1 \input texinfo @c -*- texinfo -*-
2 @setfilename gdbint.info
4 @dircategory Software development
6 * Gdb-Internals: (gdbint). The GNU debugger's internals.
10 This file documents the internals of the GNU debugger @value{GDBN}.
11 Copyright 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001,2002,2003,2004
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 no Front-Cover Texts, and with no Back-Cover
20 Texts. A copy of the license is included in the section entitled ``GNU
21 Free Documentation License''.
24 @setchapternewpage off
25 @settitle @value{GDBN} Internals
31 @title @value{GDBN} Internals
32 @subtitle{A guide to the internals of the GNU debugger}
34 @author Cygnus Solutions
35 @author Second Edition:
37 @author Cygnus Solutions
40 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
41 \xdef\manvers{\$Revision$} % For use in headers, footers too
43 \hfill Cygnus Solutions\par
45 \hfill \TeX{}info \texinfoversion\par
49 @vskip 0pt plus 1filll
50 Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001,
51 2002, 2003, 2004 Free Software Foundation, Inc.
53 Permission is granted to copy, distribute and/or modify this document
54 under the terms of the GNU Free Documentation License, Version 1.1 or
55 any later version published by the Free Software Foundation; with no
56 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
57 Texts. A copy of the license is included in the section entitled ``GNU
58 Free Documentation License''.
64 @c Perhaps this should be the title of the document (but only for info,
65 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
66 @top Scope of this Document
68 This document documents the internals of the GNU debugger, @value{GDBN}. It
69 includes description of @value{GDBN}'s key algorithms and operations, as well
70 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
81 * Target Architecture Definition::
82 * Target Vector Definition::
91 * GDB Observers:: @value{GDBN} Currently available observers
92 * GNU Free Documentation License:: The license for this documentation
99 @cindex requirements for @value{GDBN}
101 Before diving into the internals, you should understand the formal
102 requirements and other expectations for @value{GDBN}. Although some
103 of these may seem obvious, there have been proposals for @value{GDBN}
104 that have run counter to these requirements.
106 First of all, @value{GDBN} is a debugger. It's not designed to be a
107 front panel for embedded systems. It's not a text editor. It's not a
108 shell. It's not a programming environment.
110 @value{GDBN} is an interactive tool. Although a batch mode is
111 available, @value{GDBN}'s primary role is to interact with a human
114 @value{GDBN} should be responsive to the user. A programmer hot on
115 the trail of a nasty bug, and operating under a looming deadline, is
116 going to be very impatient of everything, including the response time
117 to debugger commands.
119 @value{GDBN} should be relatively permissive, such as for expressions.
120 While the compiler should be picky (or have the option to be made
121 picky), since source code lives for a long time usually, the
122 programmer doing debugging shouldn't be spending time figuring out to
123 mollify the debugger.
125 @value{GDBN} will be called upon to deal with really large programs.
126 Executable sizes of 50 to 100 megabytes occur regularly, and we've
127 heard reports of programs approaching 1 gigabyte in size.
129 @value{GDBN} should be able to run everywhere. No other debugger is
130 available for even half as many configurations as @value{GDBN}
134 @node Overall Structure
136 @chapter Overall Structure
138 @value{GDBN} consists of three major subsystems: user interface,
139 symbol handling (the @dfn{symbol side}), and target system handling (the
142 The user interface consists of several actual interfaces, plus
145 The symbol side consists of object file readers, debugging info
146 interpreters, symbol table management, source language expression
147 parsing, type and value printing.
149 The target side consists of execution control, stack frame analysis, and
150 physical target manipulation.
152 The target side/symbol side division is not formal, and there are a
153 number of exceptions. For instance, core file support involves symbolic
154 elements (the basic core file reader is in BFD) and target elements (it
155 supplies the contents of memory and the values of registers). Instead,
156 this division is useful for understanding how the minor subsystems
159 @section The Symbol Side
161 The symbolic side of @value{GDBN} can be thought of as ``everything
162 you can do in @value{GDBN} without having a live program running''.
163 For instance, you can look at the types of variables, and evaluate
164 many kinds of expressions.
166 @section The Target Side
168 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
169 Although it may make reference to symbolic info here and there, most
170 of the target side will run with only a stripped executable
171 available---or even no executable at all, in remote debugging cases.
173 Operations such as disassembly, stack frame crawls, and register
174 display, are able to work with no symbolic info at all. In some cases,
175 such as disassembly, @value{GDBN} will use symbolic info to present addresses
176 relative to symbols rather than as raw numbers, but it will work either
179 @section Configurations
183 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
184 @dfn{Target} refers to the system where the program being debugged
185 executes. In most cases they are the same machine, in which case a
186 third type of @dfn{Native} attributes come into play.
188 Defines and include files needed to build on the host are host support.
189 Examples are tty support, system defined types, host byte order, host
192 Defines and information needed to handle the target format are target
193 dependent. Examples are the stack frame format, instruction set,
194 breakpoint instruction, registers, and how to set up and tear down the stack
197 Information that is only needed when the host and target are the same,
198 is native dependent. One example is Unix child process support; if the
199 host and target are not the same, doing a fork to start the target
200 process is a bad idea. The various macros needed for finding the
201 registers in the @code{upage}, running @code{ptrace}, and such are all
202 in the native-dependent files.
204 Another example of native-dependent code is support for features that
205 are really part of the target environment, but which require
206 @code{#include} files that are only available on the host system. Core
207 file handling and @code{setjmp} handling are two common cases.
209 When you want to make @value{GDBN} work ``native'' on a particular machine, you
210 have to include all three kinds of information.
218 @value{GDBN} uses a number of debugging-specific algorithms. They are
219 often not very complicated, but get lost in the thicket of special
220 cases and real-world issues. This chapter describes the basic
221 algorithms and mentions some of the specific target definitions that
227 @cindex call stack frame
228 A frame is a construct that @value{GDBN} uses to keep track of calling
229 and called functions.
231 @findex create_new_frame
233 @code{FRAME_FP} in the machine description has no meaning to the
234 machine-independent part of @value{GDBN}, except that it is used when
235 setting up a new frame from scratch, as follows:
238 create_new_frame (read_register (DEPRECATED_FP_REGNUM), read_pc ()));
241 @cindex frame pointer register
242 Other than that, all the meaning imparted to @code{DEPRECATED_FP_REGNUM}
243 is imparted by the machine-dependent code. So,
244 @code{DEPRECATED_FP_REGNUM} can have any value that is convenient for
245 the code that creates new frames. (@code{create_new_frame} calls
246 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} if it is defined; that is where
247 you should use the @code{DEPRECATED_FP_REGNUM} value, if your frames are
251 Given a @value{GDBN} frame, define @code{DEPRECATED_FRAME_CHAIN} to
252 determine the address of the calling function's frame. This will be
253 used to create a new @value{GDBN} frame struct, and then
254 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} and
255 @code{DEPRECATED_INIT_FRAME_PC} will be called for the new frame.
257 @section Breakpoint Handling
260 In general, a breakpoint is a user-designated location in the program
261 where the user wants to regain control if program execution ever reaches
264 There are two main ways to implement breakpoints; either as ``hardware''
265 breakpoints or as ``software'' breakpoints.
267 @cindex hardware breakpoints
268 @cindex program counter
269 Hardware breakpoints are sometimes available as a builtin debugging
270 features with some chips. Typically these work by having dedicated
271 register into which the breakpoint address may be stored. If the PC
272 (shorthand for @dfn{program counter})
273 ever matches a value in a breakpoint registers, the CPU raises an
274 exception and reports it to @value{GDBN}.
276 Another possibility is when an emulator is in use; many emulators
277 include circuitry that watches the address lines coming out from the
278 processor, and force it to stop if the address matches a breakpoint's
281 A third possibility is that the target already has the ability to do
282 breakpoints somehow; for instance, a ROM monitor may do its own
283 software breakpoints. So although these are not literally ``hardware
284 breakpoints'', from @value{GDBN}'s point of view they work the same;
285 @value{GDBN} need not do anything more than set the breakpoint and wait
286 for something to happen.
288 Since they depend on hardware resources, hardware breakpoints may be
289 limited in number; when the user asks for more, @value{GDBN} will
290 start trying to set software breakpoints. (On some architectures,
291 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
292 whether there's enough hardware resources to insert all the hardware
293 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
294 an error message only when the program being debugged is continued.)
296 @cindex software breakpoints
297 Software breakpoints require @value{GDBN} to do somewhat more work.
298 The basic theory is that @value{GDBN} will replace a program
299 instruction with a trap, illegal divide, or some other instruction
300 that will cause an exception, and then when it's encountered,
301 @value{GDBN} will take the exception and stop the program. When the
302 user says to continue, @value{GDBN} will restore the original
303 instruction, single-step, re-insert the trap, and continue on.
305 Since it literally overwrites the program being tested, the program area
306 must be writable, so this technique won't work on programs in ROM. It
307 can also distort the behavior of programs that examine themselves,
308 although such a situation would be highly unusual.
310 Also, the software breakpoint instruction should be the smallest size of
311 instruction, so it doesn't overwrite an instruction that might be a jump
312 target, and cause disaster when the program jumps into the middle of the
313 breakpoint instruction. (Strictly speaking, the breakpoint must be no
314 larger than the smallest interval between instructions that may be jump
315 targets; perhaps there is an architecture where only even-numbered
316 instructions may jumped to.) Note that it's possible for an instruction
317 set not to have any instructions usable for a software breakpoint,
318 although in practice only the ARC has failed to define such an
322 The basic definition of the software breakpoint is the macro
325 Basic breakpoint object handling is in @file{breakpoint.c}. However,
326 much of the interesting breakpoint action is in @file{infrun.c}.
328 @section Single Stepping
330 @section Signal Handling
332 @section Thread Handling
334 @section Inferior Function Calls
336 @section Longjmp Support
338 @cindex @code{longjmp} debugging
339 @value{GDBN} has support for figuring out that the target is doing a
340 @code{longjmp} and for stopping at the target of the jump, if we are
341 stepping. This is done with a few specialized internal breakpoints,
342 which are visible in the output of the @samp{maint info breakpoint}
345 @findex GET_LONGJMP_TARGET
346 To make this work, you need to define a macro called
347 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
348 structure and extract the longjmp target address. Since @code{jmp_buf}
349 is target specific, you will need to define it in the appropriate
350 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
351 @file{sparc-tdep.c} for examples of how to do this.
356 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
357 breakpoints}) which break when data is accessed rather than when some
358 instruction is executed. When you have data which changes without
359 your knowing what code does that, watchpoints are the silver bullet to
360 hunt down and kill such bugs.
362 @cindex hardware watchpoints
363 @cindex software watchpoints
364 Watchpoints can be either hardware-assisted or not; the latter type is
365 known as ``software watchpoints.'' @value{GDBN} always uses
366 hardware-assisted watchpoints if they are available, and falls back on
367 software watchpoints otherwise. Typical situations where @value{GDBN}
368 will use software watchpoints are:
372 The watched memory region is too large for the underlying hardware
373 watchpoint support. For example, each x86 debug register can watch up
374 to 4 bytes of memory, so trying to watch data structures whose size is
375 more than 16 bytes will cause @value{GDBN} to use software
379 The value of the expression to be watched depends on data held in
380 registers (as opposed to memory).
383 Too many different watchpoints requested. (On some architectures,
384 this situation is impossible to detect until the debugged program is
385 resumed.) Note that x86 debug registers are used both for hardware
386 breakpoints and for watchpoints, so setting too many hardware
387 breakpoints might cause watchpoint insertion to fail.
390 No hardware-assisted watchpoints provided by the target
394 Software watchpoints are very slow, since @value{GDBN} needs to
395 single-step the program being debugged and test the value of the
396 watched expression(s) after each instruction. The rest of this
397 section is mostly irrelevant for software watchpoints.
399 @value{GDBN} uses several macros and primitives to support hardware
403 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
404 @item TARGET_HAS_HARDWARE_WATCHPOINTS
405 If defined, the target supports hardware watchpoints.
407 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
408 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
409 Return the number of hardware watchpoints of type @var{type} that are
410 possible to be set. The value is positive if @var{count} watchpoints
411 of this type can be set, zero if setting watchpoints of this type is
412 not supported, and negative if @var{count} is more than the maximum
413 number of watchpoints of type @var{type} that can be set. @var{other}
414 is non-zero if other types of watchpoints are currently enabled (there
415 are architectures which cannot set watchpoints of different types at
418 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
419 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
420 Return non-zero if hardware watchpoints can be used to watch a region
421 whose address is @var{addr} and whose length in bytes is @var{len}.
423 @findex TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT
424 @item TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT (@var{size})
425 Return non-zero if hardware watchpoints can be used to watch a region
426 whose size is @var{size}. @value{GDBN} only uses this macro as a
427 fall-back, in case @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is not
430 @findex TARGET_DISABLE_HW_WATCHPOINTS
431 @item TARGET_DISABLE_HW_WATCHPOINTS (@var{pid})
432 Disables watchpoints in the process identified by @var{pid}. This is
433 used, e.g., on HP-UX which provides operations to disable and enable
434 the page-level memory protection that implements hardware watchpoints
437 @findex TARGET_ENABLE_HW_WATCHPOINTS
438 @item TARGET_ENABLE_HW_WATCHPOINTS (@var{pid})
439 Enables watchpoints in the process identified by @var{pid}. This is
440 used, e.g., on HP-UX which provides operations to disable and enable
441 the page-level memory protection that implements hardware watchpoints
444 @findex target_insert_watchpoint
445 @findex target_remove_watchpoint
446 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
447 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
448 Insert or remove a hardware watchpoint starting at @var{addr}, for
449 @var{len} bytes. @var{type} is the watchpoint type, one of the
450 possible values of the enumerated data type @code{target_hw_bp_type},
451 defined by @file{breakpoint.h} as follows:
454 enum target_hw_bp_type
456 hw_write = 0, /* Common (write) HW watchpoint */
457 hw_read = 1, /* Read HW watchpoint */
458 hw_access = 2, /* Access (read or write) HW watchpoint */
459 hw_execute = 3 /* Execute HW breakpoint */
464 These two macros should return 0 for success, non-zero for failure.
466 @cindex insert or remove hardware breakpoint
467 @findex target_remove_hw_breakpoint
468 @findex target_insert_hw_breakpoint
469 @item target_remove_hw_breakpoint (@var{addr}, @var{shadow})
470 @itemx target_insert_hw_breakpoint (@var{addr}, @var{shadow})
471 Insert or remove a hardware-assisted breakpoint at address @var{addr}.
472 Returns zero for success, non-zero for failure. @var{shadow} is the
473 real contents of the byte where the breakpoint has been inserted; it
474 is generally not valid when hardware breakpoints are used, but since
475 no other code touches these values, the implementations of the above
476 two macros can use them for their internal purposes.
478 @findex target_stopped_data_address
479 @item target_stopped_data_address (@var{addr_p})
480 If the inferior has some watchpoint that triggered, place the address
481 associated with the watchpoint at the location pointed to by
482 @var{addr_p} and return non-zero. Otherwise, return zero.
484 @findex HAVE_STEPPABLE_WATCHPOINT
485 @item HAVE_STEPPABLE_WATCHPOINT
486 If defined to a non-zero value, it is not necessary to disable a
487 watchpoint to step over it.
489 @findex HAVE_NONSTEPPABLE_WATCHPOINT
490 @item HAVE_NONSTEPPABLE_WATCHPOINT
491 If defined to a non-zero value, @value{GDBN} should disable a
492 watchpoint to step the inferior over it.
494 @findex HAVE_CONTINUABLE_WATCHPOINT
495 @item HAVE_CONTINUABLE_WATCHPOINT
496 If defined to a non-zero value, it is possible to continue the
497 inferior after a watchpoint has been hit.
499 @findex CANNOT_STEP_HW_WATCHPOINTS
500 @item CANNOT_STEP_HW_WATCHPOINTS
501 If this is defined to a non-zero value, @value{GDBN} will remove all
502 watchpoints before stepping the inferior.
504 @findex STOPPED_BY_WATCHPOINT
505 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
506 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
507 the type @code{struct target_waitstatus}, defined by @file{target.h}.
510 @subsection x86 Watchpoints
511 @cindex x86 debug registers
512 @cindex watchpoints, on x86
514 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
515 registers designed to facilitate debugging. @value{GDBN} provides a
516 generic library of functions that x86-based ports can use to implement
517 support for watchpoints and hardware-assisted breakpoints. This
518 subsection documents the x86 watchpoint facilities in @value{GDBN}.
520 To use the generic x86 watchpoint support, a port should do the
524 @findex I386_USE_GENERIC_WATCHPOINTS
526 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
527 target-dependent headers.
530 Include the @file{config/i386/nm-i386.h} header file @emph{after}
531 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
534 Add @file{i386-nat.o} to the value of the Make variable
535 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
536 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
539 Provide implementations for the @code{I386_DR_LOW_*} macros described
540 below. Typically, each macro should call a target-specific function
541 which does the real work.
544 The x86 watchpoint support works by maintaining mirror images of the
545 debug registers. Values are copied between the mirror images and the
546 real debug registers via a set of macros which each target needs to
550 @findex I386_DR_LOW_SET_CONTROL
551 @item I386_DR_LOW_SET_CONTROL (@var{val})
552 Set the Debug Control (DR7) register to the value @var{val}.
554 @findex I386_DR_LOW_SET_ADDR
555 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
556 Put the address @var{addr} into the debug register number @var{idx}.
558 @findex I386_DR_LOW_RESET_ADDR
559 @item I386_DR_LOW_RESET_ADDR (@var{idx})
560 Reset (i.e.@: zero out) the address stored in the debug register
563 @findex I386_DR_LOW_GET_STATUS
564 @item I386_DR_LOW_GET_STATUS
565 Return the value of the Debug Status (DR6) register. This value is
566 used immediately after it is returned by
567 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
571 For each one of the 4 debug registers (whose indices are from 0 to 3)
572 that store addresses, a reference count is maintained by @value{GDBN},
573 to allow sharing of debug registers by several watchpoints. This
574 allows users to define several watchpoints that watch the same
575 expression, but with different conditions and/or commands, without
576 wasting debug registers which are in short supply. @value{GDBN}
577 maintains the reference counts internally, targets don't have to do
578 anything to use this feature.
580 The x86 debug registers can each watch a region that is 1, 2, or 4
581 bytes long. The ia32 architecture requires that each watched region
582 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
583 region on 4-byte boundary. However, the x86 watchpoint support in
584 @value{GDBN} can watch unaligned regions and regions larger than 4
585 bytes (up to 16 bytes) by allocating several debug registers to watch
586 a single region. This allocation of several registers per a watched
587 region is also done automatically without target code intervention.
589 The generic x86 watchpoint support provides the following API for the
590 @value{GDBN}'s application code:
593 @findex i386_region_ok_for_watchpoint
594 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
595 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
596 this function. It counts the number of debug registers required to
597 watch a given region, and returns a non-zero value if that number is
598 less than 4, the number of debug registers available to x86
601 @findex i386_stopped_data_address
602 @item i386_stopped_data_address (@var{addr_p})
604 @code{target_stopped_data_address} is set to call this function.
606 function examines the breakpoint condition bits in the DR6 Debug
607 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
608 macro, and returns the address associated with the first bit that is
611 @findex i386_stopped_by_watchpoint
612 @item i386_stopped_by_watchpoint (void)
613 The macro @code{STOPPED_BY_WATCHPOINT}
614 is set to call this function. The
615 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
616 function examines the breakpoint condition bits in the DR6 Debug
617 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
618 macro, and returns true if any bit is set. Otherwise, false is
621 @findex i386_insert_watchpoint
622 @findex i386_remove_watchpoint
623 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
624 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
625 Insert or remove a watchpoint. The macros
626 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
627 are set to call these functions. @code{i386_insert_watchpoint} first
628 looks for a debug register which is already set to watch the same
629 region for the same access types; if found, it just increments the
630 reference count of that debug register, thus implementing debug
631 register sharing between watchpoints. If no such register is found,
632 the function looks for a vacant debug register, sets its mirrored
633 value to @var{addr}, sets the mirrored value of DR7 Debug Control
634 register as appropriate for the @var{len} and @var{type} parameters,
635 and then passes the new values of the debug register and DR7 to the
636 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
637 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
638 required to cover the given region, the above process is repeated for
641 @code{i386_remove_watchpoint} does the opposite: it resets the address
642 in the mirrored value of the debug register and its read/write and
643 length bits in the mirrored value of DR7, then passes these new
644 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
645 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
646 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
647 decrements the reference count, and only calls
648 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
649 the count goes to zero.
651 @findex i386_insert_hw_breakpoint
652 @findex i386_remove_hw_breakpoint
653 @item i386_insert_hw_breakpoint (@var{addr}, @var{shadow}
654 @itemx i386_remove_hw_breakpoint (@var{addr}, @var{shadow})
655 These functions insert and remove hardware-assisted breakpoints. The
656 macros @code{target_insert_hw_breakpoint} and
657 @code{target_remove_hw_breakpoint} are set to call these functions.
658 These functions work like @code{i386_insert_watchpoint} and
659 @code{i386_remove_watchpoint}, respectively, except that they set up
660 the debug registers to watch instruction execution, and each
661 hardware-assisted breakpoint always requires exactly one debug
664 @findex i386_stopped_by_hwbp
665 @item i386_stopped_by_hwbp (void)
666 This function returns non-zero if the inferior has some watchpoint or
667 hardware breakpoint that triggered. It works like
668 @code{i386_stopped_data_address}, except that it doesn't record the
669 address whose watchpoint triggered.
671 @findex i386_cleanup_dregs
672 @item i386_cleanup_dregs (void)
673 This function clears all the reference counts, addresses, and control
674 bits in the mirror images of the debug registers. It doesn't affect
675 the actual debug registers in the inferior process.
682 x86 processors support setting watchpoints on I/O reads or writes.
683 However, since no target supports this (as of March 2001), and since
684 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
685 watchpoints, this feature is not yet available to @value{GDBN} running
689 x86 processors can enable watchpoints locally, for the current task
690 only, or globally, for all the tasks. For each debug register,
691 there's a bit in the DR7 Debug Control register that determines
692 whether the associated address is watched locally or globally. The
693 current implementation of x86 watchpoint support in @value{GDBN}
694 always sets watchpoints to be locally enabled, since global
695 watchpoints might interfere with the underlying OS and are probably
696 unavailable in many platforms.
699 @section Observing changes in @value{GDBN} internals
700 @cindex observer pattern interface
701 @cindex notifications about changes in internals
703 In order to function properly, several modules need to be notified when
704 some changes occur in the @value{GDBN} internals. Traditionally, these
705 modules have relied on several paradigms, the most common ones being
706 hooks and gdb-events. Unfortunately, none of these paradigms was
707 versatile enough to become the standard notification mechanism in
708 @value{GDBN}. The fact that they only supported one ``client'' was also
711 A new paradigm, based on the Observer pattern of the @cite{Design
712 Patterns} book, has therefore been implemented. The goal was to provide
713 a new interface overcoming the issues with the notification mechanisms
714 previously available. This new interface needed to be strongly typed,
715 easy to extend, and versatile enough to be used as the standard
716 interface when adding new notifications.
718 See @ref{GDB Observers} for a brief description of the observers
719 currently implemented in GDB. The rationale for the current
720 implementation is also briefly discussed.
724 @chapter User Interface
726 @value{GDBN} has several user interfaces. Although the command-line interface
727 is the most common and most familiar, there are others.
729 @section Command Interpreter
731 @cindex command interpreter
733 The command interpreter in @value{GDBN} is fairly simple. It is designed to
734 allow for the set of commands to be augmented dynamically, and also
735 has a recursive subcommand capability, where the first argument to
736 a command may itself direct a lookup on a different command list.
738 For instance, the @samp{set} command just starts a lookup on the
739 @code{setlist} command list, while @samp{set thread} recurses
740 to the @code{set_thread_cmd_list}.
744 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
745 the main command list, and should be used for those commands. The usual
746 place to add commands is in the @code{_initialize_@var{xyz}} routines at
747 the ends of most source files.
749 @findex add_setshow_cmd
750 @findex add_setshow_cmd_full
751 To add paired @samp{set} and @samp{show} commands, use
752 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
753 a slightly simpler interface which is useful when you don't need to
754 further modify the new command structures, while the latter returns
755 the new command structures for manipulation.
757 @cindex deprecating commands
758 @findex deprecate_cmd
759 Before removing commands from the command set it is a good idea to
760 deprecate them for some time. Use @code{deprecate_cmd} on commands or
761 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
762 @code{struct cmd_list_element} as it's first argument. You can use the
763 return value from @code{add_com} or @code{add_cmd} to deprecate the
764 command immediately after it is created.
766 The first time a command is used the user will be warned and offered a
767 replacement (if one exists). Note that the replacement string passed to
768 @code{deprecate_cmd} should be the full name of the command, i.e. the
769 entire string the user should type at the command line.
771 @section UI-Independent Output---the @code{ui_out} Functions
772 @c This section is based on the documentation written by Fernando
773 @c Nasser <fnasser@redhat.com>.
775 @cindex @code{ui_out} functions
776 The @code{ui_out} functions present an abstraction level for the
777 @value{GDBN} output code. They hide the specifics of different user
778 interfaces supported by @value{GDBN}, and thus free the programmer
779 from the need to write several versions of the same code, one each for
780 every UI, to produce output.
782 @subsection Overview and Terminology
784 In general, execution of each @value{GDBN} command produces some sort
785 of output, and can even generate an input request.
787 Output can be generated for the following purposes:
791 to display a @emph{result} of an operation;
794 to convey @emph{info} or produce side-effects of a requested
798 to provide a @emph{notification} of an asynchronous event (including
799 progress indication of a prolonged asynchronous operation);
802 to display @emph{error messages} (including warnings);
805 to show @emph{debug data};
808 to @emph{query} or prompt a user for input (a special case).
812 This section mainly concentrates on how to build result output,
813 although some of it also applies to other kinds of output.
815 Generation of output that displays the results of an operation
816 involves one or more of the following:
820 output of the actual data
823 formatting the output as appropriate for console output, to make it
824 easily readable by humans
827 machine oriented formatting--a more terse formatting to allow for easy
828 parsing by programs which read @value{GDBN}'s output
831 annotation, whose purpose is to help legacy GUIs to identify interesting
835 The @code{ui_out} routines take care of the first three aspects.
836 Annotations are provided by separate annotation routines. Note that use
837 of annotations for an interface between a GUI and @value{GDBN} is
840 Output can be in the form of a single item, which we call a @dfn{field};
841 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
842 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
843 header and a body. In a BNF-like form:
846 @item <table> @expansion{}
847 @code{<header> <body>}
848 @item <header> @expansion{}
849 @code{@{ <column> @}}
850 @item <column> @expansion{}
851 @code{<width> <alignment> <title>}
852 @item <body> @expansion{}
857 @subsection General Conventions
859 Most @code{ui_out} routines are of type @code{void}, the exceptions are
860 @code{ui_out_stream_new} (which returns a pointer to the newly created
861 object) and the @code{make_cleanup} routines.
863 The first parameter is always the @code{ui_out} vector object, a pointer
864 to a @code{struct ui_out}.
866 The @var{format} parameter is like in @code{printf} family of functions.
867 When it is present, there must also be a variable list of arguments
868 sufficient used to satisfy the @code{%} specifiers in the supplied
871 When a character string argument is not used in a @code{ui_out} function
872 call, a @code{NULL} pointer has to be supplied instead.
875 @subsection Table, Tuple and List Functions
877 @cindex list output functions
878 @cindex table output functions
879 @cindex tuple output functions
880 This section introduces @code{ui_out} routines for building lists,
881 tuples and tables. The routines to output the actual data items
882 (fields) are presented in the next section.
884 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
885 containing information about an object; a @dfn{list} is a sequence of
886 fields where each field describes an identical object.
888 Use the @dfn{table} functions when your output consists of a list of
889 rows (tuples) and the console output should include a heading. Use this
890 even when you are listing just one object but you still want the header.
892 @cindex nesting level in @code{ui_out} functions
893 Tables can not be nested. Tuples and lists can be nested up to a
894 maximum of five levels.
896 The overall structure of the table output code is something like this:
911 Here is the description of table-, tuple- and list-related @code{ui_out}
914 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
915 The function @code{ui_out_table_begin} marks the beginning of the output
916 of a table. It should always be called before any other @code{ui_out}
917 function for a given table. @var{nbrofcols} is the number of columns in
918 the table. @var{nr_rows} is the number of rows in the table.
919 @var{tblid} is an optional string identifying the table. The string
920 pointed to by @var{tblid} is copied by the implementation of
921 @code{ui_out_table_begin}, so the application can free the string if it
924 The companion function @code{ui_out_table_end}, described below, marks
925 the end of the table's output.
928 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
929 @code{ui_out_table_header} provides the header information for a single
930 table column. You call this function several times, one each for every
931 column of the table, after @code{ui_out_table_begin}, but before
932 @code{ui_out_table_body}.
934 The value of @var{width} gives the column width in characters. The
935 value of @var{alignment} is one of @code{left}, @code{center}, and
936 @code{right}, and it specifies how to align the header: left-justify,
937 center, or right-justify it. @var{colhdr} points to a string that
938 specifies the column header; the implementation copies that string, so
939 column header strings in @code{malloc}ed storage can be freed after the
943 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
944 This function delimits the table header from the table body.
947 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
948 This function signals the end of a table's output. It should be called
949 after the table body has been produced by the list and field output
952 There should be exactly one call to @code{ui_out_table_end} for each
953 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
954 will signal an internal error.
957 The output of the tuples that represent the table rows must follow the
958 call to @code{ui_out_table_body} and precede the call to
959 @code{ui_out_table_end}. You build a tuple by calling
960 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
961 calls to functions which actually output fields between them.
963 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
964 This function marks the beginning of a tuple output. @var{id} points
965 to an optional string that identifies the tuple; it is copied by the
966 implementation, and so strings in @code{malloc}ed storage can be freed
970 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
971 This function signals an end of a tuple output. There should be exactly
972 one call to @code{ui_out_tuple_end} for each call to
973 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
977 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
978 This function first opens the tuple and then establishes a cleanup
979 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
980 and correct implementation of the non-portable@footnote{The function
981 cast is not portable ISO C.} code sequence:
983 struct cleanup *old_cleanup;
984 ui_out_tuple_begin (uiout, "...");
985 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
990 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
991 This function marks the beginning of a list output. @var{id} points to
992 an optional string that identifies the list; it is copied by the
993 implementation, and so strings in @code{malloc}ed storage can be freed
997 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
998 This function signals an end of a list output. There should be exactly
999 one call to @code{ui_out_list_end} for each call to
1000 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1004 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1005 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1006 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1007 that will close the list.list.
1010 @subsection Item Output Functions
1012 @cindex item output functions
1013 @cindex field output functions
1015 The functions described below produce output for the actual data
1016 items, or fields, which contain information about the object.
1018 Choose the appropriate function accordingly to your particular needs.
1020 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1021 This is the most general output function. It produces the
1022 representation of the data in the variable-length argument list
1023 according to formatting specifications in @var{format}, a
1024 @code{printf}-like format string. The optional argument @var{fldname}
1025 supplies the name of the field. The data items themselves are
1026 supplied as additional arguments after @var{format}.
1028 This generic function should be used only when it is not possible to
1029 use one of the specialized versions (see below).
1032 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1033 This function outputs a value of an @code{int} variable. It uses the
1034 @code{"%d"} output conversion specification. @var{fldname} specifies
1035 the name of the field.
1038 @deftypefun void ui_out_field_fmt_int (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{fldname}, int @var{value})
1039 This function outputs a value of an @code{int} variable. It differs from
1040 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1041 @var{fldname} specifies
1042 the name of the field.
1045 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1046 This function outputs an address.
1049 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1050 This function outputs a string using the @code{"%s"} conversion
1054 Sometimes, there's a need to compose your output piece by piece using
1055 functions that operate on a stream, such as @code{value_print} or
1056 @code{fprintf_symbol_filtered}. These functions accept an argument of
1057 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1058 used to store the data stream used for the output. When you use one
1059 of these functions, you need a way to pass their results stored in a
1060 @code{ui_file} object to the @code{ui_out} functions. To this end,
1061 you first create a @code{ui_stream} object by calling
1062 @code{ui_out_stream_new}, pass the @code{stream} member of that
1063 @code{ui_stream} object to @code{value_print} and similar functions,
1064 and finally call @code{ui_out_field_stream} to output the field you
1065 constructed. When the @code{ui_stream} object is no longer needed,
1066 you should destroy it and free its memory by calling
1067 @code{ui_out_stream_delete}.
1069 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1070 This function creates a new @code{ui_stream} object which uses the
1071 same output methods as the @code{ui_out} object whose pointer is
1072 passed in @var{uiout}. It returns a pointer to the newly created
1073 @code{ui_stream} object.
1076 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1077 This functions destroys a @code{ui_stream} object specified by
1081 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1082 This function consumes all the data accumulated in
1083 @code{streambuf->stream} and outputs it like
1084 @code{ui_out_field_string} does. After a call to
1085 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1086 the stream is still valid and may be used for producing more fields.
1089 @strong{Important:} If there is any chance that your code could bail
1090 out before completing output generation and reaching the point where
1091 @code{ui_out_stream_delete} is called, it is necessary to set up a
1092 cleanup, to avoid leaking memory and other resources. Here's a
1093 skeleton code to do that:
1096 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1097 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1102 If the function already has the old cleanup chain set (for other kinds
1103 of cleanups), you just have to add your cleanup to it:
1106 mybuf = ui_out_stream_new (uiout);
1107 make_cleanup (ui_out_stream_delete, mybuf);
1110 Note that with cleanups in place, you should not call
1111 @code{ui_out_stream_delete} directly, or you would attempt to free the
1114 @subsection Utility Output Functions
1116 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1117 This function skips a field in a table. Use it if you have to leave
1118 an empty field without disrupting the table alignment. The argument
1119 @var{fldname} specifies a name for the (missing) filed.
1122 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1123 This function outputs the text in @var{string} in a way that makes it
1124 easy to be read by humans. For example, the console implementation of
1125 this method filters the text through a built-in pager, to prevent it
1126 from scrolling off the visible portion of the screen.
1128 Use this function for printing relatively long chunks of text around
1129 the actual field data: the text it produces is not aligned according
1130 to the table's format. Use @code{ui_out_field_string} to output a
1131 string field, and use @code{ui_out_message}, described below, to
1132 output short messages.
1135 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1136 This function outputs @var{nspaces} spaces. It is handy to align the
1137 text produced by @code{ui_out_text} with the rest of the table or
1141 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1142 This function produces a formatted message, provided that the current
1143 verbosity level is at least as large as given by @var{verbosity}. The
1144 current verbosity level is specified by the user with the @samp{set
1145 verbositylevel} command.@footnote{As of this writing (April 2001),
1146 setting verbosity level is not yet implemented, and is always returned
1147 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1148 argument more than zero will cause the message to never be printed.}
1151 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1152 This function gives the console output filter (a paging filter) a hint
1153 of where to break lines which are too long. Ignored for all other
1154 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1155 be printed to indent the wrapped text on the next line; it must remain
1156 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1157 explicit newline is produced by one of the other functions. If
1158 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1161 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1162 This function flushes whatever output has been accumulated so far, if
1163 the UI buffers output.
1167 @subsection Examples of Use of @code{ui_out} functions
1169 @cindex using @code{ui_out} functions
1170 @cindex @code{ui_out} functions, usage examples
1171 This section gives some practical examples of using the @code{ui_out}
1172 functions to generalize the old console-oriented code in
1173 @value{GDBN}. The examples all come from functions defined on the
1174 @file{breakpoints.c} file.
1176 This example, from the @code{breakpoint_1} function, shows how to
1179 The original code was:
1182 if (!found_a_breakpoint++)
1184 annotate_breakpoints_headers ();
1187 printf_filtered ("Num ");
1189 printf_filtered ("Type ");
1191 printf_filtered ("Disp ");
1193 printf_filtered ("Enb ");
1197 printf_filtered ("Address ");
1200 printf_filtered ("What\n");
1202 annotate_breakpoints_table ();
1206 Here's the new version:
1209 nr_printable_breakpoints = @dots{};
1212 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1214 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1216 if (nr_printable_breakpoints > 0)
1217 annotate_breakpoints_headers ();
1218 if (nr_printable_breakpoints > 0)
1220 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1221 if (nr_printable_breakpoints > 0)
1223 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1224 if (nr_printable_breakpoints > 0)
1226 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1227 if (nr_printable_breakpoints > 0)
1229 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1232 if (nr_printable_breakpoints > 0)
1234 if (TARGET_ADDR_BIT <= 32)
1235 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1237 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1239 if (nr_printable_breakpoints > 0)
1241 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1242 ui_out_table_body (uiout);
1243 if (nr_printable_breakpoints > 0)
1244 annotate_breakpoints_table ();
1247 This example, from the @code{print_one_breakpoint} function, shows how
1248 to produce the actual data for the table whose structure was defined
1249 in the above example. The original code was:
1254 printf_filtered ("%-3d ", b->number);
1256 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1257 || ((int) b->type != bptypes[(int) b->type].type))
1258 internal_error ("bptypes table does not describe type #%d.",
1260 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1262 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1264 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1268 This is the new version:
1272 ui_out_tuple_begin (uiout, "bkpt");
1274 ui_out_field_int (uiout, "number", b->number);
1276 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1277 || ((int) b->type != bptypes[(int) b->type].type))
1278 internal_error ("bptypes table does not describe type #%d.",
1280 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1282 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1284 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1288 This example, also from @code{print_one_breakpoint}, shows how to
1289 produce a complicated output field using the @code{print_expression}
1290 functions which requires a stream to be passed. It also shows how to
1291 automate stream destruction with cleanups. The original code was:
1295 print_expression (b->exp, gdb_stdout);
1301 struct ui_stream *stb = ui_out_stream_new (uiout);
1302 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1305 print_expression (b->exp, stb->stream);
1306 ui_out_field_stream (uiout, "what", local_stream);
1309 This example, also from @code{print_one_breakpoint}, shows how to use
1310 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1315 if (b->dll_pathname == NULL)
1316 printf_filtered ("<any library> ");
1318 printf_filtered ("library \"%s\" ", b->dll_pathname);
1325 if (b->dll_pathname == NULL)
1327 ui_out_field_string (uiout, "what", "<any library>");
1328 ui_out_spaces (uiout, 1);
1332 ui_out_text (uiout, "library \"");
1333 ui_out_field_string (uiout, "what", b->dll_pathname);
1334 ui_out_text (uiout, "\" ");
1338 The following example from @code{print_one_breakpoint} shows how to
1339 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1344 if (b->forked_inferior_pid != 0)
1345 printf_filtered ("process %d ", b->forked_inferior_pid);
1352 if (b->forked_inferior_pid != 0)
1354 ui_out_text (uiout, "process ");
1355 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1356 ui_out_spaces (uiout, 1);
1360 Here's an example of using @code{ui_out_field_string}. The original
1365 if (b->exec_pathname != NULL)
1366 printf_filtered ("program \"%s\" ", b->exec_pathname);
1373 if (b->exec_pathname != NULL)
1375 ui_out_text (uiout, "program \"");
1376 ui_out_field_string (uiout, "what", b->exec_pathname);
1377 ui_out_text (uiout, "\" ");
1381 Finally, here's an example of printing an address. The original code:
1385 printf_filtered ("%s ",
1386 hex_string_custom ((unsigned long) b->address, 8));
1393 ui_out_field_core_addr (uiout, "Address", b->address);
1397 @section Console Printing
1406 @cindex @code{libgdb}
1407 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1408 to provide an API to @value{GDBN}'s functionality.
1411 @cindex @code{libgdb}
1412 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1413 better able to support graphical and other environments.
1415 Since @code{libgdb} development is on-going, its architecture is still
1416 evolving. The following components have so far been identified:
1420 Observer - @file{gdb-events.h}.
1422 Builder - @file{ui-out.h}
1424 Event Loop - @file{event-loop.h}
1426 Library - @file{gdb.h}
1429 The model that ties these components together is described below.
1431 @section The @code{libgdb} Model
1433 A client of @code{libgdb} interacts with the library in two ways.
1437 As an observer (using @file{gdb-events}) receiving notifications from
1438 @code{libgdb} of any internal state changes (break point changes, run
1441 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1442 obtain various status values from @value{GDBN}.
1445 Since @code{libgdb} could have multiple clients (e.g. a GUI supporting
1446 the existing @value{GDBN} CLI), those clients must co-operate when
1447 controlling @code{libgdb}. In particular, a client must ensure that
1448 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1449 before responding to a @file{gdb-event} by making a query.
1451 @section CLI support
1453 At present @value{GDBN}'s CLI is very much entangled in with the core of
1454 @code{libgdb}. Consequently, a client wishing to include the CLI in
1455 their interface needs to carefully co-ordinate its own and the CLI's
1458 It is suggested that the client set @code{libgdb} up to be bi-modal
1459 (alternate between CLI and client query modes). The notes below sketch
1464 The client registers itself as an observer of @code{libgdb}.
1466 The client create and install @code{cli-out} builder using its own
1467 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1468 @code{gdb_stdout} streams.
1470 The client creates a separate custom @code{ui-out} builder that is only
1471 used while making direct queries to @code{libgdb}.
1474 When the client receives input intended for the CLI, it simply passes it
1475 along. Since the @code{cli-out} builder is installed by default, all
1476 the CLI output in response to that command is routed (pronounced rooted)
1477 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1478 At the same time, the client is kept abreast of internal changes by
1479 virtue of being a @code{libgdb} observer.
1481 The only restriction on the client is that it must wait until
1482 @code{libgdb} becomes idle before initiating any queries (using the
1483 client's custom builder).
1485 @section @code{libgdb} components
1487 @subheading Observer - @file{gdb-events.h}
1488 @file{gdb-events} provides the client with a very raw mechanism that can
1489 be used to implement an observer. At present it only allows for one
1490 observer and that observer must, internally, handle the need to delay
1491 the processing of any event notifications until after @code{libgdb} has
1492 finished the current command.
1494 @subheading Builder - @file{ui-out.h}
1495 @file{ui-out} provides the infrastructure necessary for a client to
1496 create a builder. That builder is then passed down to @code{libgdb}
1497 when doing any queries.
1499 @subheading Event Loop - @file{event-loop.h}
1500 @c There could be an entire section on the event-loop
1501 @file{event-loop}, currently non-re-entrant, provides a simple event
1502 loop. A client would need to either plug its self into this loop or,
1503 implement a new event-loop that GDB would use.
1505 The event-loop will eventually be made re-entrant. This is so that
1506 @value{GDBN} can better handle the problem of some commands blocking
1507 instead of returning.
1509 @subheading Library - @file{gdb.h}
1510 @file{libgdb} is the most obvious component of this system. It provides
1511 the query interface. Each function is parameterized by a @code{ui-out}
1512 builder. The result of the query is constructed using that builder
1513 before the query function returns.
1515 @node Symbol Handling
1517 @chapter Symbol Handling
1519 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1520 functions, and types.
1522 @section Symbol Reading
1524 @cindex symbol reading
1525 @cindex reading of symbols
1526 @cindex symbol files
1527 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1528 file is the file containing the program which @value{GDBN} is
1529 debugging. @value{GDBN} can be directed to use a different file for
1530 symbols (with the @samp{symbol-file} command), and it can also read
1531 more symbols via the @samp{add-file} and @samp{load} commands, or while
1532 reading symbols from shared libraries.
1534 @findex find_sym_fns
1535 Symbol files are initially opened by code in @file{symfile.c} using
1536 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1537 of the file by examining its header. @code{find_sym_fns} then uses
1538 this identification to locate a set of symbol-reading functions.
1540 @findex add_symtab_fns
1541 @cindex @code{sym_fns} structure
1542 @cindex adding a symbol-reading module
1543 Symbol-reading modules identify themselves to @value{GDBN} by calling
1544 @code{add_symtab_fns} during their module initialization. The argument
1545 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1546 name (or name prefix) of the symbol format, the length of the prefix,
1547 and pointers to four functions. These functions are called at various
1548 times to process symbol files whose identification matches the specified
1551 The functions supplied by each module are:
1554 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1556 @cindex secondary symbol file
1557 Called from @code{symbol_file_add} when we are about to read a new
1558 symbol file. This function should clean up any internal state (possibly
1559 resulting from half-read previous files, for example) and prepare to
1560 read a new symbol file. Note that the symbol file which we are reading
1561 might be a new ``main'' symbol file, or might be a secondary symbol file
1562 whose symbols are being added to the existing symbol table.
1564 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1565 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1566 new symbol file being read. Its @code{private} field has been zeroed,
1567 and can be modified as desired. Typically, a struct of private
1568 information will be @code{malloc}'d, and a pointer to it will be placed
1569 in the @code{private} field.
1571 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1572 @code{error} if it detects an unavoidable problem.
1574 @item @var{xyz}_new_init()
1576 Called from @code{symbol_file_add} when discarding existing symbols.
1577 This function needs only handle the symbol-reading module's internal
1578 state; the symbol table data structures visible to the rest of
1579 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1580 arguments and no result. It may be called after
1581 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1582 may be called alone if all symbols are simply being discarded.
1584 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1586 Called from @code{symbol_file_add} to actually read the symbols from a
1587 symbol-file into a set of psymtabs or symtabs.
1589 @code{sf} points to the @code{struct sym_fns} originally passed to
1590 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1591 the offset between the file's specified start address and its true
1592 address in memory. @code{mainline} is 1 if this is the main symbol
1593 table being read, and 0 if a secondary symbol file (e.g. shared library
1594 or dynamically loaded file) is being read.@refill
1597 In addition, if a symbol-reading module creates psymtabs when
1598 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1599 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1600 from any point in the @value{GDBN} symbol-handling code.
1603 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1605 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1606 the psymtab has not already been read in and had its @code{pst->symtab}
1607 pointer set. The argument is the psymtab to be fleshed-out into a
1608 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1609 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1610 zero if there were no symbols in that part of the symbol file.
1613 @section Partial Symbol Tables
1615 @value{GDBN} has three types of symbol tables:
1618 @cindex full symbol table
1621 Full symbol tables (@dfn{symtabs}). These contain the main
1622 information about symbols and addresses.
1626 Partial symbol tables (@dfn{psymtabs}). These contain enough
1627 information to know when to read the corresponding part of the full
1630 @cindex minimal symbol table
1633 Minimal symbol tables (@dfn{msymtabs}). These contain information
1634 gleaned from non-debugging symbols.
1637 @cindex partial symbol table
1638 This section describes partial symbol tables.
1640 A psymtab is constructed by doing a very quick pass over an executable
1641 file's debugging information. Small amounts of information are
1642 extracted---enough to identify which parts of the symbol table will
1643 need to be re-read and fully digested later, when the user needs the
1644 information. The speed of this pass causes @value{GDBN} to start up very
1645 quickly. Later, as the detailed rereading occurs, it occurs in small
1646 pieces, at various times, and the delay therefrom is mostly invisible to
1648 @c (@xref{Symbol Reading}.)
1650 The symbols that show up in a file's psymtab should be, roughly, those
1651 visible to the debugger's user when the program is not running code from
1652 that file. These include external symbols and types, static symbols and
1653 types, and @code{enum} values declared at file scope.
1655 The psymtab also contains the range of instruction addresses that the
1656 full symbol table would represent.
1658 @cindex finding a symbol
1659 @cindex symbol lookup
1660 The idea is that there are only two ways for the user (or much of the
1661 code in the debugger) to reference a symbol:
1664 @findex find_pc_function
1665 @findex find_pc_line
1667 By its address (e.g. execution stops at some address which is inside a
1668 function in this file). The address will be noticed to be in the
1669 range of this psymtab, and the full symtab will be read in.
1670 @code{find_pc_function}, @code{find_pc_line}, and other
1671 @code{find_pc_@dots{}} functions handle this.
1673 @cindex lookup_symbol
1676 (e.g. the user asks to print a variable, or set a breakpoint on a
1677 function). Global names and file-scope names will be found in the
1678 psymtab, which will cause the symtab to be pulled in. Local names will
1679 have to be qualified by a global name, or a file-scope name, in which
1680 case we will have already read in the symtab as we evaluated the
1681 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1682 local scope, in which case the first case applies. @code{lookup_symbol}
1683 does most of the work here.
1686 The only reason that psymtabs exist is to cause a symtab to be read in
1687 at the right moment. Any symbol that can be elided from a psymtab,
1688 while still causing that to happen, should not appear in it. Since
1689 psymtabs don't have the idea of scope, you can't put local symbols in
1690 them anyway. Psymtabs don't have the idea of the type of a symbol,
1691 either, so types need not appear, unless they will be referenced by
1694 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1695 been read, and another way if the corresponding symtab has been read
1696 in. Such bugs are typically caused by a psymtab that does not contain
1697 all the visible symbols, or which has the wrong instruction address
1700 The psymtab for a particular section of a symbol file (objfile) could be
1701 thrown away after the symtab has been read in. The symtab should always
1702 be searched before the psymtab, so the psymtab will never be used (in a
1703 bug-free environment). Currently, psymtabs are allocated on an obstack,
1704 and all the psymbols themselves are allocated in a pair of large arrays
1705 on an obstack, so there is little to be gained by trying to free them
1706 unless you want to do a lot more work.
1710 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
1712 @cindex fundamental types
1713 These are the fundamental types that @value{GDBN} uses internally. Fundamental
1714 types from the various debugging formats (stabs, ELF, etc) are mapped
1715 into one of these. They are basically a union of all fundamental types
1716 that @value{GDBN} knows about for all the languages that @value{GDBN}
1719 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
1722 Each time @value{GDBN} builds an internal type, it marks it with one
1723 of these types. The type may be a fundamental type, such as
1724 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
1725 which is a pointer to another type. Typically, several @code{FT_*}
1726 types map to one @code{TYPE_CODE_*} type, and are distinguished by
1727 other members of the type struct, such as whether the type is signed
1728 or unsigned, and how many bits it uses.
1730 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
1732 These are instances of type structs that roughly correspond to
1733 fundamental types and are created as global types for @value{GDBN} to
1734 use for various ugly historical reasons. We eventually want to
1735 eliminate these. Note for example that @code{builtin_type_int}
1736 initialized in @file{gdbtypes.c} is basically the same as a
1737 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
1738 an @code{FT_INTEGER} fundamental type. The difference is that the
1739 @code{builtin_type} is not associated with any particular objfile, and
1740 only one instance exists, while @file{c-lang.c} builds as many
1741 @code{TYPE_CODE_INT} types as needed, with each one associated with
1742 some particular objfile.
1744 @section Object File Formats
1745 @cindex object file formats
1749 @cindex @code{a.out} format
1750 The @code{a.out} format is the original file format for Unix. It
1751 consists of three sections: @code{text}, @code{data}, and @code{bss},
1752 which are for program code, initialized data, and uninitialized data,
1755 The @code{a.out} format is so simple that it doesn't have any reserved
1756 place for debugging information. (Hey, the original Unix hackers used
1757 @samp{adb}, which is a machine-language debugger!) The only debugging
1758 format for @code{a.out} is stabs, which is encoded as a set of normal
1759 symbols with distinctive attributes.
1761 The basic @code{a.out} reader is in @file{dbxread.c}.
1766 The COFF format was introduced with System V Release 3 (SVR3) Unix.
1767 COFF files may have multiple sections, each prefixed by a header. The
1768 number of sections is limited.
1770 The COFF specification includes support for debugging. Although this
1771 was a step forward, the debugging information was woefully limited. For
1772 instance, it was not possible to represent code that came from an
1775 The COFF reader is in @file{coffread.c}.
1779 @cindex ECOFF format
1780 ECOFF is an extended COFF originally introduced for Mips and Alpha
1783 The basic ECOFF reader is in @file{mipsread.c}.
1787 @cindex XCOFF format
1788 The IBM RS/6000 running AIX uses an object file format called XCOFF.
1789 The COFF sections, symbols, and line numbers are used, but debugging
1790 symbols are @code{dbx}-style stabs whose strings are located in the
1791 @code{.debug} section (rather than the string table). For more
1792 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
1794 The shared library scheme has a clean interface for figuring out what
1795 shared libraries are in use, but the catch is that everything which
1796 refers to addresses (symbol tables and breakpoints at least) needs to be
1797 relocated for both shared libraries and the main executable. At least
1798 using the standard mechanism this can only be done once the program has
1799 been run (or the core file has been read).
1803 @cindex PE-COFF format
1804 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
1805 executables. PE is basically COFF with additional headers.
1807 While BFD includes special PE support, @value{GDBN} needs only the basic
1813 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
1814 to COFF in being organized into a number of sections, but it removes
1815 many of COFF's limitations.
1817 The basic ELF reader is in @file{elfread.c}.
1822 SOM is HP's object file and debug format (not to be confused with IBM's
1823 SOM, which is a cross-language ABI).
1825 The SOM reader is in @file{hpread.c}.
1827 @subsection Other File Formats
1829 @cindex Netware Loadable Module format
1830 Other file formats that have been supported by @value{GDBN} include Netware
1831 Loadable Modules (@file{nlmread.c}).
1833 @section Debugging File Formats
1835 This section describes characteristics of debugging information that
1836 are independent of the object file format.
1840 @cindex stabs debugging info
1841 @code{stabs} started out as special symbols within the @code{a.out}
1842 format. Since then, it has been encapsulated into other file
1843 formats, such as COFF and ELF.
1845 While @file{dbxread.c} does some of the basic stab processing,
1846 including for encapsulated versions, @file{stabsread.c} does
1851 @cindex COFF debugging info
1852 The basic COFF definition includes debugging information. The level
1853 of support is minimal and non-extensible, and is not often used.
1855 @subsection Mips debug (Third Eye)
1857 @cindex ECOFF debugging info
1858 ECOFF includes a definition of a special debug format.
1860 The file @file{mdebugread.c} implements reading for this format.
1864 @cindex DWARF 1 debugging info
1865 DWARF 1 is a debugging format that was originally designed to be
1866 used with ELF in SVR4 systems.
1871 @c If defined, these are the producer strings in a DWARF 1 file. All of
1872 @c these have reasonable defaults already.
1874 The DWARF 1 reader is in @file{dwarfread.c}.
1878 @cindex DWARF 2 debugging info
1879 DWARF 2 is an improved but incompatible version of DWARF 1.
1881 The DWARF 2 reader is in @file{dwarf2read.c}.
1885 @cindex SOM debugging info
1886 Like COFF, the SOM definition includes debugging information.
1888 @section Adding a New Symbol Reader to @value{GDBN}
1890 @cindex adding debugging info reader
1891 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
1892 there is probably little to be done.
1894 If you need to add a new object file format, you must first add it to
1895 BFD. This is beyond the scope of this document.
1897 You must then arrange for the BFD code to provide access to the
1898 debugging symbols. Generally @value{GDBN} will have to call swapping routines
1899 from BFD and a few other BFD internal routines to locate the debugging
1900 information. As much as possible, @value{GDBN} should not depend on the BFD
1901 internal data structures.
1903 For some targets (e.g., COFF), there is a special transfer vector used
1904 to call swapping routines, since the external data structures on various
1905 platforms have different sizes and layouts. Specialized routines that
1906 will only ever be implemented by one object file format may be called
1907 directly. This interface should be described in a file
1908 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
1911 @node Language Support
1913 @chapter Language Support
1915 @cindex language support
1916 @value{GDBN}'s language support is mainly driven by the symbol reader,
1917 although it is possible for the user to set the source language
1920 @value{GDBN} chooses the source language by looking at the extension
1921 of the file recorded in the debug info; @file{.c} means C, @file{.f}
1922 means Fortran, etc. It may also use a special-purpose language
1923 identifier if the debug format supports it, like with DWARF.
1925 @section Adding a Source Language to @value{GDBN}
1927 @cindex adding source language
1928 To add other languages to @value{GDBN}'s expression parser, follow the
1932 @item Create the expression parser.
1934 @cindex expression parser
1935 This should reside in a file @file{@var{lang}-exp.y}. Routines for
1936 building parsed expressions into a @code{union exp_element} list are in
1939 @cindex language parser
1940 Since we can't depend upon everyone having Bison, and YACC produces
1941 parsers that define a bunch of global names, the following lines
1942 @strong{must} be included at the top of the YACC parser, to prevent the
1943 various parsers from defining the same global names:
1946 #define yyparse @var{lang}_parse
1947 #define yylex @var{lang}_lex
1948 #define yyerror @var{lang}_error
1949 #define yylval @var{lang}_lval
1950 #define yychar @var{lang}_char
1951 #define yydebug @var{lang}_debug
1952 #define yypact @var{lang}_pact
1953 #define yyr1 @var{lang}_r1
1954 #define yyr2 @var{lang}_r2
1955 #define yydef @var{lang}_def
1956 #define yychk @var{lang}_chk
1957 #define yypgo @var{lang}_pgo
1958 #define yyact @var{lang}_act
1959 #define yyexca @var{lang}_exca
1960 #define yyerrflag @var{lang}_errflag
1961 #define yynerrs @var{lang}_nerrs
1964 At the bottom of your parser, define a @code{struct language_defn} and
1965 initialize it with the right values for your language. Define an
1966 @code{initialize_@var{lang}} routine and have it call
1967 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
1968 that your language exists. You'll need some other supporting variables
1969 and functions, which will be used via pointers from your
1970 @code{@var{lang}_language_defn}. See the declaration of @code{struct
1971 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
1972 for more information.
1974 @item Add any evaluation routines, if necessary
1976 @cindex expression evaluation routines
1977 @findex evaluate_subexp
1978 @findex prefixify_subexp
1979 @findex length_of_subexp
1980 If you need new opcodes (that represent the operations of the language),
1981 add them to the enumerated type in @file{expression.h}. Add support
1982 code for these operations in the @code{evaluate_subexp} function
1983 defined in the file @file{eval.c}. Add cases
1984 for new opcodes in two functions from @file{parse.c}:
1985 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
1986 the number of @code{exp_element}s that a given operation takes up.
1988 @item Update some existing code
1990 Add an enumerated identifier for your language to the enumerated type
1991 @code{enum language} in @file{defs.h}.
1993 Update the routines in @file{language.c} so your language is included.
1994 These routines include type predicates and such, which (in some cases)
1995 are language dependent. If your language does not appear in the switch
1996 statement, an error is reported.
1998 @vindex current_language
1999 Also included in @file{language.c} is the code that updates the variable
2000 @code{current_language}, and the routines that translate the
2001 @code{language_@var{lang}} enumerated identifier into a printable
2004 @findex _initialize_language
2005 Update the function @code{_initialize_language} to include your
2006 language. This function picks the default language upon startup, so is
2007 dependent upon which languages that @value{GDBN} is built for.
2009 @findex allocate_symtab
2010 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2011 code so that the language of each symtab (source file) is set properly.
2012 This is used to determine the language to use at each stack frame level.
2013 Currently, the language is set based upon the extension of the source
2014 file. If the language can be better inferred from the symbol
2015 information, please set the language of the symtab in the symbol-reading
2018 @findex print_subexp
2019 @findex op_print_tab
2020 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2021 expression opcodes you have added to @file{expression.h}. Also, add the
2022 printed representations of your operators to @code{op_print_tab}.
2024 @item Add a place of call
2027 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2028 @code{parse_exp_1} (defined in @file{parse.c}).
2030 @item Use macros to trim code
2032 @cindex trimming language-dependent code
2033 The user has the option of building @value{GDBN} for some or all of the
2034 languages. If the user decides to build @value{GDBN} for the language
2035 @var{lang}, then every file dependent on @file{language.h} will have the
2036 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2037 leave out large routines that the user won't need if he or she is not
2038 using your language.
2040 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2041 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2042 compiled form of your parser) is not linked into @value{GDBN} at all.
2044 See the file @file{configure.in} for how @value{GDBN} is configured
2045 for different languages.
2047 @item Edit @file{Makefile.in}
2049 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2050 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2051 not get linked in, or, worse yet, it may not get @code{tar}red into the
2056 @node Host Definition
2058 @chapter Host Definition
2060 With the advent of Autoconf, it's rarely necessary to have host
2061 definition machinery anymore. The following information is provided,
2062 mainly, as an historical reference.
2064 @section Adding a New Host
2066 @cindex adding a new host
2067 @cindex host, adding
2068 @value{GDBN}'s host configuration support normally happens via Autoconf.
2069 New host-specific definitions should not be needed. Older hosts
2070 @value{GDBN} still use the host-specific definitions and files listed
2071 below, but these mostly exist for historical reasons, and will
2072 eventually disappear.
2075 @item gdb/config/@var{arch}/@var{xyz}.mh
2076 This file once contained both host and native configuration information
2077 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2078 configuration information is now handed by Autoconf.
2080 Host configuration information included a definition of
2081 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2082 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2083 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2085 New host only configurations do not need this file.
2087 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2088 This file once contained definitions and includes required when hosting
2089 gdb on machine @var{xyz}. Those definitions and includes are now
2090 handled by Autoconf.
2092 New host and native configurations do not need this file.
2094 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2095 file to define the macros @var{HOST_FLOAT_FORMAT},
2096 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2097 also needs to be replaced with either an Autoconf or run-time test.}
2101 @subheading Generic Host Support Files
2103 @cindex generic host support
2104 There are some ``generic'' versions of routines that can be used by
2105 various systems. These can be customized in various ways by macros
2106 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2107 the @var{xyz} host, you can just include the generic file's name (with
2108 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2110 Otherwise, if your machine needs custom support routines, you will need
2111 to write routines that perform the same functions as the generic file.
2112 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2113 into @code{XDEPFILES}.
2116 @cindex remote debugging support
2117 @cindex serial line support
2119 This contains serial line support for Unix systems. This is always
2120 included, via the makefile variable @code{SER_HARDWIRE}; override this
2121 variable in the @file{.mh} file to avoid it.
2124 This contains serial line support for 32-bit programs running under DOS,
2125 using the DJGPP (a.k.a.@: GO32) execution environment.
2127 @cindex TCP remote support
2129 This contains generic TCP support using sockets.
2132 @section Host Conditionals
2134 When @value{GDBN} is configured and compiled, various macros are
2135 defined or left undefined, to control compilation based on the
2136 attributes of the host system. These macros and their meanings (or if
2137 the meaning is not documented here, then one of the source files where
2138 they are used is indicated) are:
2141 @item @value{GDBN}INIT_FILENAME
2142 The default name of @value{GDBN}'s initialization file (normally
2146 This macro is deprecated.
2148 @item SIGWINCH_HANDLER
2149 If your host defines @code{SIGWINCH}, you can define this to be the name
2150 of a function to be called if @code{SIGWINCH} is received.
2152 @item SIGWINCH_HANDLER_BODY
2153 Define this to expand into code that will define the function named by
2154 the expansion of @code{SIGWINCH_HANDLER}.
2156 @item ALIGN_STACK_ON_STARTUP
2157 @cindex stack alignment
2158 Define this if your system is of a sort that will crash in
2159 @code{tgetent} if the stack happens not to be longword-aligned when
2160 @code{main} is called. This is a rare situation, but is known to occur
2161 on several different types of systems.
2163 @item CRLF_SOURCE_FILES
2164 @cindex DOS text files
2165 Define this if host files use @code{\r\n} rather than @code{\n} as a
2166 line terminator. This will cause source file listings to omit @code{\r}
2167 characters when printing and it will allow @code{\r\n} line endings of files
2168 which are ``sourced'' by gdb. It must be possible to open files in binary
2169 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2171 @item DEFAULT_PROMPT
2173 The default value of the prompt string (normally @code{"(gdb) "}).
2176 @cindex terminal device
2177 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2180 Define this if binary files are opened the same way as text files.
2184 In some cases, use the system call @code{mmap} for reading symbol
2185 tables. For some machines this allows for sharing and quick updates.
2188 Define this if the host system has @code{termio.h}.
2195 Values for host-side constants.
2198 Substitute for isatty, if not available.
2201 This is the longest integer type available on the host. If not defined,
2202 it will default to @code{long long} or @code{long}, depending on
2203 @code{CC_HAS_LONG_LONG}.
2205 @item CC_HAS_LONG_LONG
2206 @cindex @code{long long} data type
2207 Define this if the host C compiler supports @code{long long}. This is set
2208 by the @code{configure} script.
2210 @item PRINTF_HAS_LONG_LONG
2211 Define this if the host can handle printing of long long integers via
2212 the printf format conversion specifier @code{ll}. This is set by the
2213 @code{configure} script.
2215 @item HAVE_LONG_DOUBLE
2216 Define this if the host C compiler supports @code{long double}. This is
2217 set by the @code{configure} script.
2219 @item PRINTF_HAS_LONG_DOUBLE
2220 Define this if the host can handle printing of long double float-point
2221 numbers via the printf format conversion specifier @code{Lg}. This is
2222 set by the @code{configure} script.
2224 @item SCANF_HAS_LONG_DOUBLE
2225 Define this if the host can handle the parsing of long double
2226 float-point numbers via the scanf format conversion specifier
2227 @code{Lg}. This is set by the @code{configure} script.
2229 @item LSEEK_NOT_LINEAR
2230 Define this if @code{lseek (n)} does not necessarily move to byte number
2231 @code{n} in the file. This is only used when reading source files. It
2232 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2235 This macro is used as the argument to @code{lseek} (or, most commonly,
2236 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2237 which is the POSIX equivalent.
2240 If defined, this should be one or more tokens, such as @code{volatile},
2241 that can be used in both the declaration and definition of functions to
2242 indicate that they never return. The default is already set correctly
2243 if compiling with GCC. This will almost never need to be defined.
2246 If defined, this should be one or more tokens, such as
2247 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2248 of functions to indicate that they never return. The default is already
2249 set correctly if compiling with GCC. This will almost never need to be
2254 Define these to appropriate value for the system @code{lseek}, if not already
2258 This is the signal for stopping @value{GDBN}. Defaults to
2259 @code{SIGTSTP}. (Only redefined for the Convex.)
2262 Means that System V (prior to SVR4) include files are in use. (FIXME:
2263 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2264 @file{utils.c} for other things, at the moment.)
2267 Define this to help placate @code{lint} in some situations.
2270 Define this to override the defaults of @code{__volatile__} or
2275 @node Target Architecture Definition
2277 @chapter Target Architecture Definition
2279 @cindex target architecture definition
2280 @value{GDBN}'s target architecture defines what sort of
2281 machine-language programs @value{GDBN} can work with, and how it works
2284 The target architecture object is implemented as the C structure
2285 @code{struct gdbarch *}. The structure, and its methods, are generated
2286 using the Bourne shell script @file{gdbarch.sh}.
2288 @section Operating System ABI Variant Handling
2289 @cindex OS ABI variants
2291 @value{GDBN} provides a mechanism for handling variations in OS
2292 ABIs. An OS ABI variant may have influence over any number of
2293 variables in the target architecture definition. There are two major
2294 components in the OS ABI mechanism: sniffers and handlers.
2296 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2297 (the architecture may be wildcarded) in an attempt to determine the
2298 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2299 to be @dfn{generic}, while sniffers for a specific architecture are
2300 considered to be @dfn{specific}. A match from a specific sniffer
2301 overrides a match from a generic sniffer. Multiple sniffers for an
2302 architecture/flavour may exist, in order to differentiate between two
2303 different operating systems which use the same basic file format. The
2304 OS ABI framework provides a generic sniffer for ELF-format files which
2305 examines the @code{EI_OSABI} field of the ELF header, as well as note
2306 sections known to be used by several operating systems.
2308 @cindex fine-tuning @code{gdbarch} structure
2309 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2310 selected OS ABI. There may be only one handler for a given OS ABI
2311 for each BFD architecture.
2313 The following OS ABI variants are defined in @file{osabi.h}:
2317 @findex GDB_OSABI_UNKNOWN
2318 @item GDB_OSABI_UNKNOWN
2319 The ABI of the inferior is unknown. The default @code{gdbarch}
2320 settings for the architecture will be used.
2322 @findex GDB_OSABI_SVR4
2323 @item GDB_OSABI_SVR4
2324 UNIX System V Release 4
2326 @findex GDB_OSABI_HURD
2327 @item GDB_OSABI_HURD
2328 GNU using the Hurd kernel
2330 @findex GDB_OSABI_SOLARIS
2331 @item GDB_OSABI_SOLARIS
2334 @findex GDB_OSABI_OSF1
2335 @item GDB_OSABI_OSF1
2336 OSF/1, including Digital UNIX and Compaq Tru64 UNIX
2338 @findex GDB_OSABI_LINUX
2339 @item GDB_OSABI_LINUX
2340 GNU using the Linux kernel
2342 @findex GDB_OSABI_FREEBSD_AOUT
2343 @item GDB_OSABI_FREEBSD_AOUT
2344 FreeBSD using the a.out executable format
2346 @findex GDB_OSABI_FREEBSD_ELF
2347 @item GDB_OSABI_FREEBSD_ELF
2348 FreeBSD using the ELF executable format
2350 @findex GDB_OSABI_NETBSD_AOUT
2351 @item GDB_OSABI_NETBSD_AOUT
2352 NetBSD using the a.out executable format
2354 @findex GDB_OSABI_NETBSD_ELF
2355 @item GDB_OSABI_NETBSD_ELF
2356 NetBSD using the ELF executable format
2358 @findex GDB_OSABI_WINCE
2359 @item GDB_OSABI_WINCE
2362 @findex GDB_OSABI_GO32
2363 @item GDB_OSABI_GO32
2366 @findex GDB_OSABI_NETWARE
2367 @item GDB_OSABI_NETWARE
2370 @findex GDB_OSABI_ARM_EABI_V1
2371 @item GDB_OSABI_ARM_EABI_V1
2372 ARM Embedded ABI version 1
2374 @findex GDB_OSABI_ARM_EABI_V2
2375 @item GDB_OSABI_ARM_EABI_V2
2376 ARM Embedded ABI version 2
2378 @findex GDB_OSABI_ARM_APCS
2379 @item GDB_OSABI_ARM_APCS
2380 Generic ARM Procedure Call Standard
2384 Here are the functions that make up the OS ABI framework:
2386 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2387 Return the name of the OS ABI corresponding to @var{osabi}.
2390 @deftypefun void gdbarch_register_osabi (enum bfd_architecture @var{arch}, unsigned long @var{machine}, enum gdb_osabi @var{osabi}, void (*@var{init_osabi})(struct gdbarch_info @var{info}, struct gdbarch *@var{gdbarch}))
2391 Register the OS ABI handler specified by @var{init_osabi} for the
2392 architecture, machine type and OS ABI specified by @var{arch},
2393 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2394 machine type, which implies the architecture's default machine type,
2398 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2399 Register the OS ABI file sniffer specified by @var{sniffer} for the
2400 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2401 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2402 be generic, and is allowed to examine @var{flavour}-flavoured files for
2406 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2407 Examine the file described by @var{abfd} to determine its OS ABI.
2408 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2412 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2413 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2414 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2415 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2416 architecture, a warning will be issued and the debugging session will continue
2417 with the defaults already established for @var{gdbarch}.
2420 @section Registers and Memory
2422 @value{GDBN}'s model of the target machine is rather simple.
2423 @value{GDBN} assumes the machine includes a bank of registers and a
2424 block of memory. Each register may have a different size.
2426 @value{GDBN} does not have a magical way to match up with the
2427 compiler's idea of which registers are which; however, it is critical
2428 that they do match up accurately. The only way to make this work is
2429 to get accurate information about the order that the compiler uses,
2430 and to reflect that in the @code{REGISTER_NAME} and related macros.
2432 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2434 @section Pointers Are Not Always Addresses
2435 @cindex pointer representation
2436 @cindex address representation
2437 @cindex word-addressed machines
2438 @cindex separate data and code address spaces
2439 @cindex spaces, separate data and code address
2440 @cindex address spaces, separate data and code
2441 @cindex code pointers, word-addressed
2442 @cindex converting between pointers and addresses
2443 @cindex D10V addresses
2445 On almost all 32-bit architectures, the representation of a pointer is
2446 indistinguishable from the representation of some fixed-length number
2447 whose value is the byte address of the object pointed to. On such
2448 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2449 However, architectures with smaller word sizes are often cramped for
2450 address space, so they may choose a pointer representation that breaks this
2451 identity, and allows a larger code address space.
2453 For example, the Renesas D10V is a 16-bit VLIW processor whose
2454 instructions are 32 bits long@footnote{Some D10V instructions are
2455 actually pairs of 16-bit sub-instructions. However, since you can't
2456 jump into the middle of such a pair, code addresses can only refer to
2457 full 32 bit instructions, which is what matters in this explanation.}.
2458 If the D10V used ordinary byte addresses to refer to code locations,
2459 then the processor would only be able to address 64kb of instructions.
2460 However, since instructions must be aligned on four-byte boundaries, the
2461 low two bits of any valid instruction's byte address are always
2462 zero---byte addresses waste two bits. So instead of byte addresses,
2463 the D10V uses word addresses---byte addresses shifted right two bits---to
2464 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2467 However, this means that code pointers and data pointers have different
2468 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2469 @code{0xC020} when used as a data address, but refers to byte address
2470 @code{0x30080} when used as a code address.
2472 (The D10V also uses separate code and data address spaces, which also
2473 affects the correspondence between pointers and addresses, but we're
2474 going to ignore that here; this example is already too long.)
2476 To cope with architectures like this---the D10V is not the only
2477 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2478 byte numbers, and @dfn{pointers}, which are the target's representation
2479 of an address of a particular type of data. In the example above,
2480 @code{0xC020} is the pointer, which refers to one of the addresses
2481 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2482 @value{GDBN} provides functions for turning a pointer into an address
2483 and vice versa, in the appropriate way for the current architecture.
2485 Unfortunately, since addresses and pointers are identical on almost all
2486 processors, this distinction tends to bit-rot pretty quickly. Thus,
2487 each time you port @value{GDBN} to an architecture which does
2488 distinguish between pointers and addresses, you'll probably need to
2489 clean up some architecture-independent code.
2491 Here are functions which convert between pointers and addresses:
2493 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2494 Treat the bytes at @var{buf} as a pointer or reference of type
2495 @var{type}, and return the address it represents, in a manner
2496 appropriate for the current architecture. This yields an address
2497 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2498 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2501 For example, if the current architecture is the Intel x86, this function
2502 extracts a little-endian integer of the appropriate length from
2503 @var{buf} and returns it. However, if the current architecture is the
2504 D10V, this function will return a 16-bit integer extracted from
2505 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2507 If @var{type} is not a pointer or reference type, then this function
2508 will signal an internal error.
2511 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2512 Store the address @var{addr} in @var{buf}, in the proper format for a
2513 pointer of type @var{type} in the current architecture. Note that
2514 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2517 For example, if the current architecture is the Intel x86, this function
2518 stores @var{addr} unmodified as a little-endian integer of the
2519 appropriate length in @var{buf}. However, if the current architecture
2520 is the D10V, this function divides @var{addr} by four if @var{type} is
2521 a pointer to a function, and then stores it in @var{buf}.
2523 If @var{type} is not a pointer or reference type, then this function
2524 will signal an internal error.
2527 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2528 Assuming that @var{val} is a pointer, return the address it represents,
2529 as appropriate for the current architecture.
2531 This function actually works on integral values, as well as pointers.
2532 For pointers, it performs architecture-specific conversions as
2533 described above for @code{extract_typed_address}.
2536 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2537 Create and return a value representing a pointer of type @var{type} to
2538 the address @var{addr}, as appropriate for the current architecture.
2539 This function performs architecture-specific conversions as described
2540 above for @code{store_typed_address}.
2543 Here are some macros which architectures can define to indicate the
2544 relationship between pointers and addresses. These have default
2545 definitions, appropriate for architectures on which all pointers are
2546 simple unsigned byte addresses.
2548 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2549 Assume that @var{buf} holds a pointer of type @var{type}, in the
2550 appropriate format for the current architecture. Return the byte
2551 address the pointer refers to.
2553 This function may safely assume that @var{type} is either a pointer or a
2554 C@t{++} reference type.
2557 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2558 Store in @var{buf} a pointer of type @var{type} representing the address
2559 @var{addr}, in the appropriate format for the current architecture.
2561 This function may safely assume that @var{type} is either a pointer or a
2562 C@t{++} reference type.
2565 @section Address Classes
2566 @cindex address classes
2567 @cindex DW_AT_byte_size
2568 @cindex DW_AT_address_class
2570 Sometimes information about different kinds of addresses is available
2571 via the debug information. For example, some programming environments
2572 define addresses of several different sizes. If the debug information
2573 distinguishes these kinds of address classes through either the size
2574 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2575 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2576 following macros should be defined in order to disambiguate these
2577 types within @value{GDBN} as well as provide the added information to
2578 a @value{GDBN} user when printing type expressions.
2580 @deftypefn {Target Macro} int ADDRESS_CLASS_TYPE_FLAGS (int @var{byte_size}, int @var{dwarf2_addr_class})
2581 Returns the type flags needed to construct a pointer type whose size
2582 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2583 This function is normally called from within a symbol reader. See
2584 @file{dwarf2read.c}.
2587 @deftypefn {Target Macro} char *ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (int @var{type_flags})
2588 Given the type flags representing an address class qualifier, return
2591 @deftypefn {Target Macro} int ADDRESS_CLASS_NAME_to_TYPE_FLAGS (int @var{name}, int *var{type_flags_ptr})
2592 Given an address qualifier name, set the @code{int} refererenced by @var{type_flags_ptr} to the type flags
2593 for that address class qualifier.
2596 Since the need for address classes is rather rare, none of
2597 the address class macros defined by default. Predicate
2598 macros are provided to detect when they are defined.
2600 Consider a hypothetical architecture in which addresses are normally
2601 32-bits wide, but 16-bit addresses are also supported. Furthermore,
2602 suppose that the @w{DWARF 2} information for this architecture simply
2603 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
2604 of these "short" pointers. The following functions could be defined
2605 to implement the address class macros:
2608 somearch_address_class_type_flags (int byte_size,
2609 int dwarf2_addr_class)
2612 return TYPE_FLAG_ADDRESS_CLASS_1;
2618 somearch_address_class_type_flags_to_name (int type_flags)
2620 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
2627 somearch_address_class_name_to_type_flags (char *name,
2628 int *type_flags_ptr)
2630 if (strcmp (name, "short") == 0)
2632 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
2640 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
2641 to indicate the presence of one of these "short" pointers. E.g, if
2642 the debug information indicates that @code{short_ptr_var} is one of these
2643 short pointers, @value{GDBN} might show the following behavior:
2646 (gdb) ptype short_ptr_var
2647 type = int * @@short
2651 @section Raw and Virtual Register Representations
2652 @cindex raw register representation
2653 @cindex virtual register representation
2654 @cindex representations, raw and virtual registers
2656 @emph{Maintainer note: This section is pretty much obsolete. The
2657 functionality described here has largely been replaced by
2658 pseudo-registers and the mechanisms described in @ref{Target
2659 Architecture Definition, , Using Different Register and Memory Data
2660 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
2661 Bug Tracking Database} and
2662 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
2663 up-to-date information.}
2665 Some architectures use one representation for a value when it lives in a
2666 register, but use a different representation when it lives in memory.
2667 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
2668 the target registers, and the @dfn{virtual} representation is the one
2669 used in memory, and within @value{GDBN} @code{struct value} objects.
2671 @emph{Maintainer note: Notice that the same mechanism is being used to
2672 both convert a register to a @code{struct value} and alternative
2675 For almost all data types on almost all architectures, the virtual and
2676 raw representations are identical, and no special handling is needed.
2677 However, they do occasionally differ. For example:
2681 The x86 architecture supports an 80-bit @code{long double} type. However, when
2682 we store those values in memory, they occupy twelve bytes: the
2683 floating-point number occupies the first ten, and the final two bytes
2684 are unused. This keeps the values aligned on four-byte boundaries,
2685 allowing more efficient access. Thus, the x86 80-bit floating-point
2686 type is the raw representation, and the twelve-byte loosely-packed
2687 arrangement is the virtual representation.
2690 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
2691 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
2692 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
2693 raw representation, and the trimmed 32-bit representation is the
2694 virtual representation.
2697 In general, the raw representation is determined by the architecture, or
2698 @value{GDBN}'s interface to the architecture, while the virtual representation
2699 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
2700 @code{registers}, holds the register contents in raw format, and the
2701 @value{GDBN} remote protocol transmits register values in raw format.
2703 Your architecture may define the following macros to request
2704 conversions between the raw and virtual format:
2706 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
2707 Return non-zero if register number @var{reg}'s value needs different raw
2708 and virtual formats.
2710 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
2711 unless this macro returns a non-zero value for that register.
2714 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
2715 The size of register number @var{reg}'s raw value. This is the number
2716 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
2717 remote protocol packet.
2720 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
2721 The size of register number @var{reg}'s value, in its virtual format.
2722 This is the size a @code{struct value}'s buffer will have, holding that
2726 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
2727 This is the type of the virtual representation of register number
2728 @var{reg}. Note that there is no need for a macro giving a type for the
2729 register's raw form; once the register's value has been obtained, @value{GDBN}
2730 always uses the virtual form.
2733 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2734 Convert the value of register number @var{reg} to @var{type}, which
2735 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2736 at @var{from} holds the register's value in raw format; the macro should
2737 convert the value to virtual format, and place it at @var{to}.
2739 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
2740 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
2741 arguments in different orders.
2743 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
2744 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
2748 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2749 Convert the value of register number @var{reg} to @var{type}, which
2750 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2751 at @var{from} holds the register's value in raw format; the macro should
2752 convert the value to virtual format, and place it at @var{to}.
2754 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
2755 their @var{reg} and @var{type} arguments in different orders.
2759 @section Using Different Register and Memory Data Representations
2760 @cindex register representation
2761 @cindex memory representation
2762 @cindex representations, register and memory
2763 @cindex register data formats, converting
2764 @cindex @code{struct value}, converting register contents to
2766 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
2767 significant change. Many of the macros and functions refered to in this
2768 section are likely to be subject to further revision. See
2769 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
2770 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
2771 further information. cagney/2002-05-06.}
2773 Some architectures can represent a data object in a register using a
2774 form that is different to the objects more normal memory representation.
2780 The Alpha architecture can represent 32 bit integer values in
2781 floating-point registers.
2784 The x86 architecture supports 80-bit floating-point registers. The
2785 @code{long double} data type occupies 96 bits in memory but only 80 bits
2786 when stored in a register.
2790 In general, the register representation of a data type is determined by
2791 the architecture, or @value{GDBN}'s interface to the architecture, while
2792 the memory representation is determined by the Application Binary
2795 For almost all data types on almost all architectures, the two
2796 representations are identical, and no special handling is needed.
2797 However, they do occasionally differ. Your architecture may define the
2798 following macros to request conversions between the register and memory
2799 representations of a data type:
2801 @deftypefn {Target Macro} int CONVERT_REGISTER_P (int @var{reg})
2802 Return non-zero if the representation of a data value stored in this
2803 register may be different to the representation of that same data value
2804 when stored in memory.
2806 When non-zero, the macros @code{REGISTER_TO_VALUE} and
2807 @code{VALUE_TO_REGISTER} are used to perform any necessary conversion.
2810 @deftypefn {Target Macro} void REGISTER_TO_VALUE (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2811 Convert the value of register number @var{reg} to a data object of type
2812 @var{type}. The buffer at @var{from} holds the register's value in raw
2813 format; the converted value should be placed in the buffer at @var{to}.
2815 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
2816 their @var{reg} and @var{type} arguments in different orders.
2818 You should only use @code{REGISTER_TO_VALUE} with registers for which
2819 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
2822 @deftypefn {Target Macro} void VALUE_TO_REGISTER (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2823 Convert a data value of type @var{type} to register number @var{reg}'
2826 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
2827 their @var{reg} and @var{type} arguments in different orders.
2829 You should only use @code{VALUE_TO_REGISTER} with registers for which
2830 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
2833 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
2834 See @file{mips-tdep.c}. It does not do what you want.
2838 @section Frame Interpretation
2840 @section Inferior Call Setup
2842 @section Compiler Characteristics
2844 @section Target Conditionals
2846 This section describes the macros that you can use to define the target
2851 @item ADDR_BITS_REMOVE (addr)
2852 @findex ADDR_BITS_REMOVE
2853 If a raw machine instruction address includes any bits that are not
2854 really part of the address, then define this macro to expand into an
2855 expression that zeroes those bits in @var{addr}. This is only used for
2856 addresses of instructions, and even then not in all contexts.
2858 For example, the two low-order bits of the PC on the Hewlett-Packard PA
2859 2.0 architecture contain the privilege level of the corresponding
2860 instruction. Since instructions must always be aligned on four-byte
2861 boundaries, the processor masks out these bits to generate the actual
2862 address of the instruction. ADDR_BITS_REMOVE should filter out these
2863 bits with an expression such as @code{((addr) & ~3)}.
2865 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS (@var{name}, @var{type_flags_ptr})
2866 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS
2867 If @var{name} is a valid address class qualifier name, set the @code{int}
2868 referenced by @var{type_flags_ptr} to the mask representing the qualifier
2869 and return 1. If @var{name} is not a valid address class qualifier name,
2872 The value for @var{type_flags_ptr} should be one of
2873 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
2874 possibly some combination of these values or'd together.
2875 @xref{Target Architecture Definition, , Address Classes}.
2877 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P ()
2878 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P
2879 Predicate which indicates whether @code{ADDRESS_CLASS_NAME_TO_TYPE_FLAGS}
2882 @item ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
2883 @findex ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
2884 Given a pointers byte size (as described by the debug information) and
2885 the possible @code{DW_AT_address_class} value, return the type flags
2886 used by @value{GDBN} to represent this address class. The value
2887 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
2888 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
2889 values or'd together.
2890 @xref{Target Architecture Definition, , Address Classes}.
2892 @item ADDRESS_CLASS_TYPE_FLAGS_P ()
2893 @findex ADDRESS_CLASS_TYPE_FLAGS_P
2894 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS} has
2897 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (@var{type_flags})
2898 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME
2899 Return the name of the address class qualifier associated with the type
2900 flags given by @var{type_flags}.
2902 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P ()
2903 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P
2904 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS_TO_NAME} has
2906 @xref{Target Architecture Definition, , Address Classes}.
2908 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
2909 @findex ADDRESS_TO_POINTER
2910 Store in @var{buf} a pointer of type @var{type} representing the address
2911 @var{addr}, in the appropriate format for the current architecture.
2912 This macro may safely assume that @var{type} is either a pointer or a
2913 C@t{++} reference type.
2914 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
2916 @item BELIEVE_PCC_PROMOTION
2917 @findex BELIEVE_PCC_PROMOTION
2918 Define if the compiler promotes a @code{short} or @code{char}
2919 parameter to an @code{int}, but still reports the parameter as its
2920 original type, rather than the promoted type.
2922 @item BITS_BIG_ENDIAN
2923 @findex BITS_BIG_ENDIAN
2924 Define this if the numbering of bits in the targets does @strong{not} match the
2925 endianness of the target byte order. A value of 1 means that the bits
2926 are numbered in a big-endian bit order, 0 means little-endian.
2930 This is the character array initializer for the bit pattern to put into
2931 memory where a breakpoint is set. Although it's common to use a trap
2932 instruction for a breakpoint, it's not required; for instance, the bit
2933 pattern could be an invalid instruction. The breakpoint must be no
2934 longer than the shortest instruction of the architecture.
2936 @code{BREAKPOINT} has been deprecated in favor of
2937 @code{BREAKPOINT_FROM_PC}.
2939 @item BIG_BREAKPOINT
2940 @itemx LITTLE_BREAKPOINT
2941 @findex LITTLE_BREAKPOINT
2942 @findex BIG_BREAKPOINT
2943 Similar to BREAKPOINT, but used for bi-endian targets.
2945 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
2946 favor of @code{BREAKPOINT_FROM_PC}.
2948 @item DEPRECATED_REMOTE_BREAKPOINT
2949 @itemx DEPRECATED_LITTLE_REMOTE_BREAKPOINT
2950 @itemx DEPRECATED_BIG_REMOTE_BREAKPOINT
2951 @findex DEPRECATED_BIG_REMOTE_BREAKPOINT
2952 @findex DEPRECATED_LITTLE_REMOTE_BREAKPOINT
2953 @findex DEPRECATED_REMOTE_BREAKPOINT
2954 Specify the breakpoint instruction sequence for a remote target.
2955 @code{DEPRECATED_REMOTE_BREAKPOINT},
2956 @code{DEPRECATED_BIG_REMOTE_BREAKPOINT} and
2957 @code{DEPRECATED_LITTLE_REMOTE_BREAKPOINT} have been deprecated in
2958 favor of @code{BREAKPOINT_FROM_PC} (@pxref{BREAKPOINT_FROM_PC}).
2960 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
2961 @findex BREAKPOINT_FROM_PC
2962 @anchor{BREAKPOINT_FROM_PC} Use the program counter to determine the
2963 contents and size of a breakpoint instruction. It returns a pointer to
2964 a string of bytes that encode a breakpoint instruction, stores the
2965 length of the string to @code{*@var{lenptr}}, and adjusts the program
2966 counter (if necessary) to point to the actual memory location where the
2967 breakpoint should be inserted.
2969 Although it is common to use a trap instruction for a breakpoint, it's
2970 not required; for instance, the bit pattern could be an invalid
2971 instruction. The breakpoint must be no longer than the shortest
2972 instruction of the architecture.
2974 Replaces all the other @var{BREAKPOINT} macros.
2976 @item MEMORY_INSERT_BREAKPOINT (@var{addr}, @var{contents_cache})
2977 @itemx MEMORY_REMOVE_BREAKPOINT (@var{addr}, @var{contents_cache})
2978 @findex MEMORY_REMOVE_BREAKPOINT
2979 @findex MEMORY_INSERT_BREAKPOINT
2980 Insert or remove memory based breakpoints. Reasonable defaults
2981 (@code{default_memory_insert_breakpoint} and
2982 @code{default_memory_remove_breakpoint} respectively) have been
2983 provided so that it is not necessary to define these for most
2984 architectures. Architectures which may want to define
2985 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
2986 likely have instructions that are oddly sized or are not stored in a
2987 conventional manner.
2989 It may also be desirable (from an efficiency standpoint) to define
2990 custom breakpoint insertion and removal routines if
2991 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
2994 @item ADJUST_BREAKPOINT_ADDRESS (@var{address})
2995 @findex ADJUST_BREAKPOINT_ADDRESS
2996 @cindex breakpoint address adjusted
2997 Given an address at which a breakpoint is desired, return a breakpoint
2998 address adjusted to account for architectural constraints on
2999 breakpoint placement. This method is not needed by most targets.
3001 The FR-V target (see @file{frv-tdep.c}) requires this method.
3002 The FR-V is a VLIW architecture in which a number of RISC-like
3003 instructions are grouped (packed) together into an aggregate
3004 instruction or instruction bundle. When the processor executes
3005 one of these bundles, the component instructions are executed
3008 In the course of optimization, the compiler may group instructions
3009 from distinct source statements into the same bundle. The line number
3010 information associated with one of the latter statements will likely
3011 refer to some instruction other than the first one in the bundle. So,
3012 if the user attempts to place a breakpoint on one of these latter
3013 statements, @value{GDBN} must be careful to @emph{not} place the break
3014 instruction on any instruction other than the first one in the bundle.
3015 (Remember though that the instructions within a bundle execute
3016 in parallel, so the @emph{first} instruction is the instruction
3017 at the lowest address and has nothing to do with execution order.)
3019 The FR-V's @code{ADJUST_BREAKPOINT_ADDRESS} method will adjust a
3020 breakpoint's address by scanning backwards for the beginning of
3021 the bundle, returning the address of the bundle.
3023 Since the adjustment of a breakpoint may significantly alter a user's
3024 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3025 is initially set and each time that that breakpoint is hit.
3027 @item CALL_DUMMY_LOCATION
3028 @findex CALL_DUMMY_LOCATION
3029 See the file @file{inferior.h}.
3031 This method has been replaced by @code{push_dummy_code}
3032 (@pxref{push_dummy_code}).
3034 @item CANNOT_FETCH_REGISTER (@var{regno})
3035 @findex CANNOT_FETCH_REGISTER
3036 A C expression that should be nonzero if @var{regno} cannot be fetched
3037 from an inferior process. This is only relevant if
3038 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3040 @item CANNOT_STORE_REGISTER (@var{regno})
3041 @findex CANNOT_STORE_REGISTER
3042 A C expression that should be nonzero if @var{regno} should not be
3043 written to the target. This is often the case for program counters,
3044 status words, and other special registers. If this is not defined,
3045 @value{GDBN} will assume that all registers may be written.
3047 @item int CONVERT_REGISTER_P(@var{regnum})
3048 @findex CONVERT_REGISTER_P
3049 Return non-zero if register @var{regnum} can represent data values in a
3051 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3053 @item DECR_PC_AFTER_BREAK
3054 @findex DECR_PC_AFTER_BREAK
3055 Define this to be the amount by which to decrement the PC after the
3056 program encounters a breakpoint. This is often the number of bytes in
3057 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3059 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3060 @findex DISABLE_UNSETTABLE_BREAK
3061 If defined, this should evaluate to 1 if @var{addr} is in a shared
3062 library in which breakpoints cannot be set and so should be disabled.
3064 @item PRINT_FLOAT_INFO()
3065 @findex PRINT_FLOAT_INFO
3066 If defined, then the @samp{info float} command will print information about
3067 the processor's floating point unit.
3069 @item print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3070 @findex print_registers_info
3071 If defined, pretty print the value of the register @var{regnum} for the
3072 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3073 either all registers (@var{all} is non zero) or a select subset of
3074 registers (@var{all} is zero).
3076 The default method prints one register per line, and if @var{all} is
3077 zero omits floating-point registers.
3079 @item PRINT_VECTOR_INFO()
3080 @findex PRINT_VECTOR_INFO
3081 If defined, then the @samp{info vector} command will call this function
3082 to print information about the processor's vector unit.
3084 By default, the @samp{info vector} command will print all vector
3085 registers (the register's type having the vector attribute).
3087 @item DWARF_REG_TO_REGNUM
3088 @findex DWARF_REG_TO_REGNUM
3089 Convert DWARF register number into @value{GDBN} regnum. If not defined,
3090 no conversion will be performed.
3092 @item DWARF2_REG_TO_REGNUM
3093 @findex DWARF2_REG_TO_REGNUM
3094 Convert DWARF2 register number into @value{GDBN} regnum. If not
3095 defined, no conversion will be performed.
3097 @item ECOFF_REG_TO_REGNUM
3098 @findex ECOFF_REG_TO_REGNUM
3099 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
3100 no conversion will be performed.
3102 @item END_OF_TEXT_DEFAULT
3103 @findex END_OF_TEXT_DEFAULT
3104 This is an expression that should designate the end of the text section.
3107 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
3108 @findex EXTRACT_RETURN_VALUE
3109 Define this to extract a function's return value of type @var{type} from
3110 the raw register state @var{regbuf} and copy that, in virtual format,
3113 This method has been deprecated in favour of @code{gdbarch_return_value}
3114 (@pxref{gdbarch_return_value}).
3116 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
3117 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS
3118 @anchor{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}
3119 When defined, extract from the array @var{regbuf} (containing the raw
3120 register state) the @code{CORE_ADDR} at which a function should return
3121 its structure value.
3123 @xref{gdbarch_return_value}.
3125 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P()
3126 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P
3127 Predicate for @code{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}.
3129 @item DEPRECATED_FP_REGNUM
3130 @findex DEPRECATED_FP_REGNUM
3131 If the virtual frame pointer is kept in a register, then define this
3132 macro to be the number (greater than or equal to zero) of that register.
3134 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3137 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3138 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3139 Define this to an expression that returns 1 if the function invocation
3140 represented by @var{fi} does not have a stack frame associated with it.
3143 @item frame_align (@var{address})
3144 @anchor{frame_align}
3146 Define this to adjust @var{address} so that it meets the alignment
3147 requirements for the start of a new stack frame. A stack frame's
3148 alignment requirements are typically stronger than a target processors
3149 stack alignment requirements (@pxref{DEPRECATED_STACK_ALIGN}).
3151 This function is used to ensure that, when creating a dummy frame, both
3152 the initial stack pointer and (if needed) the address of the return
3153 value are correctly aligned.
3155 Unlike @code{DEPRECATED_STACK_ALIGN}, this function always adjusts the
3156 address in the direction of stack growth.
3158 By default, no frame based stack alignment is performed.
3160 @item int frame_red_zone_size
3162 The number of bytes, beyond the innermost-stack-address, reserved by the
3163 @sc{abi}. A function is permitted to use this scratch area (instead of
3164 allocating extra stack space).
3166 When performing an inferior function call, to ensure that it does not
3167 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3168 @var{frame_red_zone_size} bytes before pushing parameters onto the
3171 By default, zero bytes are allocated. The value must be aligned
3172 (@pxref{frame_align}).
3174 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3175 @emph{red zone} when describing this scratch area.
3178 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3179 @findex DEPRECATED_FRAME_CHAIN
3180 Given @var{frame}, return a pointer to the calling frame.
3182 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3183 @findex DEPRECATED_FRAME_CHAIN_VALID
3184 Define this to be an expression that returns zero if the given frame is an
3185 outermost frame, with no caller, and nonzero otherwise. Most normal
3186 situations can be handled without defining this macro, including @code{NULL}
3187 chain pointers, dummy frames, and frames whose PC values are inside the
3188 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3191 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3192 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3193 See @file{frame.h}. Determines the address of all registers in the
3194 current stack frame storing each in @code{frame->saved_regs}. Space for
3195 @code{frame->saved_regs} shall be allocated by
3196 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3197 @code{frame_saved_regs_zalloc}.
3199 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3201 @item FRAME_NUM_ARGS (@var{fi})
3202 @findex FRAME_NUM_ARGS
3203 For the frame described by @var{fi} return the number of arguments that
3204 are being passed. If the number of arguments is not known, return
3207 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3208 @findex DEPRECATED_FRAME_SAVED_PC
3209 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3210 saved there. This is the return address.
3212 This method is deprecated. @xref{unwind_pc}.
3214 @item CORE_ADDR unwind_pc (struct frame_info *@var{this_frame})
3216 @anchor{unwind_pc} Return the instruction address, in @var{this_frame}'s
3217 caller, at which execution will resume after @var{this_frame} returns.
3218 This is commonly refered to as the return address.
3220 The implementation, which must be frame agnostic (work with any frame),
3221 is typically no more than:
3225 frame_unwind_unsigned_register (this_frame, D10V_PC_REGNUM, &pc);
3226 return d10v_make_iaddr (pc);
3230 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3232 @item CORE_ADDR unwind_sp (struct frame_info *@var{this_frame})
3234 @anchor{unwind_sp} Return the frame's inner most stack address. This is
3235 commonly refered to as the frame's @dfn{stack pointer}.
3237 The implementation, which must be frame agnostic (work with any frame),
3238 is typically no more than:
3242 frame_unwind_unsigned_register (this_frame, D10V_SP_REGNUM, &sp);
3243 return d10v_make_daddr (sp);
3247 @xref{TARGET_READ_SP}, which this method replaces.
3249 @item FUNCTION_EPILOGUE_SIZE
3250 @findex FUNCTION_EPILOGUE_SIZE
3251 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3252 function end symbol is 0. For such targets, you must define
3253 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3254 function's epilogue.
3256 @item DEPRECATED_FUNCTION_START_OFFSET
3257 @findex DEPRECATED_FUNCTION_START_OFFSET
3258 An integer, giving the offset in bytes from a function's address (as
3259 used in the values of symbols, function pointers, etc.), and the
3260 function's first genuine instruction.
3262 This is zero on almost all machines: the function's address is usually
3263 the address of its first instruction. However, on the VAX, for
3264 example, each function starts with two bytes containing a bitmask
3265 indicating which registers to save upon entry to the function. The
3266 VAX @code{call} instructions check this value, and save the
3267 appropriate registers automatically. Thus, since the offset from the
3268 function's address to its first instruction is two bytes,
3269 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3271 @item GCC_COMPILED_FLAG_SYMBOL
3272 @itemx GCC2_COMPILED_FLAG_SYMBOL
3273 @findex GCC2_COMPILED_FLAG_SYMBOL
3274 @findex GCC_COMPILED_FLAG_SYMBOL
3275 If defined, these are the names of the symbols that @value{GDBN} will
3276 look for to detect that GCC compiled the file. The default symbols
3277 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3278 respectively. (Currently only defined for the Delta 68.)
3280 @item @value{GDBN}_MULTI_ARCH
3281 @findex @value{GDBN}_MULTI_ARCH
3282 If defined and non-zero, enables support for multiple architectures
3283 within @value{GDBN}.
3285 This support can be enabled at two levels. At level one, only
3286 definitions for previously undefined macros are provided; at level two,
3287 a multi-arch definition of all architecture dependent macros will be
3290 @item @value{GDBN}_TARGET_IS_HPPA
3291 @findex @value{GDBN}_TARGET_IS_HPPA
3292 This determines whether horrible kludge code in @file{dbxread.c} and
3293 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3294 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3297 @item GET_LONGJMP_TARGET
3298 @findex GET_LONGJMP_TARGET
3299 For most machines, this is a target-dependent parameter. On the
3300 DECstation and the Iris, this is a native-dependent parameter, since
3301 the header file @file{setjmp.h} is needed to define it.
3303 This macro determines the target PC address that @code{longjmp} will jump to,
3304 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3305 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3306 pointer. It examines the current state of the machine as needed.
3308 @item DEPRECATED_GET_SAVED_REGISTER
3309 @findex DEPRECATED_GET_SAVED_REGISTER
3310 Define this if you need to supply your own definition for the function
3311 @code{DEPRECATED_GET_SAVED_REGISTER}.
3313 @item DEPRECATED_IBM6000_TARGET
3314 @findex DEPRECATED_IBM6000_TARGET
3315 Shows that we are configured for an IBM RS/6000 system. This
3316 conditional should be eliminated (FIXME) and replaced by
3317 feature-specific macros. It was introduced in a haste and we are
3318 repenting at leisure.
3320 @item I386_USE_GENERIC_WATCHPOINTS
3321 An x86-based target can define this to use the generic x86 watchpoint
3322 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3324 @item SYMBOLS_CAN_START_WITH_DOLLAR
3325 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3326 Some systems have routines whose names start with @samp{$}. Giving this
3327 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3328 routines when parsing tokens that begin with @samp{$}.
3330 On HP-UX, certain system routines (millicode) have names beginning with
3331 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3332 routine that handles inter-space procedure calls on PA-RISC.
3334 @item DEPRECATED_INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3335 @findex DEPRECATED_INIT_EXTRA_FRAME_INFO
3336 If additional information about the frame is required this should be
3337 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3338 is allocated using @code{frame_extra_info_zalloc}.
3340 @item DEPRECATED_INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3341 @findex DEPRECATED_INIT_FRAME_PC
3342 This is a C statement that sets the pc of the frame pointed to by
3343 @var{prev}. [By default...]
3345 @item INNER_THAN (@var{lhs}, @var{rhs})
3347 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3348 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3349 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3352 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3353 @findex gdbarch_in_function_epilogue_p
3354 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3355 The epilogue of a function is defined as the part of a function where
3356 the stack frame of the function already has been destroyed up to the
3357 final `return from function call' instruction.
3359 @item DEPRECATED_SIGTRAMP_START (@var{pc})
3360 @findex DEPRECATED_SIGTRAMP_START
3361 @itemx DEPRECATED_SIGTRAMP_END (@var{pc})
3362 @findex DEPRECATED_SIGTRAMP_END
3363 Define these to be the start and end address of the @code{sigtramp} for the
3364 given @var{pc}. On machines where the address is just a compile time
3365 constant, the macro expansion will typically just ignore the supplied
3368 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3369 @findex IN_SOLIB_CALL_TRAMPOLINE
3370 Define this to evaluate to nonzero if the program is stopped in the
3371 trampoline that connects to a shared library.
3373 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3374 @findex IN_SOLIB_RETURN_TRAMPOLINE
3375 Define this to evaluate to nonzero if the program is stopped in the
3376 trampoline that returns from a shared library.
3378 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3379 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3380 Define this to evaluate to nonzero if the program is stopped in the
3383 @item SKIP_SOLIB_RESOLVER (@var{pc})
3384 @findex SKIP_SOLIB_RESOLVER
3385 Define this to evaluate to the (nonzero) address at which execution
3386 should continue to get past the dynamic linker's symbol resolution
3387 function. A zero value indicates that it is not important or necessary
3388 to set a breakpoint to get through the dynamic linker and that single
3389 stepping will suffice.
3391 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3392 @findex INTEGER_TO_ADDRESS
3393 @cindex converting integers to addresses
3394 Define this when the architecture needs to handle non-pointer to address
3395 conversions specially. Converts that value to an address according to
3396 the current architectures conventions.
3398 @emph{Pragmatics: When the user copies a well defined expression from
3399 their source code and passes it, as a parameter, to @value{GDBN}'s
3400 @code{print} command, they should get the same value as would have been
3401 computed by the target program. Any deviation from this rule can cause
3402 major confusion and annoyance, and needs to be justified carefully. In
3403 other words, @value{GDBN} doesn't really have the freedom to do these
3404 conversions in clever and useful ways. It has, however, been pointed
3405 out that users aren't complaining about how @value{GDBN} casts integers
3406 to pointers; they are complaining that they can't take an address from a
3407 disassembly listing and give it to @code{x/i}. Adding an architecture
3408 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3409 @value{GDBN} to ``get it right'' in all circumstances.}
3411 @xref{Target Architecture Definition, , Pointers Are Not Always
3414 @item NO_HIF_SUPPORT
3415 @findex NO_HIF_SUPPORT
3416 (Specific to the a29k.)
3418 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3419 @findex POINTER_TO_ADDRESS
3420 Assume that @var{buf} holds a pointer of type @var{type}, in the
3421 appropriate format for the current architecture. Return the byte
3422 address the pointer refers to.
3423 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3425 @item REGISTER_CONVERTIBLE (@var{reg})
3426 @findex REGISTER_CONVERTIBLE
3427 Return non-zero if @var{reg} uses different raw and virtual formats.
3428 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3430 @item REGISTER_TO_VALUE(@var{regnum}, @var{type}, @var{from}, @var{to})
3431 @findex REGISTER_TO_VALUE
3432 Convert the raw contents of register @var{regnum} into a value of type
3434 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3436 @item DEPRECATED_REGISTER_RAW_SIZE (@var{reg})
3437 @findex DEPRECATED_REGISTER_RAW_SIZE
3438 Return the raw size of @var{reg}; defaults to the size of the register's
3440 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3442 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3443 @findex register_reggroup_p
3444 @cindex register groups
3445 Return non-zero if register @var{regnum} is a member of the register
3446 group @var{reggroup}.
3448 By default, registers are grouped as follows:
3451 @item float_reggroup
3452 Any register with a valid name and a floating-point type.
3453 @item vector_reggroup
3454 Any register with a valid name and a vector type.
3455 @item general_reggroup
3456 Any register with a valid name and a type other than vector or
3457 floating-point. @samp{float_reggroup}.
3459 @itemx restore_reggroup
3461 Any register with a valid name.
3464 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3465 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3466 Return the virtual size of @var{reg}; defaults to the size of the
3467 register's virtual type.
3468 Return the virtual size of @var{reg}.
3469 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3471 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3472 @findex REGISTER_VIRTUAL_TYPE
3473 Return the virtual type of @var{reg}.
3474 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3476 @item struct type *register_type (@var{gdbarch}, @var{reg})
3477 @findex register_type
3478 If defined, return the type of register @var{reg}. This function
3479 superseeds @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3480 Definition, , Raw and Virtual Register Representations}.
3482 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3483 @findex REGISTER_CONVERT_TO_VIRTUAL
3484 Convert the value of register @var{reg} from its raw form to its virtual
3486 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3488 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3489 @findex REGISTER_CONVERT_TO_RAW
3490 Convert the value of register @var{reg} from its virtual form to its raw
3492 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3494 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3495 @findex regset_from_core_section
3496 Return the appropriate register set for a core file section with name
3497 @var{sect_name} and size @var{sect_size}.
3499 @item SOFTWARE_SINGLE_STEP_P()
3500 @findex SOFTWARE_SINGLE_STEP_P
3501 Define this as 1 if the target does not have a hardware single-step
3502 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3504 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3505 @findex SOFTWARE_SINGLE_STEP
3506 A function that inserts or removes (depending on
3507 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3508 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3511 @item SOFUN_ADDRESS_MAYBE_MISSING
3512 @findex SOFUN_ADDRESS_MAYBE_MISSING
3513 Somebody clever observed that, the more actual addresses you have in the
3514 debug information, the more time the linker has to spend relocating
3515 them. So whenever there's some other way the debugger could find the
3516 address it needs, you should omit it from the debug info, to make
3519 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3520 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3521 entries in stabs-format debugging information. @code{N_SO} stabs mark
3522 the beginning and ending addresses of compilation units in the text
3523 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3525 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3529 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3530 addresses where the function starts by taking the function name from
3531 the stab, and then looking that up in the minsyms (the
3532 linker/assembler symbol table). In other words, the stab has the
3533 name, and the linker/assembler symbol table is the only place that carries
3537 @code{N_SO} stabs have an address of zero, too. You just look at the
3538 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3539 and guess the starting and ending addresses of the compilation unit from
3543 @item PC_LOAD_SEGMENT
3544 @findex PC_LOAD_SEGMENT
3545 If defined, print information about the load segment for the program
3546 counter. (Defined only for the RS/6000.)
3550 If the program counter is kept in a register, then define this macro to
3551 be the number (greater than or equal to zero) of that register.
3553 This should only need to be defined if @code{TARGET_READ_PC} and
3554 @code{TARGET_WRITE_PC} are not defined.
3557 @findex PARM_BOUNDARY
3558 If non-zero, round arguments to a boundary of this many bits before
3559 pushing them on the stack.
3561 @item stabs_argument_has_addr (@var{gdbarch}, @var{type})
3562 @findex stabs_argument_has_addr
3563 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3564 @anchor{stabs_argument_has_addr} Define this to return nonzero if a
3565 function argument of type @var{type} is passed by reference instead of
3568 This method replaces @code{DEPRECATED_REG_STRUCT_HAS_ADDR}
3569 (@pxref{DEPRECATED_REG_STRUCT_HAS_ADDR}).
3571 @item PROCESS_LINENUMBER_HOOK
3572 @findex PROCESS_LINENUMBER_HOOK
3573 A hook defined for XCOFF reading.
3575 @item PROLOGUE_FIRSTLINE_OVERLAP
3576 @findex PROLOGUE_FIRSTLINE_OVERLAP
3577 (Only used in unsupported Convex configuration.)
3581 If defined, this is the number of the processor status register. (This
3582 definition is only used in generic code when parsing "$ps".)
3584 @item DEPRECATED_POP_FRAME
3585 @findex DEPRECATED_POP_FRAME
3587 If defined, used by @code{frame_pop} to remove a stack frame. This
3588 method has been superseeded by generic code.
3590 @item push_dummy_call (@var{gdbarch}, @var{function}, @var{regcache}, @var{pc_addr}, @var{nargs}, @var{args}, @var{sp}, @var{struct_return}, @var{struct_addr})
3591 @findex push_dummy_call
3592 @findex DEPRECATED_PUSH_ARGUMENTS.
3593 @anchor{push_dummy_call} Define this to push the dummy frame's call to
3594 the inferior function onto the stack. In addition to pushing
3595 @var{nargs}, the code should push @var{struct_addr} (when
3596 @var{struct_return}), and the return address (@var{bp_addr}).
3598 @var{function} is a pointer to a @code{struct value}; on architectures that use
3599 function descriptors, this contains the function descriptor value.
3601 Returns the updated top-of-stack pointer.
3603 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3605 @item CORE_ADDR push_dummy_code (@var{gdbarch}, @var{sp}, @var{funaddr}, @var{using_gcc}, @var{args}, @var{nargs}, @var{value_type}, @var{real_pc}, @var{bp_addr})
3606 @findex push_dummy_code
3607 @anchor{push_dummy_code} Given a stack based call dummy, push the
3608 instruction sequence (including space for a breakpoint) to which the
3609 called function should return.
3611 Set @var{bp_addr} to the address at which the breakpoint instruction
3612 should be inserted, @var{real_pc} to the resume address when starting
3613 the call sequence, and return the updated inner-most stack address.
3615 By default, the stack is grown sufficient to hold a frame-aligned
3616 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3617 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3619 This method replaces @code{CALL_DUMMY_LOCATION},
3620 @code{DEPRECATED_REGISTER_SIZE}.
3622 @item REGISTER_NAME(@var{i})
3623 @findex REGISTER_NAME
3624 Return the name of register @var{i} as a string. May return @code{NULL}
3625 or @code{NUL} to indicate that register @var{i} is not valid.
3627 @item DEPRECATED_REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3628 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3629 @anchor{DEPRECATED_REG_STRUCT_HAS_ADDR}Define this to return 1 if the
3630 given type will be passed by pointer rather than directly.
3632 This method has been replaced by @code{stabs_argument_has_addr}
3633 (@pxref{stabs_argument_has_addr}).
3635 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3636 @findex SAVE_DUMMY_FRAME_TOS
3637 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3638 notify the target dependent code of the top-of-stack value that will be
3639 passed to the the inferior code. This is the value of the @code{SP}
3640 after both the dummy frame and space for parameters/results have been
3641 allocated on the stack. @xref{unwind_dummy_id}.
3643 @item SDB_REG_TO_REGNUM
3644 @findex SDB_REG_TO_REGNUM
3645 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3646 defined, no conversion will be done.
3648 @item enum return_value_convention gdbarch_return_value (struct gdbarch *@var{gdbarch}, struct type *@var{valtype}, struct regcache *@var{regcache}, void *@var{readbuf}, const void *@var{writebuf})
3649 @findex gdbarch_return_value
3650 @anchor{gdbarch_return_value} Given a function with a return-value of
3651 type @var{rettype}, return which return-value convention that function
3654 @value{GDBN} currently recognizes two function return-value conventions:
3655 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3656 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3657 value is found in memory and the address of that memory location is
3658 passed in as the function's first parameter.
3660 If the register convention is being used, and @var{writebuf} is
3661 non-@code{NULL}, also copy the return-value in @var{writebuf} into
3664 If the register convention is being used, and @var{readbuf} is
3665 non-@code{NULL}, also copy the return value from @var{regcache} into
3666 @var{readbuf} (@var{regcache} contains a copy of the registers from the
3667 just returned function).
3669 @xref{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}, for a description of how
3670 return-values that use the struct convention are handled.
3672 @emph{Maintainer note: This method replaces separate predicate, extract,
3673 store methods. By having only one method, the logic needed to determine
3674 the return-value convention need only be implemented in one place. If
3675 @value{GDBN} were written in an @sc{oo} language, this method would
3676 instead return an object that knew how to perform the register
3677 return-value extract and store.}
3679 @emph{Maintainer note: This method does not take a @var{gcc_p}
3680 parameter, and such a parameter should not be added. If an architecture
3681 that requires per-compiler or per-function information be identified,
3682 then the replacement of @var{rettype} with @code{struct value}
3683 @var{function} should be persued.}
3685 @emph{Maintainer note: The @var{regcache} parameter limits this methods
3686 to the inner most frame. While replacing @var{regcache} with a
3687 @code{struct frame_info} @var{frame} parameter would remove that
3688 limitation there has yet to be a demonstrated need for such a change.}
3690 @item SKIP_PERMANENT_BREAKPOINT
3691 @findex SKIP_PERMANENT_BREAKPOINT
3692 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3693 steps over a breakpoint by removing it, stepping one instruction, and
3694 re-inserting the breakpoint. However, permanent breakpoints are
3695 hardwired into the inferior, and can't be removed, so this strategy
3696 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
3697 state so that execution will resume just after the breakpoint. This
3698 macro does the right thing even when the breakpoint is in the delay slot
3699 of a branch or jump.
3701 @item SKIP_PROLOGUE (@var{pc})
3702 @findex SKIP_PROLOGUE
3703 A C expression that returns the address of the ``real'' code beyond the
3704 function entry prologue found at @var{pc}.
3706 @item SKIP_TRAMPOLINE_CODE (@var{pc})
3707 @findex SKIP_TRAMPOLINE_CODE
3708 If the target machine has trampoline code that sits between callers and
3709 the functions being called, then define this macro to return a new PC
3710 that is at the start of the real function.
3714 If the stack-pointer is kept in a register, then define this macro to be
3715 the number (greater than or equal to zero) of that register, or -1 if
3716 there is no such register.
3718 @item STAB_REG_TO_REGNUM
3719 @findex STAB_REG_TO_REGNUM
3720 Define this to convert stab register numbers (as gotten from `r'
3721 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
3724 @item DEPRECATED_STACK_ALIGN (@var{addr})
3725 @anchor{DEPRECATED_STACK_ALIGN}
3726 @findex DEPRECATED_STACK_ALIGN
3727 Define this to increase @var{addr} so that it meets the alignment
3728 requirements for the processor's stack.
3730 Unlike @ref{frame_align}, this function always adjusts @var{addr}
3733 By default, no stack alignment is performed.
3735 @item STEP_SKIPS_DELAY (@var{addr})
3736 @findex STEP_SKIPS_DELAY
3737 Define this to return true if the address is of an instruction with a
3738 delay slot. If a breakpoint has been placed in the instruction's delay
3739 slot, @value{GDBN} will single-step over that instruction before resuming
3740 normally. Currently only defined for the Mips.
3742 @item STORE_RETURN_VALUE (@var{type}, @var{regcache}, @var{valbuf})
3743 @findex STORE_RETURN_VALUE
3744 A C expression that writes the function return value, found in
3745 @var{valbuf}, into the @var{regcache}. @var{type} is the type of the
3746 value that is to be returned.
3748 This method has been deprecated in favour of @code{gdbarch_return_value}
3749 (@pxref{gdbarch_return_value}).
3751 @item SYMBOL_RELOADING_DEFAULT
3752 @findex SYMBOL_RELOADING_DEFAULT
3753 The default value of the ``symbol-reloading'' variable. (Never defined in
3756 @item TARGET_CHAR_BIT
3757 @findex TARGET_CHAR_BIT
3758 Number of bits in a char; defaults to 8.
3760 @item TARGET_CHAR_SIGNED
3761 @findex TARGET_CHAR_SIGNED
3762 Non-zero if @code{char} is normally signed on this architecture; zero if
3763 it should be unsigned.
3765 The ISO C standard requires the compiler to treat @code{char} as
3766 equivalent to either @code{signed char} or @code{unsigned char}; any
3767 character in the standard execution set is supposed to be positive.
3768 Most compilers treat @code{char} as signed, but @code{char} is unsigned
3769 on the IBM S/390, RS6000, and PowerPC targets.
3771 @item TARGET_COMPLEX_BIT
3772 @findex TARGET_COMPLEX_BIT
3773 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
3775 At present this macro is not used.
3777 @item TARGET_DOUBLE_BIT
3778 @findex TARGET_DOUBLE_BIT
3779 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
3781 @item TARGET_DOUBLE_COMPLEX_BIT
3782 @findex TARGET_DOUBLE_COMPLEX_BIT
3783 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
3785 At present this macro is not used.
3787 @item TARGET_FLOAT_BIT
3788 @findex TARGET_FLOAT_BIT
3789 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
3791 @item TARGET_INT_BIT
3792 @findex TARGET_INT_BIT
3793 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3795 @item TARGET_LONG_BIT
3796 @findex TARGET_LONG_BIT
3797 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3799 @item TARGET_LONG_DOUBLE_BIT
3800 @findex TARGET_LONG_DOUBLE_BIT
3801 Number of bits in a long double float;
3802 defaults to @code{2 * TARGET_DOUBLE_BIT}.
3804 @item TARGET_LONG_LONG_BIT
3805 @findex TARGET_LONG_LONG_BIT
3806 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
3808 @item TARGET_PTR_BIT
3809 @findex TARGET_PTR_BIT
3810 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
3812 @item TARGET_SHORT_BIT
3813 @findex TARGET_SHORT_BIT
3814 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
3816 @item TARGET_READ_PC
3817 @findex TARGET_READ_PC
3818 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
3819 @findex TARGET_WRITE_PC
3820 @anchor{TARGET_WRITE_PC}
3821 @itemx TARGET_READ_SP
3822 @findex TARGET_READ_SP
3823 @itemx TARGET_READ_FP
3824 @findex TARGET_READ_FP
3829 @anchor{TARGET_READ_SP} These change the behavior of @code{read_pc},
3830 @code{write_pc}, and @code{read_sp}. For most targets, these may be
3831 left undefined. @value{GDBN} will call the read and write register
3832 functions with the relevant @code{_REGNUM} argument.
3834 These macros are useful when a target keeps one of these registers in a
3835 hard to get at place; for example, part in a segment register and part
3836 in an ordinary register.
3838 @xref{unwind_sp}, which replaces @code{TARGET_READ_SP}.
3840 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
3841 @findex TARGET_VIRTUAL_FRAME_POINTER
3842 Returns a @code{(register, offset)} pair representing the virtual frame
3843 pointer in use at the code address @var{pc}. If virtual frame pointers
3844 are not used, a default definition simply returns
3845 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
3847 @item TARGET_HAS_HARDWARE_WATCHPOINTS
3848 If non-zero, the target has support for hardware-assisted
3849 watchpoints. @xref{Algorithms, watchpoints}, for more details and
3850 other related macros.
3852 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
3853 @findex TARGET_PRINT_INSN
3854 This is the function used by @value{GDBN} to print an assembly
3855 instruction. It prints the instruction at address @var{addr} in
3856 debugged memory and returns the length of the instruction, in bytes. If
3857 a target doesn't define its own printing routine, it defaults to an
3858 accessor function for the global pointer
3859 @code{deprecated_tm_print_insn}. This usually points to a function in
3860 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
3861 @var{info} is a structure (of type @code{disassemble_info}) defined in
3862 @file{include/dis-asm.h} used to pass information to the instruction
3865 @item struct frame_id unwind_dummy_id (struct frame_info *@var{frame})
3866 @findex unwind_dummy_id
3867 @anchor{unwind_dummy_id} Given @var{frame} return a @code{struct
3868 frame_id} that uniquely identifies an inferior function call's dummy
3869 frame. The value returned must match the dummy frame stack value
3870 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
3871 @xref{SAVE_DUMMY_FRAME_TOS}.
3873 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
3874 @findex DEPRECATED_USE_STRUCT_CONVENTION
3875 If defined, this must be an expression that is nonzero if a value of the
3876 given @var{type} being returned from a function must have space
3877 allocated for it on the stack. @var{gcc_p} is true if the function
3878 being considered is known to have been compiled by GCC; this is helpful
3879 for systems where GCC is known to use different calling convention than
3882 This method has been deprecated in favour of @code{gdbarch_return_value}
3883 (@pxref{gdbarch_return_value}).
3885 @item VALUE_TO_REGISTER(@var{type}, @var{regnum}, @var{from}, @var{to})
3886 @findex VALUE_TO_REGISTER
3887 Convert a value of type @var{type} into the raw contents of register
3889 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3891 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3892 @findex VARIABLES_INSIDE_BLOCK
3893 For dbx-style debugging information, if the compiler puts variable
3894 declarations inside LBRAC/RBRAC blocks, this should be defined to be
3895 nonzero. @var{desc} is the value of @code{n_desc} from the
3896 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
3897 presence of either the @code{GCC_COMPILED_SYMBOL} or the
3898 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
3900 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3901 @findex OS9K_VARIABLES_INSIDE_BLOCK
3902 Similarly, for OS/9000. Defaults to 1.
3905 Motorola M68K target conditionals.
3909 Define this to be the 4-bit location of the breakpoint trap vector. If
3910 not defined, it will default to @code{0xf}.
3912 @item REMOTE_BPT_VECTOR
3913 Defaults to @code{1}.
3915 @item NAME_OF_MALLOC
3916 @findex NAME_OF_MALLOC
3917 A string containing the name of the function to call in order to
3918 allocate some memory in the inferior. The default value is "malloc".
3922 @section Adding a New Target
3924 @cindex adding a target
3925 The following files add a target to @value{GDBN}:
3929 @item gdb/config/@var{arch}/@var{ttt}.mt
3930 Contains a Makefile fragment specific to this target. Specifies what
3931 object files are needed for target @var{ttt}, by defining
3932 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
3933 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
3936 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
3937 but these are now deprecated, replaced by autoconf, and may go away in
3938 future versions of @value{GDBN}.
3940 @item gdb/@var{ttt}-tdep.c
3941 Contains any miscellaneous code required for this target machine. On
3942 some machines it doesn't exist at all. Sometimes the macros in
3943 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
3944 as functions here instead, and the macro is simply defined to call the
3945 function. This is vastly preferable, since it is easier to understand
3948 @item gdb/@var{arch}-tdep.c
3949 @itemx gdb/@var{arch}-tdep.h
3950 This often exists to describe the basic layout of the target machine's
3951 processor chip (registers, stack, etc.). If used, it is included by
3952 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
3955 @item gdb/config/@var{arch}/tm-@var{ttt}.h
3956 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
3957 macro definitions about the target machine's registers, stack frame
3958 format and instructions.
3960 New targets do not need this file and should not create it.
3962 @item gdb/config/@var{arch}/tm-@var{arch}.h
3963 This often exists to describe the basic layout of the target machine's
3964 processor chip (registers, stack, etc.). If used, it is included by
3965 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
3968 New targets do not need this file and should not create it.
3972 If you are adding a new operating system for an existing CPU chip, add a
3973 @file{config/tm-@var{os}.h} file that describes the operating system
3974 facilities that are unusual (extra symbol table info; the breakpoint
3975 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
3976 that just @code{#include}s @file{tm-@var{arch}.h} and
3977 @file{config/tm-@var{os}.h}.
3980 @section Converting an existing Target Architecture to Multi-arch
3981 @cindex converting targets to multi-arch
3983 This section describes the current accepted best practice for converting
3984 an existing target architecture to the multi-arch framework.
3986 The process consists of generating, testing, posting and committing a
3987 sequence of patches. Each patch must contain a single change, for
3993 Directly convert a group of functions into macros (the conversion does
3994 not change the behavior of any of the functions).
3997 Replace a non-multi-arch with a multi-arch mechanism (e.g.,
4001 Enable multi-arch level one.
4004 Delete one or more files.
4009 There isn't a size limit on a patch, however, a developer is strongly
4010 encouraged to keep the patch size down.
4012 Since each patch is well defined, and since each change has been tested
4013 and shows no regressions, the patches are considered @emph{fairly}
4014 obvious. Such patches, when submitted by developers listed in the
4015 @file{MAINTAINERS} file, do not need approval. Occasional steps in the
4016 process may be more complicated and less clear. The developer is
4017 expected to use their judgment and is encouraged to seek advice as
4020 @subsection Preparation
4022 The first step is to establish control. Build (with @option{-Werror}
4023 enabled) and test the target so that there is a baseline against which
4024 the debugger can be compared.
4026 At no stage can the test results regress or @value{GDBN} stop compiling
4027 with @option{-Werror}.
4029 @subsection Add the multi-arch initialization code
4031 The objective of this step is to establish the basic multi-arch
4032 framework. It involves
4037 The addition of a @code{@var{arch}_gdbarch_init} function@footnote{The
4038 above is from the original example and uses K&R C. @value{GDBN}
4039 has since converted to ISO C but lets ignore that.} that creates
4042 static struct gdbarch *
4043 d10v_gdbarch_init (info, arches)
4044 struct gdbarch_info info;
4045 struct gdbarch_list *arches;
4047 struct gdbarch *gdbarch;
4048 /* there is only one d10v architecture */
4050 return arches->gdbarch;
4051 gdbarch = gdbarch_alloc (&info, NULL);
4059 A per-architecture dump function to print any architecture specific
4063 mips_dump_tdep (struct gdbarch *current_gdbarch,
4064 struct ui_file *file)
4066 @dots{} code to print architecture specific info @dots{}
4071 A change to @code{_initialize_@var{arch}_tdep} to register this new
4075 _initialize_mips_tdep (void)
4077 gdbarch_register (bfd_arch_mips, mips_gdbarch_init,
4082 Add the macro @code{GDB_MULTI_ARCH}, defined as 0 (zero), to the file@*
4083 @file{config/@var{arch}/tm-@var{arch}.h}.
4087 @subsection Update multi-arch incompatible mechanisms
4089 Some mechanisms do not work with multi-arch. They include:
4092 @item FRAME_FIND_SAVED_REGS
4093 Replaced with @code{DEPRECATED_FRAME_INIT_SAVED_REGS}
4097 At this stage you could also consider converting the macros into
4100 @subsection Prepare for multi-arch level to one
4102 Temporally set @code{GDB_MULTI_ARCH} to @code{GDB_MULTI_ARCH_PARTIAL}
4103 and then build and start @value{GDBN} (the change should not be
4104 committed). @value{GDBN} may not build, and once built, it may die with
4105 an internal error listing the architecture methods that must be
4108 Fix any build problems (patch(es)).
4110 Convert all the architecture methods listed, which are only macros, into
4111 functions (patch(es)).
4113 Update @code{@var{arch}_gdbarch_init} to set all the missing
4114 architecture methods and wrap the corresponding macros in @code{#if
4115 !GDB_MULTI_ARCH} (patch(es)).
4117 @subsection Set multi-arch level one
4119 Change the value of @code{GDB_MULTI_ARCH} to GDB_MULTI_ARCH_PARTIAL (a
4122 Any problems with throwing ``the switch'' should have been fixed
4125 @subsection Convert remaining macros
4127 Suggest converting macros into functions (and setting the corresponding
4128 architecture method) in small batches.
4130 @subsection Set multi-arch level to two
4132 This should go smoothly.
4134 @subsection Delete the TM file
4136 The @file{tm-@var{arch}.h} can be deleted. @file{@var{arch}.mt} and
4137 @file{configure.in} updated.
4140 @node Target Vector Definition
4142 @chapter Target Vector Definition
4143 @cindex target vector
4145 The target vector defines the interface between @value{GDBN}'s
4146 abstract handling of target systems, and the nitty-gritty code that
4147 actually exercises control over a process or a serial port.
4148 @value{GDBN} includes some 30-40 different target vectors; however,
4149 each configuration of @value{GDBN} includes only a few of them.
4151 @section File Targets
4153 Both executables and core files have target vectors.
4155 @section Standard Protocol and Remote Stubs
4157 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4158 that runs in the target system. @value{GDBN} provides several sample
4159 @dfn{stubs} that can be integrated into target programs or operating
4160 systems for this purpose; they are named @file{*-stub.c}.
4162 The @value{GDBN} user's manual describes how to put such a stub into
4163 your target code. What follows is a discussion of integrating the
4164 SPARC stub into a complicated operating system (rather than a simple
4165 program), by Stu Grossman, the author of this stub.
4167 The trap handling code in the stub assumes the following upon entry to
4172 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4178 you are in the correct trap window.
4181 As long as your trap handler can guarantee those conditions, then there
4182 is no reason why you shouldn't be able to ``share'' traps with the stub.
4183 The stub has no requirement that it be jumped to directly from the
4184 hardware trap vector. That is why it calls @code{exceptionHandler()},
4185 which is provided by the external environment. For instance, this could
4186 set up the hardware traps to actually execute code which calls the stub
4187 first, and then transfers to its own trap handler.
4189 For the most point, there probably won't be much of an issue with
4190 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4191 and often indicate unrecoverable error conditions. Anyway, this is all
4192 controlled by a table, and is trivial to modify. The most important
4193 trap for us is for @code{ta 1}. Without that, we can't single step or
4194 do breakpoints. Everything else is unnecessary for the proper operation
4195 of the debugger/stub.
4197 From reading the stub, it's probably not obvious how breakpoints work.
4198 They are simply done by deposit/examine operations from @value{GDBN}.
4200 @section ROM Monitor Interface
4202 @section Custom Protocols
4204 @section Transport Layer
4206 @section Builtin Simulator
4209 @node Native Debugging
4211 @chapter Native Debugging
4212 @cindex native debugging
4214 Several files control @value{GDBN}'s configuration for native support:
4218 @item gdb/config/@var{arch}/@var{xyz}.mh
4219 Specifies Makefile fragments needed by a @emph{native} configuration on
4220 machine @var{xyz}. In particular, this lists the required
4221 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4222 Also specifies the header file which describes native support on
4223 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4224 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4225 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4227 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4228 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4229 on machine @var{xyz}. While the file is no longer used for this
4230 purpose, the @file{.mh} suffix remains. Perhaps someone will
4231 eventually rename these fragments so that they have a @file{.mn}
4234 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4235 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4236 macro definitions describing the native system environment, such as
4237 child process control and core file support.
4239 @item gdb/@var{xyz}-nat.c
4240 Contains any miscellaneous C code required for this native support of
4241 this machine. On some machines it doesn't exist at all.
4244 There are some ``generic'' versions of routines that can be used by
4245 various systems. These can be customized in various ways by macros
4246 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4247 the @var{xyz} host, you can just include the generic file's name (with
4248 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4250 Otherwise, if your machine needs custom support routines, you will need
4251 to write routines that perform the same functions as the generic file.
4252 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4253 into @code{NATDEPFILES}.
4257 This contains the @emph{target_ops vector} that supports Unix child
4258 processes on systems which use ptrace and wait to control the child.
4261 This contains the @emph{target_ops vector} that supports Unix child
4262 processes on systems which use /proc to control the child.
4265 This does the low-level grunge that uses Unix system calls to do a ``fork
4266 and exec'' to start up a child process.
4269 This is the low level interface to inferior processes for systems using
4270 the Unix @code{ptrace} call in a vanilla way.
4273 @section Native core file Support
4274 @cindex native core files
4277 @findex fetch_core_registers
4278 @item core-aout.c::fetch_core_registers()
4279 Support for reading registers out of a core file. This routine calls
4280 @code{register_addr()}, see below. Now that BFD is used to read core
4281 files, virtually all machines should use @code{core-aout.c}, and should
4282 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4283 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4285 @item core-aout.c::register_addr()
4286 If your @code{nm-@var{xyz}.h} file defines the macro
4287 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4288 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4289 register number @code{regno}. @code{blockend} is the offset within the
4290 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4291 @file{core-aout.c} will define the @code{register_addr()} function and
4292 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4293 you are using the standard @code{fetch_core_registers()}, you will need
4294 to define your own version of @code{register_addr()}, put it into your
4295 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4296 the @code{NATDEPFILES} list. If you have your own
4297 @code{fetch_core_registers()}, you may not need a separate
4298 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4299 implementations simply locate the registers themselves.@refill
4302 When making @value{GDBN} run native on a new operating system, to make it
4303 possible to debug core files, you will need to either write specific
4304 code for parsing your OS's core files, or customize
4305 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4306 machine uses to define the struct of registers that is accessible
4307 (possibly in the u-area) in a core file (rather than
4308 @file{machine/reg.h}), and an include file that defines whatever header
4309 exists on a core file (e.g. the u-area or a @code{struct core}). Then
4310 modify @code{trad_unix_core_file_p} to use these values to set up the
4311 section information for the data segment, stack segment, any other
4312 segments in the core file (perhaps shared library contents or control
4313 information), ``registers'' segment, and if there are two discontiguous
4314 sets of registers (e.g. integer and float), the ``reg2'' segment. This
4315 section information basically delimits areas in the core file in a
4316 standard way, which the section-reading routines in BFD know how to seek
4319 Then back in @value{GDBN}, you need a matching routine called
4320 @code{fetch_core_registers}. If you can use the generic one, it's in
4321 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4322 It will be passed a char pointer to the entire ``registers'' segment,
4323 its length, and a zero; or a char pointer to the entire ``regs2''
4324 segment, its length, and a 2. The routine should suck out the supplied
4325 register values and install them into @value{GDBN}'s ``registers'' array.
4327 If your system uses @file{/proc} to control processes, and uses ELF
4328 format core files, then you may be able to use the same routines for
4329 reading the registers out of processes and out of core files.
4337 @section shared libraries
4339 @section Native Conditionals
4340 @cindex native conditionals
4342 When @value{GDBN} is configured and compiled, various macros are
4343 defined or left undefined, to control compilation when the host and
4344 target systems are the same. These macros should be defined (or left
4345 undefined) in @file{nm-@var{system}.h}.
4349 @item CHILD_PREPARE_TO_STORE
4350 @findex CHILD_PREPARE_TO_STORE
4351 If the machine stores all registers at once in the child process, then
4352 define this to ensure that all values are correct. This usually entails
4353 a read from the child.
4355 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4358 @item FETCH_INFERIOR_REGISTERS
4359 @findex FETCH_INFERIOR_REGISTERS
4360 Define this if the native-dependent code will provide its own routines
4361 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4362 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4363 @file{infptrace.c} is included in this configuration, the default
4364 routines in @file{infptrace.c} are used for these functions.
4368 This macro is normally defined to be the number of the first floating
4369 point register, if the machine has such registers. As such, it would
4370 appear only in target-specific code. However, @file{/proc} support uses this
4371 to decide whether floats are in use on this target.
4373 @item GET_LONGJMP_TARGET
4374 @findex GET_LONGJMP_TARGET
4375 For most machines, this is a target-dependent parameter. On the
4376 DECstation and the Iris, this is a native-dependent parameter, since
4377 @file{setjmp.h} is needed to define it.
4379 This macro determines the target PC address that @code{longjmp} will jump to,
4380 assuming that we have just stopped at a longjmp breakpoint. It takes a
4381 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4382 pointer. It examines the current state of the machine as needed.
4384 @item I386_USE_GENERIC_WATCHPOINTS
4385 An x86-based machine can define this to use the generic x86 watchpoint
4386 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4389 @findex KERNEL_U_ADDR
4390 Define this to the address of the @code{u} structure (the ``user
4391 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
4392 needs to know this so that it can subtract this address from absolute
4393 addresses in the upage, that are obtained via ptrace or from core files.
4394 On systems that don't need this value, set it to zero.
4396 @item KERNEL_U_ADDR_HPUX
4397 @findex KERNEL_U_ADDR_HPUX
4398 Define this to cause @value{GDBN} to determine the address of @code{u} at
4399 runtime, by using HP-style @code{nlist} on the kernel's image in the
4402 @item ONE_PROCESS_WRITETEXT
4403 @findex ONE_PROCESS_WRITETEXT
4404 Define this to be able to, when a breakpoint insertion fails, warn the
4405 user that another process may be running with the same executable.
4408 @findex PROC_NAME_FMT
4409 Defines the format for the name of a @file{/proc} device. Should be
4410 defined in @file{nm.h} @emph{only} in order to override the default
4411 definition in @file{procfs.c}.
4413 @item PTRACE_ARG3_TYPE
4414 @findex PTRACE_ARG3_TYPE
4415 The type of the third argument to the @code{ptrace} system call, if it
4416 exists and is different from @code{int}.
4418 @item REGISTER_U_ADDR
4419 @findex REGISTER_U_ADDR
4420 Defines the offset of the registers in the ``u area''.
4422 @item SHELL_COMMAND_CONCAT
4423 @findex SHELL_COMMAND_CONCAT
4424 If defined, is a string to prefix on the shell command used to start the
4429 If defined, this is the name of the shell to use to run the inferior.
4430 Defaults to @code{"/bin/sh"}.
4432 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4434 Define this to expand into an expression that will cause the symbols in
4435 @var{filename} to be added to @value{GDBN}'s symbol table. If
4436 @var{readsyms} is zero symbols are not read but any necessary low level
4437 processing for @var{filename} is still done.
4439 @item SOLIB_CREATE_INFERIOR_HOOK
4440 @findex SOLIB_CREATE_INFERIOR_HOOK
4441 Define this to expand into any shared-library-relocation code that you
4442 want to be run just after the child process has been forked.
4444 @item START_INFERIOR_TRAPS_EXPECTED
4445 @findex START_INFERIOR_TRAPS_EXPECTED
4446 When starting an inferior, @value{GDBN} normally expects to trap
4448 the shell execs, and once when the program itself execs. If the actual
4449 number of traps is something other than 2, then define this macro to
4450 expand into the number expected.
4454 This determines whether small routines in @file{*-tdep.c}, which
4455 translate register values between @value{GDBN}'s internal
4456 representation and the @file{/proc} representation, are compiled.
4459 @findex U_REGS_OFFSET
4460 This is the offset of the registers in the upage. It need only be
4461 defined if the generic ptrace register access routines in
4462 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4463 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4464 the default value from @file{infptrace.c} is good enough, leave it
4467 The default value means that u.u_ar0 @emph{points to} the location of
4468 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4469 that @code{u.u_ar0} @emph{is} the location of the registers.
4473 See @file{objfiles.c}.
4476 @findex DEBUG_PTRACE
4477 Define this to debug @code{ptrace} calls.
4481 @node Support Libraries
4483 @chapter Support Libraries
4488 BFD provides support for @value{GDBN} in several ways:
4491 @item identifying executable and core files
4492 BFD will identify a variety of file types, including a.out, coff, and
4493 several variants thereof, as well as several kinds of core files.
4495 @item access to sections of files
4496 BFD parses the file headers to determine the names, virtual addresses,
4497 sizes, and file locations of all the various named sections in files
4498 (such as the text section or the data section). @value{GDBN} simply
4499 calls BFD to read or write section @var{x} at byte offset @var{y} for
4502 @item specialized core file support
4503 BFD provides routines to determine the failing command name stored in a
4504 core file, the signal with which the program failed, and whether a core
4505 file matches (i.e.@: could be a core dump of) a particular executable
4508 @item locating the symbol information
4509 @value{GDBN} uses an internal interface of BFD to determine where to find the
4510 symbol information in an executable file or symbol-file. @value{GDBN} itself
4511 handles the reading of symbols, since BFD does not ``understand'' debug
4512 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4517 @cindex opcodes library
4519 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4520 library because it's also used in binutils, for @file{objdump}).
4527 @cindex @code{libiberty} library
4529 The @code{libiberty} library provides a set of functions and features
4530 that integrate and improve on functionality found in modern operating
4531 systems. Broadly speaking, such features can be divided into three
4532 groups: supplemental functions (functions that may be missing in some
4533 environments and operating systems), replacement functions (providing
4534 a uniform and easier to use interface for commonly used standard
4535 functions), and extensions (which provide additional functionality
4536 beyond standard functions).
4538 @value{GDBN} uses various features provided by the @code{libiberty}
4539 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4540 floating format support functions, the input options parser
4541 @samp{getopt}, the @samp{obstack} extension, and other functions.
4543 @subsection @code{obstacks} in @value{GDBN}
4544 @cindex @code{obstacks}
4546 The obstack mechanism provides a convenient way to allocate and free
4547 chunks of memory. Each obstack is a pool of memory that is managed
4548 like a stack. Objects (of any nature, size and alignment) are
4549 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4550 @code{libiberty}'s documenatation for a more detailed explanation of
4553 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4554 object files. There is an obstack associated with each internal
4555 representation of an object file. Lots of things get allocated on
4556 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4557 symbols, minimal symbols, types, vectors of fundamental types, class
4558 fields of types, object files section lists, object files section
4559 offets lists, line tables, symbol tables, partial symbol tables,
4560 string tables, symbol table private data, macros tables, debug
4561 information sections and entries, import and export lists (som),
4562 unwind information (hppa), dwarf2 location expressions data. Plus
4563 various strings such as directory names strings, debug format strings,
4566 An essential and convenient property of all data on @code{obstacks} is
4567 that memory for it gets allocated (with @code{obstack_alloc}) at
4568 various times during a debugging sesssion, but it is released all at
4569 once using the @code{obstack_free} function. The @code{obstack_free}
4570 function takes a pointer to where in the stack it must start the
4571 deletion from (much like the cleanup chains have a pointer to where to
4572 start the cleanups). Because of the stack like structure of the
4573 @code{obstacks}, this allows to free only a top portion of the
4574 obstack. There are a few instances in @value{GDBN} where such thing
4575 happens. Calls to @code{obstack_free} are done after some local data
4576 is allocated to the obstack. Only the local data is deleted from the
4577 obstack. Of course this assumes that nothing between the
4578 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4579 else on the same obstack. For this reason it is best and safest to
4580 use temporary @code{obstacks}.
4582 Releasing the whole obstack is also not safe per se. It is safe only
4583 under the condition that we know the @code{obstacks} memory is no
4584 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4585 when we get rid of the whole objfile(s), for instance upon reading a
4589 @cindex regular expressions library
4600 @item SIGN_EXTEND_CHAR
4602 @item SWITCH_ENUM_BUG
4617 This chapter covers topics that are lower-level than the major
4618 algorithms of @value{GDBN}.
4623 Cleanups are a structured way to deal with things that need to be done
4626 When your code does something (e.g., @code{xmalloc} some memory, or
4627 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4628 the memory or @code{close} the file), it can make a cleanup. The
4629 cleanup will be done at some future point: when the command is finished
4630 and control returns to the top level; when an error occurs and the stack
4631 is unwound; or when your code decides it's time to explicitly perform
4632 cleanups. Alternatively you can elect to discard the cleanups you
4638 @item struct cleanup *@var{old_chain};
4639 Declare a variable which will hold a cleanup chain handle.
4641 @findex make_cleanup
4642 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
4643 Make a cleanup which will cause @var{function} to be called with
4644 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
4645 handle that can later be passed to @code{do_cleanups} or
4646 @code{discard_cleanups}. Unless you are going to call
4647 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
4648 from @code{make_cleanup}.
4651 @item do_cleanups (@var{old_chain});
4652 Do all cleanups added to the chain since the corresponding
4653 @code{make_cleanup} call was made.
4655 @findex discard_cleanups
4656 @item discard_cleanups (@var{old_chain});
4657 Same as @code{do_cleanups} except that it just removes the cleanups from
4658 the chain and does not call the specified functions.
4661 Cleanups are implemented as a chain. The handle returned by
4662 @code{make_cleanups} includes the cleanup passed to the call and any
4663 later cleanups appended to the chain (but not yet discarded or
4667 make_cleanup (a, 0);
4669 struct cleanup *old = make_cleanup (b, 0);
4677 will call @code{c()} and @code{b()} but will not call @code{a()}. The
4678 cleanup that calls @code{a()} will remain in the cleanup chain, and will
4679 be done later unless otherwise discarded.@refill
4681 Your function should explicitly do or discard the cleanups it creates.
4682 Failing to do this leads to non-deterministic behavior since the caller
4683 will arbitrarily do or discard your functions cleanups. This need leads
4684 to two common cleanup styles.
4686 The first style is try/finally. Before it exits, your code-block calls
4687 @code{do_cleanups} with the old cleanup chain and thus ensures that your
4688 code-block's cleanups are always performed. For instance, the following
4689 code-segment avoids a memory leak problem (even when @code{error} is
4690 called and a forced stack unwind occurs) by ensuring that the
4691 @code{xfree} will always be called:
4694 struct cleanup *old = make_cleanup (null_cleanup, 0);
4695 data = xmalloc (sizeof blah);
4696 make_cleanup (xfree, data);
4701 The second style is try/except. Before it exits, your code-block calls
4702 @code{discard_cleanups} with the old cleanup chain and thus ensures that
4703 any created cleanups are not performed. For instance, the following
4704 code segment, ensures that the file will be closed but only if there is
4708 FILE *file = fopen ("afile", "r");
4709 struct cleanup *old = make_cleanup (close_file, file);
4711 discard_cleanups (old);
4715 Some functions, e.g. @code{fputs_filtered()} or @code{error()}, specify
4716 that they ``should not be called when cleanups are not in place''. This
4717 means that any actions you need to reverse in the case of an error or
4718 interruption must be on the cleanup chain before you call these
4719 functions, since they might never return to your code (they
4720 @samp{longjmp} instead).
4722 @section Per-architecture module data
4723 @cindex per-architecture module data
4724 @cindex multi-arch data
4725 @cindex data-pointer, per-architecture/per-module
4727 The multi-arch framework includes a mechanism for adding module
4728 specific per-architecture data-pointers to the @code{struct gdbarch}
4729 architecture object.
4731 A module registers one or more per-architecture data-pointers using:
4733 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
4734 @var{pre_init} is used to, on-demand, allocate an initial value for a
4735 per-architecture data-pointer using the architecture's obstack (passed
4736 in as a parameter). Since @var{pre_init} can be called during
4737 architecture creation, it is not parameterized with the architecture.
4738 and must not call modules that use per-architecture data.
4741 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
4742 @var{post_init} is used to obtain an initial value for a
4743 per-architecture data-pointer @emph{after}. Since @var{post_init} is
4744 always called after architecture creation, it both receives the fully
4745 initialized architecture and is free to call modules that use
4746 per-architecture data (care needs to be taken to ensure that those
4747 other modules do not try to call back to this module as that will
4748 create in cycles in the initialization call graph).
4751 These functions return a @code{struct gdbarch_data} that is used to
4752 identify the per-architecture data-pointer added for that module.
4754 The per-architecture data-pointer is accessed using the function:
4756 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
4757 Given the architecture @var{arch} and module data handle
4758 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
4759 or @code{gdbarch_data_register_post_init}), this function returns the
4760 current value of the per-architecture data-pointer. If the data
4761 pointer is @code{NULL}, it is first initialized by calling the
4762 corresponding @var{pre_init} or @var{post_init} method.
4765 The examples below assume the following definitions:
4768 struct nozel @{ int total; @};
4769 static struct gdbarch_data *nozel_handle;
4772 A module can extend the architecture vector, adding additional
4773 per-architecture data, using the @var{pre_init} method. The module's
4774 per-architecture data is then initialized during architecture
4777 In the below, the module's per-architecture @emph{nozel} is added. An
4778 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
4779 from @code{gdbarch_init}.
4783 nozel_pre_init (struct obstack *obstack)
4785 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
4792 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
4794 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
4795 data->total = nozel;
4799 A module can on-demand create architecture dependant data structures
4800 using @code{post_init}.
4802 In the below, the nozel's total is computed on-demand by
4803 @code{nozel_post_init} using information obtained from the
4808 nozel_post_init (struct gdbarch *gdbarch)
4810 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
4811 nozel->total = gdbarch@dots{} (gdbarch);
4818 nozel_total (struct gdbarch *gdbarch)
4820 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
4825 @section Wrapping Output Lines
4826 @cindex line wrap in output
4829 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
4830 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
4831 added in places that would be good breaking points. The utility
4832 routines will take care of actually wrapping if the line width is
4835 The argument to @code{wrap_here} is an indentation string which is
4836 printed @emph{only} if the line breaks there. This argument is saved
4837 away and used later. It must remain valid until the next call to
4838 @code{wrap_here} or until a newline has been printed through the
4839 @code{*_filtered} functions. Don't pass in a local variable and then
4842 It is usually best to call @code{wrap_here} after printing a comma or
4843 space. If you call it before printing a space, make sure that your
4844 indentation properly accounts for the leading space that will print if
4845 the line wraps there.
4847 Any function or set of functions that produce filtered output must
4848 finish by printing a newline, to flush the wrap buffer, before switching
4849 to unfiltered (@code{printf}) output. Symbol reading routines that
4850 print warnings are a good example.
4852 @section @value{GDBN} Coding Standards
4853 @cindex coding standards
4855 @value{GDBN} follows the GNU coding standards, as described in
4856 @file{etc/standards.texi}. This file is also available for anonymous
4857 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
4858 of the standard; in general, when the GNU standard recommends a practice
4859 but does not require it, @value{GDBN} requires it.
4861 @value{GDBN} follows an additional set of coding standards specific to
4862 @value{GDBN}, as described in the following sections.
4867 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
4870 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
4873 @subsection Memory Management
4875 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
4876 @code{calloc}, @code{free} and @code{asprintf}.
4878 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
4879 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
4880 these functions do not return when the memory pool is empty. Instead,
4881 they unwind the stack using cleanups. These functions return
4882 @code{NULL} when requested to allocate a chunk of memory of size zero.
4884 @emph{Pragmatics: By using these functions, the need to check every
4885 memory allocation is removed. These functions provide portable
4888 @value{GDBN} does not use the function @code{free}.
4890 @value{GDBN} uses the function @code{xfree} to return memory to the
4891 memory pool. Consistent with ISO-C, this function ignores a request to
4892 free a @code{NULL} pointer.
4894 @emph{Pragmatics: On some systems @code{free} fails when passed a
4895 @code{NULL} pointer.}
4897 @value{GDBN} can use the non-portable function @code{alloca} for the
4898 allocation of small temporary values (such as strings).
4900 @emph{Pragmatics: This function is very non-portable. Some systems
4901 restrict the memory being allocated to no more than a few kilobytes.}
4903 @value{GDBN} uses the string function @code{xstrdup} and the print
4904 function @code{xstrprintf}.
4906 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
4907 functions such as @code{sprintf} are very prone to buffer overflow
4911 @subsection Compiler Warnings
4912 @cindex compiler warnings
4914 With few exceptions, developers should include the configuration option
4915 @samp{--enable-gdb-build-warnings=,-Werror} when building @value{GDBN}.
4916 The exceptions are listed in the file @file{gdb/MAINTAINERS}.
4918 This option causes @value{GDBN} (when built using GCC) to be compiled
4919 with a carefully selected list of compiler warning flags. Any warnings
4920 from those flags being treated as errors.
4922 The current list of warning flags includes:
4926 Since @value{GDBN} coding standard requires all functions to be declared
4927 using a prototype, the flag has the side effect of ensuring that
4928 prototyped functions are always visible with out resorting to
4929 @samp{-Wstrict-prototypes}.
4932 Such code often appears to work except on instruction set architectures
4933 that use register windows.
4940 @itemx -Wformat-nonliteral
4941 Since @value{GDBN} uses the @code{format printf} attribute on all
4942 @code{printf} like functions these check not just @code{printf} calls
4943 but also calls to functions such as @code{fprintf_unfiltered}.
4946 This warning includes uses of the assignment operator within an
4947 @code{if} statement.
4949 @item -Wpointer-arith
4951 @item -Wuninitialized
4953 @item -Wunused-label
4954 This warning has the additional benefit of detecting the absence of the
4955 @code{case} reserved word in a switch statement:
4957 enum @{ FD_SCHEDULED, NOTHING_SCHEDULED @} sched;
4970 @item -Wunused-function
4973 @emph{Pragmatics: Due to the way that @value{GDBN} is implemented most
4974 functions have unused parameters. Consequently the warning
4975 @samp{-Wunused-parameter} is precluded from the list. The macro
4976 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
4977 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
4978 is being used. The options @samp{-Wall} and @samp{-Wunused} are also
4979 precluded because they both include @samp{-Wunused-parameter}.}
4981 @emph{Pragmatics: @value{GDBN} has not simply accepted the warnings
4982 enabled by @samp{-Wall -Werror -W...}. Instead it is selecting warnings
4983 when and where their benefits can be demonstrated.}
4985 @subsection Formatting
4987 @cindex source code formatting
4988 The standard GNU recommendations for formatting must be followed
4991 A function declaration should not have its name in column zero. A
4992 function definition should have its name in column zero.
4996 static void foo (void);
5004 @emph{Pragmatics: This simplifies scripting. Function definitions can
5005 be found using @samp{^function-name}.}
5007 There must be a space between a function or macro name and the opening
5008 parenthesis of its argument list (except for macro definitions, as
5009 required by C). There must not be a space after an open paren/bracket
5010 or before a close paren/bracket.
5012 While additional whitespace is generally helpful for reading, do not use
5013 more than one blank line to separate blocks, and avoid adding whitespace
5014 after the end of a program line (as of 1/99, some 600 lines had
5015 whitespace after the semicolon). Excess whitespace causes difficulties
5016 for @code{diff} and @code{patch} utilities.
5018 Pointers are declared using the traditional K&R C style:
5032 @subsection Comments
5034 @cindex comment formatting
5035 The standard GNU requirements on comments must be followed strictly.
5037 Block comments must appear in the following form, with no @code{/*}- or
5038 @code{*/}-only lines, and no leading @code{*}:
5041 /* Wait for control to return from inferior to debugger. If inferior
5042 gets a signal, we may decide to start it up again instead of
5043 returning. That is why there is a loop in this function. When
5044 this function actually returns it means the inferior should be left
5045 stopped and @value{GDBN} should read more commands. */
5048 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5049 comment works correctly, and @kbd{M-q} fills the block consistently.)
5051 Put a blank line between the block comments preceding function or
5052 variable definitions, and the definition itself.
5054 In general, put function-body comments on lines by themselves, rather
5055 than trying to fit them into the 20 characters left at the end of a
5056 line, since either the comment or the code will inevitably get longer
5057 than will fit, and then somebody will have to move it anyhow.
5061 @cindex C data types
5062 Code must not depend on the sizes of C data types, the format of the
5063 host's floating point numbers, the alignment of anything, or the order
5064 of evaluation of expressions.
5066 @cindex function usage
5067 Use functions freely. There are only a handful of compute-bound areas
5068 in @value{GDBN} that might be affected by the overhead of a function
5069 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5070 limited by the target interface (whether serial line or system call).
5072 However, use functions with moderation. A thousand one-line functions
5073 are just as hard to understand as a single thousand-line function.
5075 @emph{Macros are bad, M'kay.}
5076 (But if you have to use a macro, make sure that the macro arguments are
5077 protected with parentheses.)
5081 Declarations like @samp{struct foo *} should be used in preference to
5082 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5085 @subsection Function Prototypes
5086 @cindex function prototypes
5088 Prototypes must be used when both @emph{declaring} and @emph{defining}
5089 a function. Prototypes for @value{GDBN} functions must include both the
5090 argument type and name, with the name matching that used in the actual
5091 function definition.
5093 All external functions should have a declaration in a header file that
5094 callers include, except for @code{_initialize_*} functions, which must
5095 be external so that @file{init.c} construction works, but shouldn't be
5096 visible to random source files.
5098 Where a source file needs a forward declaration of a static function,
5099 that declaration must appear in a block near the top of the source file.
5102 @subsection Internal Error Recovery
5104 During its execution, @value{GDBN} can encounter two types of errors.
5105 User errors and internal errors. User errors include not only a user
5106 entering an incorrect command but also problems arising from corrupt
5107 object files and system errors when interacting with the target.
5108 Internal errors include situations where @value{GDBN} has detected, at
5109 run time, a corrupt or erroneous situation.
5111 When reporting an internal error, @value{GDBN} uses
5112 @code{internal_error} and @code{gdb_assert}.
5114 @value{GDBN} must not call @code{abort} or @code{assert}.
5116 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5117 the code detected a user error, recovered from it and issued a
5118 @code{warning} or the code failed to correctly recover from the user
5119 error and issued an @code{internal_error}.}
5121 @subsection File Names
5123 Any file used when building the core of @value{GDBN} must be in lower
5124 case. Any file used when building the core of @value{GDBN} must be 8.3
5125 unique. These requirements apply to both source and generated files.
5127 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5128 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5129 is introduced to the build process both @file{Makefile.in} and
5130 @file{configure.in} need to be modified accordingly. Compare the
5131 convoluted conversion process needed to transform @file{COPYING} into
5132 @file{copying.c} with the conversion needed to transform
5133 @file{version.in} into @file{version.c}.}
5135 Any file non 8.3 compliant file (that is not used when building the core
5136 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5138 @emph{Pragmatics: This is clearly a compromise.}
5140 When @value{GDBN} has a local version of a system header file (ex
5141 @file{string.h}) the file name based on the POSIX header prefixed with
5142 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5143 independent: they should use only macros defined by @file{configure},
5144 the compiler, or the host; they should include only system headers; they
5145 should refer only to system types. They may be shared between multiple
5146 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5148 For other files @samp{-} is used as the separator.
5151 @subsection Include Files
5153 A @file{.c} file should include @file{defs.h} first.
5155 A @file{.c} file should directly include the @code{.h} file of every
5156 declaration and/or definition it directly refers to. It cannot rely on
5159 A @file{.h} file should directly include the @code{.h} file of every
5160 declaration and/or definition it directly refers to. It cannot rely on
5161 indirect inclusion. Exception: The file @file{defs.h} does not need to
5162 be directly included.
5164 An external declaration should only appear in one include file.
5166 An external declaration should never appear in a @code{.c} file.
5167 Exception: a declaration for the @code{_initialize} function that
5168 pacifies @option{-Wmissing-declaration}.
5170 A @code{typedef} definition should only appear in one include file.
5172 An opaque @code{struct} declaration can appear in multiple @file{.h}
5173 files. Where possible, a @file{.h} file should use an opaque
5174 @code{struct} declaration instead of an include.
5176 All @file{.h} files should be wrapped in:
5179 #ifndef INCLUDE_FILE_NAME_H
5180 #define INCLUDE_FILE_NAME_H
5186 @subsection Clean Design and Portable Implementation
5189 In addition to getting the syntax right, there's the little question of
5190 semantics. Some things are done in certain ways in @value{GDBN} because long
5191 experience has shown that the more obvious ways caused various kinds of
5194 @cindex assumptions about targets
5195 You can't assume the byte order of anything that comes from a target
5196 (including @var{value}s, object files, and instructions). Such things
5197 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5198 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5199 such as @code{bfd_get_32}.
5201 You can't assume that you know what interface is being used to talk to
5202 the target system. All references to the target must go through the
5203 current @code{target_ops} vector.
5205 You can't assume that the host and target machines are the same machine
5206 (except in the ``native'' support modules). In particular, you can't
5207 assume that the target machine's header files will be available on the
5208 host machine. Target code must bring along its own header files --
5209 written from scratch or explicitly donated by their owner, to avoid
5213 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5214 to write the code portably than to conditionalize it for various
5217 @cindex system dependencies
5218 New @code{#ifdef}'s which test for specific compilers or manufacturers
5219 or operating systems are unacceptable. All @code{#ifdef}'s should test
5220 for features. The information about which configurations contain which
5221 features should be segregated into the configuration files. Experience
5222 has proven far too often that a feature unique to one particular system
5223 often creeps into other systems; and that a conditional based on some
5224 predefined macro for your current system will become worthless over
5225 time, as new versions of your system come out that behave differently
5226 with regard to this feature.
5228 Adding code that handles specific architectures, operating systems,
5229 target interfaces, or hosts, is not acceptable in generic code.
5231 @cindex portable file name handling
5232 @cindex file names, portability
5233 One particularly notorious area where system dependencies tend to
5234 creep in is handling of file names. The mainline @value{GDBN} code
5235 assumes Posix semantics of file names: absolute file names begin with
5236 a forward slash @file{/}, slashes are used to separate leading
5237 directories, case-sensitive file names. These assumptions are not
5238 necessarily true on non-Posix systems such as MS-Windows. To avoid
5239 system-dependent code where you need to take apart or construct a file
5240 name, use the following portable macros:
5243 @findex HAVE_DOS_BASED_FILE_SYSTEM
5244 @item HAVE_DOS_BASED_FILE_SYSTEM
5245 This preprocessing symbol is defined to a non-zero value on hosts
5246 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5247 symbol to write conditional code which should only be compiled for
5250 @findex IS_DIR_SEPARATOR
5251 @item IS_DIR_SEPARATOR (@var{c})
5252 Evaluates to a non-zero value if @var{c} is a directory separator
5253 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5254 such a character, but on Windows, both @file{/} and @file{\} will
5257 @findex IS_ABSOLUTE_PATH
5258 @item IS_ABSOLUTE_PATH (@var{file})
5259 Evaluates to a non-zero value if @var{file} is an absolute file name.
5260 For Unix and GNU/Linux hosts, a name which begins with a slash
5261 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5262 @file{x:\bar} are also absolute file names.
5264 @findex FILENAME_CMP
5265 @item FILENAME_CMP (@var{f1}, @var{f2})
5266 Calls a function which compares file names @var{f1} and @var{f2} as
5267 appropriate for the underlying host filesystem. For Posix systems,
5268 this simply calls @code{strcmp}; on case-insensitive filesystems it
5269 will call @code{strcasecmp} instead.
5271 @findex DIRNAME_SEPARATOR
5272 @item DIRNAME_SEPARATOR
5273 Evaluates to a character which separates directories in
5274 @code{PATH}-style lists, typically held in environment variables.
5275 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5277 @findex SLASH_STRING
5279 This evaluates to a constant string you should use to produce an
5280 absolute filename from leading directories and the file's basename.
5281 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5282 @code{"\\"} for some Windows-based ports.
5285 In addition to using these macros, be sure to use portable library
5286 functions whenever possible. For example, to extract a directory or a
5287 basename part from a file name, use the @code{dirname} and
5288 @code{basename} library functions (available in @code{libiberty} for
5289 platforms which don't provide them), instead of searching for a slash
5290 with @code{strrchr}.
5292 Another way to generalize @value{GDBN} along a particular interface is with an
5293 attribute struct. For example, @value{GDBN} has been generalized to handle
5294 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5295 by defining the @code{target_ops} structure and having a current target (as
5296 well as a stack of targets below it, for memory references). Whenever
5297 something needs to be done that depends on which remote interface we are
5298 using, a flag in the current target_ops structure is tested (e.g.,
5299 @code{target_has_stack}), or a function is called through a pointer in the
5300 current target_ops structure. In this way, when a new remote interface
5301 is added, only one module needs to be touched---the one that actually
5302 implements the new remote interface. Other examples of
5303 attribute-structs are BFD access to multiple kinds of object file
5304 formats, or @value{GDBN}'s access to multiple source languages.
5306 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5307 the code interfacing between @code{ptrace} and the rest of
5308 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5309 something was very painful. In @value{GDBN} 4.x, these have all been
5310 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5311 with variations between systems the same way any system-independent
5312 file would (hooks, @code{#if defined}, etc.), and machines which are
5313 radically different don't need to use @file{infptrace.c} at all.
5315 All debugging code must be controllable using the @samp{set debug
5316 @var{module}} command. Do not use @code{printf} to print trace
5317 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5318 @code{#ifdef DEBUG}.
5323 @chapter Porting @value{GDBN}
5324 @cindex porting to new machines
5326 Most of the work in making @value{GDBN} compile on a new machine is in
5327 specifying the configuration of the machine. This is done in a
5328 dizzying variety of header files and configuration scripts, which we
5329 hope to make more sensible soon. Let's say your new host is called an
5330 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5331 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5332 @samp{sparc-sun-sunos4}). In particular:
5336 In the top level directory, edit @file{config.sub} and add @var{arch},
5337 @var{xvend}, and @var{xos} to the lists of supported architectures,
5338 vendors, and operating systems near the bottom of the file. Also, add
5339 @var{xyz} as an alias that maps to
5340 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5344 ./config.sub @var{xyz}
5351 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5355 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5356 and no error messages.
5359 You need to port BFD, if that hasn't been done already. Porting BFD is
5360 beyond the scope of this manual.
5363 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5364 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5365 desired target is already available) also edit @file{gdb/configure.tgt},
5366 setting @code{gdb_target} to something appropriate (for instance,
5369 @emph{Maintainer's note: Work in progress. The file
5370 @file{gdb/configure.host} originally needed to be modified when either a
5371 new native target or a new host machine was being added to @value{GDBN}.
5372 Recent changes have removed this requirement. The file now only needs
5373 to be modified when adding a new native configuration. This will likely
5374 changed again in the future.}
5377 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5378 target-dependent @file{.h} and @file{.c} files used for your
5384 @chapter Releasing @value{GDBN}
5385 @cindex making a new release of gdb
5387 @section Versions and Branches
5389 @subsection Version Identifiers
5391 @value{GDBN}'s version is determined by the file @file{gdb/version.in}.
5393 @value{GDBN}'s mainline uses ISO dates to differentiate between
5394 versions. The CVS repository uses @var{YYYY}-@var{MM}-@var{DD}-cvs
5395 while the corresponding snapshot uses @var{YYYYMMDD}.
5397 @value{GDBN}'s release branch uses a slightly more complicated scheme.
5398 When the branch is first cut, the mainline version identifier is
5399 prefixed with the @var{major}.@var{minor} from of the previous release
5400 series but with .90 appended. As draft releases are drawn from the
5401 branch, the minor minor number (.90) is incremented. Once the first
5402 release (@var{M}.@var{N}) has been made, the version prefix is updated
5403 to @var{M}.@var{N}.0.90 (dot zero, dot ninety). Follow on releases have
5404 an incremented minor minor version number (.0).
5406 Using 5.1 (previous) and 5.2 (current), the example below illustrates a
5407 typical sequence of version identifiers:
5411 final release from previous branch
5412 @item 2002-03-03-cvs
5413 main-line the day the branch is cut
5414 @item 5.1.90-2002-03-03-cvs
5415 corresponding branch version
5417 first draft release candidate
5418 @item 5.1.91-2002-03-17-cvs
5419 updated branch version
5421 second draft release candidate
5422 @item 5.1.92-2002-03-31-cvs
5423 updated branch version
5425 final release candidate (see below)
5428 @item 5.2.0.90-2002-04-07-cvs
5429 updated CVS branch version
5431 second official release
5438 Minor minor minor draft release candidates such as 5.2.0.91 have been
5439 omitted from the example. Such release candidates are, typically, never
5442 For 5.1.93 the bziped tar ball @file{gdb-5.1.93.tar.bz2} is just the
5443 official @file{gdb-5.2.tar} renamed and compressed.
5446 To avoid version conflicts, vendors are expected to modify the file
5447 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5448 (an official @value{GDBN} release never uses alphabetic characters in
5449 its version identifer).
5451 Since @value{GDBN} does not make minor minor minor releases (e.g.,
5452 5.1.0.1) the conflict between that and a minor minor draft release
5453 identifier (e.g., 5.1.0.90) is avoided.
5456 @subsection Branches
5458 @value{GDBN} draws a release series (5.2, 5.2.1, @dots{}) from a single
5459 release branch (gdb_5_2-branch). Since minor minor minor releases
5460 (5.1.0.1) are not made, the need to branch the release branch is avoided
5461 (it also turns out that the effort required for such a a branch and
5462 release is significantly greater than the effort needed to create a new
5463 release from the head of the release branch).
5465 Releases 5.0 and 5.1 used branch and release tags of the form:
5468 gdb_N_M-YYYY-MM-DD-branchpoint
5469 gdb_N_M-YYYY-MM-DD-branch
5470 gdb_M_N-YYYY-MM-DD-release
5473 Release 5.2 is trialing the branch and release tags:
5476 gdb_N_M-YYYY-MM-DD-branchpoint
5478 gdb_M_N-YYYY-MM-DD-release
5481 @emph{Pragmatics: The branchpoint and release tags need to identify when
5482 a branch and release are made. The branch tag, denoting the head of the
5483 branch, does not have this criteria.}
5486 @section Branch Commit Policy
5488 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
5489 5.1 and 5.2 all used the below:
5493 The @file{gdb/MAINTAINERS} file still holds.
5495 Don't fix something on the branch unless/until it is also fixed in the
5496 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
5497 file is better than committing a hack.
5499 When considering a patch for the branch, suggested criteria include:
5500 Does it fix a build? Does it fix the sequence @kbd{break main; run}
5501 when debugging a static binary?
5503 The further a change is from the core of @value{GDBN}, the less likely
5504 the change will worry anyone (e.g., target specific code).
5506 Only post a proposal to change the core of @value{GDBN} after you've
5507 sent individual bribes to all the people listed in the
5508 @file{MAINTAINERS} file @t{;-)}
5511 @emph{Pragmatics: Provided updates are restricted to non-core
5512 functionality there is little chance that a broken change will be fatal.
5513 This means that changes such as adding a new architectures or (within
5514 reason) support for a new host are considered acceptable.}
5517 @section Obsoleting code
5519 Before anything else, poke the other developers (and around the source
5520 code) to see if there is anything that can be removed from @value{GDBN}
5521 (an old target, an unused file).
5523 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
5524 line. Doing this means that it is easy to identify something that has
5525 been obsoleted when greping through the sources.
5527 The process is done in stages --- this is mainly to ensure that the
5528 wider @value{GDBN} community has a reasonable opportunity to respond.
5529 Remember, everything on the Internet takes a week.
5533 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
5534 list} Creating a bug report to track the task's state, is also highly
5539 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
5540 Announcement mailing list}.
5544 Go through and edit all relevant files and lines so that they are
5545 prefixed with the word @code{OBSOLETE}.
5547 Wait until the next GDB version, containing this obsolete code, has been
5550 Remove the obsolete code.
5554 @emph{Maintainer note: While removing old code is regrettable it is
5555 hopefully better for @value{GDBN}'s long term development. Firstly it
5556 helps the developers by removing code that is either no longer relevant
5557 or simply wrong. Secondly since it removes any history associated with
5558 the file (effectively clearing the slate) the developer has a much freer
5559 hand when it comes to fixing broken files.}
5563 @section Before the Branch
5565 The most important objective at this stage is to find and fix simple
5566 changes that become a pain to track once the branch is created. For
5567 instance, configuration problems that stop @value{GDBN} from even
5568 building. If you can't get the problem fixed, document it in the
5569 @file{gdb/PROBLEMS} file.
5571 @subheading Prompt for @file{gdb/NEWS}
5573 People always forget. Send a post reminding them but also if you know
5574 something interesting happened add it yourself. The @code{schedule}
5575 script will mention this in its e-mail.
5577 @subheading Review @file{gdb/README}
5579 Grab one of the nightly snapshots and then walk through the
5580 @file{gdb/README} looking for anything that can be improved. The
5581 @code{schedule} script will mention this in its e-mail.
5583 @subheading Refresh any imported files.
5585 A number of files are taken from external repositories. They include:
5589 @file{texinfo/texinfo.tex}
5591 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
5594 @file{etc/standards.texi}, @file{etc/make-stds.texi}
5597 @subheading Check the ARI
5599 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
5600 (Awk Regression Index ;-) that checks for a number of errors and coding
5601 conventions. The checks include things like using @code{malloc} instead
5602 of @code{xmalloc} and file naming problems. There shouldn't be any
5605 @subsection Review the bug data base
5607 Close anything obviously fixed.
5609 @subsection Check all cross targets build
5611 The targets are listed in @file{gdb/MAINTAINERS}.
5614 @section Cut the Branch
5616 @subheading Create the branch
5621 $ V=`echo $v | sed 's/\./_/g'`
5622 $ D=`date -u +%Y-%m-%d`
5625 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5626 -D $D-gmt gdb_$V-$D-branchpoint insight+dejagnu
5627 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
5628 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight+dejagnu
5631 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5632 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight+dejagnu
5633 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5634 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight+dejagnu
5642 by using @kbd{-D YYYY-MM-DD-gmt} the branch is forced to an exact
5645 the trunk is first taged so that the branch point can easily be found
5647 Insight (which includes GDB) and dejagnu are all tagged at the same time
5649 @file{version.in} gets bumped to avoid version number conflicts
5651 the reading of @file{.cvsrc} is disabled using @file{-f}
5654 @subheading Update @file{version.in}
5659 $ V=`echo $v | sed 's/\./_/g'`
5663 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
5664 -r gdb_$V-branch src/gdb/version.in
5665 cvs -f -d :ext:sources.redhat.com:/cvs/src co
5666 -r gdb_5_2-branch src/gdb/version.in
5668 U src/gdb/version.in
5670 $ echo $u.90-0000-00-00-cvs > version.in
5672 5.1.90-0000-00-00-cvs
5673 $ cvs -f commit version.in
5678 @file{0000-00-00} is used as a date to pump prime the version.in update
5681 @file{.90} and the previous branch version are used as fairly arbitrary
5682 initial branch version number
5686 @subheading Update the web and news pages
5690 @subheading Tweak cron to track the new branch
5692 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
5693 This file needs to be updated so that:
5697 a daily timestamp is added to the file @file{version.in}
5699 the new branch is included in the snapshot process
5703 See the file @file{gdbadmin/cron/README} for how to install the updated
5706 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
5707 any changes. That file is copied to both the branch/ and current/
5708 snapshot directories.
5711 @subheading Update the NEWS and README files
5713 The @file{NEWS} file needs to be updated so that on the branch it refers
5714 to @emph{changes in the current release} while on the trunk it also
5715 refers to @emph{changes since the current release}.
5717 The @file{README} file needs to be updated so that it refers to the
5720 @subheading Post the branch info
5722 Send an announcement to the mailing lists:
5726 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
5728 @email{gdb@@sources.redhat.com, GDB Discsussion mailing list} and
5729 @email{gdb-testers@@sources.redhat.com, GDB Discsussion mailing list}
5732 @emph{Pragmatics: The branch creation is sent to the announce list to
5733 ensure that people people not subscribed to the higher volume discussion
5736 The announcement should include:
5742 how to check out the branch using CVS
5744 the date/number of weeks until the release
5746 the branch commit policy
5750 @section Stabilize the branch
5752 Something goes here.
5754 @section Create a Release
5756 The process of creating and then making available a release is broken
5757 down into a number of stages. The first part addresses the technical
5758 process of creating a releasable tar ball. The later stages address the
5759 process of releasing that tar ball.
5761 When making a release candidate just the first section is needed.
5763 @subsection Create a release candidate
5765 The objective at this stage is to create a set of tar balls that can be
5766 made available as a formal release (or as a less formal release
5769 @subsubheading Freeze the branch
5771 Send out an e-mail notifying everyone that the branch is frozen to
5772 @email{gdb-patches@@sources.redhat.com}.
5774 @subsubheading Establish a few defaults.
5779 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
5781 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
5785 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
5787 /home/gdbadmin/bin/autoconf
5796 Check the @code{autoconf} version carefully. You want to be using the
5797 version taken from the @file{binutils} snapshot directory, which can be
5798 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
5799 unlikely that a system installed version of @code{autoconf} (e.g.,
5800 @file{/usr/bin/autoconf}) is correct.
5803 @subsubheading Check out the relevant modules:
5806 $ for m in gdb insight dejagnu
5808 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
5818 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
5819 any confusion between what is written here and what your local
5820 @code{cvs} really does.
5823 @subsubheading Update relevant files.
5829 Major releases get their comments added as part of the mainline. Minor
5830 releases should probably mention any significant bugs that were fixed.
5832 Don't forget to include the @file{ChangeLog} entry.
5835 $ emacs gdb/src/gdb/NEWS
5840 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
5841 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5846 You'll need to update:
5858 $ emacs gdb/src/gdb/README
5863 $ cp gdb/src/gdb/README insight/src/gdb/README
5864 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5867 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
5868 before the initial branch was cut so just a simple substitute is needed
5871 @emph{Maintainer note: Other projects generate @file{README} and
5872 @file{INSTALL} from the core documentation. This might be worth
5875 @item gdb/version.in
5878 $ echo $v > gdb/src/gdb/version.in
5879 $ cat gdb/src/gdb/version.in
5881 $ emacs gdb/src/gdb/version.in
5884 ... Bump to version ...
5886 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
5887 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5890 @item dejagnu/src/dejagnu/configure.in
5892 Dejagnu is more complicated. The version number is a parameter to
5893 @code{AM_INIT_AUTOMAKE}. Tweak it to read something like gdb-5.1.91.
5895 Don't forget to re-generate @file{configure}.
5897 Don't forget to include a @file{ChangeLog} entry.
5900 $ emacs dejagnu/src/dejagnu/configure.in
5905 $ ( cd dejagnu/src/dejagnu && autoconf )
5910 @subsubheading Do the dirty work
5912 This is identical to the process used to create the daily snapshot.
5915 $ for m in gdb insight
5917 ( cd $m/src && gmake -f src-release $m.tar )
5919 $ ( m=dejagnu; cd $m/src && gmake -f src-release $m.tar.bz2 )
5922 If the top level source directory does not have @file{src-release}
5923 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
5926 $ for m in gdb insight
5928 ( cd $m/src && gmake -f Makefile.in $m.tar )
5930 $ ( m=dejagnu; cd $m/src && gmake -f Makefile.in $m.tar.bz2 )
5933 @subsubheading Check the source files
5935 You're looking for files that have mysteriously disappeared.
5936 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
5937 for the @file{version.in} update @kbd{cronjob}.
5940 $ ( cd gdb/src && cvs -f -q -n update )
5944 @dots{} lots of generated files @dots{}
5949 @dots{} lots of generated files @dots{}
5954 @emph{Don't worry about the @file{gdb.info-??} or
5955 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
5956 was also generated only something strange with CVS means that they
5957 didn't get supressed). Fixing it would be nice though.}
5959 @subsubheading Create compressed versions of the release
5965 dejagnu/ dejagnu-gdb-5.2.tar.bz2 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
5966 $ for m in gdb insight
5968 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
5969 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
5979 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
5980 in that mode, @code{gzip} does not know the name of the file and, hence,
5981 can not include it in the compressed file. This is also why the release
5982 process runs @code{tar} and @code{bzip2} as separate passes.
5985 @subsection Sanity check the tar ball
5987 Pick a popular machine (Solaris/PPC?) and try the build on that.
5990 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
5995 $ ./gdb/gdb ./gdb/gdb
5999 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6001 Starting program: /tmp/gdb-5.2/gdb/gdb
6003 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6004 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6006 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6010 @subsection Make a release candidate available
6012 If this is a release candidate then the only remaining steps are:
6016 Commit @file{version.in} and @file{ChangeLog}
6018 Tweak @file{version.in} (and @file{ChangeLog} to read
6019 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6020 process can restart.
6022 Make the release candidate available in
6023 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6025 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6026 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6029 @subsection Make a formal release available
6031 (And you thought all that was required was to post an e-mail.)
6033 @subsubheading Install on sware
6035 Copy the new files to both the release and the old release directory:
6038 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6039 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6043 Clean up the releases directory so that only the most recent releases
6044 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6047 $ cd ~ftp/pub/gdb/releases
6052 Update the file @file{README} and @file{.message} in the releases
6059 $ ln README .message
6062 @subsubheading Update the web pages.
6066 @item htdocs/download/ANNOUNCEMENT
6067 This file, which is posted as the official announcement, includes:
6070 General announcement
6072 News. If making an @var{M}.@var{N}.1 release, retain the news from
6073 earlier @var{M}.@var{N} release.
6078 @item htdocs/index.html
6079 @itemx htdocs/news/index.html
6080 @itemx htdocs/download/index.html
6081 These files include:
6084 announcement of the most recent release
6086 news entry (remember to update both the top level and the news directory).
6088 These pages also need to be regenerate using @code{index.sh}.
6090 @item download/onlinedocs/
6091 You need to find the magic command that is used to generate the online
6092 docs from the @file{.tar.bz2}. The best way is to look in the output
6093 from one of the nightly @code{cron} jobs and then just edit accordingly.
6097 $ ~/ss/update-web-docs \
6098 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6100 /www/sourceware/htdocs/gdb/download/onlinedocs \
6105 Just like the online documentation. Something like:
6108 $ /bin/sh ~/ss/update-web-ari \
6109 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6111 /www/sourceware/htdocs/gdb/download/ari \
6117 @subsubheading Shadow the pages onto gnu
6119 Something goes here.
6122 @subsubheading Install the @value{GDBN} tar ball on GNU
6124 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6125 @file{~ftp/gnu/gdb}.
6127 @subsubheading Make the @file{ANNOUNCEMENT}
6129 Post the @file{ANNOUNCEMENT} file you created above to:
6133 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6135 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6136 day or so to let things get out)
6138 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6143 The release is out but you're still not finished.
6145 @subsubheading Commit outstanding changes
6147 In particular you'll need to commit any changes to:
6151 @file{gdb/ChangeLog}
6153 @file{gdb/version.in}
6160 @subsubheading Tag the release
6165 $ d=`date -u +%Y-%m-%d`
6168 $ ( cd insight/src/gdb && cvs -f -q update )
6169 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6172 Insight is used since that contains more of the release than
6173 @value{GDBN} (@code{dejagnu} doesn't get tagged but I think we can live
6176 @subsubheading Mention the release on the trunk
6178 Just put something in the @file{ChangeLog} so that the trunk also
6179 indicates when the release was made.
6181 @subsubheading Restart @file{gdb/version.in}
6183 If @file{gdb/version.in} does not contain an ISO date such as
6184 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6185 committed all the release changes it can be set to
6186 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6187 is important - it affects the snapshot process).
6189 Don't forget the @file{ChangeLog}.
6191 @subsubheading Merge into trunk
6193 The files committed to the branch may also need changes merged into the
6196 @subsubheading Revise the release schedule
6198 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6199 Discussion List} with an updated announcement. The schedule can be
6200 generated by running:
6203 $ ~/ss/schedule `date +%s` schedule
6207 The first parameter is approximate date/time in seconds (from the epoch)
6208 of the most recent release.
6210 Also update the schedule @code{cronjob}.
6212 @section Post release
6214 Remove any @code{OBSOLETE} code.
6221 The testsuite is an important component of the @value{GDBN} package.
6222 While it is always worthwhile to encourage user testing, in practice
6223 this is rarely sufficient; users typically use only a small subset of
6224 the available commands, and it has proven all too common for a change
6225 to cause a significant regression that went unnoticed for some time.
6227 The @value{GDBN} testsuite uses the DejaGNU testing framework.
6228 DejaGNU is built using @code{Tcl} and @code{expect}. The tests
6229 themselves are calls to various @code{Tcl} procs; the framework runs all the
6230 procs and summarizes the passes and fails.
6232 @section Using the Testsuite
6234 @cindex running the test suite
6235 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6236 testsuite's objdir) and type @code{make check}. This just sets up some
6237 environment variables and invokes DejaGNU's @code{runtest} script. While
6238 the testsuite is running, you'll get mentions of which test file is in use,
6239 and a mention of any unexpected passes or fails. When the testsuite is
6240 finished, you'll get a summary that looks like this:
6245 # of expected passes 6016
6246 # of unexpected failures 58
6247 # of unexpected successes 5
6248 # of expected failures 183
6249 # of unresolved testcases 3
6250 # of untested testcases 5
6253 The ideal test run consists of expected passes only; however, reality
6254 conspires to keep us from this ideal. Unexpected failures indicate
6255 real problems, whether in @value{GDBN} or in the testsuite. Expected
6256 failures are still failures, but ones which have been decided are too
6257 hard to deal with at the time; for instance, a test case might work
6258 everywhere except on AIX, and there is no prospect of the AIX case
6259 being fixed in the near future. Expected failures should not be added
6260 lightly, since you may be masking serious bugs in @value{GDBN}.
6261 Unexpected successes are expected fails that are passing for some
6262 reason, while unresolved and untested cases often indicate some minor
6263 catastrophe, such as the compiler being unable to deal with a test
6266 When making any significant change to @value{GDBN}, you should run the
6267 testsuite before and after the change, to confirm that there are no
6268 regressions. Note that truly complete testing would require that you
6269 run the testsuite with all supported configurations and a variety of
6270 compilers; however this is more than really necessary. In many cases
6271 testing with a single configuration is sufficient. Other useful
6272 options are to test one big-endian (Sparc) and one little-endian (x86)
6273 host, a cross config with a builtin simulator (powerpc-eabi,
6274 mips-elf), or a 64-bit host (Alpha).
6276 If you add new functionality to @value{GDBN}, please consider adding
6277 tests for it as well; this way future @value{GDBN} hackers can detect
6278 and fix their changes that break the functionality you added.
6279 Similarly, if you fix a bug that was not previously reported as a test
6280 failure, please add a test case for it. Some cases are extremely
6281 difficult to test, such as code that handles host OS failures or bugs
6282 in particular versions of compilers, and it's OK not to try to write
6283 tests for all of those.
6285 DejaGNU supports separate build, host, and target machines. However,
6286 some @value{GDBN} test scripts do not work if the build machine and
6287 the host machine are not the same. In such an environment, these scripts
6288 will give a result of ``UNRESOLVED'', like this:
6291 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6294 @section Testsuite Organization
6296 @cindex test suite organization
6297 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6298 testsuite includes some makefiles and configury, these are very minimal,
6299 and used for little besides cleaning up, since the tests themselves
6300 handle the compilation of the programs that @value{GDBN} will run. The file
6301 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6302 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6303 configuration-specific files, typically used for special-purpose
6304 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6306 The tests themselves are to be found in @file{testsuite/gdb.*} and
6307 subdirectories of those. The names of the test files must always end
6308 with @file{.exp}. DejaGNU collects the test files by wildcarding
6309 in the test directories, so both subdirectories and individual files
6310 get chosen and run in alphabetical order.
6312 The following table lists the main types of subdirectories and what they
6313 are for. Since DejaGNU finds test files no matter where they are
6314 located, and since each test file sets up its own compilation and
6315 execution environment, this organization is simply for convenience and
6320 This is the base testsuite. The tests in it should apply to all
6321 configurations of @value{GDBN} (but generic native-only tests may live here).
6322 The test programs should be in the subset of C that is valid K&R,
6323 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6326 @item gdb.@var{lang}
6327 Language-specific tests for any language @var{lang} besides C. Examples are
6328 @file{gdb.cp} and @file{gdb.java}.
6330 @item gdb.@var{platform}
6331 Non-portable tests. The tests are specific to a specific configuration
6332 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6335 @item gdb.@var{compiler}
6336 Tests specific to a particular compiler. As of this writing (June
6337 1999), there aren't currently any groups of tests in this category that
6338 couldn't just as sensibly be made platform-specific, but one could
6339 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6342 @item gdb.@var{subsystem}
6343 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6344 instance, @file{gdb.disasm} exercises various disassemblers, while
6345 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6348 @section Writing Tests
6349 @cindex writing tests
6351 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6352 should be able to copy existing tests to handle new cases.
6354 You should try to use @code{gdb_test} whenever possible, since it
6355 includes cases to handle all the unexpected errors that might happen.
6356 However, it doesn't cost anything to add new test procedures; for
6357 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6358 calls @code{gdb_test} multiple times.
6360 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6361 necessary, such as when @value{GDBN} has several valid responses to a command.
6363 The source language programs do @emph{not} need to be in a consistent
6364 style. Since @value{GDBN} is used to debug programs written in many different
6365 styles, it's worth having a mix of styles in the testsuite; for
6366 instance, some @value{GDBN} bugs involving the display of source lines would
6367 never manifest themselves if the programs used GNU coding style
6374 Check the @file{README} file, it often has useful information that does not
6375 appear anywhere else in the directory.
6378 * Getting Started:: Getting started working on @value{GDBN}
6379 * Debugging GDB:: Debugging @value{GDBN} with itself
6382 @node Getting Started,,, Hints
6384 @section Getting Started
6386 @value{GDBN} is a large and complicated program, and if you first starting to
6387 work on it, it can be hard to know where to start. Fortunately, if you
6388 know how to go about it, there are ways to figure out what is going on.
6390 This manual, the @value{GDBN} Internals manual, has information which applies
6391 generally to many parts of @value{GDBN}.
6393 Information about particular functions or data structures are located in
6394 comments with those functions or data structures. If you run across a
6395 function or a global variable which does not have a comment correctly
6396 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6397 free to submit a bug report, with a suggested comment if you can figure
6398 out what the comment should say. If you find a comment which is
6399 actually wrong, be especially sure to report that.
6401 Comments explaining the function of macros defined in host, target, or
6402 native dependent files can be in several places. Sometimes they are
6403 repeated every place the macro is defined. Sometimes they are where the
6404 macro is used. Sometimes there is a header file which supplies a
6405 default definition of the macro, and the comment is there. This manual
6406 also documents all the available macros.
6407 @c (@pxref{Host Conditionals}, @pxref{Target
6408 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
6411 Start with the header files. Once you have some idea of how
6412 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
6413 @file{gdbtypes.h}), you will find it much easier to understand the
6414 code which uses and creates those symbol tables.
6416 You may wish to process the information you are getting somehow, to
6417 enhance your understanding of it. Summarize it, translate it to another
6418 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
6419 the code to predict what a test case would do and write the test case
6420 and verify your prediction, etc. If you are reading code and your eyes
6421 are starting to glaze over, this is a sign you need to use a more active
6424 Once you have a part of @value{GDBN} to start with, you can find more
6425 specifically the part you are looking for by stepping through each
6426 function with the @code{next} command. Do not use @code{step} or you
6427 will quickly get distracted; when the function you are stepping through
6428 calls another function try only to get a big-picture understanding
6429 (perhaps using the comment at the beginning of the function being
6430 called) of what it does. This way you can identify which of the
6431 functions being called by the function you are stepping through is the
6432 one which you are interested in. You may need to examine the data
6433 structures generated at each stage, with reference to the comments in
6434 the header files explaining what the data structures are supposed to
6437 Of course, this same technique can be used if you are just reading the
6438 code, rather than actually stepping through it. The same general
6439 principle applies---when the code you are looking at calls something
6440 else, just try to understand generally what the code being called does,
6441 rather than worrying about all its details.
6443 @cindex command implementation
6444 A good place to start when tracking down some particular area is with
6445 a command which invokes that feature. Suppose you want to know how
6446 single-stepping works. As a @value{GDBN} user, you know that the
6447 @code{step} command invokes single-stepping. The command is invoked
6448 via command tables (see @file{command.h}); by convention the function
6449 which actually performs the command is formed by taking the name of
6450 the command and adding @samp{_command}, or in the case of an
6451 @code{info} subcommand, @samp{_info}. For example, the @code{step}
6452 command invokes the @code{step_command} function and the @code{info
6453 display} command invokes @code{display_info}. When this convention is
6454 not followed, you might have to use @code{grep} or @kbd{M-x
6455 tags-search} in emacs, or run @value{GDBN} on itself and set a
6456 breakpoint in @code{execute_command}.
6458 @cindex @code{bug-gdb} mailing list
6459 If all of the above fail, it may be appropriate to ask for information
6460 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
6461 wondering if anyone could give me some tips about understanding
6462 @value{GDBN}''---if we had some magic secret we would put it in this manual.
6463 Suggestions for improving the manual are always welcome, of course.
6465 @node Debugging GDB,,,Hints
6467 @section Debugging @value{GDBN} with itself
6468 @cindex debugging @value{GDBN}
6470 If @value{GDBN} is limping on your machine, this is the preferred way to get it
6471 fully functional. Be warned that in some ancient Unix systems, like
6472 Ultrix 4.2, a program can't be running in one process while it is being
6473 debugged in another. Rather than typing the command @kbd{@w{./gdb
6474 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
6475 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
6477 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
6478 @file{.gdbinit} file that sets up some simple things to make debugging
6479 gdb easier. The @code{info} command, when executed without a subcommand
6480 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
6481 gdb. See @file{.gdbinit} for details.
6483 If you use emacs, you will probably want to do a @code{make TAGS} after
6484 you configure your distribution; this will put the machine dependent
6485 routines for your local machine where they will be accessed first by
6488 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
6489 have run @code{fixincludes} if you are compiling with gcc.
6491 @section Submitting Patches
6493 @cindex submitting patches
6494 Thanks for thinking of offering your changes back to the community of
6495 @value{GDBN} users. In general we like to get well designed enhancements.
6496 Thanks also for checking in advance about the best way to transfer the
6499 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
6500 This manual summarizes what we believe to be clean design for @value{GDBN}.
6502 If the maintainers don't have time to put the patch in when it arrives,
6503 or if there is any question about a patch, it goes into a large queue
6504 with everyone else's patches and bug reports.
6506 @cindex legal papers for code contributions
6507 The legal issue is that to incorporate substantial changes requires a
6508 copyright assignment from you and/or your employer, granting ownership
6509 of the changes to the Free Software Foundation. You can get the
6510 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
6511 and asking for it. We recommend that people write in "All programs
6512 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
6513 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
6515 contributed with only one piece of legalese pushed through the
6516 bureaucracy and filed with the FSF. We can't start merging changes until
6517 this paperwork is received by the FSF (their rules, which we follow
6518 since we maintain it for them).
6520 Technically, the easiest way to receive changes is to receive each
6521 feature as a small context diff or unidiff, suitable for @code{patch}.
6522 Each message sent to me should include the changes to C code and
6523 header files for a single feature, plus @file{ChangeLog} entries for
6524 each directory where files were modified, and diffs for any changes
6525 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
6526 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
6527 single feature, they can be split down into multiple messages.
6529 In this way, if we read and like the feature, we can add it to the
6530 sources with a single patch command, do some testing, and check it in.
6531 If you leave out the @file{ChangeLog}, we have to write one. If you leave
6532 out the doc, we have to puzzle out what needs documenting. Etc., etc.
6534 The reason to send each change in a separate message is that we will not
6535 install some of the changes. They'll be returned to you with questions
6536 or comments. If we're doing our job correctly, the message back to you
6537 will say what you have to fix in order to make the change acceptable.
6538 The reason to have separate messages for separate features is so that
6539 the acceptable changes can be installed while one or more changes are
6540 being reworked. If multiple features are sent in a single message, we
6541 tend to not put in the effort to sort out the acceptable changes from
6542 the unacceptable, so none of the features get installed until all are
6545 If this sounds painful or authoritarian, well, it is. But we get a lot
6546 of bug reports and a lot of patches, and many of them don't get
6547 installed because we don't have the time to finish the job that the bug
6548 reporter or the contributor could have done. Patches that arrive
6549 complete, working, and well designed, tend to get installed on the day
6550 they arrive. The others go into a queue and get installed as time
6551 permits, which, since the maintainers have many demands to meet, may not
6552 be for quite some time.
6554 Please send patches directly to
6555 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
6557 @section Obsolete Conditionals
6558 @cindex obsolete code
6560 Fragments of old code in @value{GDBN} sometimes reference or set the following
6561 configuration macros. They should not be used by new code, and old uses
6562 should be removed as those parts of the debugger are otherwise touched.
6565 @item STACK_END_ADDR
6566 This macro used to define where the end of the stack appeared, for use
6567 in interpreting core file formats that don't record this address in the
6568 core file itself. This information is now configured in BFD, and @value{GDBN}
6569 gets the info portably from there. The values in @value{GDBN}'s configuration
6570 files should be moved into BFD configuration files (if needed there),
6571 and deleted from all of @value{GDBN}'s config files.
6573 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
6574 is so old that it has never been converted to use BFD. Now that's old!
6578 @include observer.texi