command, i.e. the
737 entire string the user should type at the command line.
739 @section UI-Independent Output---the @code{ui_out} Functions
740 @c This section is based on the documentation written by Fernando
741 @c Nasser <fnasser@redhat.com>.
743 @cindex @code{ui_out} functions
744 The @code{ui_out} functions present an abstraction level for the
745 @value{GDBN} output code. They hide the specifics of different user
746 interfaces supported by @value{GDBN}, and thus free the programmer
747 from the need to write several versions of the same code, one each for
748 every UI, to produce output.
750 @subsection Overview and Terminology
752 In general, execution of each @value{GDBN} command produces some sort
753 of output, and can even generate an input request.
755 Output can be generated for the following purposes:
759 to display a @emph{result} of an operation;
762 to convey @emph{info} or produce side-effects of a requested
766 to provide a @emph{notification} of an asynchronous event (including
767 progress indication of a prolonged asynchronous operation);
770 to display @emph{error messages} (including warnings);
773 to show @emph{debug data};
776 to @emph{query} or prompt a user for input (a special case).
780 This section mainly concentrates on how to build result output,
781 although some of it also applies to other kinds of output.
783 Generation of output that displays the results of an operation
784 involves one or more of the following:
788 output of the actual data
791 formatting the output as appropriate for console output, to make it
792 easily readable by humans
795 machine oriented formatting--a more terse formatting to allow for easy
796 parsing by programs which read @value{GDBN}'s output
799 annotation, whose purpose is to help legacy GUIs to identify interesting
803 The @code{ui_out} routines take care of the first three aspects.
804 Annotations are provided by separate annotation routines. Note that use
805 of annotations for an interface between a GUI and @value{GDBN} is
808 Output can be in the form of a single item, which we call a @dfn{field};
809 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
810 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
811 header and a body. In a BNF-like form:
814 @item <table> @expansion{}
815 @code{<header> <body>}
816 @item <header> @expansion{}
817 @code{@{ <column> @}}
818 @item <column> @expansion{}
819 @code{<width> <alignment> <title>}
820 @item <body> @expansion{}
825 @subsection General Conventions
827 Most @code{ui_out} routines are of type @code{void}, the exceptions are
828 @code{ui_out_stream_new} (which returns a pointer to the newly created
829 object) and the @code{make_cleanup} routines.
831 The first parameter is always the @code{ui_out} vector object, a pointer
832 to a @code{struct ui_out}.
834 The @var{format} parameter is like in @code{printf} family of functions.
835 When it is present, there must also be a variable list of arguments
836 sufficient used to satisfy the @code{%} specifiers in the supplied
839 When a character string argument is not used in a @code{ui_out} function
840 call, a @code{NULL} pointer has to be supplied instead.
843 @subsection Table, Tuple and List Functions
845 @cindex list output functions
846 @cindex table output functions
847 @cindex tuple output functions
848 This section introduces @code{ui_out} routines for building lists,
849 tuples and tables. The routines to output the actual data items
850 (fields) are presented in the next section.
852 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
853 containing information about an object; a @dfn{list} is a sequence of
854 fields where each field describes an identical object.
856 Use the @dfn{table} functions when your output consists of a list of
857 rows (tuples) and the console output should include a heading. Use this
858 even when you are listing just one object but you still want the header.
860 @cindex nesting level in @code{ui_out} functions
861 Tables can not be nested. Tuples and lists can be nested up to a
862 maximum of five levels.
864 The overall structure of the table output code is something like this:
879 Here is the description of table-, tuple- and list-related @code{ui_out}
882 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
883 The function @code{ui_out_table_begin} marks the beginning of the output
884 of a table. It should always be called before any other @code{ui_out}
885 function for a given table. @var{nbrofcols} is the number of columns in
886 the table. @var{nr_rows} is the number of rows in the table.
887 @var{tblid} is an optional string identifying the table. The string
888 pointed to by @var{tblid} is copied by the implementation of
889 @code{ui_out_table_begin}, so the application can free the string if it
892 The companion function @code{ui_out_table_end}, described below, marks
893 the end of the table's output.
896 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
897 @code{ui_out_table_header} provides the header information for a single
898 table column. You call this function several times, one each for every
899 column of the table, after @code{ui_out_table_begin}, but before
900 @code{ui_out_table_body}.
902 The value of @var{width} gives the column width in characters. The
903 value of @var{alignment} is one of @code{left}, @code{center}, and
904 @code{right}, and it specifies how to align the header: left-justify,
905 center, or right-justify it. @var{colhdr} points to a string that
906 specifies the column header; the implementation copies that string, so
907 column header strings in @code{malloc}ed storage can be freed after the
911 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
912 This function delimits the table header from the table body.
915 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
916 This function signals the end of a table's output. It should be called
917 after the table body has been produced by the list and field output
920 There should be exactly one call to @code{ui_out_table_end} for each
921 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
922 will signal an internal error.
925 The output of the tuples that represent the table rows must follow the
926 call to @code{ui_out_table_body} and precede the call to
927 @code{ui_out_table_end}. You build a tuple by calling
928 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
929 calls to functions which actually output fields between them.
931 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
932 This function marks the beginning of a tuple output. @var{id} points
933 to an optional string that identifies the tuple; it is copied by the
934 implementation, and so strings in @code{malloc}ed storage can be freed
938 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
939 This function signals an end of a tuple output. There should be exactly
940 one call to @code{ui_out_tuple_end} for each call to
941 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
945 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
946 This function first opens the tuple and then establishes a cleanup
947 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
948 and correct implementation of the non-portable@footnote{The function
949 cast is not portable ISO-C.} code sequence:
951 struct cleanup *old_cleanup;
952 ui_out_tuple_begin (uiout, "...");
953 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
958 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
959 This function marks the beginning of a list output. @var{id} points to
960 an optional string that identifies the list; it is copied by the
961 implementation, and so strings in @code{malloc}ed storage can be freed
965 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
966 This function signals an end of a list output. There should be exactly
967 one call to @code{ui_out_list_end} for each call to
968 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
972 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
973 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
974 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
975 that will close the list.list.
978 @subsection Item Output Functions
980 @cindex item output functions
981 @cindex field output functions
983 The functions described below produce output for the actual data
984 items, or fields, which contain information about the object.
986 Choose the appropriate function accordingly to your particular needs.
988 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
989 This is the most general output function. It produces the
990 representation of the data in the variable-length argument list
991 according to formatting specifications in @var{format}, a
992 @code{printf}-like format string. The optional argument @var{fldname}
993 supplies the name of the field. The data items themselves are
994 supplied as additional arguments after @var{format}.
996 This generic function should be used only when it is not possible to
997 use one of the specialized versions (see below).
1000 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1001 This function outputs a value of an @code{int} variable. It uses the
1002 @code{"%d"} output conversion specification. @var{fldname} specifies
1003 the name of the field.
1006 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1007 This function outputs an address.
1010 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1011 This function outputs a string using the @code{"%s"} conversion
1015 Sometimes, there's a need to compose your output piece by piece using
1016 functions that operate on a stream, such as @code{value_print} or
1017 @code{fprintf_symbol_filtered}. These functions accept an argument of
1018 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1019 used to store the data stream used for the output. When you use one
1020 of these functions, you need a way to pass their results stored in a
1021 @code{ui_file} object to the @code{ui_out} functions. To this end,
1022 you first create a @code{ui_stream} object by calling
1023 @code{ui_out_stream_new}, pass the @code{stream} member of that
1024 @code{ui_stream} object to @code{value_print} and similar functions,
1025 and finally call @code{ui_out_field_stream} to output the field you
1026 constructed. When the @code{ui_stream} object is no longer needed,
1027 you should destroy it and free its memory by calling
1028 @code{ui_out_stream_delete}.
1030 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1031 This function creates a new @code{ui_stream} object which uses the
1032 same output methods as the @code{ui_out} object whose pointer is
1033 passed in @var{uiout}. It returns a pointer to the newly created
1034 @code{ui_stream} object.
1037 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1038 This functions destroys a @code{ui_stream} object specified by
1042 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1043 This function consumes all the data accumulated in
1044 @code{streambuf->stream} and outputs it like
1045 @code{ui_out_field_string} does. After a call to
1046 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1047 the stream is still valid and may be used for producing more fields.
1050 @strong{Important:} If there is any chance that your code could bail
1051 out before completing output generation and reaching the point where
1052 @code{ui_out_stream_delete} is called, it is necessary to set up a
1053 cleanup, to avoid leaking memory and other resources. Here's a
1054 skeleton code to do that:
1057 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1058 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1063 If the function already has the old cleanup chain set (for other kinds
1064 of cleanups), you just have to add your cleanup to it:
1067 mybuf = ui_out_stream_new (uiout);
1068 make_cleanup (ui_out_stream_delete, mybuf);
1071 Note that with cleanups in place, you should not call
1072 @code{ui_out_stream_delete} directly, or you would attempt to free the
1075 @subsection Utility Output Functions
1077 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1078 This function skips a field in a table. Use it if you have to leave
1079 an empty field without disrupting the table alignment. The argument
1080 @var{fldname} specifies a name for the (missing) filed.
1083 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1084 This function outputs the text in @var{string} in a way that makes it
1085 easy to be read by humans. For example, the console implementation of
1086 this method filters the text through a built-in pager, to prevent it
1087 from scrolling off the visible portion of the screen.
1089 Use this function for printing relatively long chunks of text around
1090 the actual field data: the text it produces is not aligned according
1091 to the table's format. Use @code{ui_out_field_string} to output a
1092 string field, and use @code{ui_out_message}, described below, to
1093 output short messages.
1096 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1097 This function outputs @var{nspaces} spaces. It is handy to align the
1098 text produced by @code{ui_out_text} with the rest of the table or
1102 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1103 This function produces a formatted message, provided that the current
1104 verbosity level is at least as large as given by @var{verbosity}. The
1105 current verbosity level is specified by the user with the @samp{set
1106 verbositylevel} command.@footnote{As of this writing (April 2001),
1107 setting verbosity level is not yet implemented, and is always returned
1108 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1109 argument more than zero will cause the message to never be printed.}
1112 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1113 This function gives the console output filter (a paging filter) a hint
1114 of where to break lines which are too long. Ignored for all other
1115 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1116 be printed to indent the wrapped text on the next line; it must remain
1117 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1118 explicit newline is produced by one of the other functions. If
1119 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1122 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1123 This function flushes whatever output has been accumulated so far, if
1124 the UI buffers output.
1128 @subsection Examples of Use of @code{ui_out} functions
1130 @cindex using @code{ui_out} functions
1131 @cindex @code{ui_out} functions, usage examples
1132 This section gives some practical examples of using the @code{ui_out}
1133 functions to generalize the old console-oriented code in
1134 @value{GDBN}. The examples all come from functions defined on the
1135 @file{breakpoints.c} file.
1137 This example, from the @code{breakpoint_1} function, shows how to
1140 The original code was:
1143 if (!found_a_breakpoint++)
1145 annotate_breakpoints_headers ();
1148 printf_filtered ("Num ");
1150 printf_filtered ("Type ");
1152 printf_filtered ("Disp ");
1154 printf_filtered ("Enb ");
1158 printf_filtered ("Address ");
1161 printf_filtered ("What\n");
1163 annotate_breakpoints_table ();
1167 Here's the new version:
1170 nr_printable_breakpoints = @dots{};
1173 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1175 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1177 if (nr_printable_breakpoints > 0)
1178 annotate_breakpoints_headers ();
1179 if (nr_printable_breakpoints > 0)
1181 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1182 if (nr_printable_breakpoints > 0)
1184 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1185 if (nr_printable_breakpoints > 0)
1187 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1188 if (nr_printable_breakpoints > 0)
1190 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1193 if (nr_printable_breakpoints > 0)
1195 if (TARGET_ADDR_BIT <= 32)
1196 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1198 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1200 if (nr_printable_breakpoints > 0)
1202 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1203 ui_out_table_body (uiout);
1204 if (nr_printable_breakpoints > 0)
1205 annotate_breakpoints_table ();
1208 This example, from the @code{print_one_breakpoint} function, shows how
1209 to produce the actual data for the table whose structure was defined
1210 in the above example. The original code was:
1215 printf_filtered ("%-3d ", b->number);
1217 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1218 || ((int) b->type != bptypes[(int) b->type].type))
1219 internal_error ("bptypes table does not describe type #%d.",
1221 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1223 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1225 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1229 This is the new version:
1233 ui_out_tuple_begin (uiout, "bkpt");
1235 ui_out_field_int (uiout, "number", b->number);
1237 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1238 || ((int) b->type != bptypes[(int) b->type].type))
1239 internal_error ("bptypes table does not describe type #%d.",
1241 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1243 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1245 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1249 This example, also from @code{print_one_breakpoint}, shows how to
1250 produce a complicated output field using the @code{print_expression}
1251 functions which requires a stream to be passed. It also shows how to
1252 automate stream destruction with cleanups. The original code was:
1256 print_expression (b->exp, gdb_stdout);
1262 struct ui_stream *stb = ui_out_stream_new (uiout);
1263 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1266 print_expression (b->exp, stb->stream);
1267 ui_out_field_stream (uiout, "what", local_stream);
1270 This example, also from @code{print_one_breakpoint}, shows how to use
1271 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1276 if (b->dll_pathname == NULL)
1277 printf_filtered ("<any library> ");
1279 printf_filtered ("library \"%s\" ", b->dll_pathname);
1286 if (b->dll_pathname == NULL)
1288 ui_out_field_string (uiout, "what", "<any library>");
1289 ui_out_spaces (uiout, 1);
1293 ui_out_text (uiout, "library \"");
1294 ui_out_field_string (uiout, "what", b->dll_pathname);
1295 ui_out_text (uiout, "\" ");
1299 The following example from @code{print_one_breakpoint} shows how to
1300 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1305 if (b->forked_inferior_pid != 0)
1306 printf_filtered ("process %d ", b->forked_inferior_pid);
1313 if (b->forked_inferior_pid != 0)
1315 ui_out_text (uiout, "process ");
1316 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1317 ui_out_spaces (uiout, 1);
1321 Here's an example of using @code{ui_out_field_string}. The original
1326 if (b->exec_pathname != NULL)
1327 printf_filtered ("program \"%s\" ", b->exec_pathname);
1334 if (b->exec_pathname != NULL)
1336 ui_out_text (uiout, "program \"");
1337 ui_out_field_string (uiout, "what", b->exec_pathname);
1338 ui_out_text (uiout, "\" ");
1342 Finally, here's an example of printing an address. The original code:
1346 printf_filtered ("%s ",
1347 local_hex_string_custom ((unsigned long) b->address, "08l"));
1354 ui_out_field_core_addr (uiout, "Address", b->address);
1358 @section Console Printing
1367 @cindex @code{libgdb}
1368 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1369 to provide an API to @value{GDBN}'s functionality.
1372 @cindex @code{libgdb}
1373 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1374 better able to support graphical and other environments.
1376 Since @code{libgdb} development is on-going, its architecture is still
1377 evolving. The following components have so far been identified:
1381 Observer - @file{gdb-events.h}.
1383 Builder - @file{ui-out.h}
1385 Event Loop - @file{event-loop.h}
1387 Library - @file{gdb.h}
1390 The model that ties these components together is described below.
1392 @section The @code{libgdb} Model
1394 A client of @code{libgdb} interacts with the library in two ways.
1398 As an observer (using @file{gdb-events}) receiving notifications from
1399 @code{libgdb} of any internal state changes (break point changes, run
1402 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1403 obtain various status values from @value{GDBN}.
1406 Since @code{libgdb} could have multiple clients (e.g. a GUI supporting
1407 the existing @value{GDBN} CLI), those clients must co-operate when
1408 controlling @code{libgdb}. In particular, a client must ensure that
1409 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1410 before responding to a @file{gdb-event} by making a query.
1412 @section CLI support
1414 At present @value{GDBN}'s CLI is very much entangled in with the core of
1415 @code{libgdb}. Consequently, a client wishing to include the CLI in
1416 their interface needs to carefully co-ordinate its own and the CLI's
1419 It is suggested that the client set @code{libgdb} up to be bi-modal
1420 (alternate between CLI and client query modes). The notes below sketch
1425 The client registers itself as an observer of @code{libgdb}.
1427 The client create and install @code{cli-out} builder using its own
1428 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1429 @code{gdb_stdout} streams.
1431 The client creates a separate custom @code{ui-out} builder that is only
1432 used while making direct queries to @code{libgdb}.
1435 When the client receives input intended for the CLI, it simply passes it
1436 along. Since the @code{cli-out} builder is installed by default, all
1437 the CLI output in response to that command is routed (pronounced rooted)
1438 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1439 At the same time, the client is kept abreast of internal changes by
1440 virtue of being a @code{libgdb} observer.
1442 The only restriction on the client is that it must wait until
1443 @code{libgdb} becomes idle before initiating any queries (using the
1444 client's custom builder).
1446 @section @code{libgdb} components
1448 @subheading Observer - @file{gdb-events.h}
1449 @file{gdb-events} provides the client with a very raw mechanism that can
1450 be used to implement an observer. At present it only allows for one
1451 observer and that observer must, internally, handle the need to delay
1452 the processing of any event notifications until after @code{libgdb} has
1453 finished the current command.
1455 @subheading Builder - @file{ui-out.h}
1456 @file{ui-out} provides the infrastructure necessary for a client to
1457 create a builder. That builder is then passed down to @code{libgdb}
1458 when doing any queries.
1460 @subheading Event Loop - @file{event-loop.h}
1461 @c There could be an entire section on the event-loop
1462 @file{event-loop}, currently non-re-entrant, provides a simple event
1463 loop. A client would need to either plug its self into this loop or,
1464 implement a new event-loop that GDB would use.
1466 The event-loop will eventually be made re-entrant. This is so that
1467 @value{GDB} can better handle the problem of some commands blocking
1468 instead of returning.
1470 @subheading Library - @file{gdb.h}
1471 @file{libgdb} is the most obvious component of this system. It provides
1472 the query interface. Each function is parameterized by a @code{ui-out}
1473 builder. The result of the query is constructed using that builder
1474 before the query function returns.
1476 @node Symbol Handling
1478 @chapter Symbol Handling
1480 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1481 functions, and types.
1483 @section Symbol Reading
1485 @cindex symbol reading
1486 @cindex reading of symbols
1487 @cindex symbol files
1488 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1489 file is the file containing the program which @value{GDBN} is
1490 debugging. @value{GDBN} can be directed to use a different file for
1491 symbols (with the @samp{symbol-file} command), and it can also read
1492 more symbols via the @samp{add-file} and @samp{load} commands, or while
1493 reading symbols from shared libraries.
1495 @findex find_sym_fns
1496 Symbol files are initially opened by code in @file{symfile.c} using
1497 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1498 of the file by examining its header. @code{find_sym_fns} then uses
1499 this identification to locate a set of symbol-reading functions.
1501 @findex add_symtab_fns
1502 @cindex @code{sym_fns} structure
1503 @cindex adding a symbol-reading module
1504 Symbol-reading modules identify themselves to @value{GDBN} by calling
1505 @code{add_symtab_fns} during their module initialization. The argument
1506 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1507 name (or name prefix) of the symbol format, the length of the prefix,
1508 and pointers to four functions. These functions are called at various
1509 times to process symbol files whose identification matches the specified
1512 The functions supplied by each module are:
1515 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1517 @cindex secondary symbol file
1518 Called from @code{symbol_file_add} when we are about to read a new
1519 symbol file. This function should clean up any internal state (possibly
1520 resulting from half-read previous files, for example) and prepare to
1521 read a new symbol file. Note that the symbol file which we are reading
1522 might be a new ``main'' symbol file, or might be a secondary symbol file
1523 whose symbols are being added to the existing symbol table.
1525 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1526 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1527 new symbol file being read. Its @code{private} field has been zeroed,
1528 and can be modified as desired. Typically, a struct of private
1529 information will be @code{malloc}'d, and a pointer to it will be placed
1530 in the @code{private} field.
1532 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1533 @code{error} if it detects an unavoidable problem.
1535 @item @var{xyz}_new_init()
1537 Called from @code{symbol_file_add} when discarding existing symbols.
1538 This function needs only handle the symbol-reading module's internal
1539 state; the symbol table data structures visible to the rest of
1540 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1541 arguments and no result. It may be called after
1542 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1543 may be called alone if all symbols are simply being discarded.
1545 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1547 Called from @code{symbol_file_add} to actually read the symbols from a
1548 symbol-file into a set of psymtabs or symtabs.
1550 @code{sf} points to the @code{struct sym_fns} originally passed to
1551 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1552 the offset between the file's specified start address and its true
1553 address in memory. @code{mainline} is 1 if this is the main symbol
1554 table being read, and 0 if a secondary symbol file (e.g. shared library
1555 or dynamically loaded file) is being read.@refill
1558 In addition, if a symbol-reading module creates psymtabs when
1559 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1560 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1561 from any point in the @value{GDBN} symbol-handling code.
1564 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1566 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1567 the psymtab has not already been read in and had its @code{pst->symtab}
1568 pointer set. The argument is the psymtab to be fleshed-out into a
1569 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1570 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1571 zero if there were no symbols in that part of the symbol file.
1574 @section Partial Symbol Tables
1576 @value{GDBN} has three types of symbol tables:
1579 @cindex full symbol table
1582 Full symbol tables (@dfn{symtabs}). These contain the main
1583 information about symbols and addresses.
1587 Partial symbol tables (@dfn{psymtabs}). These contain enough
1588 information to know when to read the corresponding part of the full
1591 @cindex minimal symbol table
1594 Minimal symbol tables (@dfn{msymtabs}). These contain information
1595 gleaned from non-debugging symbols.
1598 @cindex partial symbol table
1599 This section describes partial symbol tables.
1601 A psymtab is constructed by doing a very quick pass over an executable
1602 file's debugging information. Small amounts of information are
1603 extracted---enough to identify which parts of the symbol table will
1604 need to be re-read and fully digested later, when the user needs the
1605 information. The speed of this pass causes @value{GDBN} to start up very
1606 quickly. Later, as the detailed rereading occurs, it occurs in small
1607 pieces, at various times, and the delay therefrom is mostly invisible to
1609 @c (@xref{Symbol Reading}.)
1611 The symbols that show up in a file's psymtab should be, roughly, those
1612 visible to the debugger's user when the program is not running code from
1613 that file. These include external symbols and types, static symbols and
1614 types, and @code{enum} values declared at file scope.
1616 The psymtab also contains the range of instruction addresses that the
1617 full symbol table would represent.
1619 @cindex finding a symbol
1620 @cindex symbol lookup
1621 The idea is that there are only two ways for the user (or much of the
1622 code in the debugger) to reference a symbol:
1625 @findex find_pc_function
1626 @findex find_pc_line
1628 By its address (e.g. execution stops at some address which is inside a
1629 function in this file). The address will be noticed to be in the
1630 range of this psymtab, and the full symtab will be read in.
1631 @code{find_pc_function}, @code{find_pc_line}, and other
1632 @code{find_pc_@dots{}} functions handle this.
1634 @cindex lookup_symbol
1637 (e.g. the user asks to print a variable, or set a breakpoint on a
1638 function). Global names and file-scope names will be found in the
1639 psymtab, which will cause the symtab to be pulled in. Local names will
1640 have to be qualified by a global name, or a file-scope name, in which
1641 case we will have already read in the symtab as we evaluated the
1642 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1643 local scope, in which case the first case applies. @code{lookup_symbol}
1644 does most of the work here.
1647 The only reason that psymtabs exist is to cause a symtab to be read in
1648 at the right moment. Any symbol that can be elided from a psymtab,
1649 while still causing that to happen, should not appear in it. Since
1650 psymtabs don't have the idea of scope, you can't put local symbols in
1651 them anyway. Psymtabs don't have the idea of the type of a symbol,
1652 either, so types need not appear, unless they will be referenced by
1655 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1656 been read, and another way if the corresponding symtab has been read
1657 in. Such bugs are typically caused by a psymtab that does not contain
1658 all the visible symbols, or which has the wrong instruction address
1661 The psymtab for a particular section of a symbol file (objfile) could be
1662 thrown away after the symtab has been read in. The symtab should always
1663 be searched before the psymtab, so the psymtab will never be used (in a
1664 bug-free environment). Currently, psymtabs are allocated on an obstack,
1665 and all the psymbols themselves are allocated in a pair of large arrays
1666 on an obstack, so there is little to be gained by trying to free them
1667 unless you want to do a lot more work.
1671 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
1673 @cindex fundamental types
1674 These are the fundamental types that @value{GDBN} uses internally. Fundamental
1675 types from the various debugging formats (stabs, ELF, etc) are mapped
1676 into one of these. They are basically a union of all fundamental types
1677 that @value{GDBN} knows about for all the languages that @value{GDBN}
1680 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
1683 Each time @value{GDBN} builds an internal type, it marks it with one
1684 of these types. The type may be a fundamental type, such as
1685 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
1686 which is a pointer to another type. Typically, several @code{FT_*}
1687 types map to one @code{TYPE_CODE_*} type, and are distinguished by
1688 other members of the type struct, such as whether the type is signed
1689 or unsigned, and how many bits it uses.
1691 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
1693 These are instances of type structs that roughly correspond to
1694 fundamental types and are created as global types for @value{GDBN} to
1695 use for various ugly historical reasons. We eventually want to
1696 eliminate these. Note for example that @code{builtin_type_int}
1697 initialized in @file{gdbtypes.c} is basically the same as a
1698 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
1699 an @code{FT_INTEGER} fundamental type. The difference is that the
1700 @code{builtin_type} is not associated with any particular objfile, and
1701 only one instance exists, while @file{c-lang.c} builds as many
1702 @code{TYPE_CODE_INT} types as needed, with each one associated with
1703 some particular objfile.
1705 @section Object File Formats
1706 @cindex object file formats
1710 @cindex @code{a.out} format
1711 The @code{a.out} format is the original file format for Unix. It
1712 consists of three sections: @code{text}, @code{data}, and @code{bss},
1713 which are for program code, initialized data, and uninitialized data,
1716 The @code{a.out} format is so simple that it doesn't have any reserved
1717 place for debugging information. (Hey, the original Unix hackers used
1718 @samp{adb}, which is a machine-language debugger!) The only debugging
1719 format for @code{a.out} is stabs, which is encoded as a set of normal
1720 symbols with distinctive attributes.
1722 The basic @code{a.out} reader is in @file{dbxread.c}.
1727 The COFF format was introduced with System V Release 3 (SVR3) Unix.
1728 COFF files may have multiple sections, each prefixed by a header. The
1729 number of sections is limited.
1731 The COFF specification includes support for debugging. Although this
1732 was a step forward, the debugging information was woefully limited. For
1733 instance, it was not possible to represent code that came from an
1736 The COFF reader is in @file{coffread.c}.
1740 @cindex ECOFF format
1741 ECOFF is an extended COFF originally introduced for Mips and Alpha
1744 The basic ECOFF reader is in @file{mipsread.c}.
1748 @cindex XCOFF format
1749 The IBM RS/6000 running AIX uses an object file format called XCOFF.
1750 The COFF sections, symbols, and line numbers are used, but debugging
1751 symbols are @code{dbx}-style stabs whose strings are located in the
1752 @code{.debug} section (rather than the string table). For more
1753 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
1755 The shared library scheme has a clean interface for figuring out what
1756 shared libraries are in use, but the catch is that everything which
1757 refers to addresses (symbol tables and breakpoints at least) needs to be
1758 relocated for both shared libraries and the main executable. At least
1759 using the standard mechanism this can only be done once the program has
1760 been run (or the core file has been read).
1764 @cindex PE-COFF format
1765 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
1766 executables. PE is basically COFF with additional headers.
1768 While BFD includes special PE support, @value{GDBN} needs only the basic
1774 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
1775 to COFF in being organized into a number of sections, but it removes
1776 many of COFF's limitations.
1778 The basic ELF reader is in @file{elfread.c}.
1783 SOM is HP's object file and debug format (not to be confused with IBM's
1784 SOM, which is a cross-language ABI).
1786 The SOM reader is in @file{hpread.c}.
1788 @subsection Other File Formats
1790 @cindex Netware Loadable Module format
1791 Other file formats that have been supported by @value{GDBN} include Netware
1792 Loadable Modules (@file{nlmread.c}).
1794 @section Debugging File Formats
1796 This section describes characteristics of debugging information that
1797 are independent of the object file format.
1801 @cindex stabs debugging info
1802 @code{stabs} started out as special symbols within the @code{a.out}
1803 format. Since then, it has been encapsulated into other file
1804 formats, such as COFF and ELF.
1806 While @file{dbxread.c} does some of the basic stab processing,
1807 including for encapsulated versions, @file{stabsread.c} does
1812 @cindex COFF debugging info
1813 The basic COFF definition includes debugging information. The level
1814 of support is minimal and non-extensible, and is not often used.
1816 @subsection Mips debug (Third Eye)
1818 @cindex ECOFF debugging info
1819 ECOFF includes a definition of a special debug format.
1821 The file @file{mdebugread.c} implements reading for this format.
1825 @cindex DWARF 1 debugging info
1826 DWARF 1 is a debugging format that was originally designed to be
1827 used with ELF in SVR4 systems.
1833 @c If defined, these are the producer strings in a DWARF 1 file. All of
1834 @c these have reasonable defaults already.
1836 The DWARF 1 reader is in @file{dwarfread.c}.
1840 @cindex DWARF 2 debugging info
1841 DWARF 2 is an improved but incompatible version of DWARF 1.
1843 The DWARF 2 reader is in @file{dwarf2read.c}.
1847 @cindex SOM debugging info
1848 Like COFF, the SOM definition includes debugging information.
1850 @section Adding a New Symbol Reader to @value{GDBN}
1852 @cindex adding debugging info reader
1853 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
1854 there is probably little to be done.
1856 If you need to add a new object file format, you must first add it to
1857 BFD. This is beyond the scope of this document.
1859 You must then arrange for the BFD code to provide access to the
1860 debugging symbols. Generally @value{GDBN} will have to call swapping routines
1861 from BFD and a few other BFD internal routines to locate the debugging
1862 information. As much as possible, @value{GDBN} should not depend on the BFD
1863 internal data structures.
1865 For some targets (e.g., COFF), there is a special transfer vector used
1866 to call swapping routines, since the external data structures on various
1867 platforms have different sizes and layouts. Specialized routines that
1868 will only ever be implemented by one object file format may be called
1869 directly. This interface should be described in a file
1870 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
1873 @node Language Support
1875 @chapter Language Support
1877 @cindex language support
1878 @value{GDBN}'s language support is mainly driven by the symbol reader,
1879 although it is possible for the user to set the source language
1882 @value{GDBN} chooses the source language by looking at the extension
1883 of the file recorded in the debug info; @file{.c} means C, @file{.f}
1884 means Fortran, etc. It may also use a special-purpose language
1885 identifier if the debug format supports it, like with DWARF.
1887 @section Adding a Source Language to @value{GDBN}
1889 @cindex adding source language
1890 To add other languages to @value{GDBN}'s expression parser, follow the
1894 @item Create the expression parser.
1896 @cindex expression parser
1897 This should reside in a file @file{@var{lang}-exp.y}. Routines for
1898 building parsed expressions into a @code{union exp_element} list are in
1901 @cindex language parser
1902 Since we can't depend upon everyone having Bison, and YACC produces
1903 parsers that define a bunch of global names, the following lines
1904 @strong{must} be included at the top of the YACC parser, to prevent the
1905 various parsers from defining the same global names:
1908 #define yyparse @var{lang}_parse
1909 #define yylex @var{lang}_lex
1910 #define yyerror @var{lang}_error
1911 #define yylval @var{lang}_lval
1912 #define yychar @var{lang}_char
1913 #define yydebug @var{lang}_debug
1914 #define yypact @var{lang}_pact
1915 #define yyr1 @var{lang}_r1
1916 #define yyr2 @var{lang}_r2
1917 #define yydef @var{lang}_def
1918 #define yychk @var{lang}_chk
1919 #define yypgo @var{lang}_pgo
1920 #define yyact @var{lang}_act
1921 #define yyexca @var{lang}_exca
1922 #define yyerrflag @var{lang}_errflag
1923 #define yynerrs @var{lang}_nerrs
1926 At the bottom of your parser, define a @code{struct language_defn} and
1927 initialize it with the right values for your language. Define an
1928 @code{initialize_@var{lang}} routine and have it call
1929 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
1930 that your language exists. You'll need some other supporting variables
1931 and functions, which will be used via pointers from your
1932 @code{@var{lang}_language_defn}. See the declaration of @code{struct
1933 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
1934 for more information.
1936 @item Add any evaluation routines, if necessary
1938 @cindex expression evaluation routines
1939 @findex evaluate_subexp
1940 @findex prefixify_subexp
1941 @findex length_of_subexp
1942 If you need new opcodes (that represent the operations of the language),
1943 add them to the enumerated type in @file{expression.h}. Add support
1944 code for these operations in the @code{evaluate_subexp} function
1945 defined in the file @file{eval.c}. Add cases
1946 for new opcodes in two functions from @file{parse.c}:
1947 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
1948 the number of @code{exp_element}s that a given operation takes up.
1950 @item Update some existing code
1952 Add an enumerated identifier for your language to the enumerated type
1953 @code{enum language} in @file{defs.h}.
1955 Update the routines in @file{language.c} so your language is included.
1956 These routines include type predicates and such, which (in some cases)
1957 are language dependent. If your language does not appear in the switch
1958 statement, an error is reported.
1960 @vindex current_language
1961 Also included in @file{language.c} is the code that updates the variable
1962 @code{current_language}, and the routines that translate the
1963 @code{language_@var{lang}} enumerated identifier into a printable
1966 @findex _initialize_language
1967 Update the function @code{_initialize_language} to include your
1968 language. This function picks the default language upon startup, so is
1969 dependent upon which languages that @value{GDBN} is built for.
1971 @findex allocate_symtab
1972 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
1973 code so that the language of each symtab (source file) is set properly.
1974 This is used to determine the language to use at each stack frame level.
1975 Currently, the language is set based upon the extension of the source
1976 file. If the language can be better inferred from the symbol
1977 information, please set the language of the symtab in the symbol-reading
1980 @findex print_subexp
1981 @findex op_print_tab
1982 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
1983 expression opcodes you have added to @file{expression.h}. Also, add the
1984 printed representations of your operators to @code{op_print_tab}.
1986 @item Add a place of call
1989 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
1990 @code{parse_exp_1} (defined in @file{parse.c}).
1992 @item Use macros to trim code
1994 @cindex trimming language-dependent code
1995 The user has the option of building @value{GDBN} for some or all of the
1996 languages. If the user decides to build @value{GDBN} for the language
1997 @var{lang}, then every file dependent on @file{language.h} will have the
1998 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
1999 leave out large routines that the user won't need if he or she is not
2000 using your language.
2002 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2003 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2004 compiled form of your parser) is not linked into @value{GDBN} at all.
2006 See the file @file{configure.in} for how @value{GDBN} is configured
2007 for different languages.
2009 @item Edit @file{Makefile.in}
2011 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2012 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2013 not get linked in, or, worse yet, it may not get @code{tar}red into the
2018 @node Host Definition
2020 @chapter Host Definition
2022 With the advent of Autoconf, it's rarely necessary to have host
2023 definition machinery anymore. The following information is provided,
2024 mainly, as an historical reference.
2026 @section Adding a New Host
2028 @cindex adding a new host
2029 @cindex host, adding
2030 @value{GDBN}'s host configuration support normally happens via Autoconf.
2031 New host-specific definitions should not be needed. Older hosts
2032 @value{GDBN} still use the host-specific definitions and files listed
2033 below, but these mostly exist for historical reasons, and will
2034 eventually disappear.
2037 @item gdb/config/@var{arch}/@var{xyz}.mh
2038 This file once contained both host and native configuration information
2039 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2040 configuration information is now handed by Autoconf.
2042 Host configuration information included a definition of
2043 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2044 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2045 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2047 New host only configurations do not need this file.
2049 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2050 This file once contained definitions and includes required when hosting
2051 gdb on machine @var{xyz}. Those definitions and includes are now
2052 handled by Autoconf.
2054 New host and native configurations do not need this file.
2056 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2057 file to define the macros @var{HOST_FLOAT_FORMAT},
2058 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2059 also needs to be replaced with either an Autoconf or run-time test.}
2063 @subheading Generic Host Support Files
2065 @cindex generic host support
2066 There are some ``generic'' versions of routines that can be used by
2067 various systems. These can be customized in various ways by macros
2068 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2069 the @var{xyz} host, you can just include the generic file's name (with
2070 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2072 Otherwise, if your machine needs custom support routines, you will need
2073 to write routines that perform the same functions as the generic file.
2074 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2075 into @code{XDEPFILES}.
2078 @cindex remote debugging support
2079 @cindex serial line support
2081 This contains serial line support for Unix systems. This is always
2082 included, via the makefile variable @code{SER_HARDWIRE}; override this
2083 variable in the @file{.mh} file to avoid it.
2086 This contains serial line support for 32-bit programs running under DOS,
2087 using the DJGPP (a.k.a.@: GO32) execution environment.
2089 @cindex TCP remote support
2091 This contains generic TCP support using sockets.
2094 @section Host Conditionals
2096 When @value{GDBN} is configured and compiled, various macros are
2097 defined or left undefined, to control compilation based on the
2098 attributes of the host system. These macros and their meanings (or if
2099 the meaning is not documented here, then one of the source files where
2100 they are used is indicated) are:
2103 @item @value{GDBN}INIT_FILENAME
2104 The default name of @value{GDBN}'s initialization file (normally
2108 This macro is deprecated.
2111 Define this if your system does not have a @code{<sys/file.h>}.
2113 @item SIGWINCH_HANDLER
2114 If your host defines @code{SIGWINCH}, you can define this to be the name
2115 of a function to be called if @code{SIGWINCH} is received.
2117 @item SIGWINCH_HANDLER_BODY
2118 Define this to expand into code that will define the function named by
2119 the expansion of @code{SIGWINCH_HANDLER}.
2121 @item ALIGN_STACK_ON_STARTUP
2122 @cindex stack alignment
2123 Define this if your system is of a sort that will crash in
2124 @code{tgetent} if the stack happens not to be longword-aligned when
2125 @code{main} is called. This is a rare situation, but is known to occur
2126 on several different types of systems.
2128 @item CRLF_SOURCE_FILES
2129 @cindex DOS text files
2130 Define this if host files use @code{\r\n} rather than @code{\n} as a
2131 line terminator. This will cause source file listings to omit @code{\r}
2132 characters when printing and it will allow @code{\r\n} line endings of files
2133 which are ``sourced'' by gdb. It must be possible to open files in binary
2134 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2136 @item DEFAULT_PROMPT
2138 The default value of the prompt string (normally @code{"(gdb) "}).
2141 @cindex terminal device
2142 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2144 @item FCLOSE_PROVIDED
2145 Define this if the system declares @code{fclose} in the headers included
2146 in @code{defs.h}. This isn't needed unless your compiler is unusually
2150 Define this if binary files are opened the same way as text files.
2152 @item GETENV_PROVIDED
2153 Define this if the system declares @code{getenv} in its headers included
2154 in @code{defs.h}. This isn't needed unless your compiler is unusually
2159 In some cases, use the system call @code{mmap} for reading symbol
2160 tables. For some machines this allows for sharing and quick updates.
2163 Define this if the host system has @code{termio.h}.
2170 Values for host-side constants.
2173 Substitute for isatty, if not available.
2176 This is the longest integer type available on the host. If not defined,
2177 it will default to @code{long long} or @code{long}, depending on
2178 @code{CC_HAS_LONG_LONG}.
2180 @item CC_HAS_LONG_LONG
2181 @cindex @code{long long} data type
2182 Define this if the host C compiler supports @code{long long}. This is set
2183 by the @code{configure} script.
2185 @item PRINTF_HAS_LONG_LONG
2186 Define this if the host can handle printing of long long integers via
2187 the printf format conversion specifier @code{ll}. This is set by the
2188 @code{configure} script.
2190 @item HAVE_LONG_DOUBLE
2191 Define this if the host C compiler supports @code{long double}. This is
2192 set by the @code{configure} script.
2194 @item PRINTF_HAS_LONG_DOUBLE
2195 Define this if the host can handle printing of long double float-point
2196 numbers via the printf format conversion specifier @code{Lg}. This is
2197 set by the @code{configure} script.
2199 @item SCANF_HAS_LONG_DOUBLE
2200 Define this if the host can handle the parsing of long double
2201 float-point numbers via the scanf format conversion specifier
2202 @code{Lg}. This is set by the @code{configure} script.
2204 @item LSEEK_NOT_LINEAR
2205 Define this if @code{lseek (n)} does not necessarily move to byte number
2206 @code{n} in the file. This is only used when reading source files. It
2207 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2210 This macro is used as the argument to @code{lseek} (or, most commonly,
2211 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2212 which is the POSIX equivalent.
2214 @item MMAP_BASE_ADDRESS
2215 When using HAVE_MMAP, the first mapping should go at this address.
2217 @item MMAP_INCREMENT
2218 when using HAVE_MMAP, this is the increment between mappings.
2221 If defined, this should be one or more tokens, such as @code{volatile},
2222 that can be used in both the declaration and definition of functions to
2223 indicate that they never return. The default is already set correctly
2224 if compiling with GCC. This will almost never need to be defined.
2227 If defined, this should be one or more tokens, such as
2228 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2229 of functions to indicate that they never return. The default is already
2230 set correctly if compiling with GCC. This will almost never need to be
2233 @item USE_GENERIC_DUMMY_FRAMES
2234 @cindex generic dummy frames
2235 Define this to 1 if the target is using the generic inferior function
2236 call code. See @code{blockframe.c} for more information.
2240 @value{GDBN} will use the @code{mmalloc} library for memory allocation
2241 for symbol reading if this symbol is defined. Be careful defining it
2242 since there are systems on which @code{mmalloc} does not work for some
2243 reason. One example is the DECstation, where its RPC library can't
2244 cope with our redefinition of @code{malloc} to call @code{mmalloc}.
2245 When defining @code{USE_MMALLOC}, you will also have to set
2246 @code{MMALLOC} in the Makefile, to point to the @code{mmalloc} library. This
2247 define is set when you configure with @samp{--with-mmalloc}.
2251 Define this if you are using @code{mmalloc}, but don't want the overhead
2252 of checking the heap with @code{mmcheck}. Note that on some systems,
2253 the C runtime makes calls to @code{malloc} prior to calling @code{main}, and if
2254 @code{free} is ever called with these pointers after calling
2255 @code{mmcheck} to enable checking, a memory corruption abort is certain
2256 to occur. These systems can still use @code{mmalloc}, but must define
2260 Define this to 1 if the C runtime allocates memory prior to
2261 @code{mmcheck} being called, but that memory is never freed so we don't
2262 have to worry about it triggering a memory corruption abort. The
2263 default is 0, which means that @code{mmcheck} will only install the heap
2264 checking functions if there has not yet been any memory allocation
2265 calls, and if it fails to install the functions, @value{GDBN} will issue a
2266 warning. This is currently defined if you configure using
2267 @samp{--with-mmalloc}.
2269 @item NO_SIGINTERRUPT
2270 @findex siginterrupt
2271 Define this to indicate that @code{siginterrupt} is not available.
2275 Define these to appropriate value for the system @code{lseek}, if not already
2279 This is the signal for stopping @value{GDBN}. Defaults to
2280 @code{SIGTSTP}. (Only redefined for the Convex.)
2283 Define this if the interior's tty should be opened with the @code{O_NOCTTY}
2284 flag. (FIXME: This should be a native-only flag, but @file{inflow.c} is
2288 Means that System V (prior to SVR4) include files are in use. (FIXME:
2289 This symbol is abused in @file{infrun.c}, @file{regex.c},
2290 @file{remote-nindy.c}, and @file{utils.c} for other things, at the
2294 Define this to help placate @code{lint} in some situations.
2297 Define this to override the defaults of @code{__volatile__} or
2302 @node Target Architecture Definition
2304 @chapter Target Architecture Definition
2306 @cindex target architecture definition
2307 @value{GDBN}'s target architecture defines what sort of
2308 machine-language programs @value{GDBN} can work with, and how it works
2311 The target architecture object is implemented as the C structure
2312 @code{struct gdbarch *}. The structure, and its methods, are generated
2313 using the Bourne shell script @file{gdbarch.sh}.
2315 @section Registers and Memory
2317 @value{GDBN}'s model of the target machine is rather simple.
2318 @value{GDBN} assumes the machine includes a bank of registers and a
2319 block of memory. Each register may have a different size.
2321 @value{GDBN} does not have a magical way to match up with the
2322 compiler's idea of which registers are which; however, it is critical
2323 that they do match up accurately. The only way to make this work is
2324 to get accurate information about the order that the compiler uses,
2325 and to reflect that in the @code{REGISTER_NAME} and related macros.
2327 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2329 @section Pointers Are Not Always Addresses
2330 @cindex pointer representation
2331 @cindex address representation
2332 @cindex word-addressed machines
2333 @cindex separate data and code address spaces
2334 @cindex spaces, separate data and code address
2335 @cindex address spaces, separate data and code
2336 @cindex code pointers, word-addressed
2337 @cindex converting between pointers and addresses
2338 @cindex D10V addresses
2340 On almost all 32-bit architectures, the representation of a pointer is
2341 indistinguishable from the representation of some fixed-length number
2342 whose value is the byte address of the object pointed to. On such
2343 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2344 However, architectures with smaller word sizes are often cramped for
2345 address space, so they may choose a pointer representation that breaks this
2346 identity, and allows a larger code address space.
2348 For example, the Mitsubishi D10V is a 16-bit VLIW processor whose
2349 instructions are 32 bits long@footnote{Some D10V instructions are
2350 actually pairs of 16-bit sub-instructions. However, since you can't
2351 jump into the middle of such a pair, code addresses can only refer to
2352 full 32 bit instructions, which is what matters in this explanation.}.
2353 If the D10V used ordinary byte addresses to refer to code locations,
2354 then the processor would only be able to address 64kb of instructions.
2355 However, since instructions must be aligned on four-byte boundaries, the
2356 low two bits of any valid instruction's byte address are always
2357 zero---byte addresses waste two bits. So instead of byte addresses,
2358 the D10V uses word addresses---byte addresses shifted right two bits---to
2359 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2362 However, this means that code pointers and data pointers have different
2363 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2364 @code{0xC020} when used as a data address, but refers to byte address
2365 @code{0x30080} when used as a code address.
2367 (The D10V also uses separate code and data address spaces, which also
2368 affects the correspondence between pointers and addresses, but we're
2369 going to ignore that here; this example is already too long.)
2371 To cope with architectures like this---the D10V is not the only
2372 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2373 byte numbers, and @dfn{pointers}, which are the target's representation
2374 of an address of a particular type of data. In the example above,
2375 @code{0xC020} is the pointer, which refers to one of the addresses
2376 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2377 @value{GDBN} provides functions for turning a pointer into an address
2378 and vice versa, in the appropriate way for the current architecture.
2380 Unfortunately, since addresses and pointers are identical on almost all
2381 processors, this distinction tends to bit-rot pretty quickly. Thus,
2382 each time you port @value{GDBN} to an architecture which does
2383 distinguish between pointers and addresses, you'll probably need to
2384 clean up some architecture-independent code.
2386 Here are functions which convert between pointers and addresses:
2388 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2389 Treat the bytes at @var{buf} as a pointer or reference of type
2390 @var{type}, and return the address it represents, in a manner
2391 appropriate for the current architecture. This yields an address
2392 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2393 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2396 For example, if the current architecture is the Intel x86, this function
2397 extracts a little-endian integer of the appropriate length from
2398 @var{buf} and returns it. However, if the current architecture is the
2399 D10V, this function will return a 16-bit integer extracted from
2400 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2402 If @var{type} is not a pointer or reference type, then this function
2403 will signal an internal error.
2406 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2407 Store the address @var{addr} in @var{buf}, in the proper format for a
2408 pointer of type @var{type} in the current architecture. Note that
2409 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2412 For example, if the current architecture is the Intel x86, this function
2413 stores @var{addr} unmodified as a little-endian integer of the
2414 appropriate length in @var{buf}. However, if the current architecture
2415 is the D10V, this function divides @var{addr} by four if @var{type} is
2416 a pointer to a function, and then stores it in @var{buf}.
2418 If @var{type} is not a pointer or reference type, then this function
2419 will signal an internal error.
2422 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2423 Assuming that @var{val} is a pointer, return the address it represents,
2424 as appropriate for the current architecture.
2426 This function actually works on integral values, as well as pointers.
2427 For pointers, it performs architecture-specific conversions as
2428 described above for @code{extract_typed_address}.
2431 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2432 Create and return a value representing a pointer of type @var{type} to
2433 the address @var{addr}, as appropriate for the current architecture.
2434 This function performs architecture-specific conversions as described
2435 above for @code{store_typed_address}.
2439 @value{GDBN} also provides functions that do the same tasks, but assume
2440 that pointers are simply byte addresses; they aren't sensitive to the
2441 current architecture, beyond knowing the appropriate endianness.
2443 @deftypefun CORE_ADDR extract_address (void *@var{addr}, int len)
2444 Extract a @var{len}-byte number from @var{addr} in the appropriate
2445 endianness for the current architecture, and return it. Note that
2446 @var{addr} refers to @value{GDBN}'s memory, not the inferior's.
2448 This function should only be used in architecture-specific code; it
2449 doesn't have enough information to turn bits into a true address in the
2450 appropriate way for the current architecture. If you can, use
2451 @code{extract_typed_address} instead.
2454 @deftypefun void store_address (void *@var{addr}, int @var{len}, LONGEST @var{val})
2455 Store @var{val} at @var{addr} as a @var{len}-byte integer, in the
2456 appropriate endianness for the current architecture. Note that
2457 @var{addr} refers to a buffer in @value{GDBN}'s memory, not the
2460 This function should only be used in architecture-specific code; it
2461 doesn't have enough information to turn a true address into bits in the
2462 appropriate way for the current architecture. If you can, use
2463 @code{store_typed_address} instead.
2467 Here are some macros which architectures can define to indicate the
2468 relationship between pointers and addresses. These have default
2469 definitions, appropriate for architectures on which all pointers are
2470 simple unsigned byte addresses.
2472 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2473 Assume that @var{buf} holds a pointer of type @var{type}, in the
2474 appropriate format for the current architecture. Return the byte
2475 address the pointer refers to.
2477 This function may safely assume that @var{type} is either a pointer or a
2478 C@t{++} reference type.
2481 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2482 Store in @var{buf} a pointer of type @var{type} representing the address
2483 @var{addr}, in the appropriate format for the current architecture.
2485 This function may safely assume that @var{type} is either a pointer or a
2486 C@t{++} reference type.
2490 @section Using Different Register and Memory Data Representations
2491 @cindex raw representation
2492 @cindex virtual representation
2493 @cindex representations, raw and virtual
2494 @cindex register data formats, converting
2495 @cindex @code{struct value}, converting register contents to
2497 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
2498 significant change. Many of the macros and functions refered to in the
2499 sections below are likely to be made obsolete. See the file @file{TODO}
2500 for more up-to-date information.}
2502 Some architectures use one representation for a value when it lives in a
2503 register, but use a different representation when it lives in memory.
2504 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
2505 the target registers, and the @dfn{virtual} representation is the one
2506 used in memory, and within @value{GDBN} @code{struct value} objects.
2508 For almost all data types on almost all architectures, the virtual and
2509 raw representations are identical, and no special handling is needed.
2510 However, they do occasionally differ. For example:
2514 The x86 architecture supports an 80-bit @code{long double} type. However, when
2515 we store those values in memory, they occupy twelve bytes: the
2516 floating-point number occupies the first ten, and the final two bytes
2517 are unused. This keeps the values aligned on four-byte boundaries,
2518 allowing more efficient access. Thus, the x86 80-bit floating-point
2519 type is the raw representation, and the twelve-byte loosely-packed
2520 arrangement is the virtual representation.
2523 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
2524 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
2525 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
2526 raw representation, and the trimmed 32-bit representation is the
2527 virtual representation.
2530 In general, the raw representation is determined by the architecture, or
2531 @value{GDBN}'s interface to the architecture, while the virtual representation
2532 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
2533 @code{registers}, holds the register contents in raw format, and the
2534 @value{GDBN} remote protocol transmits register values in raw format.
2536 Your architecture may define the following macros to request
2537 conversions between the raw and virtual format:
2539 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
2540 Return non-zero if register number @var{reg}'s value needs different raw
2541 and virtual formats.
2543 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
2544 unless this macro returns a non-zero value for that register.
2547 @deftypefn {Target Macro} int REGISTER_RAW_SIZE (int @var{reg})
2548 The size of register number @var{reg}'s raw value. This is the number
2549 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
2550 remote protocol packet.
2553 @deftypefn {Target Macro} int REGISTER_VIRTUAL_SIZE (int @var{reg})
2554 The size of register number @var{reg}'s value, in its virtual format.
2555 This is the size a @code{struct value}'s buffer will have, holding that
2559 @deftypefn {Target Macro} struct type *REGISTER_VIRTUAL_TYPE (int @var{reg})
2560 This is the type of the virtual representation of register number
2561 @var{reg}. Note that there is no need for a macro giving a type for the
2562 register's raw form; once the register's value has been obtained, @value{GDBN}
2563 always uses the virtual form.
2566 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2567 Convert the value of register number @var{reg} to @var{type}, which
2568 should always be @code{REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2569 at @var{from} holds the register's value in raw format; the macro should
2570 convert the value to virtual format, and place it at @var{to}.
2572 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
2573 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
2574 arguments in different orders.
2576 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
2577 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
2581 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2582 Convert the value of register number @var{reg} to @var{type}, which
2583 should always be @code{REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2584 at @var{from} holds the register's value in raw format; the macro should
2585 convert the value to virtual format, and place it at @var{to}.
2587 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
2588 their @var{reg} and @var{type} arguments in different orders.
2592 @section Frame Interpretation
2594 @section Inferior Call Setup
2596 @section Compiler Characteristics
2598 @section Target Conditionals
2600 This section describes the macros that you can use to define the target
2605 @item ADDITIONAL_OPTIONS
2606 @itemx ADDITIONAL_OPTION_CASES
2607 @itemx ADDITIONAL_OPTION_HANDLER
2608 @itemx ADDITIONAL_OPTION_HELP
2609 @findex ADDITIONAL_OPTION_HELP
2610 @findex ADDITIONAL_OPTION_HANDLER
2611 @findex ADDITIONAL_OPTION_CASES
2612 @findex ADDITIONAL_OPTIONS
2613 These are a set of macros that allow the addition of additional command
2614 line options to @value{GDBN}. They are currently used only for the unsupported
2615 i960 Nindy target, and should not be used in any other configuration.
2617 @item ADDR_BITS_REMOVE (addr)
2618 @findex ADDR_BITS_REMOVE
2619 If a raw machine instruction address includes any bits that are not
2620 really part of the address, then define this macro to expand into an
2621 expression that zeroes those bits in @var{addr}. This is only used for
2622 addresses of instructions, and even then not in all contexts.
2624 For example, the two low-order bits of the PC on the Hewlett-Packard PA
2625 2.0 architecture contain the privilege level of the corresponding
2626 instruction. Since instructions must always be aligned on four-byte
2627 boundaries, the processor masks out these bits to generate the actual
2628 address of the instruction. ADDR_BITS_REMOVE should filter out these
2629 bits with an expression such as @code{((addr) & ~3)}.
2631 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
2632 @findex ADDRESS_TO_POINTER
2633 Store in @var{buf} a pointer of type @var{type} representing the address
2634 @var{addr}, in the appropriate format for the current architecture.
2635 This macro may safely assume that @var{type} is either a pointer or a
2636 C@t{++} reference type.
2637 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
2639 @item BEFORE_MAIN_LOOP_HOOK
2640 @findex BEFORE_MAIN_LOOP_HOOK
2641 Define this to expand into any code that you want to execute before the
2642 main loop starts. Although this is not, strictly speaking, a target
2643 conditional, that is how it is currently being used. Note that if a
2644 configuration were to define it one way for a host and a different way
2645 for the target, @value{GDBN} will probably not compile, let alone run
2646 correctly. This macro is currently used only for the unsupported i960 Nindy
2647 target, and should not be used in any other configuration.
2649 @item BELIEVE_PCC_PROMOTION
2650 @findex BELIEVE_PCC_PROMOTION
2651 Define if the compiler promotes a @code{short} or @code{char}
2652 parameter to an @code{int}, but still reports the parameter as its
2653 original type, rather than the promoted type.
2655 @item BELIEVE_PCC_PROMOTION_TYPE
2656 @findex BELIEVE_PCC_PROMOTION_TYPE
2657 Define this if @value{GDBN} should believe the type of a @code{short}
2658 argument when compiled by @code{pcc}, but look within a full int space to get
2659 its value. Only defined for Sun-3 at present.
2661 @item BITS_BIG_ENDIAN
2662 @findex BITS_BIG_ENDIAN
2663 Define this if the numbering of bits in the targets does @strong{not} match the
2664 endianness of the target byte order. A value of 1 means that the bits
2665 are numbered in a big-endian bit order, 0 means little-endian.
2669 This is the character array initializer for the bit pattern to put into
2670 memory where a breakpoint is set. Although it's common to use a trap
2671 instruction for a breakpoint, it's not required; for instance, the bit
2672 pattern could be an invalid instruction. The breakpoint must be no
2673 longer than the shortest instruction of the architecture.
2675 @code{BREAKPOINT} has been deprecated in favor of
2676 @code{BREAKPOINT_FROM_PC}.
2678 @item BIG_BREAKPOINT
2679 @itemx LITTLE_BREAKPOINT
2680 @findex LITTLE_BREAKPOINT
2681 @findex BIG_BREAKPOINT
2682 Similar to BREAKPOINT, but used for bi-endian targets.
2684 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
2685 favor of @code{BREAKPOINT_FROM_PC}.
2687 @item REMOTE_BREAKPOINT
2688 @itemx LITTLE_REMOTE_BREAKPOINT
2689 @itemx BIG_REMOTE_BREAKPOINT
2690 @findex BIG_REMOTE_BREAKPOINT
2691 @findex LITTLE_REMOTE_BREAKPOINT
2692 @findex REMOTE_BREAKPOINT
2693 Similar to BREAKPOINT, but used for remote targets.
2695 @code{BIG_REMOTE_BREAKPOINT} and @code{LITTLE_REMOTE_BREAKPOINT} have been
2696 deprecated in favor of @code{BREAKPOINT_FROM_PC}.
2698 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
2699 @findex BREAKPOINT_FROM_PC
2700 Use the program counter to determine the contents and size of a
2701 breakpoint instruction. It returns a pointer to a string of bytes
2702 that encode a breakpoint instruction, stores the length of the string
2703 to *@var{lenptr}, and adjusts pc (if necessary) to point to the actual
2704 memory location where the breakpoint should be inserted.
2706 Although it is common to use a trap instruction for a breakpoint, it's
2707 not required; for instance, the bit pattern could be an invalid
2708 instruction. The breakpoint must be no longer than the shortest
2709 instruction of the architecture.
2711 Replaces all the other @var{BREAKPOINT} macros.
2713 @item MEMORY_INSERT_BREAKPOINT (@var{addr}, @var{contents_cache})
2714 @itemx MEMORY_REMOVE_BREAKPOINT (@var{addr}, @var{contents_cache})
2715 @findex MEMORY_REMOVE_BREAKPOINT
2716 @findex MEMORY_INSERT_BREAKPOINT
2717 Insert or remove memory based breakpoints. Reasonable defaults
2718 (@code{default_memory_insert_breakpoint} and
2719 @code{default_memory_remove_breakpoint} respectively) have been
2720 provided so that it is not necessary to define these for most
2721 architectures. Architectures which may want to define
2722 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
2723 likely have instructions that are oddly sized or are not stored in a
2724 conventional manner.
2726 It may also be desirable (from an efficiency standpoint) to define
2727 custom breakpoint insertion and removal routines if
2728 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
2732 @findex CALL_DUMMY_P
2733 A C expression that is non-zero when the target supports inferior function
2736 @item CALL_DUMMY_WORDS
2737 @findex CALL_DUMMY_WORDS
2738 Pointer to an array of @code{LONGEST} words of data containing
2739 host-byte-ordered @code{REGISTER_BYTES} sized values that partially
2740 specify the sequence of instructions needed for an inferior function
2743 Should be deprecated in favor of a macro that uses target-byte-ordered
2746 @item SIZEOF_CALL_DUMMY_WORDS
2747 @findex SIZEOF_CALL_DUMMY_WORDS
2748 The size of @code{CALL_DUMMY_WORDS}. When @code{CALL_DUMMY_P} this must
2749 return a positive value. See also @code{CALL_DUMMY_LENGTH}.
2753 A static initializer for @code{CALL_DUMMY_WORDS}. Deprecated.
2755 @item CALL_DUMMY_LOCATION
2756 @findex CALL_DUMMY_LOCATION
2757 See the file @file{inferior.h}.
2759 @item CALL_DUMMY_STACK_ADJUST
2760 @findex CALL_DUMMY_STACK_ADJUST
2761 Stack adjustment needed when performing an inferior function call.
2763 Should be deprecated in favor of something like @code{STACK_ALIGN}.
2765 @item CALL_DUMMY_STACK_ADJUST_P
2766 @findex CALL_DUMMY_STACK_ADJUST_P
2767 Predicate for use of @code{CALL_DUMMY_STACK_ADJUST}.
2769 Should be deprecated in favor of something like @code{STACK_ALIGN}.
2771 @item CANNOT_FETCH_REGISTER (@var{regno})
2772 @findex CANNOT_FETCH_REGISTER
2773 A C expression that should be nonzero if @var{regno} cannot be fetched
2774 from an inferior process. This is only relevant if
2775 @code{FETCH_INFERIOR_REGISTERS} is not defined.
2777 @item CANNOT_STORE_REGISTER (@var{regno})
2778 @findex CANNOT_STORE_REGISTER
2779 A C expression that should be nonzero if @var{regno} should not be
2780 written to the target. This is often the case for program counters,
2781 status words, and other special registers. If this is not defined,
2782 @value{GDBN} will assume that all registers may be written.
2784 @item DO_DEFERRED_STORES
2785 @itemx CLEAR_DEFERRED_STORES
2786 @findex CLEAR_DEFERRED_STORES
2787 @findex DO_DEFERRED_STORES
2788 Define this to execute any deferred stores of registers into the inferior,
2789 and to cancel any deferred stores.
2791 Currently only implemented correctly for native Sparc configurations?
2793 @item COERCE_FLOAT_TO_DOUBLE (@var{formal}, @var{actual})
2794 @findex COERCE_FLOAT_TO_DOUBLE
2795 @cindex promotion to @code{double}
2796 @cindex @code{float} arguments
2797 @cindex prototyped functions, passing arguments to
2798 @cindex passing arguments to prototyped functions
2799 Return non-zero if GDB should promote @code{float} values to
2800 @code{double} when calling a non-prototyped function. The argument
2801 @var{actual} is the type of the value we want to pass to the function.
2802 The argument @var{formal} is the type of this argument, as it appears in
2803 the function's definition. Note that @var{formal} may be zero if we
2804 have no debugging information for the function, or if we're passing more
2805 arguments than are officially declared (for example, varargs). This
2806 macro is never invoked if the function definitely has a prototype.
2808 How you should pass arguments to a function depends on whether it was
2809 defined in K&R style or prototype style. If you define a function using
2810 the K&R syntax that takes a @code{float} argument, then callers must
2811 pass that argument as a @code{double}. If you define the function using
2812 the prototype syntax, then you must pass the argument as a @code{float},
2815 Unfortunately, on certain older platforms, the debug info doesn't
2816 indicate reliably how each function was defined. A function type's
2817 @code{TYPE_FLAG_PROTOTYPED} flag may be unset, even if the function was
2818 defined in prototype style. When calling a function whose
2819 @code{TYPE_FLAG_PROTOTYPED} flag is unset, GDB consults the
2820 @code{COERCE_FLOAT_TO_DOUBLE} macro to decide what to do.
2822 @findex standard_coerce_float_to_double
2823 For modern targets, it is proper to assume that, if the prototype flag
2824 is unset, that can be trusted: @code{float} arguments should be promoted
2825 to @code{double}. You should use the function
2826 @code{standard_coerce_float_to_double} to get this behavior.
2828 @findex default_coerce_float_to_double
2829 For some older targets, if the prototype flag is unset, that doesn't
2830 tell us anything. So we guess that, if we don't have a type for the
2831 formal parameter (@i{i.e.}, the first argument to
2832 @code{COERCE_FLOAT_TO_DOUBLE} is null), then we should promote it;
2833 otherwise, we should leave it alone. The function
2834 @code{default_coerce_float_to_double} provides this behavior; it is the
2835 default value, for compatibility with older configurations.
2838 @findex CPLUS_MARKERz
2839 Define this to expand into the character that G@t{++} uses to distinguish
2840 compiler-generated identifiers from programmer-specified identifiers.
2841 By default, this expands into @code{'$'}. Most System V targets should
2842 define this to @code{'.'}.
2844 @item DBX_PARM_SYMBOL_CLASS
2845 @findex DBX_PARM_SYMBOL_CLASS
2846 Hook for the @code{SYMBOL_CLASS} of a parameter when decoding DBX symbol
2847 information. In the i960, parameters can be stored as locals or as
2848 args, depending on the type of the debug record.
2850 @item DECR_PC_AFTER_BREAK
2851 @findex DECR_PC_AFTER_BREAK
2852 Define this to be the amount by which to decrement the PC after the
2853 program encounters a breakpoint. This is often the number of bytes in
2854 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
2856 @item DECR_PC_AFTER_HW_BREAK
2857 @findex DECR_PC_AFTER_HW_BREAK
2858 Similarly, for hardware breakpoints.
2860 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
2861 @findex DISABLE_UNSETTABLE_BREAK
2862 If defined, this should evaluate to 1 if @var{addr} is in a shared
2863 library in which breakpoints cannot be set and so should be disabled.
2865 @item DO_REGISTERS_INFO
2866 @findex DO_REGISTERS_INFO
2867 If defined, use this to print the value of a register or all registers.
2869 @item PRINT_FLOAT_INFO()
2870 #findex PRINT_FLOAT_INFO
2871 If defined, then the @samp{info float} command will print information about
2872 the processor's floating point unit.
2874 @item DWARF_REG_TO_REGNUM
2875 @findex DWARF_REG_TO_REGNUM
2876 Convert DWARF register number into @value{GDBN} regnum. If not defined,
2877 no conversion will be performed.
2879 @item DWARF2_REG_TO_REGNUM
2880 @findex DWARF2_REG_TO_REGNUM
2881 Convert DWARF2 register number into @value{GDBN} regnum. If not
2882 defined, no conversion will be performed.
2884 @item ECOFF_REG_TO_REGNUM
2885 @findex ECOFF_REG_TO_REGNUM
2886 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
2887 no conversion will be performed.
2889 @item END_OF_TEXT_DEFAULT
2890 @findex END_OF_TEXT_DEFAULT
2891 This is an expression that should designate the end of the text section.
2894 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
2895 @findex EXTRACT_RETURN_VALUE
2896 Define this to extract a function's return value of type @var{type} from
2897 the raw register state @var{regbuf} and copy that, in virtual format,
2900 @item EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
2901 @findex EXTRACT_STRUCT_VALUE_ADDRESS
2902 When defined, extract from the array @var{regbuf} (containing the raw
2903 register state) the @code{CORE_ADDR} at which a function should return
2904 its structure value.
2906 If not defined, @code{EXTRACT_RETURN_VALUE} is used.
2908 @item EXTRACT_STRUCT_VALUE_ADDRESS_P()
2909 @findex EXTRACT_STRUCT_VALUE_ADDRESS_P
2910 Predicate for @code{EXTRACT_STRUCT_VALUE_ADDRESS}.
2914 Deprecated in favor of @code{PRINT_FLOAT_INFO}.
2918 If the virtual frame pointer is kept in a register, then define this
2919 macro to be the number (greater than or equal to zero) of that register.
2921 This should only need to be defined if @code{TARGET_READ_FP} is not
2924 @item FRAMELESS_FUNCTION_INVOCATION(@var{fi})
2925 @findex FRAMELESS_FUNCTION_INVOCATION
2926 Define this to an expression that returns 1 if the function invocation
2927 represented by @var{fi} does not have a stack frame associated with it.
2930 @item FRAME_ARGS_ADDRESS_CORRECT
2931 @findex FRAME_ARGS_ADDRESS_CORRECT
2934 @item FRAME_CHAIN(@var{frame})
2936 Given @var{frame}, return a pointer to the calling frame.
2938 @item FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
2939 @findex FRAME_CHAIN_VALID
2940 Define this to be an expression that returns zero if the given frame is
2941 an outermost frame, with no caller, and nonzero otherwise. Several
2942 common definitions are available:
2946 @code{file_frame_chain_valid} is nonzero if the chain pointer is nonzero
2947 and given frame's PC is not inside the startup file (such as
2951 @code{func_frame_chain_valid} is nonzero if the chain
2952 pointer is nonzero and the given frame's PC is not in @code{main} or a
2953 known entry point function (such as @code{_start}).
2956 @code{generic_file_frame_chain_valid} and
2957 @code{generic_func_frame_chain_valid} are equivalent implementations for
2958 targets using generic dummy frames.
2961 @item FRAME_INIT_SAVED_REGS(@var{frame})
2962 @findex FRAME_INIT_SAVED_REGS
2963 See @file{frame.h}. Determines the address of all registers in the
2964 current stack frame storing each in @code{frame->saved_regs}. Space for
2965 @code{frame->saved_regs} shall be allocated by
2966 @code{FRAME_INIT_SAVED_REGS} using either
2967 @code{frame_saved_regs_zalloc} or @code{frame_obstack_alloc}.
2969 @code{FRAME_FIND_SAVED_REGS} and @code{EXTRA_FRAME_INFO} are deprecated.
2971 @item FRAME_NUM_ARGS (@var{fi})
2972 @findex FRAME_NUM_ARGS
2973 For the frame described by @var{fi} return the number of arguments that
2974 are being passed. If the number of arguments is not known, return
2977 @item FRAME_SAVED_PC(@var{frame})
2978 @findex FRAME_SAVED_PC
2979 Given @var{frame}, return the pc saved there. This is the return
2982 @item FUNCTION_EPILOGUE_SIZE
2983 @findex FUNCTION_EPILOGUE_SIZE
2984 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
2985 function end symbol is 0. For such targets, you must define
2986 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
2987 function's epilogue.
2989 @item FUNCTION_START_OFFSET
2990 @findex FUNCTION_START_OFFSET
2991 An integer, giving the offset in bytes from a function's address (as
2992 used in the values of symbols, function pointers, etc.), and the
2993 function's first genuine instruction.
2995 This is zero on almost all machines: the function's address is usually
2996 the address of its first instruction. However, on the VAX, for example,
2997 each function starts with two bytes containing a bitmask indicating
2998 which registers to save upon entry to the function. The VAX @code{call}
2999 instructions check this value, and save the appropriate registers
3000 automatically. Thus, since the offset from the function's address to
3001 its first instruction is two bytes, @code{FUNCTION_START_OFFSET} would
3004 @item GCC_COMPILED_FLAG_SYMBOL
3005 @itemx GCC2_COMPILED_FLAG_SYMBOL
3006 @findex GCC2_COMPILED_FLAG_SYMBOL
3007 @findex GCC_COMPILED_FLAG_SYMBOL
3008 If defined, these are the names of the symbols that @value{GDBN} will
3009 look for to detect that GCC compiled the file. The default symbols
3010 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3011 respectively. (Currently only defined for the Delta 68.)
3013 @item @value{GDBN}_MULTI_ARCH
3014 @findex @value{GDBN}_MULTI_ARCH
3015 If defined and non-zero, enables support for multiple architectures
3016 within @value{GDBN}.
3018 This support can be enabled at two levels. At level one, only
3019 definitions for previously undefined macros are provided; at level two,
3020 a multi-arch definition of all architecture dependent macros will be
3023 @item @value{GDBN}_TARGET_IS_HPPA
3024 @findex @value{GDBN}_TARGET_IS_HPPA
3025 This determines whether horrible kludge code in @file{dbxread.c} and
3026 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3027 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3030 @item GET_LONGJMP_TARGET
3031 @findex GET_LONGJMP_TARGET
3032 For most machines, this is a target-dependent parameter. On the
3033 DECstation and the Iris, this is a native-dependent parameter, since
3034 the header file @file{setjmp.h} is needed to define it.
3036 This macro determines the target PC address that @code{longjmp} will jump to,
3037 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3038 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3039 pointer. It examines the current state of the machine as needed.
3041 @item GET_SAVED_REGISTER
3042 @findex GET_SAVED_REGISTER
3043 @findex get_saved_register
3044 Define this if you need to supply your own definition for the function
3045 @code{get_saved_register}.
3047 @item IBM6000_TARGET
3048 @findex IBM6000_TARGET
3049 Shows that we are configured for an IBM RS/6000 target. This
3050 conditional should be eliminated (FIXME) and replaced by
3051 feature-specific macros. It was introduced in a haste and we are
3052 repenting at leisure.
3054 @item I386_USE_GENERIC_WATCHPOINTS
3055 An x86-based target can define this to use the generic x86 watchpoint
3056 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3058 @item SYMBOLS_CAN_START_WITH_DOLLAR
3059 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3060 Some systems have routines whose names start with @samp{$}. Giving this
3061 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3062 routines when parsing tokens that begin with @samp{$}.
3064 On HP-UX, certain system routines (millicode) have names beginning with
3065 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3066 routine that handles inter-space procedure calls on PA-RISC.
3068 @item INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3069 @findex INIT_EXTRA_FRAME_INFO
3070 If additional information about the frame is required this should be
3071 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3072 is allocated using @code{frame_obstack_alloc}.
3074 @item INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3075 @findex INIT_FRAME_PC
3076 This is a C statement that sets the pc of the frame pointed to by
3077 @var{prev}. [By default...]
3079 @item INNER_THAN (@var{lhs}, @var{rhs})
3081 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3082 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3083 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3086 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3087 @findex gdbarch_in_function_epilogue_p
3088 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3089 The epilogue of a function is defined as the part of a function where
3090 the stack frame of the function already has been destroyed up to the
3091 final `return from function call' instruction.
3093 @item IN_SIGTRAMP (@var{pc}, @var{name})
3095 Define this to return non-zero if the given @var{pc} and/or @var{name}
3096 indicates that the current function is a @code{sigtramp}.
3098 @item SIGTRAMP_START (@var{pc})
3099 @findex SIGTRAMP_START
3100 @itemx SIGTRAMP_END (@var{pc})
3101 @findex SIGTRAMP_END
3102 Define these to be the start and end address of the @code{sigtramp} for the
3103 given @var{pc}. On machines where the address is just a compile time
3104 constant, the macro expansion will typically just ignore the supplied
3107 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3108 @findex IN_SOLIB_CALL_TRAMPOLINE
3109 Define this to evaluate to nonzero if the program is stopped in the
3110 trampoline that connects to a shared library.
3112 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3113 @findex IN_SOLIB_RETURN_TRAMPOLINE
3114 Define this to evaluate to nonzero if the program is stopped in the
3115 trampoline that returns from a shared library.
3117 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3118 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3119 Define this to evaluate to nonzero if the program is stopped in the
3122 @item SKIP_SOLIB_RESOLVER (@var{pc})
3123 @findex SKIP_SOLIB_RESOLVER
3124 Define this to evaluate to the (nonzero) address at which execution
3125 should continue to get past the dynamic linker's symbol resolution
3126 function. A zero value indicates that it is not important or necessary
3127 to set a breakpoint to get through the dynamic linker and that single
3128 stepping will suffice.
3130 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3131 @findex INTEGER_TO_ADDRESS
3132 @cindex converting integers to addresses
3133 Define this when the architecture needs to handle non-pointer to address
3134 conversions specially. Converts that value to an address according to
3135 the current architectures conventions.
3137 @emph{Pragmatics: When the user copies a well defined expression from
3138 their source code and passes it, as a parameter, to @value{GDBN}'s
3139 @code{print} command, they should get the same value as would have been
3140 computed by the target program. Any deviation from this rule can cause
3141 major confusion and annoyance, and needs to be justified carefully. In
3142 other words, @value{GDBN} doesn't really have the freedom to do these
3143 conversions in clever and useful ways. It has, however, been pointed
3144 out that users aren't complaining about how @value{GDBN} casts integers
3145 to pointers; they are complaining that they can't take an address from a
3146 disassembly listing and give it to @code{x/i}. Adding an architecture
3147 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3148 @value{GDBN} to ``get it right'' in all circumstances.}
3150 @xref{Target Architecture Definition, , Pointers Are Not Always
3153 @item IS_TRAPPED_INTERNALVAR (@var{name})
3154 @findex IS_TRAPPED_INTERNALVAR
3155 This is an ugly hook to allow the specification of special actions that
3156 should occur as a side-effect of setting the value of a variable
3157 internal to @value{GDBN}. Currently only used by the h8500. Note that this
3158 could be either a host or target conditional.
3160 @item NEED_TEXT_START_END
3161 @findex NEED_TEXT_START_END
3162 Define this if @value{GDBN} should determine the start and end addresses of the
3163 text section. (Seems dubious.)
3165 @item NO_HIF_SUPPORT
3166 @findex NO_HIF_SUPPORT
3167 (Specific to the a29k.)
3169 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3170 @findex POINTER_TO_ADDRESS
3171 Assume that @var{buf} holds a pointer of type @var{type}, in the
3172 appropriate format for the current architecture. Return the byte
3173 address the pointer refers to.
3174 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3176 @item REGISTER_CONVERTIBLE (@var{reg})
3177 @findex REGISTER_CONVERTIBLE
3178 Return non-zero if @var{reg} uses different raw and virtual formats.
3179 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3181 @item REGISTER_RAW_SIZE (@var{reg})
3182 @findex REGISTER_RAW_SIZE
3183 Return the raw size of @var{reg}.
3184 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3186 @item REGISTER_VIRTUAL_SIZE (@var{reg})
3187 @findex REGISTER_VIRTUAL_SIZE
3188 Return the virtual size of @var{reg}.
3189 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3191 @item REGISTER_VIRTUAL_TYPE (@var{reg})
3192 @findex REGISTER_VIRTUAL_TYPE
3193 Return the virtual type of @var{reg}.
3194 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3196 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3197 @findex REGISTER_CONVERT_TO_VIRTUAL
3198 Convert the value of register @var{reg} from its raw form to its virtual
3200 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3202 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3203 @findex REGISTER_CONVERT_TO_RAW
3204 Convert the value of register @var{reg} from its virtual form to its raw
3206 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3208 @item RETURN_VALUE_ON_STACK(@var{type})
3209 @findex RETURN_VALUE_ON_STACK
3210 @cindex returning structures by value
3211 @cindex structures, returning by value
3213 Return non-zero if values of type TYPE are returned on the stack, using
3214 the ``struct convention'' (i.e., the caller provides a pointer to a
3215 buffer in which the callee should store the return value). This
3216 controls how the @samp{finish} command finds a function's return value,
3217 and whether an inferior function call reserves space on the stack for
3220 The full logic @value{GDBN} uses here is kind of odd.
3224 If the type being returned by value is not a structure, union, or array,
3225 and @code{RETURN_VALUE_ON_STACK} returns zero, then @value{GDBN}
3226 concludes the value is not returned using the struct convention.
3229 Otherwise, @value{GDBN} calls @code{USE_STRUCT_CONVENTION} (see below).
3230 If that returns non-zero, @value{GDBN} assumes the struct convention is
3234 In other words, to indicate that a given type is returned by value using
3235 the struct convention, that type must be either a struct, union, array,
3236 or something @code{RETURN_VALUE_ON_STACK} likes, @emph{and} something
3237 that @code{USE_STRUCT_CONVENTION} likes.
3239 Note that, in C and C@t{++}, arrays are never returned by value. In those
3240 languages, these predicates will always see a pointer type, never an
3241 array type. All the references above to arrays being returned by value
3242 apply only to other languages.
3244 @item SOFTWARE_SINGLE_STEP_P()
3245 @findex SOFTWARE_SINGLE_STEP_P
3246 Define this as 1 if the target does not have a hardware single-step
3247 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3249 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3250 @findex SOFTWARE_SINGLE_STEP
3251 A function that inserts or removes (depending on
3252 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3253 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3256 @item SOFUN_ADDRESS_MAYBE_MISSING
3257 @findex SOFUN_ADDRESS_MAYBE_MISSING
3258 Somebody clever observed that, the more actual addresses you have in the
3259 debug information, the more time the linker has to spend relocating
3260 them. So whenever there's some other way the debugger could find the
3261 address it needs, you should omit it from the debug info, to make
3264 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3265 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3266 entries in stabs-format debugging information. @code{N_SO} stabs mark
3267 the beginning and ending addresses of compilation units in the text
3268 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3270 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3274 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3275 addresses where the function starts by taking the function name from
3276 the stab, and then looking that up in the minsyms (the
3277 linker/assembler symbol table). In other words, the stab has the
3278 name, and the linker/assembler symbol table is the only place that carries
3282 @code{N_SO} stabs have an address of zero, too. You just look at the
3283 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3284 and guess the starting and ending addresses of the compilation unit from
3288 @item PCC_SOL_BROKEN
3289 @findex PCC_SOL_BROKEN
3290 (Used only in the Convex target.)
3292 @item PC_IN_CALL_DUMMY
3293 @findex PC_IN_CALL_DUMMY
3294 See @file{inferior.h}.
3296 @item PC_LOAD_SEGMENT
3297 @findex PC_LOAD_SEGMENT
3298 If defined, print information about the load segment for the program
3299 counter. (Defined only for the RS/6000.)
3303 If the program counter is kept in a register, then define this macro to
3304 be the number (greater than or equal to zero) of that register.
3306 This should only need to be defined if @code{TARGET_READ_PC} and
3307 @code{TARGET_WRITE_PC} are not defined.
3311 The number of the ``next program counter'' register, if defined.
3315 The number of the ``next next program counter'' register, if defined.
3316 Currently, this is only defined for the Motorola 88K.
3319 @findex PARM_BOUNDARY
3320 If non-zero, round arguments to a boundary of this many bits before
3321 pushing them on the stack.
3323 @item PRINT_REGISTER_HOOK (@var{regno})
3324 @findex PRINT_REGISTER_HOOK
3325 If defined, this must be a function that prints the contents of the
3326 given register to standard output.
3328 @item PRINT_TYPELESS_INTEGER
3329 @findex PRINT_TYPELESS_INTEGER
3330 This is an obscure substitute for @code{print_longest} that seems to
3331 have been defined for the Convex target.
3333 @item PROCESS_LINENUMBER_HOOK
3334 @findex PROCESS_LINENUMBER_HOOK
3335 A hook defined for XCOFF reading.
3337 @item PROLOGUE_FIRSTLINE_OVERLAP
3338 @findex PROLOGUE_FIRSTLINE_OVERLAP
3339 (Only used in unsupported Convex configuration.)
3343 If defined, this is the number of the processor status register. (This
3344 definition is only used in generic code when parsing "$ps".)
3348 @findex call_function_by_hand
3349 @findex return_command
3350 Used in @samp{call_function_by_hand} to remove an artificial stack
3351 frame and in @samp{return_command} to remove a real stack frame.
3353 @item PUSH_ARGUMENTS (@var{nargs}, @var{args}, @var{sp}, @var{struct_return}, @var{struct_addr})
3354 @findex PUSH_ARGUMENTS
3355 Define this to push arguments onto the stack for inferior function
3356 call. Returns the updated stack pointer value.
3358 @item PUSH_DUMMY_FRAME
3359 @findex PUSH_DUMMY_FRAME
3360 Used in @samp{call_function_by_hand} to create an artificial stack frame.
3362 @item REGISTER_BYTES
3363 @findex REGISTER_BYTES
3364 The total amount of space needed to store @value{GDBN}'s copy of the machine's
3367 @item REGISTER_NAME(@var{i})
3368 @findex REGISTER_NAME
3369 Return the name of register @var{i} as a string. May return @code{NULL}
3370 or @code{NUL} to indicate that register @var{i} is not valid.
3372 @item REGISTER_NAMES
3373 @findex REGISTER_NAMES
3374 Deprecated in favor of @code{REGISTER_NAME}.
3376 @item REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3377 @findex REG_STRUCT_HAS_ADDR
3378 Define this to return 1 if the given type will be passed by pointer
3379 rather than directly.
3381 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3382 @findex SAVE_DUMMY_FRAME_TOS
3383 Used in @samp{call_function_by_hand} to notify the target dependent code
3384 of the top-of-stack value that will be passed to the the inferior code.
3385 This is the value of the @code{SP} after both the dummy frame and space
3386 for parameters/results have been allocated on the stack.
3388 @item SDB_REG_TO_REGNUM
3389 @findex SDB_REG_TO_REGNUM
3390 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3391 defined, no conversion will be done.
3393 @item SHIFT_INST_REGS
3394 @findex SHIFT_INST_REGS
3395 (Only used for m88k targets.)
3397 @item SKIP_PERMANENT_BREAKPOINT
3398 @findex SKIP_PERMANENT_BREAKPOINT
3399 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3400 steps over a breakpoint by removing it, stepping one instruction, and
3401 re-inserting the breakpoint. However, permanent breakpoints are
3402 hardwired into the inferior, and can't be removed, so this strategy
3403 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
3404 state so that execution will resume just after the breakpoint. This
3405 macro does the right thing even when the breakpoint is in the delay slot
3406 of a branch or jump.
3408 @item SKIP_PROLOGUE (@var{pc})
3409 @findex SKIP_PROLOGUE
3410 A C expression that returns the address of the ``real'' code beyond the
3411 function entry prologue found at @var{pc}.
3413 @item SKIP_TRAMPOLINE_CODE (@var{pc})
3414 @findex SKIP_TRAMPOLINE_CODE
3415 If the target machine has trampoline code that sits between callers and
3416 the functions being called, then define this macro to return a new PC
3417 that is at the start of the real function.
3421 If the stack-pointer is kept in a register, then define this macro to be
3422 the number (greater than or equal to zero) of that register.
3424 This should only need to be defined if @code{TARGET_WRITE_SP} and
3425 @code{TARGET_WRITE_SP} are not defined.
3427 @item STAB_REG_TO_REGNUM
3428 @findex STAB_REG_TO_REGNUM
3429 Define this to convert stab register numbers (as gotten from `r'
3430 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
3433 @item STACK_ALIGN (@var{addr})
3435 Define this to adjust the address to the alignment required for the
3438 @item STEP_SKIPS_DELAY (@var{addr})
3439 @findex STEP_SKIPS_DELAY
3440 Define this to return true if the address is of an instruction with a
3441 delay slot. If a breakpoint has been placed in the instruction's delay
3442 slot, @value{GDBN} will single-step over that instruction before resuming
3443 normally. Currently only defined for the Mips.
3445 @item STORE_RETURN_VALUE (@var{type}, @var{valbuf})
3446 @findex STORE_RETURN_VALUE
3447 A C expression that stores a function return value of type @var{type},
3448 where @var{valbuf} is the address of the value to be stored.
3450 @item SUN_FIXED_LBRAC_BUG
3451 @findex SUN_FIXED_LBRAC_BUG
3452 (Used only for Sun-3 and Sun-4 targets.)
3454 @item SYMBOL_RELOADING_DEFAULT
3455 @findex SYMBOL_RELOADING_DEFAULT
3456 The default value of the ``symbol-reloading'' variable. (Never defined in
3459 @item TARGET_CHAR_BIT
3460 @findex TARGET_CHAR_BIT
3461 Number of bits in a char; defaults to 8.
3463 @item TARGET_CHAR_SIGNED
3464 @findex TARGET_CHAR_SIGNED
3465 Non-zero if @code{char} is normally signed on this architecture; zero if
3466 it should be unsigned.
3468 The ISO C standard requires the compiler to treat @code{char} as
3469 equivalent to either @code{signed char} or @code{unsigned char}; any
3470 character in the standard execution set is supposed to be positive.
3471 Most compilers treat @code{char} as signed, but @code{char} is unsigned
3472 on the IBM S/390, RS6000, and PowerPC targets.
3474 @item TARGET_COMPLEX_BIT
3475 @findex TARGET_COMPLEX_BIT
3476 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
3478 At present this macro is not used.
3480 @item TARGET_DOUBLE_BIT
3481 @findex TARGET_DOUBLE_BIT
3482 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
3484 @item TARGET_DOUBLE_COMPLEX_BIT
3485 @findex TARGET_DOUBLE_COMPLEX_BIT
3486 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
3488 At present this macro is not used.
3490 @item TARGET_FLOAT_BIT
3491 @findex TARGET_FLOAT_BIT
3492 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
3494 @item TARGET_INT_BIT
3495 @findex TARGET_INT_BIT
3496 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3498 @item TARGET_LONG_BIT
3499 @findex TARGET_LONG_BIT
3500 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3502 @item TARGET_LONG_DOUBLE_BIT
3503 @findex TARGET_LONG_DOUBLE_BIT
3504 Number of bits in a long double float;
3505 defaults to @code{2 * TARGET_DOUBLE_BIT}.
3507 @item TARGET_LONG_LONG_BIT
3508 @findex TARGET_LONG_LONG_BIT
3509 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
3511 @item TARGET_PTR_BIT
3512 @findex TARGET_PTR_BIT
3513 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
3515 @item TARGET_SHORT_BIT
3516 @findex TARGET_SHORT_BIT
3517 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
3519 @item TARGET_READ_PC
3520 @findex TARGET_READ_PC
3521 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
3522 @findex TARGET_WRITE_PC
3523 @itemx TARGET_READ_SP
3524 @findex TARGET_READ_SP
3525 @itemx TARGET_WRITE_SP
3526 @findex TARGET_WRITE_SP
3527 @itemx TARGET_READ_FP
3528 @findex TARGET_READ_FP
3534 These change the behavior of @code{read_pc}, @code{write_pc},
3535 @code{read_sp}, @code{write_sp} and @code{read_fp}. For most targets,
3536 these may be left undefined. @value{GDBN} will call the read and write
3537 register functions with the relevant @code{_REGNUM} argument.
3539 These macros are useful when a target keeps one of these registers in a
3540 hard to get at place; for example, part in a segment register and part
3541 in an ordinary register.
3543 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
3544 @findex TARGET_VIRTUAL_FRAME_POINTER
3545 Returns a @code{(register, offset)} pair representing the virtual
3546 frame pointer in use at the code address @var{pc}. If virtual
3547 frame pointers are not used, a default definition simply returns
3548 @code{FP_REGNUM}, with an offset of zero.
3550 @item TARGET_HAS_HARDWARE_WATCHPOINTS
3551 If non-zero, the target has support for hardware-assisted
3552 watchpoints. @xref{Algorithms, watchpoints}, for more details and
3553 other related macros.
3555 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
3556 @findex TARGET_PRINT_INSN
3557 This is the function used by @value{GDBN} to print an assembly
3558 instruction. It prints the instruction at address @var{addr} in
3559 debugged memory and returns the length of the instruction, in bytes. If
3560 a target doesn't define its own printing routine, it defaults to an
3561 accessor function for the global pointer @code{tm_print_insn}. This
3562 usually points to a function in the @code{opcodes} library (@pxref{Support
3563 Libraries, ,Opcodes}). @var{info} is a structure (of type
3564 @code{disassemble_info}) defined in @file{include/dis-asm.h} used to
3565 pass information to the instruction decoding routine.
3567 @item USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
3568 @findex USE_STRUCT_CONVENTION
3569 If defined, this must be an expression that is nonzero if a value of the
3570 given @var{type} being returned from a function must have space
3571 allocated for it on the stack. @var{gcc_p} is true if the function
3572 being considered is known to have been compiled by GCC; this is helpful
3573 for systems where GCC is known to use different calling convention than
3576 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3577 @findex VARIABLES_INSIDE_BLOCK
3578 For dbx-style debugging information, if the compiler puts variable
3579 declarations inside LBRAC/RBRAC blocks, this should be defined to be
3580 nonzero. @var{desc} is the value of @code{n_desc} from the
3581 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
3582 presence of either the @code{GCC_COMPILED_SYMBOL} or the
3583 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
3585 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3586 @findex OS9K_VARIABLES_INSIDE_BLOCK
3587 Similarly, for OS/9000. Defaults to 1.
3590 Motorola M68K target conditionals.
3594 Define this to be the 4-bit location of the breakpoint trap vector. If
3595 not defined, it will default to @code{0xf}.
3597 @item REMOTE_BPT_VECTOR
3598 Defaults to @code{1}.
3601 @section Adding a New Target
3603 @cindex adding a target
3604 The following files add a target to @value{GDBN}:
3608 @item gdb/config/@var{arch}/@var{ttt}.mt
3609 Contains a Makefile fragment specific to this target. Specifies what
3610 object files are needed for target @var{ttt}, by defining
3611 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
3612 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
3615 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
3616 but these are now deprecated, replaced by autoconf, and may go away in
3617 future versions of @value{GDBN}.
3619 @item gdb/@var{ttt}-tdep.c
3620 Contains any miscellaneous code required for this target machine. On
3621 some machines it doesn't exist at all. Sometimes the macros in
3622 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
3623 as functions here instead, and the macro is simply defined to call the
3624 function. This is vastly preferable, since it is easier to understand
3627 @item gdb/@var{arch}-tdep.c
3628 @itemx gdb/@var{arch}-tdep.h
3629 This often exists to describe the basic layout of the target machine's
3630 processor chip (registers, stack, etc.). If used, it is included by
3631 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
3634 @item gdb/config/@var{arch}/tm-@var{ttt}.h
3635 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
3636 macro definitions about the target machine's registers, stack frame
3637 format and instructions.
3639 New targets do not need this file and should not create it.
3641 @item gdb/config/@var{arch}/tm-@var{arch}.h
3642 This often exists to describe the basic layout of the target machine's
3643 processor chip (registers, stack, etc.). If used, it is included by
3644 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
3647 New targets do not need this file and should not create it.
3651 If you are adding a new operating system for an existing CPU chip, add a
3652 @file{config/tm-@var{os}.h} file that describes the operating system
3653 facilities that are unusual (extra symbol table info; the breakpoint
3654 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
3655 that just @code{#include}s @file{tm-@var{arch}.h} and
3656 @file{config/tm-@var{os}.h}.
3659 @node Target Vector Definition
3661 @chapter Target Vector Definition
3662 @cindex target vector
3664 The target vector defines the interface between @value{GDBN}'s
3665 abstract handling of target systems, and the nitty-gritty code that
3666 actually exercises control over a process or a serial port.
3667 @value{GDBN} includes some 30-40 different target vectors; however,
3668 each configuration of @value{GDBN} includes only a few of them.
3670 @section File Targets
3672 Both executables and core files have target vectors.
3674 @section Standard Protocol and Remote Stubs
3676 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
3677 that runs in the target system. @value{GDBN} provides several sample
3678 @dfn{stubs} that can be integrated into target programs or operating
3679 systems for this purpose; they are named @file{*-stub.c}.
3681 The @value{GDBN} user's manual describes how to put such a stub into
3682 your target code. What follows is a discussion of integrating the
3683 SPARC stub into a complicated operating system (rather than a simple
3684 program), by Stu Grossman, the author of this stub.
3686 The trap handling code in the stub assumes the following upon entry to
3691 %l1 and %l2 contain pc and npc respectively at the time of the trap;
3697 you are in the correct trap window.
3700 As long as your trap handler can guarantee those conditions, then there
3701 is no reason why you shouldn't be able to ``share'' traps with the stub.
3702 The stub has no requirement that it be jumped to directly from the
3703 hardware trap vector. That is why it calls @code{exceptionHandler()},
3704 which is provided by the external environment. For instance, this could
3705 set up the hardware traps to actually execute code which calls the stub
3706 first, and then transfers to its own trap handler.
3708 For the most point, there probably won't be much of an issue with
3709 ``sharing'' traps, as the traps we use are usually not used by the kernel,
3710 and often indicate unrecoverable error conditions. Anyway, this is all
3711 controlled by a table, and is trivial to modify. The most important
3712 trap for us is for @code{ta 1}. Without that, we can't single step or
3713 do breakpoints. Everything else is unnecessary for the proper operation
3714 of the debugger/stub.
3716 From reading the stub, it's probably not obvious how breakpoints work.
3717 They are simply done by deposit/examine operations from @value{GDBN}.
3719 @section ROM Monitor Interface
3721 @section Custom Protocols
3723 @section Transport Layer
3725 @section Builtin Simulator
3728 @node Native Debugging
3730 @chapter Native Debugging
3731 @cindex native debugging
3733 Several files control @value{GDBN}'s configuration for native support:
3737 @item gdb/config/@var{arch}/@var{xyz}.mh
3738 Specifies Makefile fragments needed by a @emph{native} configuration on
3739 machine @var{xyz}. In particular, this lists the required
3740 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
3741 Also specifies the header file which describes native support on
3742 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
3743 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
3744 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
3746 @emph{Maintainer's note: The @file{.mh} suffix is because this file
3747 originally contained @file{Makefile} fragments for hosting @value{GDBN}
3748 on machine @var{xyz}. While the file is no longer used for this
3749 purpose, the @file{.mh} suffix remains. Perhaps someone will
3750 eventually rename these fragments so that they have a @file{.mn}
3753 @item gdb/config/@var{arch}/nm-@var{xyz}.h
3754 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
3755 macro definitions describing the native system environment, such as
3756 child process control and core file support.
3758 @item gdb/@var{xyz}-nat.c
3759 Contains any miscellaneous C code required for this native support of
3760 this machine. On some machines it doesn't exist at all.
3763 There are some ``generic'' versions of routines that can be used by
3764 various systems. These can be customized in various ways by macros
3765 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
3766 the @var{xyz} host, you can just include the generic file's name (with
3767 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
3769 Otherwise, if your machine needs custom support routines, you will need
3770 to write routines that perform the same functions as the generic file.
3771 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
3772 into @code{NATDEPFILES}.
3776 This contains the @emph{target_ops vector} that supports Unix child
3777 processes on systems which use ptrace and wait to control the child.
3780 This contains the @emph{target_ops vector} that supports Unix child
3781 processes on systems which use /proc to control the child.
3784 This does the low-level grunge that uses Unix system calls to do a ``fork
3785 and exec'' to start up a child process.
3788 This is the low level interface to inferior processes for systems using
3789 the Unix @code{ptrace} call in a vanilla way.
3792 @section Native core file Support
3793 @cindex native core files
3796 @findex fetch_core_registers
3797 @item core-aout.c::fetch_core_registers()
3798 Support for reading registers out of a core file. This routine calls
3799 @code{register_addr()}, see below. Now that BFD is used to read core
3800 files, virtually all machines should use @code{core-aout.c}, and should
3801 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
3802 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
3804 @item core-aout.c::register_addr()
3805 If your @code{nm-@var{xyz}.h} file defines the macro
3806 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
3807 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
3808 register number @code{regno}. @code{blockend} is the offset within the
3809 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
3810 @file{core-aout.c} will define the @code{register_addr()} function and
3811 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
3812 you are using the standard @code{fetch_core_registers()}, you will need
3813 to define your own version of @code{register_addr()}, put it into your
3814 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
3815 the @code{NATDEPFILES} list. If you have your own
3816 @code{fetch_core_registers()}, you may not need a separate
3817 @code{register_addr()}. Many custom @code{fetch_core_registers()}
3818 implementations simply locate the registers themselves.@refill
3821 When making @value{GDBN} run native on a new operating system, to make it
3822 possible to debug core files, you will need to either write specific
3823 code for parsing your OS's core files, or customize
3824 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
3825 machine uses to define the struct of registers that is accessible
3826 (possibly in the u-area) in a core file (rather than
3827 @file{machine/reg.h}), and an include file that defines whatever header
3828 exists on a core file (e.g. the u-area or a @code{struct core}). Then
3829 modify @code{trad_unix_core_file_p} to use these values to set up the
3830 section information for the data segment, stack segment, any other
3831 segments in the core file (perhaps shared library contents or control
3832 information), ``registers'' segment, and if there are two discontiguous
3833 sets of registers (e.g. integer and float), the ``reg2'' segment. This
3834 section information basically delimits areas in the core file in a
3835 standard way, which the section-reading routines in BFD know how to seek
3838 Then back in @value{GDBN}, you need a matching routine called
3839 @code{fetch_core_registers}. If you can use the generic one, it's in
3840 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
3841 It will be passed a char pointer to the entire ``registers'' segment,
3842 its length, and a zero; or a char pointer to the entire ``regs2''
3843 segment, its length, and a 2. The routine should suck out the supplied
3844 register values and install them into @value{GDBN}'s ``registers'' array.
3846 If your system uses @file{/proc} to control processes, and uses ELF
3847 format core files, then you may be able to use the same routines for
3848 reading the registers out of processes and out of core files.
3856 @section shared libraries
3858 @section Native Conditionals
3859 @cindex native conditionals
3861 When @value{GDBN} is configured and compiled, various macros are
3862 defined or left undefined, to control compilation when the host and
3863 target systems are the same. These macros should be defined (or left
3864 undefined) in @file{nm-@var{system}.h}.
3868 @findex ATTACH_DETACH
3869 If defined, then @value{GDBN} will include support for the @code{attach} and
3870 @code{detach} commands.
3872 @item CHILD_PREPARE_TO_STORE
3873 @findex CHILD_PREPARE_TO_STORE
3874 If the machine stores all registers at once in the child process, then
3875 define this to ensure that all values are correct. This usually entails
3876 a read from the child.
3878 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
3881 @item FETCH_INFERIOR_REGISTERS
3882 @findex FETCH_INFERIOR_REGISTERS
3883 Define this if the native-dependent code will provide its own routines
3884 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
3885 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
3886 @file{infptrace.c} is included in this configuration, the default
3887 routines in @file{infptrace.c} are used for these functions.
3889 @item FILES_INFO_HOOK
3890 @findex FILES_INFO_HOOK
3891 (Only defined for Convex.)
3895 This macro is normally defined to be the number of the first floating
3896 point register, if the machine has such registers. As such, it would
3897 appear only in target-specific code. However, @file{/proc} support uses this
3898 to decide whether floats are in use on this target.
3900 @item GET_LONGJMP_TARGET
3901 @findex GET_LONGJMP_TARGET
3902 For most machines, this is a target-dependent parameter. On the
3903 DECstation and the Iris, this is a native-dependent parameter, since
3904 @file{setjmp.h} is needed to define it.
3906 This macro determines the target PC address that @code{longjmp} will jump to,
3907 assuming that we have just stopped at a longjmp breakpoint. It takes a
3908 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3909 pointer. It examines the current state of the machine as needed.
3911 @item I386_USE_GENERIC_WATCHPOINTS
3912 An x86-based machine can define this to use the generic x86 watchpoint
3913 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3916 @findex KERNEL_U_ADDR
3917 Define this to the address of the @code{u} structure (the ``user
3918 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
3919 needs to know this so that it can subtract this address from absolute
3920 addresses in the upage, that are obtained via ptrace or from core files.
3921 On systems that don't need this value, set it to zero.
3923 @item KERNEL_U_ADDR_BSD
3924 @findex KERNEL_U_ADDR_BSD
3925 Define this to cause @value{GDBN} to determine the address of @code{u} at
3926 runtime, by using Berkeley-style @code{nlist} on the kernel's image in
3929 @item KERNEL_U_ADDR_HPUX
3930 @findex KERNEL_U_ADDR_HPUX
3931 Define this to cause @value{GDBN} to determine the address of @code{u} at
3932 runtime, by using HP-style @code{nlist} on the kernel's image in the
3935 @item ONE_PROCESS_WRITETEXT
3936 @findex ONE_PROCESS_WRITETEXT
3937 Define this to be able to, when a breakpoint insertion fails, warn the
3938 user that another process may be running with the same executable.
3940 @item PREPARE_TO_PROCEED (@var{select_it})
3941 @findex PREPARE_TO_PROCEED
3942 This (ugly) macro allows a native configuration to customize the way the
3943 @code{proceed} function in @file{infrun.c} deals with switching between
3946 In a multi-threaded task we may select another thread and then continue
3947 or step. But if the old thread was stopped at a breakpoint, it will
3948 immediately cause another breakpoint stop without any execution (i.e. it
3949 will report a breakpoint hit incorrectly). So @value{GDBN} must step over it
3952 If defined, @code{PREPARE_TO_PROCEED} should check the current thread
3953 against the thread that reported the most recent event. If a step-over
3954 is required, it returns TRUE. If @var{select_it} is non-zero, it should
3955 reselect the old thread.
3958 @findex PROC_NAME_FMT
3959 Defines the format for the name of a @file{/proc} device. Should be
3960 defined in @file{nm.h} @emph{only} in order to override the default
3961 definition in @file{procfs.c}.
3964 @findex PTRACE_FP_BUG
3965 See @file{mach386-xdep.c}.
3967 @item PTRACE_ARG3_TYPE
3968 @findex PTRACE_ARG3_TYPE
3969 The type of the third argument to the @code{ptrace} system call, if it
3970 exists and is different from @code{int}.
3972 @item REGISTER_U_ADDR
3973 @findex REGISTER_U_ADDR
3974 Defines the offset of the registers in the ``u area''.
3976 @item SHELL_COMMAND_CONCAT
3977 @findex SHELL_COMMAND_CONCAT
3978 If defined, is a string to prefix on the shell command used to start the
3983 If defined, this is the name of the shell to use to run the inferior.
3984 Defaults to @code{"/bin/sh"}.
3986 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
3988 Define this to expand into an expression that will cause the symbols in
3989 @var{filename} to be added to @value{GDBN}'s symbol table. If
3990 @var{readsyms} is zero symbols are not read but any necessary low level
3991 processing for @var{filename} is still done.
3993 @item SOLIB_CREATE_INFERIOR_HOOK
3994 @findex SOLIB_CREATE_INFERIOR_HOOK
3995 Define this to expand into any shared-library-relocation code that you
3996 want to be run just after the child process has been forked.
3998 @item START_INFERIOR_TRAPS_EXPECTED
3999 @findex START_INFERIOR_TRAPS_EXPECTED
4000 When starting an inferior, @value{GDBN} normally expects to trap
4002 the shell execs, and once when the program itself execs. If the actual
4003 number of traps is something other than 2, then define this macro to
4004 expand into the number expected.
4006 @item SVR4_SHARED_LIBS
4007 @findex SVR4_SHARED_LIBS
4008 Define this to indicate that SVR4-style shared libraries are in use.
4012 This determines whether small routines in @file{*-tdep.c}, which
4013 translate register values between @value{GDBN}'s internal
4014 representation and the @file{/proc} representation, are compiled.
4017 @findex U_REGS_OFFSET
4018 This is the offset of the registers in the upage. It need only be
4019 defined if the generic ptrace register access routines in
4020 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4021 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4022 the default value from @file{infptrace.c} is good enough, leave it
4025 The default value means that u.u_ar0 @emph{points to} the location of
4026 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4027 that @code{u.u_ar0} @emph{is} the location of the registers.
4031 See @file{objfiles.c}.
4034 @findex DEBUG_PTRACE
4035 Define this to debug @code{ptrace} calls.
4039 @node Support Libraries
4041 @chapter Support Libraries
4046 BFD provides support for @value{GDBN} in several ways:
4049 @item identifying executable and core files
4050 BFD will identify a variety of file types, including a.out, coff, and
4051 several variants thereof, as well as several kinds of core files.
4053 @item access to sections of files
4054 BFD parses the file headers to determine the names, virtual addresses,
4055 sizes, and file locations of all the various named sections in files
4056 (such as the text section or the data section). @value{GDBN} simply
4057 calls BFD to read or write section @var{x} at byte offset @var{y} for
4060 @item specialized core file support
4061 BFD provides routines to determine the failing command name stored in a
4062 core file, the signal with which the program failed, and whether a core
4063 file matches (i.e.@: could be a core dump of) a particular executable
4066 @item locating the symbol information
4067 @value{GDBN} uses an internal interface of BFD to determine where to find the
4068 symbol information in an executable file or symbol-file. @value{GDBN} itself
4069 handles the reading of symbols, since BFD does not ``understand'' debug
4070 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4075 @cindex opcodes library
4077 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4078 library because it's also used in binutils, for @file{objdump}).
4087 @cindex regular expressions library
4098 @item SIGN_EXTEND_CHAR
4100 @item SWITCH_ENUM_BUG
4115 This chapter covers topics that are lower-level than the major
4116 algorithms of @value{GDBN}.
4121 Cleanups are a structured way to deal with things that need to be done
4124 When your code does something (e.g., @code{xmalloc} some memory, or
4125 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4126 the memory or @code{close} the file), it can make a cleanup. The
4127 cleanup will be done at some future point: when the command is finished
4128 and control returns to the top level; when an error occurs and the stack
4129 is unwound; or when your code decides it's time to explicitly perform
4130 cleanups. Alternatively you can elect to discard the cleanups you
4136 @item struct cleanup *@var{old_chain};
4137 Declare a variable which will hold a cleanup chain handle.
4139 @findex make_cleanup
4140 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
4141 Make a cleanup which will cause @var{function} to be called with
4142 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
4143 handle that can later be passed to @code{do_cleanups} or
4144 @code{discard_cleanups}. Unless you are going to call
4145 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
4146 from @code{make_cleanup}.
4149 @item do_cleanups (@var{old_chain});
4150 Do all cleanups added to the chain since the corresponding
4151 @code{make_cleanup} call was made.
4153 @findex discard_cleanups
4154 @item discard_cleanups (@var{old_chain});
4155 Same as @code{do_cleanups} except that it just removes the cleanups from
4156 the chain and does not call the specified functions.
4159 Cleanups are implemented as a chain. The handle returned by
4160 @code{make_cleanups} includes the cleanup passed to the call and any
4161 later cleanups appended to the chain (but not yet discarded or
4165 make_cleanup (a, 0);
4167 struct cleanup *old = make_cleanup (b, 0);
4175 will call @code{c()} and @code{b()} but will not call @code{a()}. The
4176 cleanup that calls @code{a()} will remain in the cleanup chain, and will
4177 be done later unless otherwise discarded.@refill
4179 Your function should explicitly do or discard the cleanups it creates.
4180 Failing to do this leads to non-deterministic behavior since the caller
4181 will arbitrarily do or discard your functions cleanups. This need leads
4182 to two common cleanup styles.
4184 The first style is try/finally. Before it exits, your code-block calls
4185 @code{do_cleanups} with the old cleanup chain and thus ensures that your
4186 code-block's cleanups are always performed. For instance, the following
4187 code-segment avoids a memory leak problem (even when @code{error} is
4188 called and a forced stack unwind occurs) by ensuring that the
4189 @code{xfree} will always be called:
4192 struct cleanup *old = make_cleanup (null_cleanup, 0);
4193 data = xmalloc (sizeof blah);
4194 make_cleanup (xfree, data);
4199 The second style is try/except. Before it exits, your code-block calls
4200 @code{discard_cleanups} with the old cleanup chain and thus ensures that
4201 any created cleanups are not performed. For instance, the following
4202 code segment, ensures that the file will be closed but only if there is
4206 FILE *file = fopen ("afile", "r");
4207 struct cleanup *old = make_cleanup (close_file, file);
4209 discard_cleanups (old);
4213 Some functions, e.g. @code{fputs_filtered()} or @code{error()}, specify
4214 that they ``should not be called when cleanups are not in place''. This
4215 means that any actions you need to reverse in the case of an error or
4216 interruption must be on the cleanup chain before you call these
4217 functions, since they might never return to your code (they
4218 @samp{longjmp} instead).
4220 @section Wrapping Output Lines
4221 @cindex line wrap in output
4224 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
4225 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
4226 added in places that would be good breaking points. The utility
4227 routines will take care of actually wrapping if the line width is
4230 The argument to @code{wrap_here} is an indentation string which is
4231 printed @emph{only} if the line breaks there. This argument is saved
4232 away and used later. It must remain valid until the next call to
4233 @code{wrap_here} or until a newline has been printed through the
4234 @code{*_filtered} functions. Don't pass in a local variable and then
4237 It is usually best to call @code{wrap_here} after printing a comma or
4238 space. If you call it before printing a space, make sure that your
4239 indentation properly accounts for the leading space that will print if
4240 the line wraps there.
4242 Any function or set of functions that produce filtered output must
4243 finish by printing a newline, to flush the wrap buffer, before switching
4244 to unfiltered (@code{printf}) output. Symbol reading routines that
4245 print warnings are a good example.
4247 @section @value{GDBN} Coding Standards
4248 @cindex coding standards
4250 @value{GDBN} follows the GNU coding standards, as described in
4251 @file{etc/standards.texi}. This file is also available for anonymous
4252 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
4253 of the standard; in general, when the GNU standard recommends a practice
4254 but does not require it, @value{GDBN} requires it.
4256 @value{GDBN} follows an additional set of coding standards specific to
4257 @value{GDBN}, as described in the following sections.
4262 @value{GDBN} assumes an ISO-C compliant compiler.
4264 @value{GDBN} does not assume an ISO-C or POSIX compliant C library.
4267 @subsection Memory Management
4269 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
4270 @code{calloc}, @code{free} and @code{asprintf}.
4272 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
4273 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
4274 these functions do not return when the memory pool is empty. Instead,
4275 they unwind the stack using cleanups. These functions return
4276 @code{NULL} when requested to allocate a chunk of memory of size zero.
4278 @emph{Pragmatics: By using these functions, the need to check every
4279 memory allocation is removed. These functions provide portable
4282 @value{GDBN} does not use the function @code{free}.
4284 @value{GDBN} uses the function @code{xfree} to return memory to the
4285 memory pool. Consistent with ISO-C, this function ignores a request to
4286 free a @code{NULL} pointer.
4288 @emph{Pragmatics: On some systems @code{free} fails when passed a
4289 @code{NULL} pointer.}
4291 @value{GDBN} can use the non-portable function @code{alloca} for the
4292 allocation of small temporary values (such as strings).
4294 @emph{Pragmatics: This function is very non-portable. Some systems
4295 restrict the memory being allocated to no more than a few kilobytes.}
4297 @value{GDBN} uses the string function @code{xstrdup} and the print
4298 function @code{xasprintf}.
4300 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
4301 functions such as @code{sprintf} are very prone to buffer overflow
4305 @subsection Compiler Warnings
4306 @cindex compiler warnings
4308 With few exceptions, developers should include the configuration option
4309 @samp{--enable-gdb-build-warnings=,-Werror} when building @value{GDBN}.
4310 The exceptions are listed in the file @file{gdb/MAINTAINERS}.
4312 This option causes @value{GDBN} (when built using GCC) to be compiled
4313 with a carefully selected list of compiler warning flags. Any warnings
4314 from those flags being treated as errors.
4316 The current list of warning flags includes:
4320 Since @value{GDBN} coding standard requires all functions to be declared
4321 using a prototype, the flag has the side effect of ensuring that
4322 prototyped functions are always visible with out resorting to
4323 @samp{-Wstrict-prototypes}.
4326 Such code often appears to work except on instruction set architectures
4327 that use register windows.
4334 Since @value{GDBN} uses the @code{format printf} attribute on all
4335 @code{printf} like functions this checks not just @code{printf} calls
4336 but also calls to functions such as @code{fprintf_unfiltered}.
4339 This warning includes uses of the assignment operator within an
4340 @code{if} statement.
4342 @item -Wpointer-arith
4344 @item -Wuninitialized
4347 @emph{Pragmatics: Due to the way that @value{GDBN} is implemented most
4348 functions have unused parameters. Consequently the warning
4349 @samp{-Wunused-parameter} is precluded from the list. The macro
4350 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
4351 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
4352 is being used. The options @samp{-Wall} and @samp{-Wunused} are also
4353 precluded because they both include @samp{-Wunused-parameter}.}
4355 @emph{Pragmatics: @value{GDBN} has not simply accepted the warnings
4356 enabled by @samp{-Wall -Werror -W...}. Instead it is selecting warnings
4357 when and where their benefits can be demonstrated.}
4359 @subsection Formatting
4361 @cindex source code formatting
4362 The standard GNU recommendations for formatting must be followed
4365 A function declaration should not have its name in column zero. A
4366 function definition should have its name in column zero.
4370 static void foo (void);
4378 @emph{Pragmatics: This simplifies scripting. Function definitions can
4379 be found using @samp{^function-name}.}
4381 There must be a space between a function or macro name and the opening
4382 parenthesis of its argument list (except for macro definitions, as
4383 required by C). There must not be a space after an open paren/bracket
4384 or before a close paren/bracket.
4386 While additional whitespace is generally helpful for reading, do not use
4387 more than one blank line to separate blocks, and avoid adding whitespace
4388 after the end of a program line (as of 1/99, some 600 lines had
4389 whitespace after the semicolon). Excess whitespace causes difficulties
4390 for @code{diff} and @code{patch} utilities.
4392 Pointers are declared using the traditional K&R C style:
4406 @subsection Comments
4408 @cindex comment formatting
4409 The standard GNU requirements on comments must be followed strictly.
4411 Block comments must appear in the following form, with no @code{/*}- or
4412 @code{*/}-only lines, and no leading @code{*}:
4415 /* Wait for control to return from inferior to debugger. If inferior
4416 gets a signal, we may decide to start it up again instead of
4417 returning. That is why there is a loop in this function. When
4418 this function actually returns it means the inferior should be left
4419 stopped and @value{GDBN} should read more commands. */
4422 (Note that this format is encouraged by Emacs; tabbing for a multi-line
4423 comment works correctly, and @kbd{M-q} fills the block consistently.)
4425 Put a blank line between the block comments preceding function or
4426 variable definitions, and the definition itself.
4428 In general, put function-body comments on lines by themselves, rather
4429 than trying to fit them into the 20 characters left at the end of a
4430 line, since either the comment or the code will inevitably get longer
4431 than will fit, and then somebody will have to move it anyhow.
4435 @cindex C data types
4436 Code must not depend on the sizes of C data types, the format of the
4437 host's floating point numbers, the alignment of anything, or the order
4438 of evaluation of expressions.
4440 @cindex function usage
4441 Use functions freely. There are only a handful of compute-bound areas
4442 in @value{GDBN} that might be affected by the overhead of a function
4443 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
4444 limited by the target interface (whether serial line or system call).
4446 However, use functions with moderation. A thousand one-line functions
4447 are just as hard to understand as a single thousand-line function.
4449 @emph{Macros are bad, M'kay.}
4450 (But if you have to use a macro, make sure that the macro arguments are
4451 protected with parentheses.)
4455 Declarations like @samp{struct foo *} should be used in preference to
4456 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
4459 @subsection Function Prototypes
4460 @cindex function prototypes
4462 Prototypes must be used when both @emph{declaring} and @emph{defining}
4463 a function. Prototypes for @value{GDBN} functions must include both the
4464 argument type and name, with the name matching that used in the actual
4465 function definition.
4467 All external functions should have a declaration in a header file that
4468 callers include, except for @code{_initialize_*} functions, which must
4469 be external so that @file{init.c} construction works, but shouldn't be
4470 visible to random source files.
4472 Where a source file needs a forward declaration of a static function,
4473 that declaration must appear in a block near the top of the source file.
4476 @subsection Internal Error Recovery
4478 During its execution, @value{GDBN} can encounter two types of errors.
4479 User errors and internal errors. User errors include not only a user
4480 entering an incorrect command but also problems arising from corrupt
4481 object files and system errors when interacting with the target.
4482 Internal errors include situations where @value{GDBN} has detected, at
4483 run time, a corrupt or erroneous situation.
4485 When reporting an internal error, @value{GDBN} uses
4486 @code{internal_error} and @code{gdb_assert}.
4488 @value{GDBN} must not call @code{abort} or @code{assert}.
4490 @emph{Pragmatics: There is no @code{internal_warning} function. Either
4491 the code detected a user error, recovered from it and issued a
4492 @code{warning} or the code failed to correctly recover from the user
4493 error and issued an @code{internal_error}.}
4495 @subsection File Names
4497 Any file used when building the core of @value{GDBN} must be in lower
4498 case. Any file used when building the core of @value{GDBN} must be 8.3
4499 unique. These requirements apply to both source and generated files.
4501 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
4502 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
4503 is introduced to the build process both @file{Makefile.in} and
4504 @file{configure.in} need to be modified accordingly. Compare the
4505 convoluted conversion process needed to transform @file{COPYING} into
4506 @file{copying.c} with the conversion needed to transform
4507 @file{version.in} into @file{version.c}.}
4509 Any file non 8.3 compliant file (that is not used when building the core
4510 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
4512 @emph{Pragmatics: This is clearly a compromise.}
4514 When @value{GDBN} has a local version of a system header file (ex
4515 @file{string.h}) the file name based on the POSIX header prefixed with
4516 @file{gdb_} (@file{gdb_string.h}).
4518 For other files @samp{-} is used as the separator.
4521 @subsection Include Files
4523 All @file{.c} files should include @file{defs.h} first.
4525 All @file{.c} files should explicitly include the headers for any
4526 declarations they refer to. They should not rely on files being
4527 included indirectly.
4529 With the exception of the global definitions supplied by @file{defs.h},
4530 a header file should explicitly include the header declaring any
4531 @code{typedefs} et.al.@: it refers to.
4533 @code{extern} declarations should never appear in @code{.c} files.
4535 All include files should be wrapped in:
4538 #ifndef INCLUDE_FILE_NAME_H
4539 #define INCLUDE_FILE_NAME_H
4545 @subsection Clean Design and Portable Implementation
4548 In addition to getting the syntax right, there's the little question of
4549 semantics. Some things are done in certain ways in @value{GDBN} because long
4550 experience has shown that the more obvious ways caused various kinds of
4553 @cindex assumptions about targets
4554 You can't assume the byte order of anything that comes from a target
4555 (including @var{value}s, object files, and instructions). Such things
4556 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
4557 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
4558 such as @code{bfd_get_32}.
4560 You can't assume that you know what interface is being used to talk to
4561 the target system. All references to the target must go through the
4562 current @code{target_ops} vector.
4564 You can't assume that the host and target machines are the same machine
4565 (except in the ``native'' support modules). In particular, you can't
4566 assume that the target machine's header files will be available on the
4567 host machine. Target code must bring along its own header files --
4568 written from scratch or explicitly donated by their owner, to avoid
4572 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
4573 to write the code portably than to conditionalize it for various
4576 @cindex system dependencies
4577 New @code{#ifdef}'s which test for specific compilers or manufacturers
4578 or operating systems are unacceptable. All @code{#ifdef}'s should test
4579 for features. The information about which configurations contain which
4580 features should be segregated into the configuration files. Experience
4581 has proven far too often that a feature unique to one particular system
4582 often creeps into other systems; and that a conditional based on some
4583 predefined macro for your current system will become worthless over
4584 time, as new versions of your system come out that behave differently
4585 with regard to this feature.
4587 Adding code that handles specific architectures, operating systems,
4588 target interfaces, or hosts, is not acceptable in generic code.
4590 @cindex portable file name handling
4591 @cindex file names, portability
4592 One particularly notorious area where system dependencies tend to
4593 creep in is handling of file names. The mainline @value{GDBN} code
4594 assumes Posix semantics of file names: absolute file names begin with
4595 a forward slash @file{/}, slashes are used to separate leading
4596 directories, case-sensitive file names. These assumptions are not
4597 necessarily true on non-Posix systems such as MS-Windows. To avoid
4598 system-dependent code where you need to take apart or construct a file
4599 name, use the following portable macros:
4602 @findex HAVE_DOS_BASED_FILE_SYSTEM
4603 @item HAVE_DOS_BASED_FILE_SYSTEM
4604 This preprocessing symbol is defined to a non-zero value on hosts
4605 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
4606 symbol to write conditional code which should only be compiled for
4609 @findex IS_DIR_SEPARATOR
4610 @item IS_DIR_SEPARATOR (@var{c})
4611 Evaluates to a non-zero value if @var{c} is a directory separator
4612 character. On Unix and GNU/Linux systems, only a slash @file{/} is
4613 such a character, but on Windows, both @file{/} and @file{\} will
4616 @findex IS_ABSOLUTE_PATH
4617 @item IS_ABSOLUTE_PATH (@var{file})
4618 Evaluates to a non-zero value if @var{file} is an absolute file name.
4619 For Unix and GNU/Linux hosts, a name which begins with a slash
4620 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
4621 @file{x:\bar} are also absolute file names.
4623 @findex FILENAME_CMP
4624 @item FILENAME_CMP (@var{f1}, @var{f2})
4625 Calls a function which compares file names @var{f1} and @var{f2} as
4626 appropriate for the underlying host filesystem. For Posix systems,
4627 this simply calls @code{strcmp}; on case-insensitive filesystems it
4628 will call @code{strcasecmp} instead.
4630 @findex DIRNAME_SEPARATOR
4631 @item DIRNAME_SEPARATOR
4632 Evaluates to a character which separates directories in
4633 @code{PATH}-style lists, typically held in environment variables.
4634 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
4636 @findex SLASH_STRING
4638 This evaluates to a constant string you should use to produce an
4639 absolute filename from leading directories and the file's basename.
4640 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
4641 @code{"\\"} for some Windows-based ports.
4644 In addition to using these macros, be sure to use portable library
4645 functions whenever possible. For example, to extract a directory or a
4646 basename part from a file name, use the @code{dirname} and
4647 @code{basename} library functions (available in @code{libiberty} for
4648 platforms which don't provide them), instead of searching for a slash
4649 with @code{strrchr}.
4651 Another way to generalize @value{GDBN} along a particular interface is with an
4652 attribute struct. For example, @value{GDBN} has been generalized to handle
4653 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
4654 by defining the @code{target_ops} structure and having a current target (as
4655 well as a stack of targets below it, for memory references). Whenever
4656 something needs to be done that depends on which remote interface we are
4657 using, a flag in the current target_ops structure is tested (e.g.,
4658 @code{target_has_stack}), or a function is called through a pointer in the
4659 current target_ops structure. In this way, when a new remote interface
4660 is added, only one module needs to be touched---the one that actually
4661 implements the new remote interface. Other examples of
4662 attribute-structs are BFD access to multiple kinds of object file
4663 formats, or @value{GDBN}'s access to multiple source languages.
4665 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
4666 the code interfacing between @code{ptrace} and the rest of
4667 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
4668 something was very painful. In @value{GDBN} 4.x, these have all been
4669 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
4670 with variations between systems the same way any system-independent
4671 file would (hooks, @code{#if defined}, etc.), and machines which are
4672 radically different don't need to use @file{infptrace.c} at all.
4674 All debugging code must be controllable using the @samp{set debug
4675 @var{module}} command. Do not use @code{printf} to print trace
4676 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
4677 @code{#ifdef DEBUG}.
4682 @chapter Porting @value{GDBN}
4683 @cindex porting to new machines
4685 Most of the work in making @value{GDBN} compile on a new machine is in
4686 specifying the configuration of the machine. This is done in a
4687 dizzying variety of header files and configuration scripts, which we
4688 hope to make more sensible soon. Let's say your new host is called an
4689 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
4690 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
4691 @samp{sparc-sun-sunos4}). In particular:
4695 In the top level directory, edit @file{config.sub} and add @var{arch},
4696 @var{xvend}, and @var{xos} to the lists of supported architectures,
4697 vendors, and operating systems near the bottom of the file. Also, add
4698 @var{xyz} as an alias that maps to
4699 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
4703 ./config.sub @var{xyz}
4710 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
4714 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
4715 and no error messages.
4718 You need to port BFD, if that hasn't been done already. Porting BFD is
4719 beyond the scope of this manual.
4722 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
4723 your system and set @code{gdb_host} to @var{xyz}, and (unless your
4724 desired target is already available) also edit @file{gdb/configure.tgt},
4725 setting @code{gdb_target} to something appropriate (for instance,
4728 @emph{Maintainer's note: Work in progress. The file
4729 @file{gdb/configure.host} originally needed to be modified when either a
4730 new native target or a new host machine was being added to @value{GDBN}.
4731 Recent changes have removed this requirement. The file now only needs
4732 to be modified when adding a new native configuration. This will likely
4733 changed again in the future.}
4736 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
4737 target-dependent @file{.h} and @file{.c} files used for your
4741 @section Configuring @value{GDBN} for Release
4743 @cindex preparing a release
4744 @cindex making a distribution tarball
4745 From the top level directory (containing @file{gdb}, @file{bfd},
4746 @file{libiberty}, and so on):
4749 make -f Makefile.in gdb.tar.gz
4753 This will properly configure, clean, rebuild any files that are
4754 distributed pre-built (e.g. @file{c-exp.tab.c} or @file{refcard.ps}),
4755 and will then make a tarfile. (If the top level directory has already
4756 been configured, you can just do @code{make gdb.tar.gz} instead.)
4758 This procedure requires:
4766 @code{makeinfo} (texinfo2 level);
4775 @code{yacc} or @code{bison}.
4779 @dots{} and the usual slew of utilities (@code{sed}, @code{tar}, etc.).
4781 @subheading TEMPORARY RELEASE PROCEDURE FOR DOCUMENTATION
4783 @file{gdb.texinfo} is currently marked up using the texinfo-2 macros,
4784 which are not yet a default for anything (but we have to start using
4787 For making paper, the only thing this implies is the right generation of
4788 @file{texinfo.tex} needs to be included in the distribution.
4790 For making info files, however, rather than duplicating the texinfo2
4791 distribution, generate @file{gdb-all.texinfo} locally, and include the
4792 files @file{gdb.info*} in the distribution. Note the plural;
4793 @code{makeinfo} will split the document into one overall file and five
4794 or so included files.
4799 @chapter Releasing @value{GDBN}
4800 @cindex making a new release of gdb
4802 @section Versions and Branches
4804 @subsection Version Identifiers
4806 @value{GDBN}'s version is determined by the file @file{gdb/version.in}.
4808 @value{GDBN}'s mainline uses ISO dates to differentiate between
4809 versions. The CVS repository uses @var{YYYY}-@var{MM}-@var{DD}-cvs
4810 while the corresponding snapshot uses @var{YYYYMMDD}.
4812 @value{GDBN}'s release branch uses a slightly more complicated scheme.
4813 When the branch is first cut, the mainline version identifier is
4814 prefixed with the @var{major}.@var{minor} from of the previous release
4815 series but with .90 appended. As draft releases are drawn from the
4816 branch, the minor minor number (.90) is incremented. Once the first
4817 release (@var{M}.@var{N}) has been made, the version prefix is updated
4818 to @var{M}.@var{N}.0.90 (dot zero, dot ninety). Follow on releases have
4819 an incremented minor minor version number (.0).
4821 Using 5.1 (previous) and 5.2 (current), the example below illustrates a
4822 typical sequence of version identifiers:
4826 final release from previous branch
4827 @item 2002-03-03-cvs
4828 main-line the day the branch is cut
4829 @item 5.1.90-2002-03-03-cvs
4830 corresponding branch version
4832 first draft release candidate
4833 @item 5.1.91-2002-03-17-cvs
4834 updated branch version
4836 second draft release candidate
4837 @item 5.1.92-2002-03-31-cvs
4838 updated branch version
4840 final release candidate (see below)
4843 @item 5.2.0.90-2002-04-07-cvs
4844 updated CVS branch version
4846 second official release
4853 Minor minor minor draft release candidates such as 5.2.0.91 have been
4854 omitted from the example. Such release candidates are, typically, never
4857 For 5.1.93 the bziped tar ball @file{gdb-5.1.93.tar.bz2} is just the
4858 official @file{gdb-5.2.tar} renamed and compressed.
4861 To avoid version conflicts, vendors are expected to modify the file
4862 @file{gdb/version.in} to include a vendor unique alphabetic identifier
4863 (an official @value{GDBN} release never uses alphabetic characters in
4864 its version identifer).
4866 Since @value{GDBN} does not make minor minor minor releases (e.g.,
4867 5.1.0.1) the conflict between that and a minor minor draft release
4868 identifier (e.g., 5.1.0.90) is avoided.
4871 @subsection Branches
4873 @value{GDBN} draws a release series (5.2, 5.2.1, @dots{}) from a single
4874 release branch (gdb_5_2-branch). Since minor minor minor releases
4875 (5.1.0.1) are not made, the need to branch the release branch is avoided
4876 (it also turns out that the effort required for such a a branch and
4877 release is significantly greater than the effort needed to create a new
4878 release from the head of the release branch).
4880 Releases 5.0 and 5.1 used branch and release tags of the form:
4883 gdb_N_M-YYYY-MM-DD-branchpoint
4884 gdb_N_M-YYYY-MM-DD-branch
4885 gdb_M_N-YYYY-MM-DD-release
4888 Release 5.2 is trialing the branch and release tags:
4891 gdb_N_M-YYYY-MM-DD-branchpoint
4893 gdb_M_N-YYYY-MM-DD-release
4896 @emph{Pragmatics: The branchpoint and release tags need to identify when
4897 a branch and release are made. The branch tag, denoting the head of the
4898 branch, does not have this criteria.}
4901 @section Branch Commit Policy
4903 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
4904 5.1 and 5.2 all used the below:
4908 The @file{gdb/MAINTAINERS} file still holds.
4910 Don't fix something on the branch unless/until it is also fixed in the
4911 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
4912 file is better than committing a hack.
4914 When considering a patch for the branch, suggested criteria include:
4915 Does it fix a build? Does it fix the sequence @kbd{break main; run}
4916 when debugging a static binary?
4918 The further a change is from the core of @value{GDBN}, the less likely
4919 the change will worry anyone (e.g., target specific code).
4921 Only post a proposal to change the core of @value{GDBN} after you've
4922 sent individual bribes to all the people listed in the
4923 @file{MAINTAINERS} file @t{;-)}
4926 @emph{Pragmatics: Provided updates are restricted to non-core
4927 functionality there is little chance that a broken change will be fatal.
4928 This means that changes such as adding a new architectures or (within
4929 reason) support for a new host are considered acceptable.}
4932 @section Obsolete any code
4934 Before anything else, poke the other developers (and around the source
4935 code) to see if there is anything that can be removed from @value{GDBN}
4936 (an old target, an unused file).
4938 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
4939 line. Doing this means that it is easy to identify obsolete code when
4940 grepping through the sources.
4942 The process has a number of steps and is intentionally slow --- this is
4943 to mainly ensure that people have had a reasonable chance to respond.
4944 Remember, everything on the internet takes a week.
4948 announce the change on @email{gdb@@sources.redhat.com, GDB mailing list}
4952 announce the change on @email{gdb-announce@@sources.redhat.com, GDB
4953 Announcement mailing list}
4957 go through and edit all relevant files and lines (e.g., in
4958 @file{configure.tgt}) so that they are prefixed with the word
4962 @emph{Maintainer note: Removing old code, while regrettable, is a good
4963 thing. Firstly it helps the developers by removing code that is either
4964 no longer relevant or simply wrong. Secondly since it removes any
4965 history associated with the file (effectively clearing the slate) the
4966 developer has a much freer hand when it comes to fixing broken files.}
4969 @section Before the Branch
4971 The most important objective at this stage is to find and fix simple
4972 changes that become a pain to track once the branch is created. For
4973 instance, configuration problems that stop @value{GDBN} from even
4974 building. If you can't get the problem fixed, document it in the
4975 @file{gdb/PROBLEMS} file.
4977 @subheading Prompt for @file{gdb/NEWS}
4979 People always forget. Send a post reminding them but also if you know
4980 something interesting happened add it yourself. The @code{schedule}
4981 script will mention this in its e-mail.
4983 @subheading Review @file{gdb/README}
4985 Grab one of the nightly snapshots and then walk through the
4986 @file{gdb/README} looking for anything that can be improved. The
4987 @code{schedule} script will mention this in its e-mail.
4989 @subheading Refresh any imported files.
4991 A number of files are taken from external repositories. They include:
4995 @file{texinfo/texinfo.tex}
4997 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
5000 @file{etc/standards.texi}, @file{etc/make-stds.texi}
5003 @subheading Check the ARI
5005 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
5006 (Awk Regression Index ;-) that checks for a number of errors and coding
5007 conventions. The checks include things like using @code{malloc} instead
5008 of @code{xmalloc} and file naming problems. There shouldn't be any
5011 @subsection Review the bug data base
5013 Close anything obviously fixed.
5015 @subsection Check all cross targets build
5017 The targets are listed in @file{gdb/MAINTAINERS}.
5020 @section Cut the branch
5022 @subheading The dirty work
5024 I think something like the below was used:
5027 $ d=`date -u +%Y-%m-%d`
5030 $ cvs -f -d /cvs/src rtag -D $d-gmt \
5031 gdb_5_1-$d-branchpoint insight+dejagnu
5032 $ cvs -f -d /cvs/src rtag -b -r gdb_V_V-$d-branchpoint \
5033 gdb_5_1-$d-branch insight+dejagnu
5039 the @kbd{-D YYYY-MM-DD-gmt} forces the branch to an exact date/time.
5041 the trunk is first tagged so that the branch point can easily be found
5043 Insight (which includes GDB) and dejagnu are tagged at the same time
5046 @subheading Post the branch info
5048 @subheading Update the web and news pages
5050 @subheading Tweak cron to track the new branch
5052 @section Stabilize the branch
5054 Something goes here.
5056 @section Create a Release
5058 This procedure can be followed when creating beta and final final
5059 releases. With a beta many of the steps can be skipped.
5061 @subheading Establish a few defaults.
5064 $ b=gdb_5_1-2001-07-29-branch
5066 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
5071 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_1-2001-07-29-branch/5.1.1
5073 /home/gdbadmin/bin/autoconf
5077 NB: Check the autoconf version carefully. You want to be using the
5078 version taken from the binutils snapshot directory. It is most likely
5079 that your system's installed version (@file{/usr/bin}?) is probably
5082 @subheading Check out the relevant modules:
5085 $ for m in gdb insight dejagnu
5087 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
5092 NB: The reading of @file{.cvsrc} is disabled (@file{-f}) so that there
5093 isn't any confusion between what is written here and what your local CVS
5096 @subheading Update relevant files.
5098 @subsubheading @file{gdb/NEWS}
5100 Major releases get their comments added as part of the mainline. Minor
5101 releases should probably mention any significant bugs that were fixed.
5103 Don't forget to update the ChangeLog.
5106 $ emacs gdb/src/gdb/NEWS
5111 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
5112 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5115 @subsubheading @file{gdb/README}
5117 You'll need to update: the version, the update date, and who did it.
5120 $ emacs gdb/src/gdb/README
5125 $ cp gdb/src/gdb/README insight/src/gdb/README
5126 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5129 @emph{Maintainer note: Hopefully the README file was reviewed before the
5130 initial branch was cut so just a simple substitute is needed to get it
5133 @emph{Maintainer note: Other projects generate @file{README} and
5134 @file{INSTALL} from the core documentation. This might be worth
5137 @subsubheading @file{gdb/version.in}
5140 $ echo $v > gdb/src/gdb/version.in
5141 $ emacs gdb/src/gdb/version.in
5146 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
5147 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5150 @subsubheading @file{dejagnu/src/dejagnu/configure.in}
5152 Dejagnu is more complicated. The version number is a parameter to
5153 @var{AM_INIT_AUTOMAKE}. Tweak it to read something like
5156 Re-generate configure.
5160 @subheading Do the dirty work
5162 This is identical to the process used when creating the daily snapshot.
5165 $ for m in gdb insight dejagnu
5167 ( cd $m/src && gmake -f Makefile.in $m.tar.bz2 )
5171 @subheading Check the source files
5173 You're looking for files that have mysteriously disappeared as the
5174 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
5175 for the @file{version.in} update @kbd{cronjob}.
5178 $ ( cd gdb/src && cvs -f -q -n update )
5187 ? gdb/doc/gdb.info-11
5188 ? gdb/doc/gdb.info-12
5189 ? gdb/doc/gdb.info-13
5190 ? gdb/doc/gdb.info-14
5191 ? gdb/doc/gdb.info-15
5192 ? gdb/doc/gdbint.info-4
5193 ? gdb/doc/gdbint.info-5
5197 @emph{Don't worry about the @file{gdb.info-??} or
5198 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
5199 was also generated only something strange with CVS means that they
5200 didn't get supressed). Fixing it would be nice though.}
5202 @subheading Re-pack the release with @code{gzip}
5206 $ bunzip2 -k -v *.bz2
5210 NB: A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} shouldn't
5211 be used since, in that mode, gzip doesn't know the file name information
5212 and consequently can't include it. This is also why the release process
5213 runs @code{tar} and @code{bzip2} as separate passes.
5215 @emph{Maintainer note: The release process could be changed to create
5216 @file{.tar} rather than @file{.tar.bz2} files.}
5218 @section Check the release
5220 Grab the @file{gdb.tar.bz2}, copy it to your local machine and then try
5221 a simple build using it.
5223 If this is a pre-release just copy the @file{.bz2} files to the snapshot
5224 directory and skip the remaining steps.
5226 If it is a final release, also make it available under a bogus name so
5227 that others can download and check it.
5229 @emph{Maintainer note: This adds an extra day to the release process but
5230 is very much worth it. Other developers are given the opportunity to
5231 check that things like your @file{NEWS} entries are correct or that
5232 other changes actually work.}
5234 @section Release the tar ball
5236 This is where, unfortunately, the notes just get vague.
5238 @subheading Install on sware
5241 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
5244 @subheading Create and update the web pages.
5250 create the directory @file{htdocs/@var{version}} (e.g., @file{htdocs/5.1.1}
5252 copy @file{index.html} and @file{ANNOUNCE} from the previous release
5253 into the @file{htdocs/@var{version}} directory and edit for content.
5254 Things like the MD5 sums, @kbd{ls -l} output, the version number and so
5255 on will need updating. Add NEWS entries to the @file{ANNOUNCE}. This
5256 ensures that the previous announcement is kept somewhere handy.
5258 copy the @file{NEWS} from the distro into the
5259 @file{htdocs/@var{version}} directory, trim down to just the most recent
5262 Add a short (identical) announcement to both @file{htdocs/index.html}
5263 and @file{htdocs/news/index.html}
5265 edit the script @file{htdocs/index.sh} to link in the new release
5266 number. Run it across all @file{index.html} files vis @kbd{./index.sh
5267 index.html */index.html}.
5269 grep the @file{htdocs} tree for references to the previous release
5270 version (@file{htdocs/download/index.html})
5273 @emph{Maintainer note: This step is too fragile --- it is too easy to
5274 mis one of the entries and forget to update it.}
5276 @subheading Generate online docs
5278 You need to find the magic command that is used to generate the online
5279 docs from the @file{.tar.bz2}. The best way is to look in the output
5280 from one of the nightly cronjobs and then just edit accordingly.
5284 $ ~/ss/update-web-docs \
5285 ~ftp/pub/gdb/releases/gdb-5.1.1.tar.bz2 \
5287 /www/sourceware/htdocs/gdb/5.1.1/onlinedocs \
5291 @subheading Something about @file{ANNOUNCEMENT}
5293 Send the @file{ANNOUNCEMENT} file you created above to:
5297 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
5299 The gnu announce list (but delay it a day or so to let things get out).
5302 @subheading Install it on GNU
5304 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
5305 @file{~ftp/gnu/gdb} (I think, I'm still waiting for it to copy into my
5310 @subheading Commit outstanding changes
5312 In particular you'll need to commit the changes to:
5316 @file{gdb/ChangeLog}
5318 @file{gdb/version.in}
5325 @subheading Tag the release
5330 $ d=`date -u +%Y-%m-%d`
5333 $ ( cd insight/src/gdb && cvs -f -q update )
5334 $ ( cd insight/src && cvs -f -q tag gdb_5_1_1-$d-release )
5337 Insight is used since that contains more of the release than GDB (yes
5338 dejagnu doesn't get tagged but I think we can live with that.).
5340 @subheading Restart @file{gdb/version.in}
5342 If @file{gdb/version.in} does not contain an ISO date such as
5343 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
5344 committed all the release changes it can be set to
5345 @file{5.1.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
5346 is important - it affects the snapshot process).
5348 Don't forget the @file{ChangeLog}.
5350 @subheading Merge into trunk
5352 The files committed to the branch may also need changes merged into the
5355 @section Post release
5357 Remove any @code{OBSOLETE} code.
5364 The testsuite is an important component of the @value{GDBN} package.
5365 While it is always worthwhile to encourage user testing, in practice
5366 this is rarely sufficient; users typically use only a small subset of
5367 the available commands, and it has proven all too common for a change
5368 to cause a significant regression that went unnoticed for some time.
5370 The @value{GDBN} testsuite uses the DejaGNU testing framework.
5371 DejaGNU is built using @code{Tcl} and @code{expect}. The tests
5372 themselves are calls to various @code{Tcl} procs; the framework runs all the
5373 procs and summarizes the passes and fails.
5375 @section Using the Testsuite
5377 @cindex running the test suite
5378 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
5379 testsuite's objdir) and type @code{make check}. This just sets up some
5380 environment variables and invokes DejaGNU's @code{runtest} script. While
5381 the testsuite is running, you'll get mentions of which test file is in use,
5382 and a mention of any unexpected passes or fails. When the testsuite is
5383 finished, you'll get a summary that looks like this:
5388 # of expected passes 6016
5389 # of unexpected failures 58
5390 # of unexpected successes 5
5391 # of expected failures 183
5392 # of unresolved testcases 3
5393 # of untested testcases 5
5396 The ideal test run consists of expected passes only; however, reality
5397 conspires to keep us from this ideal. Unexpected failures indicate
5398 real problems, whether in @value{GDBN} or in the testsuite. Expected
5399 failures are still failures, but ones which have been decided are too
5400 hard to deal with at the time; for instance, a test case might work
5401 everywhere except on AIX, and there is no prospect of the AIX case
5402 being fixed in the near future. Expected failures should not be added
5403 lightly, since you may be masking serious bugs in @value{GDBN}.
5404 Unexpected successes are expected fails that are passing for some
5405 reason, while unresolved and untested cases often indicate some minor
5406 catastrophe, such as the compiler being unable to deal with a test
5409 When making any significant change to @value{GDBN}, you should run the
5410 testsuite before and after the change, to confirm that there are no
5411 regressions. Note that truly complete testing would require that you
5412 run the testsuite with all supported configurations and a variety of
5413 compilers; however this is more than really necessary. In many cases
5414 testing with a single configuration is sufficient. Other useful
5415 options are to test one big-endian (Sparc) and one little-endian (x86)
5416 host, a cross config with a builtin simulator (powerpc-eabi,
5417 mips-elf), or a 64-bit host (Alpha).
5419 If you add new functionality to @value{GDBN}, please consider adding
5420 tests for it as well; this way future @value{GDBN} hackers can detect
5421 and fix their changes that break the functionality you added.
5422 Similarly, if you fix a bug that was not previously reported as a test
5423 failure, please add a test case for it. Some cases are extremely
5424 difficult to test, such as code that handles host OS failures or bugs
5425 in particular versions of compilers, and it's OK not to try to write
5426 tests for all of those.
5428 @section Testsuite Organization
5430 @cindex test suite organization
5431 The testsuite is entirely contained in @file{gdb/testsuite}. While the
5432 testsuite includes some makefiles and configury, these are very minimal,
5433 and used for little besides cleaning up, since the tests themselves
5434 handle the compilation of the programs that @value{GDBN} will run. The file
5435 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
5436 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
5437 configuration-specific files, typically used for special-purpose
5438 definitions of procs like @code{gdb_load} and @code{gdb_start}.
5440 The tests themselves are to be found in @file{testsuite/gdb.*} and
5441 subdirectories of those. The names of the test files must always end
5442 with @file{.exp}. DejaGNU collects the test files by wildcarding
5443 in the test directories, so both subdirectories and individual files
5444 get chosen and run in alphabetical order.
5446 The following table lists the main types of subdirectories and what they
5447 are for. Since DejaGNU finds test files no matter where they are
5448 located, and since each test file sets up its own compilation and
5449 execution environment, this organization is simply for convenience and
5454 This is the base testsuite. The tests in it should apply to all
5455 configurations of @value{GDBN} (but generic native-only tests may live here).
5456 The test programs should be in the subset of C that is valid K&R,
5457 ANSI/ISO, and C++ (@code{#ifdef}s are allowed if necessary, for instance
5460 @item gdb.@var{lang}
5461 Language-specific tests for any language @var{lang} besides C. Examples are
5462 @file{gdb.c++} and @file{gdb.java}.
5464 @item gdb.@var{platform}
5465 Non-portable tests. The tests are specific to a specific configuration
5466 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
5469 @item gdb.@var{compiler}
5470 Tests specific to a particular compiler. As of this writing (June
5471 1999), there aren't currently any groups of tests in this category that
5472 couldn't just as sensibly be made platform-specific, but one could
5473 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
5476 @item gdb.@var{subsystem}
5477 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
5478 instance, @file{gdb.disasm} exercises various disassemblers, while
5479 @file{gdb.stabs} tests pathways through the stabs symbol reader.
5482 @section Writing Tests
5483 @cindex writing tests
5485 In many areas, the @value{GDBN} tests are already quite comprehensive; you
5486 should be able to copy existing tests to handle new cases.
5488 You should try to use @code{gdb_test} whenever possible, since it
5489 includes cases to handle all the unexpected errors that might happen.
5490 However, it doesn't cost anything to add new test procedures; for
5491 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
5492 calls @code{gdb_test} multiple times.
5494 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
5495 necessary, such as when @value{GDBN} has several valid responses to a command.
5497 The source language programs do @emph{not} need to be in a consistent
5498 style. Since @value{GDBN} is used to debug programs written in many different
5499 styles, it's worth having a mix of styles in the testsuite; for
5500 instance, some @value{GDBN} bugs involving the display of source lines would
5501 never manifest themselves if the programs used GNU coding style
5508 Check the @file{README} file, it often has useful information that does not
5509 appear anywhere else in the directory.
5512 * Getting Started:: Getting started working on @value{GDBN}
5513 * Debugging GDB:: Debugging @value{GDBN} with itself
5516 @node Getting Started,,, Hints
5518 @section Getting Started
5520 @value{GDBN} is a large and complicated program, and if you first starting to
5521 work on it, it can be hard to know where to start. Fortunately, if you
5522 know how to go about it, there are ways to figure out what is going on.
5524 This manual, the @value{GDBN} Internals manual, has information which applies
5525 generally to many parts of @value{GDBN}.
5527 Information about particular functions or data structures are located in
5528 comments with those functions or data structures. If you run across a
5529 function or a global variable which does not have a comment correctly
5530 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
5531 free to submit a bug report, with a suggested comment if you can figure
5532 out what the comment should say. If you find a comment which is
5533 actually wrong, be especially sure to report that.
5535 Comments explaining the function of macros defined in host, target, or
5536 native dependent files can be in several places. Sometimes they are
5537 repeated every place the macro is defined. Sometimes they are where the
5538 macro is used. Sometimes there is a header file which supplies a
5539 default definition of the macro, and the comment is there. This manual
5540 also documents all the available macros.
5541 @c (@pxref{Host Conditionals}, @pxref{Target
5542 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
5545 Start with the header files. Once you have some idea of how
5546 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
5547 @file{gdbtypes.h}), you will find it much easier to understand the
5548 code which uses and creates those symbol tables.
5550 You may wish to process the information you are getting somehow, to
5551 enhance your understanding of it. Summarize it, translate it to another
5552 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
5553 the code to predict what a test case would do and write the test case
5554 and verify your prediction, etc. If you are reading code and your eyes
5555 are starting to glaze over, this is a sign you need to use a more active
5558 Once you have a part of @value{GDBN} to start with, you can find more
5559 specifically the part you are looking for by stepping through each
5560 function with the @code{next} command. Do not use @code{step} or you
5561 will quickly get distracted; when the function you are stepping through
5562 calls another function try only to get a big-picture understanding
5563 (perhaps using the comment at the beginning of the function being
5564 called) of what it does. This way you can identify which of the
5565 functions being called by the function you are stepping through is the
5566 one which you are interested in. You may need to examine the data
5567 structures generated at each stage, with reference to the comments in
5568 the header files explaining what the data structures are supposed to
5571 Of course, this same technique can be used if you are just reading the
5572 code, rather than actually stepping through it. The same general
5573 principle applies---when the code you are looking at calls something
5574 else, just try to understand generally what the code being called does,
5575 rather than worrying about all its details.
5577 @cindex command implementation
5578 A good place to start when tracking down some particular area is with
5579 a command which invokes that feature. Suppose you want to know how
5580 single-stepping works. As a @value{GDBN} user, you know that the
5581 @code{step} command invokes single-stepping. The command is invoked
5582 via command tables (see @file{command.h}); by convention the function
5583 which actually performs the command is formed by taking the name of
5584 the command and adding @samp{_command}, or in the case of an
5585 @code{info} subcommand, @samp{_info}. For example, the @code{step}
5586 command invokes the @code{step_command} function and the @code{info
5587 display} command invokes @code{display_info}. When this convention is
5588 not followed, you might have to use @code{grep} or @kbd{M-x
5589 tags-search} in emacs, or run @value{GDBN} on itself and set a
5590 breakpoint in @code{execute_command}.
5592 @cindex @code{bug-gdb} mailing list
5593 If all of the above fail, it may be appropriate to ask for information
5594 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
5595 wondering if anyone could give me some tips about understanding
5596 @value{GDBN}''---if we had some magic secret we would put it in this manual.
5597 Suggestions for improving the manual are always welcome, of course.
5599 @node Debugging GDB,,,Hints
5601 @section Debugging @value{GDBN} with itself
5602 @cindex debugging @value{GDBN}
5604 If @value{GDBN} is limping on your machine, this is the preferred way to get it
5605 fully functional. Be warned that in some ancient Unix systems, like
5606 Ultrix 4.2, a program can't be running in one process while it is being
5607 debugged in another. Rather than typing the command @kbd{@w{./gdb
5608 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
5609 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
5611 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
5612 @file{.gdbinit} file that sets up some simple things to make debugging
5613 gdb easier. The @code{info} command, when executed without a subcommand
5614 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
5615 gdb. See @file{.gdbinit} for details.
5617 If you use emacs, you will probably want to do a @code{make TAGS} after
5618 you configure your distribution; this will put the machine dependent
5619 routines for your local machine where they will be accessed first by
5622 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
5623 have run @code{fixincludes} if you are compiling with gcc.
5625 @section Submitting Patches
5627 @cindex submitting patches
5628 Thanks for thinking of offering your changes back to the community of
5629 @value{GDBN} users. In general we like to get well designed enhancements.
5630 Thanks also for checking in advance about the best way to transfer the
5633 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
5634 This manual summarizes what we believe to be clean design for @value{GDBN}.
5636 If the maintainers don't have time to put the patch in when it arrives,
5637 or if there is any question about a patch, it goes into a large queue
5638 with everyone else's patches and bug reports.
5640 @cindex legal papers for code contributions
5641 The legal issue is that to incorporate substantial changes requires a
5642 copyright assignment from you and/or your employer, granting ownership
5643 of the changes to the Free Software Foundation. You can get the
5644 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
5645 and asking for it. We recommend that people write in "All programs
5646 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
5647 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
5649 contributed with only one piece of legalese pushed through the
5650 bureaucracy and filed with the FSF. We can't start merging changes until
5651 this paperwork is received by the FSF (their rules, which we follow
5652 since we maintain it for them).
5654 Technically, the easiest way to receive changes is to receive each
5655 feature as a small context diff or unidiff, suitable for @code{patch}.
5656 Each message sent to me should include the changes to C code and
5657 header files for a single feature, plus @file{ChangeLog} entries for
5658 each directory where files were modified, and diffs for any changes
5659 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
5660 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
5661 single feature, they can be split down into multiple messages.
5663 In this way, if we read and like the feature, we can add it to the
5664 sources with a single patch command, do some testing, and check it in.
5665 If you leave out the @file{ChangeLog}, we have to write one. If you leave
5666 out the doc, we have to puzzle out what needs documenting. Etc., etc.
5668 The reason to send each change in a separate message is that we will not
5669 install some of the changes. They'll be returned to you with questions
5670 or comments. If we're doing our job correctly, the message back to you
5671 will say what you have to fix in order to make the change acceptable.
5672 The reason to have separate messages for separate features is so that
5673 the acceptable changes can be installed while one or more changes are
5674 being reworked. If multiple features are sent in a single message, we
5675 tend to not put in the effort to sort out the acceptable changes from
5676 the unacceptable, so none of the features get installed until all are
5679 If this sounds painful or authoritarian, well, it is. But we get a lot
5680 of bug reports and a lot of patches, and many of them don't get
5681 installed because we don't have the time to finish the job that the bug
5682 reporter or the contributor could have done. Patches that arrive
5683 complete, working, and well designed, tend to get installed on the day
5684 they arrive. The others go into a queue and get installed as time
5685 permits, which, since the maintainers have many demands to meet, may not
5686 be for quite some time.
5688 Please send patches directly to
5689 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
5691 @section Obsolete Conditionals
5692 @cindex obsolete code
5694 Fragments of old code in @value{GDBN} sometimes reference or set the following
5695 configuration macros. They should not be used by new code, and old uses
5696 should be removed as those parts of the debugger are otherwise touched.
5699 @item STACK_END_ADDR
5700 This macro used to define where the end of the stack appeared, for use
5701 in interpreting core file formats that don't record this address in the
5702 core file itself. This information is now configured in BFD, and @value{GDBN}
5703 gets the info portably from there. The values in @value{GDBN}'s configuration
5704 files should be moved into BFD configuration files (if needed there),
5705 and deleted from all of @value{GDBN}'s config files.
5707 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
5708 is so old that it has never been converted to use BFD. Now that's old!