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::
87 * Versions and Branches::
92 * GDB Observers:: @value{GDBN} Currently available observers
93 * GNU Free Documentation License:: The license for this documentation
100 @cindex requirements for @value{GDBN}
102 Before diving into the internals, you should understand the formal
103 requirements and other expectations for @value{GDBN}. Although some
104 of these may seem obvious, there have been proposals for @value{GDBN}
105 that have run counter to these requirements.
107 First of all, @value{GDBN} is a debugger. It's not designed to be a
108 front panel for embedded systems. It's not a text editor. It's not a
109 shell. It's not a programming environment.
111 @value{GDBN} is an interactive tool. Although a batch mode is
112 available, @value{GDBN}'s primary role is to interact with a human
115 @value{GDBN} should be responsive to the user. A programmer hot on
116 the trail of a nasty bug, and operating under a looming deadline, is
117 going to be very impatient of everything, including the response time
118 to debugger commands.
120 @value{GDBN} should be relatively permissive, such as for expressions.
121 While the compiler should be picky (or have the option to be made
122 picky), since source code lives for a long time usually, the
123 programmer doing debugging shouldn't be spending time figuring out to
124 mollify the debugger.
126 @value{GDBN} will be called upon to deal with really large programs.
127 Executable sizes of 50 to 100 megabytes occur regularly, and we've
128 heard reports of programs approaching 1 gigabyte in size.
130 @value{GDBN} should be able to run everywhere. No other debugger is
131 available for even half as many configurations as @value{GDBN}
135 @node Overall Structure
137 @chapter Overall Structure
139 @value{GDBN} consists of three major subsystems: user interface,
140 symbol handling (the @dfn{symbol side}), and target system handling (the
143 The user interface consists of several actual interfaces, plus
146 The symbol side consists of object file readers, debugging info
147 interpreters, symbol table management, source language expression
148 parsing, type and value printing.
150 The target side consists of execution control, stack frame analysis, and
151 physical target manipulation.
153 The target side/symbol side division is not formal, and there are a
154 number of exceptions. For instance, core file support involves symbolic
155 elements (the basic core file reader is in BFD) and target elements (it
156 supplies the contents of memory and the values of registers). Instead,
157 this division is useful for understanding how the minor subsystems
160 @section The Symbol Side
162 The symbolic side of @value{GDBN} can be thought of as ``everything
163 you can do in @value{GDBN} without having a live program running''.
164 For instance, you can look at the types of variables, and evaluate
165 many kinds of expressions.
167 @section The Target Side
169 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
170 Although it may make reference to symbolic info here and there, most
171 of the target side will run with only a stripped executable
172 available---or even no executable at all, in remote debugging cases.
174 Operations such as disassembly, stack frame crawls, and register
175 display, are able to work with no symbolic info at all. In some cases,
176 such as disassembly, @value{GDBN} will use symbolic info to present addresses
177 relative to symbols rather than as raw numbers, but it will work either
180 @section Configurations
184 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
185 @dfn{Target} refers to the system where the program being debugged
186 executes. In most cases they are the same machine, in which case a
187 third type of @dfn{Native} attributes come into play.
189 Defines and include files needed to build on the host are host support.
190 Examples are tty support, system defined types, host byte order, host
193 Defines and information needed to handle the target format are target
194 dependent. Examples are the stack frame format, instruction set,
195 breakpoint instruction, registers, and how to set up and tear down the stack
198 Information that is only needed when the host and target are the same,
199 is native dependent. One example is Unix child process support; if the
200 host and target are not the same, doing a fork to start the target
201 process is a bad idea. The various macros needed for finding the
202 registers in the @code{upage}, running @code{ptrace}, and such are all
203 in the native-dependent files.
205 Another example of native-dependent code is support for features that
206 are really part of the target environment, but which require
207 @code{#include} files that are only available on the host system. Core
208 file handling and @code{setjmp} handling are two common cases.
210 When you want to make @value{GDBN} work ``native'' on a particular machine, you
211 have to include all three kinds of information.
219 @value{GDBN} uses a number of debugging-specific algorithms. They are
220 often not very complicated, but get lost in the thicket of special
221 cases and real-world issues. This chapter describes the basic
222 algorithms and mentions some of the specific target definitions that
228 @cindex call stack frame
229 A frame is a construct that @value{GDBN} uses to keep track of calling
230 and called functions.
232 @findex create_new_frame
234 @code{FRAME_FP} in the machine description has no meaning to the
235 machine-independent part of @value{GDBN}, except that it is used when
236 setting up a new frame from scratch, as follows:
239 create_new_frame (read_register (DEPRECATED_FP_REGNUM), read_pc ()));
242 @cindex frame pointer register
243 Other than that, all the meaning imparted to @code{DEPRECATED_FP_REGNUM}
244 is imparted by the machine-dependent code. So,
245 @code{DEPRECATED_FP_REGNUM} can have any value that is convenient for
246 the code that creates new frames. (@code{create_new_frame} calls
247 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} if it is defined; that is where
248 you should use the @code{DEPRECATED_FP_REGNUM} value, if your frames are
252 Given a @value{GDBN} frame, define @code{DEPRECATED_FRAME_CHAIN} to
253 determine the address of the calling function's frame. This will be
254 used to create a new @value{GDBN} frame struct, and then
255 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} and
256 @code{DEPRECATED_INIT_FRAME_PC} will be called for the new frame.
258 @section Breakpoint Handling
261 In general, a breakpoint is a user-designated location in the program
262 where the user wants to regain control if program execution ever reaches
265 There are two main ways to implement breakpoints; either as ``hardware''
266 breakpoints or as ``software'' breakpoints.
268 @cindex hardware breakpoints
269 @cindex program counter
270 Hardware breakpoints are sometimes available as a builtin debugging
271 features with some chips. Typically these work by having dedicated
272 register into which the breakpoint address may be stored. If the PC
273 (shorthand for @dfn{program counter})
274 ever matches a value in a breakpoint registers, the CPU raises an
275 exception and reports it to @value{GDBN}.
277 Another possibility is when an emulator is in use; many emulators
278 include circuitry that watches the address lines coming out from the
279 processor, and force it to stop if the address matches a breakpoint's
282 A third possibility is that the target already has the ability to do
283 breakpoints somehow; for instance, a ROM monitor may do its own
284 software breakpoints. So although these are not literally ``hardware
285 breakpoints'', from @value{GDBN}'s point of view they work the same;
286 @value{GDBN} need not do anything more than set the breakpoint and wait
287 for something to happen.
289 Since they depend on hardware resources, hardware breakpoints may be
290 limited in number; when the user asks for more, @value{GDBN} will
291 start trying to set software breakpoints. (On some architectures,
292 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
293 whether there's enough hardware resources to insert all the hardware
294 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
295 an error message only when the program being debugged is continued.)
297 @cindex software breakpoints
298 Software breakpoints require @value{GDBN} to do somewhat more work.
299 The basic theory is that @value{GDBN} will replace a program
300 instruction with a trap, illegal divide, or some other instruction
301 that will cause an exception, and then when it's encountered,
302 @value{GDBN} will take the exception and stop the program. When the
303 user says to continue, @value{GDBN} will restore the original
304 instruction, single-step, re-insert the trap, and continue on.
306 Since it literally overwrites the program being tested, the program area
307 must be writable, so this technique won't work on programs in ROM. It
308 can also distort the behavior of programs that examine themselves,
309 although such a situation would be highly unusual.
311 Also, the software breakpoint instruction should be the smallest size of
312 instruction, so it doesn't overwrite an instruction that might be a jump
313 target, and cause disaster when the program jumps into the middle of the
314 breakpoint instruction. (Strictly speaking, the breakpoint must be no
315 larger than the smallest interval between instructions that may be jump
316 targets; perhaps there is an architecture where only even-numbered
317 instructions may jumped to.) Note that it's possible for an instruction
318 set not to have any instructions usable for a software breakpoint,
319 although in practice only the ARC has failed to define such an
323 The basic definition of the software breakpoint is the macro
326 Basic breakpoint object handling is in @file{breakpoint.c}. However,
327 much of the interesting breakpoint action is in @file{infrun.c}.
329 @section Single Stepping
331 @section Signal Handling
333 @section Thread Handling
335 @section Inferior Function Calls
337 @section Longjmp Support
339 @cindex @code{longjmp} debugging
340 @value{GDBN} has support for figuring out that the target is doing a
341 @code{longjmp} and for stopping at the target of the jump, if we are
342 stepping. This is done with a few specialized internal breakpoints,
343 which are visible in the output of the @samp{maint info breakpoint}
346 @findex GET_LONGJMP_TARGET
347 To make this work, you need to define a macro called
348 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
349 structure and extract the longjmp target address. Since @code{jmp_buf}
350 is target specific, you will need to define it in the appropriate
351 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
352 @file{sparc-tdep.c} for examples of how to do this.
357 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
358 breakpoints}) which break when data is accessed rather than when some
359 instruction is executed. When you have data which changes without
360 your knowing what code does that, watchpoints are the silver bullet to
361 hunt down and kill such bugs.
363 @cindex hardware watchpoints
364 @cindex software watchpoints
365 Watchpoints can be either hardware-assisted or not; the latter type is
366 known as ``software watchpoints.'' @value{GDBN} always uses
367 hardware-assisted watchpoints if they are available, and falls back on
368 software watchpoints otherwise. Typical situations where @value{GDBN}
369 will use software watchpoints are:
373 The watched memory region is too large for the underlying hardware
374 watchpoint support. For example, each x86 debug register can watch up
375 to 4 bytes of memory, so trying to watch data structures whose size is
376 more than 16 bytes will cause @value{GDBN} to use software
380 The value of the expression to be watched depends on data held in
381 registers (as opposed to memory).
384 Too many different watchpoints requested. (On some architectures,
385 this situation is impossible to detect until the debugged program is
386 resumed.) Note that x86 debug registers are used both for hardware
387 breakpoints and for watchpoints, so setting too many hardware
388 breakpoints might cause watchpoint insertion to fail.
391 No hardware-assisted watchpoints provided by the target
395 Software watchpoints are very slow, since @value{GDBN} needs to
396 single-step the program being debugged and test the value of the
397 watched expression(s) after each instruction. The rest of this
398 section is mostly irrelevant for software watchpoints.
400 When the inferior stops, @value{GDBN} tries to establish, among other
401 possible reasons, whether it stopped due to a watchpoint being hit.
402 For a data-write watchpoint, it does so by evaluating, for each
403 watchpoint, the expression whose value is being watched, and testing
404 whether the watched value has changed. For data-read and data-access
405 watchpoints, @value{GDBN} needs the target to supply a primitive that
406 returns the address of the data that was accessed or read (see the
407 description of @code{target_stopped_data_address} below): if this
408 primitive returns a valid address, @value{GDBN} infers that a
409 watchpoint triggered if it watches an expression whose evaluation uses
412 @value{GDBN} uses several macros and primitives to support hardware
416 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
417 @item TARGET_HAS_HARDWARE_WATCHPOINTS
418 If defined, the target supports hardware watchpoints.
420 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
421 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
422 Return the number of hardware watchpoints of type @var{type} that are
423 possible to be set. The value is positive if @var{count} watchpoints
424 of this type can be set, zero if setting watchpoints of this type is
425 not supported, and negative if @var{count} is more than the maximum
426 number of watchpoints of type @var{type} that can be set. @var{other}
427 is non-zero if other types of watchpoints are currently enabled (there
428 are architectures which cannot set watchpoints of different types at
431 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
432 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
433 Return non-zero if hardware watchpoints can be used to watch a region
434 whose address is @var{addr} and whose length in bytes is @var{len}.
436 @findex TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT
437 @item TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT (@var{size})
438 Return non-zero if hardware watchpoints can be used to watch a region
439 whose size is @var{size}. @value{GDBN} only uses this macro as a
440 fall-back, in case @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is not
443 @findex TARGET_DISABLE_HW_WATCHPOINTS
444 @item TARGET_DISABLE_HW_WATCHPOINTS (@var{pid})
445 Disables watchpoints in the process identified by @var{pid}. This is
446 used, e.g., on HPPA-RISC machines running HP-UX, which provide
447 operations to disable and enable the page-level memory protection that
448 implements hardware watchpoints on that platform.
450 @findex TARGET_ENABLE_HW_WATCHPOINTS
451 @item TARGET_ENABLE_HW_WATCHPOINTS (@var{pid})
452 Enables watchpoints in the process identified by @var{pid}. This is
453 used, e.g., on HPPA-RISC machines running HP-UX, which provide
454 operations to disable and enable the page-level memory protection that
455 implements hardware watchpoints on that platform.
457 @cindex insert or remove hardware watchpoint
458 @findex target_insert_watchpoint
459 @findex target_remove_watchpoint
460 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
461 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
462 Insert or remove a hardware watchpoint starting at @var{addr}, for
463 @var{len} bytes. @var{type} is the watchpoint type, one of the
464 possible values of the enumerated data type @code{target_hw_bp_type},
465 defined by @file{breakpoint.h} as follows:
468 enum target_hw_bp_type
470 hw_write = 0, /* Common (write) HW watchpoint */
471 hw_read = 1, /* Read HW watchpoint */
472 hw_access = 2, /* Access (read or write) HW watchpoint */
473 hw_execute = 3 /* Execute HW breakpoint */
478 These two macros should return 0 for success, non-zero for failure.
480 @cindex insert or remove hardware breakpoint
481 @findex target_remove_hw_breakpoint
482 @findex target_insert_hw_breakpoint
483 @item target_remove_hw_breakpoint (@var{addr}, @var{shadow})
484 @itemx target_insert_hw_breakpoint (@var{addr}, @var{shadow})
485 Insert or remove a hardware-assisted breakpoint at address @var{addr}.
486 Returns zero for success, non-zero for failure. @var{shadow} is the
487 real contents of the byte where the breakpoint has been inserted; it
488 is generally not valid when hardware breakpoints are used, but since
489 no other code touches these values, the implementations of the above
490 two macros can use them for their internal purposes.
492 @findex target_stopped_data_address
493 @item target_stopped_data_address (@var{addr_p})
494 If the inferior has some watchpoint that triggered, place the address
495 associated with the watchpoint at the location pointed to by
496 @var{addr_p} and return non-zero. Otherwise, return zero. Note that
497 this primitive is used by @value{GDBN} only on targets that support
498 data-read or data-access type watchpoints, so targets that have
499 support only for data-write watchpoints need not implement these
502 @findex HAVE_STEPPABLE_WATCHPOINT
503 @item HAVE_STEPPABLE_WATCHPOINT
504 If defined to a non-zero value, it is not necessary to disable a
505 watchpoint to step over it.
507 @findex HAVE_NONSTEPPABLE_WATCHPOINT
508 @item HAVE_NONSTEPPABLE_WATCHPOINT
509 If defined to a non-zero value, @value{GDBN} should disable a
510 watchpoint to step the inferior over it.
512 @findex HAVE_CONTINUABLE_WATCHPOINT
513 @item HAVE_CONTINUABLE_WATCHPOINT
514 If defined to a non-zero value, it is possible to continue the
515 inferior after a watchpoint has been hit.
517 @findex CANNOT_STEP_HW_WATCHPOINTS
518 @item CANNOT_STEP_HW_WATCHPOINTS
519 If this is defined to a non-zero value, @value{GDBN} will remove all
520 watchpoints before stepping the inferior.
522 @findex STOPPED_BY_WATCHPOINT
523 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
524 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
525 the type @code{struct target_waitstatus}, defined by @file{target.h}.
526 Normally, this macro is defined to invoke the function pointed to by
527 the @code{to_stopped_by_watchpoint} member of the structure (of the
528 type @code{target_ops}, defined on @file{target.h}) that describes the
529 target-specific operations; @code{to_stopped_by_watchpoint} ignores
530 the @var{wait_status} argument.
532 @value{GDBN} does not require the non-zero value returned by
533 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
534 determine for sure whether the inferior stopped due to a watchpoint,
535 it could return non-zero ``just in case''.
538 @subsection x86 Watchpoints
539 @cindex x86 debug registers
540 @cindex watchpoints, on x86
542 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
543 registers designed to facilitate debugging. @value{GDBN} provides a
544 generic library of functions that x86-based ports can use to implement
545 support for watchpoints and hardware-assisted breakpoints. This
546 subsection documents the x86 watchpoint facilities in @value{GDBN}.
548 To use the generic x86 watchpoint support, a port should do the
552 @findex I386_USE_GENERIC_WATCHPOINTS
554 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
555 target-dependent headers.
558 Include the @file{config/i386/nm-i386.h} header file @emph{after}
559 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
562 Add @file{i386-nat.o} to the value of the Make variable
563 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
564 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
567 Provide implementations for the @code{I386_DR_LOW_*} macros described
568 below. Typically, each macro should call a target-specific function
569 which does the real work.
572 The x86 watchpoint support works by maintaining mirror images of the
573 debug registers. Values are copied between the mirror images and the
574 real debug registers via a set of macros which each target needs to
578 @findex I386_DR_LOW_SET_CONTROL
579 @item I386_DR_LOW_SET_CONTROL (@var{val})
580 Set the Debug Control (DR7) register to the value @var{val}.
582 @findex I386_DR_LOW_SET_ADDR
583 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
584 Put the address @var{addr} into the debug register number @var{idx}.
586 @findex I386_DR_LOW_RESET_ADDR
587 @item I386_DR_LOW_RESET_ADDR (@var{idx})
588 Reset (i.e.@: zero out) the address stored in the debug register
591 @findex I386_DR_LOW_GET_STATUS
592 @item I386_DR_LOW_GET_STATUS
593 Return the value of the Debug Status (DR6) register. This value is
594 used immediately after it is returned by
595 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
599 For each one of the 4 debug registers (whose indices are from 0 to 3)
600 that store addresses, a reference count is maintained by @value{GDBN},
601 to allow sharing of debug registers by several watchpoints. This
602 allows users to define several watchpoints that watch the same
603 expression, but with different conditions and/or commands, without
604 wasting debug registers which are in short supply. @value{GDBN}
605 maintains the reference counts internally, targets don't have to do
606 anything to use this feature.
608 The x86 debug registers can each watch a region that is 1, 2, or 4
609 bytes long. The ia32 architecture requires that each watched region
610 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
611 region on 4-byte boundary. However, the x86 watchpoint support in
612 @value{GDBN} can watch unaligned regions and regions larger than 4
613 bytes (up to 16 bytes) by allocating several debug registers to watch
614 a single region. This allocation of several registers per a watched
615 region is also done automatically without target code intervention.
617 The generic x86 watchpoint support provides the following API for the
618 @value{GDBN}'s application code:
621 @findex i386_region_ok_for_watchpoint
622 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
623 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
624 this function. It counts the number of debug registers required to
625 watch a given region, and returns a non-zero value if that number is
626 less than 4, the number of debug registers available to x86
629 @findex i386_stopped_data_address
630 @item i386_stopped_data_address (@var{addr_p})
632 @code{target_stopped_data_address} is set to call this function.
634 function examines the breakpoint condition bits in the DR6 Debug
635 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
636 macro, and returns the address associated with the first bit that is
639 @findex i386_stopped_by_watchpoint
640 @item i386_stopped_by_watchpoint (void)
641 The macro @code{STOPPED_BY_WATCHPOINT}
642 is set to call this function. The
643 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
644 function examines the breakpoint condition bits in the DR6 Debug
645 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
646 macro, and returns true if any bit is set. Otherwise, false is
649 @findex i386_insert_watchpoint
650 @findex i386_remove_watchpoint
651 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
652 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
653 Insert or remove a watchpoint. The macros
654 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
655 are set to call these functions. @code{i386_insert_watchpoint} first
656 looks for a debug register which is already set to watch the same
657 region for the same access types; if found, it just increments the
658 reference count of that debug register, thus implementing debug
659 register sharing between watchpoints. If no such register is found,
660 the function looks for a vacant debug register, sets its mirrored
661 value to @var{addr}, sets the mirrored value of DR7 Debug Control
662 register as appropriate for the @var{len} and @var{type} parameters,
663 and then passes the new values of the debug register and DR7 to the
664 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
665 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
666 required to cover the given region, the above process is repeated for
669 @code{i386_remove_watchpoint} does the opposite: it resets the address
670 in the mirrored value of the debug register and its read/write and
671 length bits in the mirrored value of DR7, then passes these new
672 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
673 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
674 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
675 decrements the reference count, and only calls
676 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
677 the count goes to zero.
679 @findex i386_insert_hw_breakpoint
680 @findex i386_remove_hw_breakpoint
681 @item i386_insert_hw_breakpoint (@var{addr}, @var{shadow}
682 @itemx i386_remove_hw_breakpoint (@var{addr}, @var{shadow})
683 These functions insert and remove hardware-assisted breakpoints. The
684 macros @code{target_insert_hw_breakpoint} and
685 @code{target_remove_hw_breakpoint} are set to call these functions.
686 These functions work like @code{i386_insert_watchpoint} and
687 @code{i386_remove_watchpoint}, respectively, except that they set up
688 the debug registers to watch instruction execution, and each
689 hardware-assisted breakpoint always requires exactly one debug
692 @findex i386_stopped_by_hwbp
693 @item i386_stopped_by_hwbp (void)
694 This function returns non-zero if the inferior has some watchpoint or
695 hardware breakpoint that triggered. It works like
696 @code{i386_stopped_data_address}, except that it doesn't record the
697 address whose watchpoint triggered.
699 @findex i386_cleanup_dregs
700 @item i386_cleanup_dregs (void)
701 This function clears all the reference counts, addresses, and control
702 bits in the mirror images of the debug registers. It doesn't affect
703 the actual debug registers in the inferior process.
710 x86 processors support setting watchpoints on I/O reads or writes.
711 However, since no target supports this (as of March 2001), and since
712 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
713 watchpoints, this feature is not yet available to @value{GDBN} running
717 x86 processors can enable watchpoints locally, for the current task
718 only, or globally, for all the tasks. For each debug register,
719 there's a bit in the DR7 Debug Control register that determines
720 whether the associated address is watched locally or globally. The
721 current implementation of x86 watchpoint support in @value{GDBN}
722 always sets watchpoints to be locally enabled, since global
723 watchpoints might interfere with the underlying OS and are probably
724 unavailable in many platforms.
727 @section Observing changes in @value{GDBN} internals
728 @cindex observer pattern interface
729 @cindex notifications about changes in internals
731 In order to function properly, several modules need to be notified when
732 some changes occur in the @value{GDBN} internals. Traditionally, these
733 modules have relied on several paradigms, the most common ones being
734 hooks and gdb-events. Unfortunately, none of these paradigms was
735 versatile enough to become the standard notification mechanism in
736 @value{GDBN}. The fact that they only supported one ``client'' was also
739 A new paradigm, based on the Observer pattern of the @cite{Design
740 Patterns} book, has therefore been implemented. The goal was to provide
741 a new interface overcoming the issues with the notification mechanisms
742 previously available. This new interface needed to be strongly typed,
743 easy to extend, and versatile enough to be used as the standard
744 interface when adding new notifications.
746 See @ref{GDB Observers} for a brief description of the observers
747 currently implemented in GDB. The rationale for the current
748 implementation is also briefly discussed.
752 @chapter User Interface
754 @value{GDBN} has several user interfaces. Although the command-line interface
755 is the most common and most familiar, there are others.
757 @section Command Interpreter
759 @cindex command interpreter
761 The command interpreter in @value{GDBN} is fairly simple. It is designed to
762 allow for the set of commands to be augmented dynamically, and also
763 has a recursive subcommand capability, where the first argument to
764 a command may itself direct a lookup on a different command list.
766 For instance, the @samp{set} command just starts a lookup on the
767 @code{setlist} command list, while @samp{set thread} recurses
768 to the @code{set_thread_cmd_list}.
772 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
773 the main command list, and should be used for those commands. The usual
774 place to add commands is in the @code{_initialize_@var{xyz}} routines at
775 the ends of most source files.
777 @findex add_setshow_cmd
778 @findex add_setshow_cmd_full
779 To add paired @samp{set} and @samp{show} commands, use
780 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
781 a slightly simpler interface which is useful when you don't need to
782 further modify the new command structures, while the latter returns
783 the new command structures for manipulation.
785 @cindex deprecating commands
786 @findex deprecate_cmd
787 Before removing commands from the command set it is a good idea to
788 deprecate them for some time. Use @code{deprecate_cmd} on commands or
789 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
790 @code{struct cmd_list_element} as it's first argument. You can use the
791 return value from @code{add_com} or @code{add_cmd} to deprecate the
792 command immediately after it is created.
794 The first time a command is used the user will be warned and offered a
795 replacement (if one exists). Note that the replacement string passed to
796 @code{deprecate_cmd} should be the full name of the command, i.e. the
797 entire string the user should type at the command line.
799 @section UI-Independent Output---the @code{ui_out} Functions
800 @c This section is based on the documentation written by Fernando
801 @c Nasser <fnasser@redhat.com>.
803 @cindex @code{ui_out} functions
804 The @code{ui_out} functions present an abstraction level for the
805 @value{GDBN} output code. They hide the specifics of different user
806 interfaces supported by @value{GDBN}, and thus free the programmer
807 from the need to write several versions of the same code, one each for
808 every UI, to produce output.
810 @subsection Overview and Terminology
812 In general, execution of each @value{GDBN} command produces some sort
813 of output, and can even generate an input request.
815 Output can be generated for the following purposes:
819 to display a @emph{result} of an operation;
822 to convey @emph{info} or produce side-effects of a requested
826 to provide a @emph{notification} of an asynchronous event (including
827 progress indication of a prolonged asynchronous operation);
830 to display @emph{error messages} (including warnings);
833 to show @emph{debug data};
836 to @emph{query} or prompt a user for input (a special case).
840 This section mainly concentrates on how to build result output,
841 although some of it also applies to other kinds of output.
843 Generation of output that displays the results of an operation
844 involves one or more of the following:
848 output of the actual data
851 formatting the output as appropriate for console output, to make it
852 easily readable by humans
855 machine oriented formatting--a more terse formatting to allow for easy
856 parsing by programs which read @value{GDBN}'s output
859 annotation, whose purpose is to help legacy GUIs to identify interesting
863 The @code{ui_out} routines take care of the first three aspects.
864 Annotations are provided by separate annotation routines. Note that use
865 of annotations for an interface between a GUI and @value{GDBN} is
868 Output can be in the form of a single item, which we call a @dfn{field};
869 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
870 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
871 header and a body. In a BNF-like form:
874 @item <table> @expansion{}
875 @code{<header> <body>}
876 @item <header> @expansion{}
877 @code{@{ <column> @}}
878 @item <column> @expansion{}
879 @code{<width> <alignment> <title>}
880 @item <body> @expansion{}
885 @subsection General Conventions
887 Most @code{ui_out} routines are of type @code{void}, the exceptions are
888 @code{ui_out_stream_new} (which returns a pointer to the newly created
889 object) and the @code{make_cleanup} routines.
891 The first parameter is always the @code{ui_out} vector object, a pointer
892 to a @code{struct ui_out}.
894 The @var{format} parameter is like in @code{printf} family of functions.
895 When it is present, there must also be a variable list of arguments
896 sufficient used to satisfy the @code{%} specifiers in the supplied
899 When a character string argument is not used in a @code{ui_out} function
900 call, a @code{NULL} pointer has to be supplied instead.
903 @subsection Table, Tuple and List Functions
905 @cindex list output functions
906 @cindex table output functions
907 @cindex tuple output functions
908 This section introduces @code{ui_out} routines for building lists,
909 tuples and tables. The routines to output the actual data items
910 (fields) are presented in the next section.
912 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
913 containing information about an object; a @dfn{list} is a sequence of
914 fields where each field describes an identical object.
916 Use the @dfn{table} functions when your output consists of a list of
917 rows (tuples) and the console output should include a heading. Use this
918 even when you are listing just one object but you still want the header.
920 @cindex nesting level in @code{ui_out} functions
921 Tables can not be nested. Tuples and lists can be nested up to a
922 maximum of five levels.
924 The overall structure of the table output code is something like this:
939 Here is the description of table-, tuple- and list-related @code{ui_out}
942 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
943 The function @code{ui_out_table_begin} marks the beginning of the output
944 of a table. It should always be called before any other @code{ui_out}
945 function for a given table. @var{nbrofcols} is the number of columns in
946 the table. @var{nr_rows} is the number of rows in the table.
947 @var{tblid} is an optional string identifying the table. The string
948 pointed to by @var{tblid} is copied by the implementation of
949 @code{ui_out_table_begin}, so the application can free the string if it
952 The companion function @code{ui_out_table_end}, described below, marks
953 the end of the table's output.
956 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
957 @code{ui_out_table_header} provides the header information for a single
958 table column. You call this function several times, one each for every
959 column of the table, after @code{ui_out_table_begin}, but before
960 @code{ui_out_table_body}.
962 The value of @var{width} gives the column width in characters. The
963 value of @var{alignment} is one of @code{left}, @code{center}, and
964 @code{right}, and it specifies how to align the header: left-justify,
965 center, or right-justify it. @var{colhdr} points to a string that
966 specifies the column header; the implementation copies that string, so
967 column header strings in @code{malloc}ed storage can be freed after the
971 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
972 This function delimits the table header from the table body.
975 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
976 This function signals the end of a table's output. It should be called
977 after the table body has been produced by the list and field output
980 There should be exactly one call to @code{ui_out_table_end} for each
981 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
982 will signal an internal error.
985 The output of the tuples that represent the table rows must follow the
986 call to @code{ui_out_table_body} and precede the call to
987 @code{ui_out_table_end}. You build a tuple by calling
988 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
989 calls to functions which actually output fields between them.
991 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
992 This function marks the beginning of a tuple output. @var{id} points
993 to an optional string that identifies the tuple; it is copied by the
994 implementation, and so strings in @code{malloc}ed storage can be freed
998 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
999 This function signals an end of a tuple output. There should be exactly
1000 one call to @code{ui_out_tuple_end} for each call to
1001 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1005 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1006 This function first opens the tuple and then establishes a cleanup
1007 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1008 and correct implementation of the non-portable@footnote{The function
1009 cast is not portable ISO C.} code sequence:
1011 struct cleanup *old_cleanup;
1012 ui_out_tuple_begin (uiout, "...");
1013 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1018 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1019 This function marks the beginning of a list output. @var{id} points to
1020 an optional string that identifies the list; it is copied by the
1021 implementation, and so strings in @code{malloc}ed storage can be freed
1025 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1026 This function signals an end of a list output. There should be exactly
1027 one call to @code{ui_out_list_end} for each call to
1028 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1032 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1033 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1034 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1035 that will close the list.list.
1038 @subsection Item Output Functions
1040 @cindex item output functions
1041 @cindex field output functions
1043 The functions described below produce output for the actual data
1044 items, or fields, which contain information about the object.
1046 Choose the appropriate function accordingly to your particular needs.
1048 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1049 This is the most general output function. It produces the
1050 representation of the data in the variable-length argument list
1051 according to formatting specifications in @var{format}, a
1052 @code{printf}-like format string. The optional argument @var{fldname}
1053 supplies the name of the field. The data items themselves are
1054 supplied as additional arguments after @var{format}.
1056 This generic function should be used only when it is not possible to
1057 use one of the specialized versions (see below).
1060 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1061 This function outputs a value of an @code{int} variable. It uses the
1062 @code{"%d"} output conversion specification. @var{fldname} specifies
1063 the name of the field.
1066 @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})
1067 This function outputs a value of an @code{int} variable. It differs from
1068 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1069 @var{fldname} specifies
1070 the name of the field.
1073 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1074 This function outputs an address.
1077 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1078 This function outputs a string using the @code{"%s"} conversion
1082 Sometimes, there's a need to compose your output piece by piece using
1083 functions that operate on a stream, such as @code{value_print} or
1084 @code{fprintf_symbol_filtered}. These functions accept an argument of
1085 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1086 used to store the data stream used for the output. When you use one
1087 of these functions, you need a way to pass their results stored in a
1088 @code{ui_file} object to the @code{ui_out} functions. To this end,
1089 you first create a @code{ui_stream} object by calling
1090 @code{ui_out_stream_new}, pass the @code{stream} member of that
1091 @code{ui_stream} object to @code{value_print} and similar functions,
1092 and finally call @code{ui_out_field_stream} to output the field you
1093 constructed. When the @code{ui_stream} object is no longer needed,
1094 you should destroy it and free its memory by calling
1095 @code{ui_out_stream_delete}.
1097 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1098 This function creates a new @code{ui_stream} object which uses the
1099 same output methods as the @code{ui_out} object whose pointer is
1100 passed in @var{uiout}. It returns a pointer to the newly created
1101 @code{ui_stream} object.
1104 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1105 This functions destroys a @code{ui_stream} object specified by
1109 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1110 This function consumes all the data accumulated in
1111 @code{streambuf->stream} and outputs it like
1112 @code{ui_out_field_string} does. After a call to
1113 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1114 the stream is still valid and may be used for producing more fields.
1117 @strong{Important:} If there is any chance that your code could bail
1118 out before completing output generation and reaching the point where
1119 @code{ui_out_stream_delete} is called, it is necessary to set up a
1120 cleanup, to avoid leaking memory and other resources. Here's a
1121 skeleton code to do that:
1124 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1125 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1130 If the function already has the old cleanup chain set (for other kinds
1131 of cleanups), you just have to add your cleanup to it:
1134 mybuf = ui_out_stream_new (uiout);
1135 make_cleanup (ui_out_stream_delete, mybuf);
1138 Note that with cleanups in place, you should not call
1139 @code{ui_out_stream_delete} directly, or you would attempt to free the
1142 @subsection Utility Output Functions
1144 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1145 This function skips a field in a table. Use it if you have to leave
1146 an empty field without disrupting the table alignment. The argument
1147 @var{fldname} specifies a name for the (missing) filed.
1150 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1151 This function outputs the text in @var{string} in a way that makes it
1152 easy to be read by humans. For example, the console implementation of
1153 this method filters the text through a built-in pager, to prevent it
1154 from scrolling off the visible portion of the screen.
1156 Use this function for printing relatively long chunks of text around
1157 the actual field data: the text it produces is not aligned according
1158 to the table's format. Use @code{ui_out_field_string} to output a
1159 string field, and use @code{ui_out_message}, described below, to
1160 output short messages.
1163 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1164 This function outputs @var{nspaces} spaces. It is handy to align the
1165 text produced by @code{ui_out_text} with the rest of the table or
1169 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1170 This function produces a formatted message, provided that the current
1171 verbosity level is at least as large as given by @var{verbosity}. The
1172 current verbosity level is specified by the user with the @samp{set
1173 verbositylevel} command.@footnote{As of this writing (April 2001),
1174 setting verbosity level is not yet implemented, and is always returned
1175 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1176 argument more than zero will cause the message to never be printed.}
1179 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1180 This function gives the console output filter (a paging filter) a hint
1181 of where to break lines which are too long. Ignored for all other
1182 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1183 be printed to indent the wrapped text on the next line; it must remain
1184 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1185 explicit newline is produced by one of the other functions. If
1186 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1189 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1190 This function flushes whatever output has been accumulated so far, if
1191 the UI buffers output.
1195 @subsection Examples of Use of @code{ui_out} functions
1197 @cindex using @code{ui_out} functions
1198 @cindex @code{ui_out} functions, usage examples
1199 This section gives some practical examples of using the @code{ui_out}
1200 functions to generalize the old console-oriented code in
1201 @value{GDBN}. The examples all come from functions defined on the
1202 @file{breakpoints.c} file.
1204 This example, from the @code{breakpoint_1} function, shows how to
1207 The original code was:
1210 if (!found_a_breakpoint++)
1212 annotate_breakpoints_headers ();
1215 printf_filtered ("Num ");
1217 printf_filtered ("Type ");
1219 printf_filtered ("Disp ");
1221 printf_filtered ("Enb ");
1225 printf_filtered ("Address ");
1228 printf_filtered ("What\n");
1230 annotate_breakpoints_table ();
1234 Here's the new version:
1237 nr_printable_breakpoints = @dots{};
1240 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1242 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1244 if (nr_printable_breakpoints > 0)
1245 annotate_breakpoints_headers ();
1246 if (nr_printable_breakpoints > 0)
1248 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1249 if (nr_printable_breakpoints > 0)
1251 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1252 if (nr_printable_breakpoints > 0)
1254 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1255 if (nr_printable_breakpoints > 0)
1257 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1260 if (nr_printable_breakpoints > 0)
1262 if (TARGET_ADDR_BIT <= 32)
1263 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1265 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1267 if (nr_printable_breakpoints > 0)
1269 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1270 ui_out_table_body (uiout);
1271 if (nr_printable_breakpoints > 0)
1272 annotate_breakpoints_table ();
1275 This example, from the @code{print_one_breakpoint} function, shows how
1276 to produce the actual data for the table whose structure was defined
1277 in the above example. The original code was:
1282 printf_filtered ("%-3d ", b->number);
1284 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1285 || ((int) b->type != bptypes[(int) b->type].type))
1286 internal_error ("bptypes table does not describe type #%d.",
1288 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1290 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1292 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1296 This is the new version:
1300 ui_out_tuple_begin (uiout, "bkpt");
1302 ui_out_field_int (uiout, "number", b->number);
1304 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1305 || ((int) b->type != bptypes[(int) b->type].type))
1306 internal_error ("bptypes table does not describe type #%d.",
1308 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1310 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1312 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1316 This example, also from @code{print_one_breakpoint}, shows how to
1317 produce a complicated output field using the @code{print_expression}
1318 functions which requires a stream to be passed. It also shows how to
1319 automate stream destruction with cleanups. The original code was:
1323 print_expression (b->exp, gdb_stdout);
1329 struct ui_stream *stb = ui_out_stream_new (uiout);
1330 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1333 print_expression (b->exp, stb->stream);
1334 ui_out_field_stream (uiout, "what", local_stream);
1337 This example, also from @code{print_one_breakpoint}, shows how to use
1338 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1343 if (b->dll_pathname == NULL)
1344 printf_filtered ("<any library> ");
1346 printf_filtered ("library \"%s\" ", b->dll_pathname);
1353 if (b->dll_pathname == NULL)
1355 ui_out_field_string (uiout, "what", "<any library>");
1356 ui_out_spaces (uiout, 1);
1360 ui_out_text (uiout, "library \"");
1361 ui_out_field_string (uiout, "what", b->dll_pathname);
1362 ui_out_text (uiout, "\" ");
1366 The following example from @code{print_one_breakpoint} shows how to
1367 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1372 if (b->forked_inferior_pid != 0)
1373 printf_filtered ("process %d ", b->forked_inferior_pid);
1380 if (b->forked_inferior_pid != 0)
1382 ui_out_text (uiout, "process ");
1383 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1384 ui_out_spaces (uiout, 1);
1388 Here's an example of using @code{ui_out_field_string}. The original
1393 if (b->exec_pathname != NULL)
1394 printf_filtered ("program \"%s\" ", b->exec_pathname);
1401 if (b->exec_pathname != NULL)
1403 ui_out_text (uiout, "program \"");
1404 ui_out_field_string (uiout, "what", b->exec_pathname);
1405 ui_out_text (uiout, "\" ");
1409 Finally, here's an example of printing an address. The original code:
1413 printf_filtered ("%s ",
1414 hex_string_custom ((unsigned long) b->address, 8));
1421 ui_out_field_core_addr (uiout, "Address", b->address);
1425 @section Console Printing
1434 @cindex @code{libgdb}
1435 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1436 to provide an API to @value{GDBN}'s functionality.
1439 @cindex @code{libgdb}
1440 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1441 better able to support graphical and other environments.
1443 Since @code{libgdb} development is on-going, its architecture is still
1444 evolving. The following components have so far been identified:
1448 Observer - @file{gdb-events.h}.
1450 Builder - @file{ui-out.h}
1452 Event Loop - @file{event-loop.h}
1454 Library - @file{gdb.h}
1457 The model that ties these components together is described below.
1459 @section The @code{libgdb} Model
1461 A client of @code{libgdb} interacts with the library in two ways.
1465 As an observer (using @file{gdb-events}) receiving notifications from
1466 @code{libgdb} of any internal state changes (break point changes, run
1469 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1470 obtain various status values from @value{GDBN}.
1473 Since @code{libgdb} could have multiple clients (e.g. a GUI supporting
1474 the existing @value{GDBN} CLI), those clients must co-operate when
1475 controlling @code{libgdb}. In particular, a client must ensure that
1476 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1477 before responding to a @file{gdb-event} by making a query.
1479 @section CLI support
1481 At present @value{GDBN}'s CLI is very much entangled in with the core of
1482 @code{libgdb}. Consequently, a client wishing to include the CLI in
1483 their interface needs to carefully co-ordinate its own and the CLI's
1486 It is suggested that the client set @code{libgdb} up to be bi-modal
1487 (alternate between CLI and client query modes). The notes below sketch
1492 The client registers itself as an observer of @code{libgdb}.
1494 The client create and install @code{cli-out} builder using its own
1495 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1496 @code{gdb_stdout} streams.
1498 The client creates a separate custom @code{ui-out} builder that is only
1499 used while making direct queries to @code{libgdb}.
1502 When the client receives input intended for the CLI, it simply passes it
1503 along. Since the @code{cli-out} builder is installed by default, all
1504 the CLI output in response to that command is routed (pronounced rooted)
1505 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1506 At the same time, the client is kept abreast of internal changes by
1507 virtue of being a @code{libgdb} observer.
1509 The only restriction on the client is that it must wait until
1510 @code{libgdb} becomes idle before initiating any queries (using the
1511 client's custom builder).
1513 @section @code{libgdb} components
1515 @subheading Observer - @file{gdb-events.h}
1516 @file{gdb-events} provides the client with a very raw mechanism that can
1517 be used to implement an observer. At present it only allows for one
1518 observer and that observer must, internally, handle the need to delay
1519 the processing of any event notifications until after @code{libgdb} has
1520 finished the current command.
1522 @subheading Builder - @file{ui-out.h}
1523 @file{ui-out} provides the infrastructure necessary for a client to
1524 create a builder. That builder is then passed down to @code{libgdb}
1525 when doing any queries.
1527 @subheading Event Loop - @file{event-loop.h}
1528 @c There could be an entire section on the event-loop
1529 @file{event-loop}, currently non-re-entrant, provides a simple event
1530 loop. A client would need to either plug its self into this loop or,
1531 implement a new event-loop that GDB would use.
1533 The event-loop will eventually be made re-entrant. This is so that
1534 @value{GDBN} can better handle the problem of some commands blocking
1535 instead of returning.
1537 @subheading Library - @file{gdb.h}
1538 @file{libgdb} is the most obvious component of this system. It provides
1539 the query interface. Each function is parameterized by a @code{ui-out}
1540 builder. The result of the query is constructed using that builder
1541 before the query function returns.
1543 @node Symbol Handling
1545 @chapter Symbol Handling
1547 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1548 functions, and types.
1550 @section Symbol Reading
1552 @cindex symbol reading
1553 @cindex reading of symbols
1554 @cindex symbol files
1555 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1556 file is the file containing the program which @value{GDBN} is
1557 debugging. @value{GDBN} can be directed to use a different file for
1558 symbols (with the @samp{symbol-file} command), and it can also read
1559 more symbols via the @samp{add-file} and @samp{load} commands, or while
1560 reading symbols from shared libraries.
1562 @findex find_sym_fns
1563 Symbol files are initially opened by code in @file{symfile.c} using
1564 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1565 of the file by examining its header. @code{find_sym_fns} then uses
1566 this identification to locate a set of symbol-reading functions.
1568 @findex add_symtab_fns
1569 @cindex @code{sym_fns} structure
1570 @cindex adding a symbol-reading module
1571 Symbol-reading modules identify themselves to @value{GDBN} by calling
1572 @code{add_symtab_fns} during their module initialization. The argument
1573 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1574 name (or name prefix) of the symbol format, the length of the prefix,
1575 and pointers to four functions. These functions are called at various
1576 times to process symbol files whose identification matches the specified
1579 The functions supplied by each module are:
1582 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1584 @cindex secondary symbol file
1585 Called from @code{symbol_file_add} when we are about to read a new
1586 symbol file. This function should clean up any internal state (possibly
1587 resulting from half-read previous files, for example) and prepare to
1588 read a new symbol file. Note that the symbol file which we are reading
1589 might be a new ``main'' symbol file, or might be a secondary symbol file
1590 whose symbols are being added to the existing symbol table.
1592 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1593 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1594 new symbol file being read. Its @code{private} field has been zeroed,
1595 and can be modified as desired. Typically, a struct of private
1596 information will be @code{malloc}'d, and a pointer to it will be placed
1597 in the @code{private} field.
1599 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1600 @code{error} if it detects an unavoidable problem.
1602 @item @var{xyz}_new_init()
1604 Called from @code{symbol_file_add} when discarding existing symbols.
1605 This function needs only handle the symbol-reading module's internal
1606 state; the symbol table data structures visible to the rest of
1607 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1608 arguments and no result. It may be called after
1609 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1610 may be called alone if all symbols are simply being discarded.
1612 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1614 Called from @code{symbol_file_add} to actually read the symbols from a
1615 symbol-file into a set of psymtabs or symtabs.
1617 @code{sf} points to the @code{struct sym_fns} originally passed to
1618 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1619 the offset between the file's specified start address and its true
1620 address in memory. @code{mainline} is 1 if this is the main symbol
1621 table being read, and 0 if a secondary symbol file (e.g. shared library
1622 or dynamically loaded file) is being read.@refill
1625 In addition, if a symbol-reading module creates psymtabs when
1626 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1627 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1628 from any point in the @value{GDBN} symbol-handling code.
1631 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1633 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1634 the psymtab has not already been read in and had its @code{pst->symtab}
1635 pointer set. The argument is the psymtab to be fleshed-out into a
1636 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1637 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1638 zero if there were no symbols in that part of the symbol file.
1641 @section Partial Symbol Tables
1643 @value{GDBN} has three types of symbol tables:
1646 @cindex full symbol table
1649 Full symbol tables (@dfn{symtabs}). These contain the main
1650 information about symbols and addresses.
1654 Partial symbol tables (@dfn{psymtabs}). These contain enough
1655 information to know when to read the corresponding part of the full
1658 @cindex minimal symbol table
1661 Minimal symbol tables (@dfn{msymtabs}). These contain information
1662 gleaned from non-debugging symbols.
1665 @cindex partial symbol table
1666 This section describes partial symbol tables.
1668 A psymtab is constructed by doing a very quick pass over an executable
1669 file's debugging information. Small amounts of information are
1670 extracted---enough to identify which parts of the symbol table will
1671 need to be re-read and fully digested later, when the user needs the
1672 information. The speed of this pass causes @value{GDBN} to start up very
1673 quickly. Later, as the detailed rereading occurs, it occurs in small
1674 pieces, at various times, and the delay therefrom is mostly invisible to
1676 @c (@xref{Symbol Reading}.)
1678 The symbols that show up in a file's psymtab should be, roughly, those
1679 visible to the debugger's user when the program is not running code from
1680 that file. These include external symbols and types, static symbols and
1681 types, and @code{enum} values declared at file scope.
1683 The psymtab also contains the range of instruction addresses that the
1684 full symbol table would represent.
1686 @cindex finding a symbol
1687 @cindex symbol lookup
1688 The idea is that there are only two ways for the user (or much of the
1689 code in the debugger) to reference a symbol:
1692 @findex find_pc_function
1693 @findex find_pc_line
1695 By its address (e.g. execution stops at some address which is inside a
1696 function in this file). The address will be noticed to be in the
1697 range of this psymtab, and the full symtab will be read in.
1698 @code{find_pc_function}, @code{find_pc_line}, and other
1699 @code{find_pc_@dots{}} functions handle this.
1701 @cindex lookup_symbol
1704 (e.g. the user asks to print a variable, or set a breakpoint on a
1705 function). Global names and file-scope names will be found in the
1706 psymtab, which will cause the symtab to be pulled in. Local names will
1707 have to be qualified by a global name, or a file-scope name, in which
1708 case we will have already read in the symtab as we evaluated the
1709 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1710 local scope, in which case the first case applies. @code{lookup_symbol}
1711 does most of the work here.
1714 The only reason that psymtabs exist is to cause a symtab to be read in
1715 at the right moment. Any symbol that can be elided from a psymtab,
1716 while still causing that to happen, should not appear in it. Since
1717 psymtabs don't have the idea of scope, you can't put local symbols in
1718 them anyway. Psymtabs don't have the idea of the type of a symbol,
1719 either, so types need not appear, unless they will be referenced by
1722 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1723 been read, and another way if the corresponding symtab has been read
1724 in. Such bugs are typically caused by a psymtab that does not contain
1725 all the visible symbols, or which has the wrong instruction address
1728 The psymtab for a particular section of a symbol file (objfile) could be
1729 thrown away after the symtab has been read in. The symtab should always
1730 be searched before the psymtab, so the psymtab will never be used (in a
1731 bug-free environment). Currently, psymtabs are allocated on an obstack,
1732 and all the psymbols themselves are allocated in a pair of large arrays
1733 on an obstack, so there is little to be gained by trying to free them
1734 unless you want to do a lot more work.
1738 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
1740 @cindex fundamental types
1741 These are the fundamental types that @value{GDBN} uses internally. Fundamental
1742 types from the various debugging formats (stabs, ELF, etc) are mapped
1743 into one of these. They are basically a union of all fundamental types
1744 that @value{GDBN} knows about for all the languages that @value{GDBN}
1747 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
1750 Each time @value{GDBN} builds an internal type, it marks it with one
1751 of these types. The type may be a fundamental type, such as
1752 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
1753 which is a pointer to another type. Typically, several @code{FT_*}
1754 types map to one @code{TYPE_CODE_*} type, and are distinguished by
1755 other members of the type struct, such as whether the type is signed
1756 or unsigned, and how many bits it uses.
1758 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
1760 These are instances of type structs that roughly correspond to
1761 fundamental types and are created as global types for @value{GDBN} to
1762 use for various ugly historical reasons. We eventually want to
1763 eliminate these. Note for example that @code{builtin_type_int}
1764 initialized in @file{gdbtypes.c} is basically the same as a
1765 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
1766 an @code{FT_INTEGER} fundamental type. The difference is that the
1767 @code{builtin_type} is not associated with any particular objfile, and
1768 only one instance exists, while @file{c-lang.c} builds as many
1769 @code{TYPE_CODE_INT} types as needed, with each one associated with
1770 some particular objfile.
1772 @section Object File Formats
1773 @cindex object file formats
1777 @cindex @code{a.out} format
1778 The @code{a.out} format is the original file format for Unix. It
1779 consists of three sections: @code{text}, @code{data}, and @code{bss},
1780 which are for program code, initialized data, and uninitialized data,
1783 The @code{a.out} format is so simple that it doesn't have any reserved
1784 place for debugging information. (Hey, the original Unix hackers used
1785 @samp{adb}, which is a machine-language debugger!) The only debugging
1786 format for @code{a.out} is stabs, which is encoded as a set of normal
1787 symbols with distinctive attributes.
1789 The basic @code{a.out} reader is in @file{dbxread.c}.
1794 The COFF format was introduced with System V Release 3 (SVR3) Unix.
1795 COFF files may have multiple sections, each prefixed by a header. The
1796 number of sections is limited.
1798 The COFF specification includes support for debugging. Although this
1799 was a step forward, the debugging information was woefully limited. For
1800 instance, it was not possible to represent code that came from an
1803 The COFF reader is in @file{coffread.c}.
1807 @cindex ECOFF format
1808 ECOFF is an extended COFF originally introduced for Mips and Alpha
1811 The basic ECOFF reader is in @file{mipsread.c}.
1815 @cindex XCOFF format
1816 The IBM RS/6000 running AIX uses an object file format called XCOFF.
1817 The COFF sections, symbols, and line numbers are used, but debugging
1818 symbols are @code{dbx}-style stabs whose strings are located in the
1819 @code{.debug} section (rather than the string table). For more
1820 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
1822 The shared library scheme has a clean interface for figuring out what
1823 shared libraries are in use, but the catch is that everything which
1824 refers to addresses (symbol tables and breakpoints at least) needs to be
1825 relocated for both shared libraries and the main executable. At least
1826 using the standard mechanism this can only be done once the program has
1827 been run (or the core file has been read).
1831 @cindex PE-COFF format
1832 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
1833 executables. PE is basically COFF with additional headers.
1835 While BFD includes special PE support, @value{GDBN} needs only the basic
1841 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
1842 to COFF in being organized into a number of sections, but it removes
1843 many of COFF's limitations.
1845 The basic ELF reader is in @file{elfread.c}.
1850 SOM is HP's object file and debug format (not to be confused with IBM's
1851 SOM, which is a cross-language ABI).
1853 The SOM reader is in @file{hpread.c}.
1855 @subsection Other File Formats
1857 @cindex Netware Loadable Module format
1858 Other file formats that have been supported by @value{GDBN} include Netware
1859 Loadable Modules (@file{nlmread.c}).
1861 @section Debugging File Formats
1863 This section describes characteristics of debugging information that
1864 are independent of the object file format.
1868 @cindex stabs debugging info
1869 @code{stabs} started out as special symbols within the @code{a.out}
1870 format. Since then, it has been encapsulated into other file
1871 formats, such as COFF and ELF.
1873 While @file{dbxread.c} does some of the basic stab processing,
1874 including for encapsulated versions, @file{stabsread.c} does
1879 @cindex COFF debugging info
1880 The basic COFF definition includes debugging information. The level
1881 of support is minimal and non-extensible, and is not often used.
1883 @subsection Mips debug (Third Eye)
1885 @cindex ECOFF debugging info
1886 ECOFF includes a definition of a special debug format.
1888 The file @file{mdebugread.c} implements reading for this format.
1892 @cindex DWARF 1 debugging info
1893 DWARF 1 is a debugging format that was originally designed to be
1894 used with ELF in SVR4 systems.
1899 @c If defined, these are the producer strings in a DWARF 1 file. All of
1900 @c these have reasonable defaults already.
1902 The DWARF 1 reader is in @file{dwarfread.c}.
1906 @cindex DWARF 2 debugging info
1907 DWARF 2 is an improved but incompatible version of DWARF 1.
1909 The DWARF 2 reader is in @file{dwarf2read.c}.
1913 @cindex SOM debugging info
1914 Like COFF, the SOM definition includes debugging information.
1916 @section Adding a New Symbol Reader to @value{GDBN}
1918 @cindex adding debugging info reader
1919 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
1920 there is probably little to be done.
1922 If you need to add a new object file format, you must first add it to
1923 BFD. This is beyond the scope of this document.
1925 You must then arrange for the BFD code to provide access to the
1926 debugging symbols. Generally @value{GDBN} will have to call swapping routines
1927 from BFD and a few other BFD internal routines to locate the debugging
1928 information. As much as possible, @value{GDBN} should not depend on the BFD
1929 internal data structures.
1931 For some targets (e.g., COFF), there is a special transfer vector used
1932 to call swapping routines, since the external data structures on various
1933 platforms have different sizes and layouts. Specialized routines that
1934 will only ever be implemented by one object file format may be called
1935 directly. This interface should be described in a file
1936 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
1939 @node Language Support
1941 @chapter Language Support
1943 @cindex language support
1944 @value{GDBN}'s language support is mainly driven by the symbol reader,
1945 although it is possible for the user to set the source language
1948 @value{GDBN} chooses the source language by looking at the extension
1949 of the file recorded in the debug info; @file{.c} means C, @file{.f}
1950 means Fortran, etc. It may also use a special-purpose language
1951 identifier if the debug format supports it, like with DWARF.
1953 @section Adding a Source Language to @value{GDBN}
1955 @cindex adding source language
1956 To add other languages to @value{GDBN}'s expression parser, follow the
1960 @item Create the expression parser.
1962 @cindex expression parser
1963 This should reside in a file @file{@var{lang}-exp.y}. Routines for
1964 building parsed expressions into a @code{union exp_element} list are in
1967 @cindex language parser
1968 Since we can't depend upon everyone having Bison, and YACC produces
1969 parsers that define a bunch of global names, the following lines
1970 @strong{must} be included at the top of the YACC parser, to prevent the
1971 various parsers from defining the same global names:
1974 #define yyparse @var{lang}_parse
1975 #define yylex @var{lang}_lex
1976 #define yyerror @var{lang}_error
1977 #define yylval @var{lang}_lval
1978 #define yychar @var{lang}_char
1979 #define yydebug @var{lang}_debug
1980 #define yypact @var{lang}_pact
1981 #define yyr1 @var{lang}_r1
1982 #define yyr2 @var{lang}_r2
1983 #define yydef @var{lang}_def
1984 #define yychk @var{lang}_chk
1985 #define yypgo @var{lang}_pgo
1986 #define yyact @var{lang}_act
1987 #define yyexca @var{lang}_exca
1988 #define yyerrflag @var{lang}_errflag
1989 #define yynerrs @var{lang}_nerrs
1992 At the bottom of your parser, define a @code{struct language_defn} and
1993 initialize it with the right values for your language. Define an
1994 @code{initialize_@var{lang}} routine and have it call
1995 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
1996 that your language exists. You'll need some other supporting variables
1997 and functions, which will be used via pointers from your
1998 @code{@var{lang}_language_defn}. See the declaration of @code{struct
1999 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2000 for more information.
2002 @item Add any evaluation routines, if necessary
2004 @cindex expression evaluation routines
2005 @findex evaluate_subexp
2006 @findex prefixify_subexp
2007 @findex length_of_subexp
2008 If you need new opcodes (that represent the operations of the language),
2009 add them to the enumerated type in @file{expression.h}. Add support
2010 code for these operations in the @code{evaluate_subexp} function
2011 defined in the file @file{eval.c}. Add cases
2012 for new opcodes in two functions from @file{parse.c}:
2013 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2014 the number of @code{exp_element}s that a given operation takes up.
2016 @item Update some existing code
2018 Add an enumerated identifier for your language to the enumerated type
2019 @code{enum language} in @file{defs.h}.
2021 Update the routines in @file{language.c} so your language is included.
2022 These routines include type predicates and such, which (in some cases)
2023 are language dependent. If your language does not appear in the switch
2024 statement, an error is reported.
2026 @vindex current_language
2027 Also included in @file{language.c} is the code that updates the variable
2028 @code{current_language}, and the routines that translate the
2029 @code{language_@var{lang}} enumerated identifier into a printable
2032 @findex _initialize_language
2033 Update the function @code{_initialize_language} to include your
2034 language. This function picks the default language upon startup, so is
2035 dependent upon which languages that @value{GDBN} is built for.
2037 @findex allocate_symtab
2038 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2039 code so that the language of each symtab (source file) is set properly.
2040 This is used to determine the language to use at each stack frame level.
2041 Currently, the language is set based upon the extension of the source
2042 file. If the language can be better inferred from the symbol
2043 information, please set the language of the symtab in the symbol-reading
2046 @findex print_subexp
2047 @findex op_print_tab
2048 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2049 expression opcodes you have added to @file{expression.h}. Also, add the
2050 printed representations of your operators to @code{op_print_tab}.
2052 @item Add a place of call
2055 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2056 @code{parse_exp_1} (defined in @file{parse.c}).
2058 @item Use macros to trim code
2060 @cindex trimming language-dependent code
2061 The user has the option of building @value{GDBN} for some or all of the
2062 languages. If the user decides to build @value{GDBN} for the language
2063 @var{lang}, then every file dependent on @file{language.h} will have the
2064 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2065 leave out large routines that the user won't need if he or she is not
2066 using your language.
2068 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2069 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2070 compiled form of your parser) is not linked into @value{GDBN} at all.
2072 See the file @file{configure.in} for how @value{GDBN} is configured
2073 for different languages.
2075 @item Edit @file{Makefile.in}
2077 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2078 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2079 not get linked in, or, worse yet, it may not get @code{tar}red into the
2084 @node Host Definition
2086 @chapter Host Definition
2088 With the advent of Autoconf, it's rarely necessary to have host
2089 definition machinery anymore. The following information is provided,
2090 mainly, as an historical reference.
2092 @section Adding a New Host
2094 @cindex adding a new host
2095 @cindex host, adding
2096 @value{GDBN}'s host configuration support normally happens via Autoconf.
2097 New host-specific definitions should not be needed. Older hosts
2098 @value{GDBN} still use the host-specific definitions and files listed
2099 below, but these mostly exist for historical reasons, and will
2100 eventually disappear.
2103 @item gdb/config/@var{arch}/@var{xyz}.mh
2104 This file once contained both host and native configuration information
2105 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2106 configuration information is now handed by Autoconf.
2108 Host configuration information included a definition of
2109 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2110 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2111 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2113 New host only configurations do not need this file.
2115 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2116 This file once contained definitions and includes required when hosting
2117 gdb on machine @var{xyz}. Those definitions and includes are now
2118 handled by Autoconf.
2120 New host and native configurations do not need this file.
2122 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2123 file to define the macros @var{HOST_FLOAT_FORMAT},
2124 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2125 also needs to be replaced with either an Autoconf or run-time test.}
2129 @subheading Generic Host Support Files
2131 @cindex generic host support
2132 There are some ``generic'' versions of routines that can be used by
2133 various systems. These can be customized in various ways by macros
2134 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2135 the @var{xyz} host, you can just include the generic file's name (with
2136 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2138 Otherwise, if your machine needs custom support routines, you will need
2139 to write routines that perform the same functions as the generic file.
2140 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2141 into @code{XDEPFILES}.
2144 @cindex remote debugging support
2145 @cindex serial line support
2147 This contains serial line support for Unix systems. This is always
2148 included, via the makefile variable @code{SER_HARDWIRE}; override this
2149 variable in the @file{.mh} file to avoid it.
2152 This contains serial line support for 32-bit programs running under DOS,
2153 using the DJGPP (a.k.a.@: GO32) execution environment.
2155 @cindex TCP remote support
2157 This contains generic TCP support using sockets.
2160 @section Host Conditionals
2162 When @value{GDBN} is configured and compiled, various macros are
2163 defined or left undefined, to control compilation based on the
2164 attributes of the host system. These macros and their meanings (or if
2165 the meaning is not documented here, then one of the source files where
2166 they are used is indicated) are:
2169 @item @value{GDBN}INIT_FILENAME
2170 The default name of @value{GDBN}'s initialization file (normally
2174 This macro is deprecated.
2176 @item SIGWINCH_HANDLER
2177 If your host defines @code{SIGWINCH}, you can define this to be the name
2178 of a function to be called if @code{SIGWINCH} is received.
2180 @item SIGWINCH_HANDLER_BODY
2181 Define this to expand into code that will define the function named by
2182 the expansion of @code{SIGWINCH_HANDLER}.
2184 @item ALIGN_STACK_ON_STARTUP
2185 @cindex stack alignment
2186 Define this if your system is of a sort that will crash in
2187 @code{tgetent} if the stack happens not to be longword-aligned when
2188 @code{main} is called. This is a rare situation, but is known to occur
2189 on several different types of systems.
2191 @item CRLF_SOURCE_FILES
2192 @cindex DOS text files
2193 Define this if host files use @code{\r\n} rather than @code{\n} as a
2194 line terminator. This will cause source file listings to omit @code{\r}
2195 characters when printing and it will allow @code{\r\n} line endings of files
2196 which are ``sourced'' by gdb. It must be possible to open files in binary
2197 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2199 @item DEFAULT_PROMPT
2201 The default value of the prompt string (normally @code{"(gdb) "}).
2204 @cindex terminal device
2205 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2208 Define this if binary files are opened the same way as text files.
2212 In some cases, use the system call @code{mmap} for reading symbol
2213 tables. For some machines this allows for sharing and quick updates.
2216 Define this if the host system has @code{termio.h}.
2223 Values for host-side constants.
2226 Substitute for isatty, if not available.
2229 This is the longest integer type available on the host. If not defined,
2230 it will default to @code{long long} or @code{long}, depending on
2231 @code{CC_HAS_LONG_LONG}.
2233 @item CC_HAS_LONG_LONG
2234 @cindex @code{long long} data type
2235 Define this if the host C compiler supports @code{long long}. This is set
2236 by the @code{configure} script.
2238 @item PRINTF_HAS_LONG_LONG
2239 Define this if the host can handle printing of long long integers via
2240 the printf format conversion specifier @code{ll}. This is set by the
2241 @code{configure} script.
2243 @item HAVE_LONG_DOUBLE
2244 Define this if the host C compiler supports @code{long double}. This is
2245 set by the @code{configure} script.
2247 @item PRINTF_HAS_LONG_DOUBLE
2248 Define this if the host can handle printing of long double float-point
2249 numbers via the printf format conversion specifier @code{Lg}. This is
2250 set by the @code{configure} script.
2252 @item SCANF_HAS_LONG_DOUBLE
2253 Define this if the host can handle the parsing of long double
2254 float-point numbers via the scanf format conversion specifier
2255 @code{Lg}. This is set by the @code{configure} script.
2257 @item LSEEK_NOT_LINEAR
2258 Define this if @code{lseek (n)} does not necessarily move to byte number
2259 @code{n} in the file. This is only used when reading source files. It
2260 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2263 This macro is used as the argument to @code{lseek} (or, most commonly,
2264 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2265 which is the POSIX equivalent.
2268 If defined, this should be one or more tokens, such as @code{volatile},
2269 that can be used in both the declaration and definition of functions to
2270 indicate that they never return. The default is already set correctly
2271 if compiling with GCC. This will almost never need to be defined.
2274 If defined, this should be one or more tokens, such as
2275 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2276 of functions to indicate that they never return. The default is already
2277 set correctly if compiling with GCC. This will almost never need to be
2282 Define these to appropriate value for the system @code{lseek}, if not already
2286 This is the signal for stopping @value{GDBN}. Defaults to
2287 @code{SIGTSTP}. (Only redefined for the Convex.)
2290 Means that System V (prior to SVR4) include files are in use. (FIXME:
2291 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2292 @file{utils.c} for other things, at the moment.)
2295 Define this to help placate @code{lint} in some situations.
2298 Define this to override the defaults of @code{__volatile__} or
2303 @node Target Architecture Definition
2305 @chapter Target Architecture Definition
2307 @cindex target architecture definition
2308 @value{GDBN}'s target architecture defines what sort of
2309 machine-language programs @value{GDBN} can work with, and how it works
2312 The target architecture object is implemented as the C structure
2313 @code{struct gdbarch *}. The structure, and its methods, are generated
2314 using the Bourne shell script @file{gdbarch.sh}.
2316 @section Operating System ABI Variant Handling
2317 @cindex OS ABI variants
2319 @value{GDBN} provides a mechanism for handling variations in OS
2320 ABIs. An OS ABI variant may have influence over any number of
2321 variables in the target architecture definition. There are two major
2322 components in the OS ABI mechanism: sniffers and handlers.
2324 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2325 (the architecture may be wildcarded) in an attempt to determine the
2326 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2327 to be @dfn{generic}, while sniffers for a specific architecture are
2328 considered to be @dfn{specific}. A match from a specific sniffer
2329 overrides a match from a generic sniffer. Multiple sniffers for an
2330 architecture/flavour may exist, in order to differentiate between two
2331 different operating systems which use the same basic file format. The
2332 OS ABI framework provides a generic sniffer for ELF-format files which
2333 examines the @code{EI_OSABI} field of the ELF header, as well as note
2334 sections known to be used by several operating systems.
2336 @cindex fine-tuning @code{gdbarch} structure
2337 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2338 selected OS ABI. There may be only one handler for a given OS ABI
2339 for each BFD architecture.
2341 The following OS ABI variants are defined in @file{osabi.h}:
2345 @findex GDB_OSABI_UNKNOWN
2346 @item GDB_OSABI_UNKNOWN
2347 The ABI of the inferior is unknown. The default @code{gdbarch}
2348 settings for the architecture will be used.
2350 @findex GDB_OSABI_SVR4
2351 @item GDB_OSABI_SVR4
2352 UNIX System V Release 4
2354 @findex GDB_OSABI_HURD
2355 @item GDB_OSABI_HURD
2356 GNU using the Hurd kernel
2358 @findex GDB_OSABI_SOLARIS
2359 @item GDB_OSABI_SOLARIS
2362 @findex GDB_OSABI_OSF1
2363 @item GDB_OSABI_OSF1
2364 OSF/1, including Digital UNIX and Compaq Tru64 UNIX
2366 @findex GDB_OSABI_LINUX
2367 @item GDB_OSABI_LINUX
2368 GNU using the Linux kernel
2370 @findex GDB_OSABI_FREEBSD_AOUT
2371 @item GDB_OSABI_FREEBSD_AOUT
2372 FreeBSD using the a.out executable format
2374 @findex GDB_OSABI_FREEBSD_ELF
2375 @item GDB_OSABI_FREEBSD_ELF
2376 FreeBSD using the ELF executable format
2378 @findex GDB_OSABI_NETBSD_AOUT
2379 @item GDB_OSABI_NETBSD_AOUT
2380 NetBSD using the a.out executable format
2382 @findex GDB_OSABI_NETBSD_ELF
2383 @item GDB_OSABI_NETBSD_ELF
2384 NetBSD using the ELF executable format
2386 @findex GDB_OSABI_WINCE
2387 @item GDB_OSABI_WINCE
2390 @findex GDB_OSABI_GO32
2391 @item GDB_OSABI_GO32
2394 @findex GDB_OSABI_NETWARE
2395 @item GDB_OSABI_NETWARE
2398 @findex GDB_OSABI_ARM_EABI_V1
2399 @item GDB_OSABI_ARM_EABI_V1
2400 ARM Embedded ABI version 1
2402 @findex GDB_OSABI_ARM_EABI_V2
2403 @item GDB_OSABI_ARM_EABI_V2
2404 ARM Embedded ABI version 2
2406 @findex GDB_OSABI_ARM_APCS
2407 @item GDB_OSABI_ARM_APCS
2408 Generic ARM Procedure Call Standard
2412 Here are the functions that make up the OS ABI framework:
2414 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2415 Return the name of the OS ABI corresponding to @var{osabi}.
2418 @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}))
2419 Register the OS ABI handler specified by @var{init_osabi} for the
2420 architecture, machine type and OS ABI specified by @var{arch},
2421 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2422 machine type, which implies the architecture's default machine type,
2426 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2427 Register the OS ABI file sniffer specified by @var{sniffer} for the
2428 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2429 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2430 be generic, and is allowed to examine @var{flavour}-flavoured files for
2434 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2435 Examine the file described by @var{abfd} to determine its OS ABI.
2436 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2440 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2441 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2442 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2443 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2444 architecture, a warning will be issued and the debugging session will continue
2445 with the defaults already established for @var{gdbarch}.
2448 @section Registers and Memory
2450 @value{GDBN}'s model of the target machine is rather simple.
2451 @value{GDBN} assumes the machine includes a bank of registers and a
2452 block of memory. Each register may have a different size.
2454 @value{GDBN} does not have a magical way to match up with the
2455 compiler's idea of which registers are which; however, it is critical
2456 that they do match up accurately. The only way to make this work is
2457 to get accurate information about the order that the compiler uses,
2458 and to reflect that in the @code{REGISTER_NAME} and related macros.
2460 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2462 @section Pointers Are Not Always Addresses
2463 @cindex pointer representation
2464 @cindex address representation
2465 @cindex word-addressed machines
2466 @cindex separate data and code address spaces
2467 @cindex spaces, separate data and code address
2468 @cindex address spaces, separate data and code
2469 @cindex code pointers, word-addressed
2470 @cindex converting between pointers and addresses
2471 @cindex D10V addresses
2473 On almost all 32-bit architectures, the representation of a pointer is
2474 indistinguishable from the representation of some fixed-length number
2475 whose value is the byte address of the object pointed to. On such
2476 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2477 However, architectures with smaller word sizes are often cramped for
2478 address space, so they may choose a pointer representation that breaks this
2479 identity, and allows a larger code address space.
2481 For example, the Renesas D10V is a 16-bit VLIW processor whose
2482 instructions are 32 bits long@footnote{Some D10V instructions are
2483 actually pairs of 16-bit sub-instructions. However, since you can't
2484 jump into the middle of such a pair, code addresses can only refer to
2485 full 32 bit instructions, which is what matters in this explanation.}.
2486 If the D10V used ordinary byte addresses to refer to code locations,
2487 then the processor would only be able to address 64kb of instructions.
2488 However, since instructions must be aligned on four-byte boundaries, the
2489 low two bits of any valid instruction's byte address are always
2490 zero---byte addresses waste two bits. So instead of byte addresses,
2491 the D10V uses word addresses---byte addresses shifted right two bits---to
2492 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2495 However, this means that code pointers and data pointers have different
2496 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2497 @code{0xC020} when used as a data address, but refers to byte address
2498 @code{0x30080} when used as a code address.
2500 (The D10V also uses separate code and data address spaces, which also
2501 affects the correspondence between pointers and addresses, but we're
2502 going to ignore that here; this example is already too long.)
2504 To cope with architectures like this---the D10V is not the only
2505 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2506 byte numbers, and @dfn{pointers}, which are the target's representation
2507 of an address of a particular type of data. In the example above,
2508 @code{0xC020} is the pointer, which refers to one of the addresses
2509 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2510 @value{GDBN} provides functions for turning a pointer into an address
2511 and vice versa, in the appropriate way for the current architecture.
2513 Unfortunately, since addresses and pointers are identical on almost all
2514 processors, this distinction tends to bit-rot pretty quickly. Thus,
2515 each time you port @value{GDBN} to an architecture which does
2516 distinguish between pointers and addresses, you'll probably need to
2517 clean up some architecture-independent code.
2519 Here are functions which convert between pointers and addresses:
2521 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2522 Treat the bytes at @var{buf} as a pointer or reference of type
2523 @var{type}, and return the address it represents, in a manner
2524 appropriate for the current architecture. This yields an address
2525 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2526 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2529 For example, if the current architecture is the Intel x86, this function
2530 extracts a little-endian integer of the appropriate length from
2531 @var{buf} and returns it. However, if the current architecture is the
2532 D10V, this function will return a 16-bit integer extracted from
2533 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2535 If @var{type} is not a pointer or reference type, then this function
2536 will signal an internal error.
2539 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2540 Store the address @var{addr} in @var{buf}, in the proper format for a
2541 pointer of type @var{type} in the current architecture. Note that
2542 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2545 For example, if the current architecture is the Intel x86, this function
2546 stores @var{addr} unmodified as a little-endian integer of the
2547 appropriate length in @var{buf}. However, if the current architecture
2548 is the D10V, this function divides @var{addr} by four if @var{type} is
2549 a pointer to a function, and then stores it in @var{buf}.
2551 If @var{type} is not a pointer or reference type, then this function
2552 will signal an internal error.
2555 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2556 Assuming that @var{val} is a pointer, return the address it represents,
2557 as appropriate for the current architecture.
2559 This function actually works on integral values, as well as pointers.
2560 For pointers, it performs architecture-specific conversions as
2561 described above for @code{extract_typed_address}.
2564 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2565 Create and return a value representing a pointer of type @var{type} to
2566 the address @var{addr}, as appropriate for the current architecture.
2567 This function performs architecture-specific conversions as described
2568 above for @code{store_typed_address}.
2571 Here are some macros which architectures can define to indicate the
2572 relationship between pointers and addresses. These have default
2573 definitions, appropriate for architectures on which all pointers are
2574 simple unsigned byte addresses.
2576 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2577 Assume that @var{buf} holds a pointer of type @var{type}, in the
2578 appropriate format for the current architecture. Return the byte
2579 address the pointer refers to.
2581 This function may safely assume that @var{type} is either a pointer or a
2582 C@t{++} reference type.
2585 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2586 Store in @var{buf} a pointer of type @var{type} representing the address
2587 @var{addr}, in the appropriate format for the current architecture.
2589 This function may safely assume that @var{type} is either a pointer or a
2590 C@t{++} reference type.
2593 @section Address Classes
2594 @cindex address classes
2595 @cindex DW_AT_byte_size
2596 @cindex DW_AT_address_class
2598 Sometimes information about different kinds of addresses is available
2599 via the debug information. For example, some programming environments
2600 define addresses of several different sizes. If the debug information
2601 distinguishes these kinds of address classes through either the size
2602 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2603 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2604 following macros should be defined in order to disambiguate these
2605 types within @value{GDBN} as well as provide the added information to
2606 a @value{GDBN} user when printing type expressions.
2608 @deftypefn {Target Macro} int ADDRESS_CLASS_TYPE_FLAGS (int @var{byte_size}, int @var{dwarf2_addr_class})
2609 Returns the type flags needed to construct a pointer type whose size
2610 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2611 This function is normally called from within a symbol reader. See
2612 @file{dwarf2read.c}.
2615 @deftypefn {Target Macro} char *ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (int @var{type_flags})
2616 Given the type flags representing an address class qualifier, return
2619 @deftypefn {Target Macro} int ADDRESS_CLASS_NAME_to_TYPE_FLAGS (int @var{name}, int *var{type_flags_ptr})
2620 Given an address qualifier name, set the @code{int} refererenced by @var{type_flags_ptr} to the type flags
2621 for that address class qualifier.
2624 Since the need for address classes is rather rare, none of
2625 the address class macros defined by default. Predicate
2626 macros are provided to detect when they are defined.
2628 Consider a hypothetical architecture in which addresses are normally
2629 32-bits wide, but 16-bit addresses are also supported. Furthermore,
2630 suppose that the @w{DWARF 2} information for this architecture simply
2631 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
2632 of these "short" pointers. The following functions could be defined
2633 to implement the address class macros:
2636 somearch_address_class_type_flags (int byte_size,
2637 int dwarf2_addr_class)
2640 return TYPE_FLAG_ADDRESS_CLASS_1;
2646 somearch_address_class_type_flags_to_name (int type_flags)
2648 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
2655 somearch_address_class_name_to_type_flags (char *name,
2656 int *type_flags_ptr)
2658 if (strcmp (name, "short") == 0)
2660 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
2668 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
2669 to indicate the presence of one of these "short" pointers. E.g, if
2670 the debug information indicates that @code{short_ptr_var} is one of these
2671 short pointers, @value{GDBN} might show the following behavior:
2674 (gdb) ptype short_ptr_var
2675 type = int * @@short
2679 @section Raw and Virtual Register Representations
2680 @cindex raw register representation
2681 @cindex virtual register representation
2682 @cindex representations, raw and virtual registers
2684 @emph{Maintainer note: This section is pretty much obsolete. The
2685 functionality described here has largely been replaced by
2686 pseudo-registers and the mechanisms described in @ref{Target
2687 Architecture Definition, , Using Different Register and Memory Data
2688 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
2689 Bug Tracking Database} and
2690 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
2691 up-to-date information.}
2693 Some architectures use one representation for a value when it lives in a
2694 register, but use a different representation when it lives in memory.
2695 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
2696 the target registers, and the @dfn{virtual} representation is the one
2697 used in memory, and within @value{GDBN} @code{struct value} objects.
2699 @emph{Maintainer note: Notice that the same mechanism is being used to
2700 both convert a register to a @code{struct value} and alternative
2703 For almost all data types on almost all architectures, the virtual and
2704 raw representations are identical, and no special handling is needed.
2705 However, they do occasionally differ. For example:
2709 The x86 architecture supports an 80-bit @code{long double} type. However, when
2710 we store those values in memory, they occupy twelve bytes: the
2711 floating-point number occupies the first ten, and the final two bytes
2712 are unused. This keeps the values aligned on four-byte boundaries,
2713 allowing more efficient access. Thus, the x86 80-bit floating-point
2714 type is the raw representation, and the twelve-byte loosely-packed
2715 arrangement is the virtual representation.
2718 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
2719 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
2720 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
2721 raw representation, and the trimmed 32-bit representation is the
2722 virtual representation.
2725 In general, the raw representation is determined by the architecture, or
2726 @value{GDBN}'s interface to the architecture, while the virtual representation
2727 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
2728 @code{registers}, holds the register contents in raw format, and the
2729 @value{GDBN} remote protocol transmits register values in raw format.
2731 Your architecture may define the following macros to request
2732 conversions between the raw and virtual format:
2734 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
2735 Return non-zero if register number @var{reg}'s value needs different raw
2736 and virtual formats.
2738 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
2739 unless this macro returns a non-zero value for that register.
2742 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
2743 The size of register number @var{reg}'s raw value. This is the number
2744 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
2745 remote protocol packet.
2748 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
2749 The size of register number @var{reg}'s value, in its virtual format.
2750 This is the size a @code{struct value}'s buffer will have, holding that
2754 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
2755 This is the type of the virtual representation of register number
2756 @var{reg}. Note that there is no need for a macro giving a type for the
2757 register's raw form; once the register's value has been obtained, @value{GDBN}
2758 always uses the virtual form.
2761 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2762 Convert the value of register number @var{reg} to @var{type}, which
2763 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2764 at @var{from} holds the register's value in raw format; the macro should
2765 convert the value to virtual format, and place it at @var{to}.
2767 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
2768 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
2769 arguments in different orders.
2771 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
2772 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
2776 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2777 Convert the value of register number @var{reg} to @var{type}, which
2778 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2779 at @var{from} holds the register's value in raw format; the macro should
2780 convert the value to virtual format, and place it at @var{to}.
2782 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
2783 their @var{reg} and @var{type} arguments in different orders.
2787 @section Using Different Register and Memory Data Representations
2788 @cindex register representation
2789 @cindex memory representation
2790 @cindex representations, register and memory
2791 @cindex register data formats, converting
2792 @cindex @code{struct value}, converting register contents to
2794 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
2795 significant change. Many of the macros and functions refered to in this
2796 section are likely to be subject to further revision. See
2797 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
2798 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
2799 further information. cagney/2002-05-06.}
2801 Some architectures can represent a data object in a register using a
2802 form that is different to the objects more normal memory representation.
2808 The Alpha architecture can represent 32 bit integer values in
2809 floating-point registers.
2812 The x86 architecture supports 80-bit floating-point registers. The
2813 @code{long double} data type occupies 96 bits in memory but only 80 bits
2814 when stored in a register.
2818 In general, the register representation of a data type is determined by
2819 the architecture, or @value{GDBN}'s interface to the architecture, while
2820 the memory representation is determined by the Application Binary
2823 For almost all data types on almost all architectures, the two
2824 representations are identical, and no special handling is needed.
2825 However, they do occasionally differ. Your architecture may define the
2826 following macros to request conversions between the register and memory
2827 representations of a data type:
2829 @deftypefn {Target Macro} int CONVERT_REGISTER_P (int @var{reg})
2830 Return non-zero if the representation of a data value stored in this
2831 register may be different to the representation of that same data value
2832 when stored in memory.
2834 When non-zero, the macros @code{REGISTER_TO_VALUE} and
2835 @code{VALUE_TO_REGISTER} are used to perform any necessary conversion.
2838 @deftypefn {Target Macro} void REGISTER_TO_VALUE (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2839 Convert the value of register number @var{reg} to a data object of type
2840 @var{type}. The buffer at @var{from} holds the register's value in raw
2841 format; the converted value should be placed in the buffer at @var{to}.
2843 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
2844 their @var{reg} and @var{type} arguments in different orders.
2846 You should only use @code{REGISTER_TO_VALUE} with registers for which
2847 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
2850 @deftypefn {Target Macro} void VALUE_TO_REGISTER (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2851 Convert a data value of type @var{type} to register number @var{reg}'
2854 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
2855 their @var{reg} and @var{type} arguments in different orders.
2857 You should only use @code{VALUE_TO_REGISTER} with registers for which
2858 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
2861 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
2862 See @file{mips-tdep.c}. It does not do what you want.
2866 @section Frame Interpretation
2868 @section Inferior Call Setup
2870 @section Compiler Characteristics
2872 @section Target Conditionals
2874 This section describes the macros that you can use to define the target
2879 @item ADDR_BITS_REMOVE (addr)
2880 @findex ADDR_BITS_REMOVE
2881 If a raw machine instruction address includes any bits that are not
2882 really part of the address, then define this macro to expand into an
2883 expression that zeroes those bits in @var{addr}. This is only used for
2884 addresses of instructions, and even then not in all contexts.
2886 For example, the two low-order bits of the PC on the Hewlett-Packard PA
2887 2.0 architecture contain the privilege level of the corresponding
2888 instruction. Since instructions must always be aligned on four-byte
2889 boundaries, the processor masks out these bits to generate the actual
2890 address of the instruction. ADDR_BITS_REMOVE should filter out these
2891 bits with an expression such as @code{((addr) & ~3)}.
2893 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS (@var{name}, @var{type_flags_ptr})
2894 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS
2895 If @var{name} is a valid address class qualifier name, set the @code{int}
2896 referenced by @var{type_flags_ptr} to the mask representing the qualifier
2897 and return 1. If @var{name} is not a valid address class qualifier name,
2900 The value for @var{type_flags_ptr} should be one of
2901 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
2902 possibly some combination of these values or'd together.
2903 @xref{Target Architecture Definition, , Address Classes}.
2905 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P ()
2906 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P
2907 Predicate which indicates whether @code{ADDRESS_CLASS_NAME_TO_TYPE_FLAGS}
2910 @item ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
2911 @findex ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
2912 Given a pointers byte size (as described by the debug information) and
2913 the possible @code{DW_AT_address_class} value, return the type flags
2914 used by @value{GDBN} to represent this address class. The value
2915 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
2916 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
2917 values or'd together.
2918 @xref{Target Architecture Definition, , Address Classes}.
2920 @item ADDRESS_CLASS_TYPE_FLAGS_P ()
2921 @findex ADDRESS_CLASS_TYPE_FLAGS_P
2922 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS} has
2925 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (@var{type_flags})
2926 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME
2927 Return the name of the address class qualifier associated with the type
2928 flags given by @var{type_flags}.
2930 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P ()
2931 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P
2932 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS_TO_NAME} has
2934 @xref{Target Architecture Definition, , Address Classes}.
2936 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
2937 @findex ADDRESS_TO_POINTER
2938 Store in @var{buf} a pointer of type @var{type} representing the address
2939 @var{addr}, in the appropriate format for the current architecture.
2940 This macro may safely assume that @var{type} is either a pointer or a
2941 C@t{++} reference type.
2942 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
2944 @item BELIEVE_PCC_PROMOTION
2945 @findex BELIEVE_PCC_PROMOTION
2946 Define if the compiler promotes a @code{short} or @code{char}
2947 parameter to an @code{int}, but still reports the parameter as its
2948 original type, rather than the promoted type.
2950 @item BITS_BIG_ENDIAN
2951 @findex BITS_BIG_ENDIAN
2952 Define this if the numbering of bits in the targets does @strong{not} match the
2953 endianness of the target byte order. A value of 1 means that the bits
2954 are numbered in a big-endian bit order, 0 means little-endian.
2958 This is the character array initializer for the bit pattern to put into
2959 memory where a breakpoint is set. Although it's common to use a trap
2960 instruction for a breakpoint, it's not required; for instance, the bit
2961 pattern could be an invalid instruction. The breakpoint must be no
2962 longer than the shortest instruction of the architecture.
2964 @code{BREAKPOINT} has been deprecated in favor of
2965 @code{BREAKPOINT_FROM_PC}.
2967 @item BIG_BREAKPOINT
2968 @itemx LITTLE_BREAKPOINT
2969 @findex LITTLE_BREAKPOINT
2970 @findex BIG_BREAKPOINT
2971 Similar to BREAKPOINT, but used for bi-endian targets.
2973 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
2974 favor of @code{BREAKPOINT_FROM_PC}.
2976 @item DEPRECATED_REMOTE_BREAKPOINT
2977 @itemx DEPRECATED_LITTLE_REMOTE_BREAKPOINT
2978 @itemx DEPRECATED_BIG_REMOTE_BREAKPOINT
2979 @findex DEPRECATED_BIG_REMOTE_BREAKPOINT
2980 @findex DEPRECATED_LITTLE_REMOTE_BREAKPOINT
2981 @findex DEPRECATED_REMOTE_BREAKPOINT
2982 Specify the breakpoint instruction sequence for a remote target.
2983 @code{DEPRECATED_REMOTE_BREAKPOINT},
2984 @code{DEPRECATED_BIG_REMOTE_BREAKPOINT} and
2985 @code{DEPRECATED_LITTLE_REMOTE_BREAKPOINT} have been deprecated in
2986 favor of @code{BREAKPOINT_FROM_PC} (@pxref{BREAKPOINT_FROM_PC}).
2988 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
2989 @findex BREAKPOINT_FROM_PC
2990 @anchor{BREAKPOINT_FROM_PC} Use the program counter to determine the
2991 contents and size of a breakpoint instruction. It returns a pointer to
2992 a string of bytes that encode a breakpoint instruction, stores the
2993 length of the string to @code{*@var{lenptr}}, and adjusts the program
2994 counter (if necessary) to point to the actual memory location where the
2995 breakpoint should be inserted.
2997 Although it is common to use a trap instruction for a breakpoint, it's
2998 not required; for instance, the bit pattern could be an invalid
2999 instruction. The breakpoint must be no longer than the shortest
3000 instruction of the architecture.
3002 Replaces all the other @var{BREAKPOINT} macros.
3004 @item MEMORY_INSERT_BREAKPOINT (@var{addr}, @var{contents_cache})
3005 @itemx MEMORY_REMOVE_BREAKPOINT (@var{addr}, @var{contents_cache})
3006 @findex MEMORY_REMOVE_BREAKPOINT
3007 @findex MEMORY_INSERT_BREAKPOINT
3008 Insert or remove memory based breakpoints. Reasonable defaults
3009 (@code{default_memory_insert_breakpoint} and
3010 @code{default_memory_remove_breakpoint} respectively) have been
3011 provided so that it is not necessary to define these for most
3012 architectures. Architectures which may want to define
3013 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
3014 likely have instructions that are oddly sized or are not stored in a
3015 conventional manner.
3017 It may also be desirable (from an efficiency standpoint) to define
3018 custom breakpoint insertion and removal routines if
3019 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
3022 @item ADJUST_BREAKPOINT_ADDRESS (@var{address})
3023 @findex ADJUST_BREAKPOINT_ADDRESS
3024 @cindex breakpoint address adjusted
3025 Given an address at which a breakpoint is desired, return a breakpoint
3026 address adjusted to account for architectural constraints on
3027 breakpoint placement. This method is not needed by most targets.
3029 The FR-V target (see @file{frv-tdep.c}) requires this method.
3030 The FR-V is a VLIW architecture in which a number of RISC-like
3031 instructions are grouped (packed) together into an aggregate
3032 instruction or instruction bundle. When the processor executes
3033 one of these bundles, the component instructions are executed
3036 In the course of optimization, the compiler may group instructions
3037 from distinct source statements into the same bundle. The line number
3038 information associated with one of the latter statements will likely
3039 refer to some instruction other than the first one in the bundle. So,
3040 if the user attempts to place a breakpoint on one of these latter
3041 statements, @value{GDBN} must be careful to @emph{not} place the break
3042 instruction on any instruction other than the first one in the bundle.
3043 (Remember though that the instructions within a bundle execute
3044 in parallel, so the @emph{first} instruction is the instruction
3045 at the lowest address and has nothing to do with execution order.)
3047 The FR-V's @code{ADJUST_BREAKPOINT_ADDRESS} method will adjust a
3048 breakpoint's address by scanning backwards for the beginning of
3049 the bundle, returning the address of the bundle.
3051 Since the adjustment of a breakpoint may significantly alter a user's
3052 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3053 is initially set and each time that that breakpoint is hit.
3055 @item CALL_DUMMY_LOCATION
3056 @findex CALL_DUMMY_LOCATION
3057 See the file @file{inferior.h}.
3059 This method has been replaced by @code{push_dummy_code}
3060 (@pxref{push_dummy_code}).
3062 @item CANNOT_FETCH_REGISTER (@var{regno})
3063 @findex CANNOT_FETCH_REGISTER
3064 A C expression that should be nonzero if @var{regno} cannot be fetched
3065 from an inferior process. This is only relevant if
3066 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3068 @item CANNOT_STORE_REGISTER (@var{regno})
3069 @findex CANNOT_STORE_REGISTER
3070 A C expression that should be nonzero if @var{regno} should not be
3071 written to the target. This is often the case for program counters,
3072 status words, and other special registers. If this is not defined,
3073 @value{GDBN} will assume that all registers may be written.
3075 @item int CONVERT_REGISTER_P(@var{regnum})
3076 @findex CONVERT_REGISTER_P
3077 Return non-zero if register @var{regnum} can represent data values in a
3079 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3081 @item DECR_PC_AFTER_BREAK
3082 @findex DECR_PC_AFTER_BREAK
3083 Define this to be the amount by which to decrement the PC after the
3084 program encounters a breakpoint. This is often the number of bytes in
3085 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3087 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3088 @findex DISABLE_UNSETTABLE_BREAK
3089 If defined, this should evaluate to 1 if @var{addr} is in a shared
3090 library in which breakpoints cannot be set and so should be disabled.
3092 @item PRINT_FLOAT_INFO()
3093 @findex PRINT_FLOAT_INFO
3094 If defined, then the @samp{info float} command will print information about
3095 the processor's floating point unit.
3097 @item print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3098 @findex print_registers_info
3099 If defined, pretty print the value of the register @var{regnum} for the
3100 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3101 either all registers (@var{all} is non zero) or a select subset of
3102 registers (@var{all} is zero).
3104 The default method prints one register per line, and if @var{all} is
3105 zero omits floating-point registers.
3107 @item PRINT_VECTOR_INFO()
3108 @findex PRINT_VECTOR_INFO
3109 If defined, then the @samp{info vector} command will call this function
3110 to print information about the processor's vector unit.
3112 By default, the @samp{info vector} command will print all vector
3113 registers (the register's type having the vector attribute).
3115 @item DWARF_REG_TO_REGNUM
3116 @findex DWARF_REG_TO_REGNUM
3117 Convert DWARF register number into @value{GDBN} regnum. If not defined,
3118 no conversion will be performed.
3120 @item DWARF2_REG_TO_REGNUM
3121 @findex DWARF2_REG_TO_REGNUM
3122 Convert DWARF2 register number into @value{GDBN} regnum. If not
3123 defined, no conversion will be performed.
3125 @item ECOFF_REG_TO_REGNUM
3126 @findex ECOFF_REG_TO_REGNUM
3127 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
3128 no conversion will be performed.
3130 @item END_OF_TEXT_DEFAULT
3131 @findex END_OF_TEXT_DEFAULT
3132 This is an expression that should designate the end of the text section.
3135 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
3136 @findex EXTRACT_RETURN_VALUE
3137 Define this to extract a function's return value of type @var{type} from
3138 the raw register state @var{regbuf} and copy that, in virtual format,
3141 This method has been deprecated in favour of @code{gdbarch_return_value}
3142 (@pxref{gdbarch_return_value}).
3144 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
3145 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS
3146 @anchor{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}
3147 When defined, extract from the array @var{regbuf} (containing the raw
3148 register state) the @code{CORE_ADDR} at which a function should return
3149 its structure value.
3151 @xref{gdbarch_return_value}.
3153 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P()
3154 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P
3155 Predicate for @code{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}.
3157 @item DEPRECATED_FP_REGNUM
3158 @findex DEPRECATED_FP_REGNUM
3159 If the virtual frame pointer is kept in a register, then define this
3160 macro to be the number (greater than or equal to zero) of that register.
3162 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3165 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3166 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3167 Define this to an expression that returns 1 if the function invocation
3168 represented by @var{fi} does not have a stack frame associated with it.
3171 @item frame_align (@var{address})
3172 @anchor{frame_align}
3174 Define this to adjust @var{address} so that it meets the alignment
3175 requirements for the start of a new stack frame. A stack frame's
3176 alignment requirements are typically stronger than a target processors
3177 stack alignment requirements (@pxref{DEPRECATED_STACK_ALIGN}).
3179 This function is used to ensure that, when creating a dummy frame, both
3180 the initial stack pointer and (if needed) the address of the return
3181 value are correctly aligned.
3183 Unlike @code{DEPRECATED_STACK_ALIGN}, this function always adjusts the
3184 address in the direction of stack growth.
3186 By default, no frame based stack alignment is performed.
3188 @item int frame_red_zone_size
3190 The number of bytes, beyond the innermost-stack-address, reserved by the
3191 @sc{abi}. A function is permitted to use this scratch area (instead of
3192 allocating extra stack space).
3194 When performing an inferior function call, to ensure that it does not
3195 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3196 @var{frame_red_zone_size} bytes before pushing parameters onto the
3199 By default, zero bytes are allocated. The value must be aligned
3200 (@pxref{frame_align}).
3202 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3203 @emph{red zone} when describing this scratch area.
3206 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3207 @findex DEPRECATED_FRAME_CHAIN
3208 Given @var{frame}, return a pointer to the calling frame.
3210 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3211 @findex DEPRECATED_FRAME_CHAIN_VALID
3212 Define this to be an expression that returns zero if the given frame is an
3213 outermost frame, with no caller, and nonzero otherwise. Most normal
3214 situations can be handled without defining this macro, including @code{NULL}
3215 chain pointers, dummy frames, and frames whose PC values are inside the
3216 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3219 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3220 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3221 See @file{frame.h}. Determines the address of all registers in the
3222 current stack frame storing each in @code{frame->saved_regs}. Space for
3223 @code{frame->saved_regs} shall be allocated by
3224 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3225 @code{frame_saved_regs_zalloc}.
3227 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3229 @item FRAME_NUM_ARGS (@var{fi})
3230 @findex FRAME_NUM_ARGS
3231 For the frame described by @var{fi} return the number of arguments that
3232 are being passed. If the number of arguments is not known, return
3235 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3236 @findex DEPRECATED_FRAME_SAVED_PC
3237 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3238 saved there. This is the return address.
3240 This method is deprecated. @xref{unwind_pc}.
3242 @item CORE_ADDR unwind_pc (struct frame_info *@var{this_frame})
3244 @anchor{unwind_pc} Return the instruction address, in @var{this_frame}'s
3245 caller, at which execution will resume after @var{this_frame} returns.
3246 This is commonly refered to as the return address.
3248 The implementation, which must be frame agnostic (work with any frame),
3249 is typically no more than:
3253 frame_unwind_unsigned_register (this_frame, D10V_PC_REGNUM, &pc);
3254 return d10v_make_iaddr (pc);
3258 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3260 @item CORE_ADDR unwind_sp (struct frame_info *@var{this_frame})
3262 @anchor{unwind_sp} Return the frame's inner most stack address. This is
3263 commonly refered to as the frame's @dfn{stack pointer}.
3265 The implementation, which must be frame agnostic (work with any frame),
3266 is typically no more than:
3270 frame_unwind_unsigned_register (this_frame, D10V_SP_REGNUM, &sp);
3271 return d10v_make_daddr (sp);
3275 @xref{TARGET_READ_SP}, which this method replaces.
3277 @item FUNCTION_EPILOGUE_SIZE
3278 @findex FUNCTION_EPILOGUE_SIZE
3279 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3280 function end symbol is 0. For such targets, you must define
3281 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3282 function's epilogue.
3284 @item DEPRECATED_FUNCTION_START_OFFSET
3285 @findex DEPRECATED_FUNCTION_START_OFFSET
3286 An integer, giving the offset in bytes from a function's address (as
3287 used in the values of symbols, function pointers, etc.), and the
3288 function's first genuine instruction.
3290 This is zero on almost all machines: the function's address is usually
3291 the address of its first instruction. However, on the VAX, for
3292 example, each function starts with two bytes containing a bitmask
3293 indicating which registers to save upon entry to the function. The
3294 VAX @code{call} instructions check this value, and save the
3295 appropriate registers automatically. Thus, since the offset from the
3296 function's address to its first instruction is two bytes,
3297 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3299 @item GCC_COMPILED_FLAG_SYMBOL
3300 @itemx GCC2_COMPILED_FLAG_SYMBOL
3301 @findex GCC2_COMPILED_FLAG_SYMBOL
3302 @findex GCC_COMPILED_FLAG_SYMBOL
3303 If defined, these are the names of the symbols that @value{GDBN} will
3304 look for to detect that GCC compiled the file. The default symbols
3305 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3306 respectively. (Currently only defined for the Delta 68.)
3308 @item @value{GDBN}_MULTI_ARCH
3309 @findex @value{GDBN}_MULTI_ARCH
3310 If defined and non-zero, enables support for multiple architectures
3311 within @value{GDBN}.
3313 This support can be enabled at two levels. At level one, only
3314 definitions for previously undefined macros are provided; at level two,
3315 a multi-arch definition of all architecture dependent macros will be
3318 @item @value{GDBN}_TARGET_IS_HPPA
3319 @findex @value{GDBN}_TARGET_IS_HPPA
3320 This determines whether horrible kludge code in @file{dbxread.c} and
3321 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3322 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3325 @item GET_LONGJMP_TARGET
3326 @findex GET_LONGJMP_TARGET
3327 For most machines, this is a target-dependent parameter. On the
3328 DECstation and the Iris, this is a native-dependent parameter, since
3329 the header file @file{setjmp.h} is needed to define it.
3331 This macro determines the target PC address that @code{longjmp} will jump to,
3332 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3333 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3334 pointer. It examines the current state of the machine as needed.
3336 @item DEPRECATED_GET_SAVED_REGISTER
3337 @findex DEPRECATED_GET_SAVED_REGISTER
3338 Define this if you need to supply your own definition for the function
3339 @code{DEPRECATED_GET_SAVED_REGISTER}.
3341 @item DEPRECATED_IBM6000_TARGET
3342 @findex DEPRECATED_IBM6000_TARGET
3343 Shows that we are configured for an IBM RS/6000 system. This
3344 conditional should be eliminated (FIXME) and replaced by
3345 feature-specific macros. It was introduced in a haste and we are
3346 repenting at leisure.
3348 @item I386_USE_GENERIC_WATCHPOINTS
3349 An x86-based target can define this to use the generic x86 watchpoint
3350 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3352 @item SYMBOLS_CAN_START_WITH_DOLLAR
3353 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3354 Some systems have routines whose names start with @samp{$}. Giving this
3355 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3356 routines when parsing tokens that begin with @samp{$}.
3358 On HP-UX, certain system routines (millicode) have names beginning with
3359 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3360 routine that handles inter-space procedure calls on PA-RISC.
3362 @item DEPRECATED_INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3363 @findex DEPRECATED_INIT_EXTRA_FRAME_INFO
3364 If additional information about the frame is required this should be
3365 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3366 is allocated using @code{frame_extra_info_zalloc}.
3368 @item DEPRECATED_INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3369 @findex DEPRECATED_INIT_FRAME_PC
3370 This is a C statement that sets the pc of the frame pointed to by
3371 @var{prev}. [By default...]
3373 @item INNER_THAN (@var{lhs}, @var{rhs})
3375 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3376 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3377 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3380 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3381 @findex gdbarch_in_function_epilogue_p
3382 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3383 The epilogue of a function is defined as the part of a function where
3384 the stack frame of the function already has been destroyed up to the
3385 final `return from function call' instruction.
3387 @item DEPRECATED_SIGTRAMP_START (@var{pc})
3388 @findex DEPRECATED_SIGTRAMP_START
3389 @itemx DEPRECATED_SIGTRAMP_END (@var{pc})
3390 @findex DEPRECATED_SIGTRAMP_END
3391 Define these to be the start and end address of the @code{sigtramp} for the
3392 given @var{pc}. On machines where the address is just a compile time
3393 constant, the macro expansion will typically just ignore the supplied
3396 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3397 @findex IN_SOLIB_CALL_TRAMPOLINE
3398 Define this to evaluate to nonzero if the program is stopped in the
3399 trampoline that connects to a shared library.
3401 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3402 @findex IN_SOLIB_RETURN_TRAMPOLINE
3403 Define this to evaluate to nonzero if the program is stopped in the
3404 trampoline that returns from a shared library.
3406 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3407 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3408 Define this to evaluate to nonzero if the program is stopped in the
3411 @item SKIP_SOLIB_RESOLVER (@var{pc})
3412 @findex SKIP_SOLIB_RESOLVER
3413 Define this to evaluate to the (nonzero) address at which execution
3414 should continue to get past the dynamic linker's symbol resolution
3415 function. A zero value indicates that it is not important or necessary
3416 to set a breakpoint to get through the dynamic linker and that single
3417 stepping will suffice.
3419 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3420 @findex INTEGER_TO_ADDRESS
3421 @cindex converting integers to addresses
3422 Define this when the architecture needs to handle non-pointer to address
3423 conversions specially. Converts that value to an address according to
3424 the current architectures conventions.
3426 @emph{Pragmatics: When the user copies a well defined expression from
3427 their source code and passes it, as a parameter, to @value{GDBN}'s
3428 @code{print} command, they should get the same value as would have been
3429 computed by the target program. Any deviation from this rule can cause
3430 major confusion and annoyance, and needs to be justified carefully. In
3431 other words, @value{GDBN} doesn't really have the freedom to do these
3432 conversions in clever and useful ways. It has, however, been pointed
3433 out that users aren't complaining about how @value{GDBN} casts integers
3434 to pointers; they are complaining that they can't take an address from a
3435 disassembly listing and give it to @code{x/i}. Adding an architecture
3436 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3437 @value{GDBN} to ``get it right'' in all circumstances.}
3439 @xref{Target Architecture Definition, , Pointers Are Not Always
3442 @item NO_HIF_SUPPORT
3443 @findex NO_HIF_SUPPORT
3444 (Specific to the a29k.)
3446 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3447 @findex POINTER_TO_ADDRESS
3448 Assume that @var{buf} holds a pointer of type @var{type}, in the
3449 appropriate format for the current architecture. Return the byte
3450 address the pointer refers to.
3451 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3453 @item REGISTER_CONVERTIBLE (@var{reg})
3454 @findex REGISTER_CONVERTIBLE
3455 Return non-zero if @var{reg} uses different raw and virtual formats.
3456 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3458 @item REGISTER_TO_VALUE(@var{regnum}, @var{type}, @var{from}, @var{to})
3459 @findex REGISTER_TO_VALUE
3460 Convert the raw contents of register @var{regnum} into a value of type
3462 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3464 @item DEPRECATED_REGISTER_RAW_SIZE (@var{reg})
3465 @findex DEPRECATED_REGISTER_RAW_SIZE
3466 Return the raw size of @var{reg}; defaults to the size of the register's
3468 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3470 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3471 @findex register_reggroup_p
3472 @cindex register groups
3473 Return non-zero if register @var{regnum} is a member of the register
3474 group @var{reggroup}.
3476 By default, registers are grouped as follows:
3479 @item float_reggroup
3480 Any register with a valid name and a floating-point type.
3481 @item vector_reggroup
3482 Any register with a valid name and a vector type.
3483 @item general_reggroup
3484 Any register with a valid name and a type other than vector or
3485 floating-point. @samp{float_reggroup}.
3487 @itemx restore_reggroup
3489 Any register with a valid name.
3492 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3493 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3494 Return the virtual size of @var{reg}; defaults to the size of the
3495 register's virtual type.
3496 Return the virtual size of @var{reg}.
3497 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3499 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3500 @findex REGISTER_VIRTUAL_TYPE
3501 Return the virtual type of @var{reg}.
3502 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3504 @item struct type *register_type (@var{gdbarch}, @var{reg})
3505 @findex register_type
3506 If defined, return the type of register @var{reg}. This function
3507 superseeds @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3508 Definition, , Raw and Virtual Register Representations}.
3510 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3511 @findex REGISTER_CONVERT_TO_VIRTUAL
3512 Convert the value of register @var{reg} from its raw form to its virtual
3514 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3516 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3517 @findex REGISTER_CONVERT_TO_RAW
3518 Convert the value of register @var{reg} from its virtual form to its raw
3520 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3522 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3523 @findex regset_from_core_section
3524 Return the appropriate register set for a core file section with name
3525 @var{sect_name} and size @var{sect_size}.
3527 @item SOFTWARE_SINGLE_STEP_P()
3528 @findex SOFTWARE_SINGLE_STEP_P
3529 Define this as 1 if the target does not have a hardware single-step
3530 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3532 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3533 @findex SOFTWARE_SINGLE_STEP
3534 A function that inserts or removes (depending on
3535 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3536 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3539 @item SOFUN_ADDRESS_MAYBE_MISSING
3540 @findex SOFUN_ADDRESS_MAYBE_MISSING
3541 Somebody clever observed that, the more actual addresses you have in the
3542 debug information, the more time the linker has to spend relocating
3543 them. So whenever there's some other way the debugger could find the
3544 address it needs, you should omit it from the debug info, to make
3547 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3548 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3549 entries in stabs-format debugging information. @code{N_SO} stabs mark
3550 the beginning and ending addresses of compilation units in the text
3551 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3553 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3557 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3558 addresses where the function starts by taking the function name from
3559 the stab, and then looking that up in the minsyms (the
3560 linker/assembler symbol table). In other words, the stab has the
3561 name, and the linker/assembler symbol table is the only place that carries
3565 @code{N_SO} stabs have an address of zero, too. You just look at the
3566 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3567 and guess the starting and ending addresses of the compilation unit from
3571 @item PC_LOAD_SEGMENT
3572 @findex PC_LOAD_SEGMENT
3573 If defined, print information about the load segment for the program
3574 counter. (Defined only for the RS/6000.)
3578 If the program counter is kept in a register, then define this macro to
3579 be the number (greater than or equal to zero) of that register.
3581 This should only need to be defined if @code{TARGET_READ_PC} and
3582 @code{TARGET_WRITE_PC} are not defined.
3585 @findex PARM_BOUNDARY
3586 If non-zero, round arguments to a boundary of this many bits before
3587 pushing them on the stack.
3589 @item stabs_argument_has_addr (@var{gdbarch}, @var{type})
3590 @findex stabs_argument_has_addr
3591 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3592 @anchor{stabs_argument_has_addr} Define this to return nonzero if a
3593 function argument of type @var{type} is passed by reference instead of
3596 This method replaces @code{DEPRECATED_REG_STRUCT_HAS_ADDR}
3597 (@pxref{DEPRECATED_REG_STRUCT_HAS_ADDR}).
3599 @item PROCESS_LINENUMBER_HOOK
3600 @findex PROCESS_LINENUMBER_HOOK
3601 A hook defined for XCOFF reading.
3603 @item PROLOGUE_FIRSTLINE_OVERLAP
3604 @findex PROLOGUE_FIRSTLINE_OVERLAP
3605 (Only used in unsupported Convex configuration.)
3609 If defined, this is the number of the processor status register. (This
3610 definition is only used in generic code when parsing "$ps".)
3612 @item DEPRECATED_POP_FRAME
3613 @findex DEPRECATED_POP_FRAME
3615 If defined, used by @code{frame_pop} to remove a stack frame. This
3616 method has been superseeded by generic code.
3618 @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})
3619 @findex push_dummy_call
3620 @findex DEPRECATED_PUSH_ARGUMENTS.
3621 @anchor{push_dummy_call} Define this to push the dummy frame's call to
3622 the inferior function onto the stack. In addition to pushing
3623 @var{nargs}, the code should push @var{struct_addr} (when
3624 @var{struct_return}), and the return address (@var{bp_addr}).
3626 @var{function} is a pointer to a @code{struct value}; on architectures that use
3627 function descriptors, this contains the function descriptor value.
3629 Returns the updated top-of-stack pointer.
3631 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3633 @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})
3634 @findex push_dummy_code
3635 @anchor{push_dummy_code} Given a stack based call dummy, push the
3636 instruction sequence (including space for a breakpoint) to which the
3637 called function should return.
3639 Set @var{bp_addr} to the address at which the breakpoint instruction
3640 should be inserted, @var{real_pc} to the resume address when starting
3641 the call sequence, and return the updated inner-most stack address.
3643 By default, the stack is grown sufficient to hold a frame-aligned
3644 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3645 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3647 This method replaces @code{CALL_DUMMY_LOCATION},
3648 @code{DEPRECATED_REGISTER_SIZE}.
3650 @item REGISTER_NAME(@var{i})
3651 @findex REGISTER_NAME
3652 Return the name of register @var{i} as a string. May return @code{NULL}
3653 or @code{NUL} to indicate that register @var{i} is not valid.
3655 @item DEPRECATED_REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3656 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3657 @anchor{DEPRECATED_REG_STRUCT_HAS_ADDR}Define this to return 1 if the
3658 given type will be passed by pointer rather than directly.
3660 This method has been replaced by @code{stabs_argument_has_addr}
3661 (@pxref{stabs_argument_has_addr}).
3663 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3664 @findex SAVE_DUMMY_FRAME_TOS
3665 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3666 notify the target dependent code of the top-of-stack value that will be
3667 passed to the the inferior code. This is the value of the @code{SP}
3668 after both the dummy frame and space for parameters/results have been
3669 allocated on the stack. @xref{unwind_dummy_id}.
3671 @item SDB_REG_TO_REGNUM
3672 @findex SDB_REG_TO_REGNUM
3673 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3674 defined, no conversion will be done.
3676 @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})
3677 @findex gdbarch_return_value
3678 @anchor{gdbarch_return_value} Given a function with a return-value of
3679 type @var{rettype}, return which return-value convention that function
3682 @value{GDBN} currently recognizes two function return-value conventions:
3683 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3684 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3685 value is found in memory and the address of that memory location is
3686 passed in as the function's first parameter.
3688 If the register convention is being used, and @var{writebuf} is
3689 non-@code{NULL}, also copy the return-value in @var{writebuf} into
3692 If the register convention is being used, and @var{readbuf} is
3693 non-@code{NULL}, also copy the return value from @var{regcache} into
3694 @var{readbuf} (@var{regcache} contains a copy of the registers from the
3695 just returned function).
3697 @xref{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}, for a description of how
3698 return-values that use the struct convention are handled.
3700 @emph{Maintainer note: This method replaces separate predicate, extract,
3701 store methods. By having only one method, the logic needed to determine
3702 the return-value convention need only be implemented in one place. If
3703 @value{GDBN} were written in an @sc{oo} language, this method would
3704 instead return an object that knew how to perform the register
3705 return-value extract and store.}
3707 @emph{Maintainer note: This method does not take a @var{gcc_p}
3708 parameter, and such a parameter should not be added. If an architecture
3709 that requires per-compiler or per-function information be identified,
3710 then the replacement of @var{rettype} with @code{struct value}
3711 @var{function} should be persued.}
3713 @emph{Maintainer note: The @var{regcache} parameter limits this methods
3714 to the inner most frame. While replacing @var{regcache} with a
3715 @code{struct frame_info} @var{frame} parameter would remove that
3716 limitation there has yet to be a demonstrated need for such a change.}
3718 @item SKIP_PERMANENT_BREAKPOINT
3719 @findex SKIP_PERMANENT_BREAKPOINT
3720 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3721 steps over a breakpoint by removing it, stepping one instruction, and
3722 re-inserting the breakpoint. However, permanent breakpoints are
3723 hardwired into the inferior, and can't be removed, so this strategy
3724 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
3725 state so that execution will resume just after the breakpoint. This
3726 macro does the right thing even when the breakpoint is in the delay slot
3727 of a branch or jump.
3729 @item SKIP_PROLOGUE (@var{pc})
3730 @findex SKIP_PROLOGUE
3731 A C expression that returns the address of the ``real'' code beyond the
3732 function entry prologue found at @var{pc}.
3734 @item SKIP_TRAMPOLINE_CODE (@var{pc})
3735 @findex SKIP_TRAMPOLINE_CODE
3736 If the target machine has trampoline code that sits between callers and
3737 the functions being called, then define this macro to return a new PC
3738 that is at the start of the real function.
3742 If the stack-pointer is kept in a register, then define this macro to be
3743 the number (greater than or equal to zero) of that register, or -1 if
3744 there is no such register.
3746 @item STAB_REG_TO_REGNUM
3747 @findex STAB_REG_TO_REGNUM
3748 Define this to convert stab register numbers (as gotten from `r'
3749 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
3752 @item DEPRECATED_STACK_ALIGN (@var{addr})
3753 @anchor{DEPRECATED_STACK_ALIGN}
3754 @findex DEPRECATED_STACK_ALIGN
3755 Define this to increase @var{addr} so that it meets the alignment
3756 requirements for the processor's stack.
3758 Unlike @ref{frame_align}, this function always adjusts @var{addr}
3761 By default, no stack alignment is performed.
3763 @item STEP_SKIPS_DELAY (@var{addr})
3764 @findex STEP_SKIPS_DELAY
3765 Define this to return true if the address is of an instruction with a
3766 delay slot. If a breakpoint has been placed in the instruction's delay
3767 slot, @value{GDBN} will single-step over that instruction before resuming
3768 normally. Currently only defined for the Mips.
3770 @item STORE_RETURN_VALUE (@var{type}, @var{regcache}, @var{valbuf})
3771 @findex STORE_RETURN_VALUE
3772 A C expression that writes the function return value, found in
3773 @var{valbuf}, into the @var{regcache}. @var{type} is the type of the
3774 value that is to be returned.
3776 This method has been deprecated in favour of @code{gdbarch_return_value}
3777 (@pxref{gdbarch_return_value}).
3779 @item SYMBOL_RELOADING_DEFAULT
3780 @findex SYMBOL_RELOADING_DEFAULT
3781 The default value of the ``symbol-reloading'' variable. (Never defined in
3784 @item TARGET_CHAR_BIT
3785 @findex TARGET_CHAR_BIT
3786 Number of bits in a char; defaults to 8.
3788 @item TARGET_CHAR_SIGNED
3789 @findex TARGET_CHAR_SIGNED
3790 Non-zero if @code{char} is normally signed on this architecture; zero if
3791 it should be unsigned.
3793 The ISO C standard requires the compiler to treat @code{char} as
3794 equivalent to either @code{signed char} or @code{unsigned char}; any
3795 character in the standard execution set is supposed to be positive.
3796 Most compilers treat @code{char} as signed, but @code{char} is unsigned
3797 on the IBM S/390, RS6000, and PowerPC targets.
3799 @item TARGET_COMPLEX_BIT
3800 @findex TARGET_COMPLEX_BIT
3801 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
3803 At present this macro is not used.
3805 @item TARGET_DOUBLE_BIT
3806 @findex TARGET_DOUBLE_BIT
3807 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
3809 @item TARGET_DOUBLE_COMPLEX_BIT
3810 @findex TARGET_DOUBLE_COMPLEX_BIT
3811 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
3813 At present this macro is not used.
3815 @item TARGET_FLOAT_BIT
3816 @findex TARGET_FLOAT_BIT
3817 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
3819 @item TARGET_INT_BIT
3820 @findex TARGET_INT_BIT
3821 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3823 @item TARGET_LONG_BIT
3824 @findex TARGET_LONG_BIT
3825 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3827 @item TARGET_LONG_DOUBLE_BIT
3828 @findex TARGET_LONG_DOUBLE_BIT
3829 Number of bits in a long double float;
3830 defaults to @code{2 * TARGET_DOUBLE_BIT}.
3832 @item TARGET_LONG_LONG_BIT
3833 @findex TARGET_LONG_LONG_BIT
3834 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
3836 @item TARGET_PTR_BIT
3837 @findex TARGET_PTR_BIT
3838 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
3840 @item TARGET_SHORT_BIT
3841 @findex TARGET_SHORT_BIT
3842 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
3844 @item TARGET_READ_PC
3845 @findex TARGET_READ_PC
3846 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
3847 @findex TARGET_WRITE_PC
3848 @anchor{TARGET_WRITE_PC}
3849 @itemx TARGET_READ_SP
3850 @findex TARGET_READ_SP
3851 @itemx TARGET_READ_FP
3852 @findex TARGET_READ_FP
3857 @anchor{TARGET_READ_SP} These change the behavior of @code{read_pc},
3858 @code{write_pc}, and @code{read_sp}. For most targets, these may be
3859 left undefined. @value{GDBN} will call the read and write register
3860 functions with the relevant @code{_REGNUM} argument.
3862 These macros are useful when a target keeps one of these registers in a
3863 hard to get at place; for example, part in a segment register and part
3864 in an ordinary register.
3866 @xref{unwind_sp}, which replaces @code{TARGET_READ_SP}.
3868 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
3869 @findex TARGET_VIRTUAL_FRAME_POINTER
3870 Returns a @code{(register, offset)} pair representing the virtual frame
3871 pointer in use at the code address @var{pc}. If virtual frame pointers
3872 are not used, a default definition simply returns
3873 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
3875 @item TARGET_HAS_HARDWARE_WATCHPOINTS
3876 If non-zero, the target has support for hardware-assisted
3877 watchpoints. @xref{Algorithms, watchpoints}, for more details and
3878 other related macros.
3880 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
3881 @findex TARGET_PRINT_INSN
3882 This is the function used by @value{GDBN} to print an assembly
3883 instruction. It prints the instruction at address @var{addr} in
3884 debugged memory and returns the length of the instruction, in bytes. If
3885 a target doesn't define its own printing routine, it defaults to an
3886 accessor function for the global pointer
3887 @code{deprecated_tm_print_insn}. This usually points to a function in
3888 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
3889 @var{info} is a structure (of type @code{disassemble_info}) defined in
3890 @file{include/dis-asm.h} used to pass information to the instruction
3893 @item struct frame_id unwind_dummy_id (struct frame_info *@var{frame})
3894 @findex unwind_dummy_id
3895 @anchor{unwind_dummy_id} Given @var{frame} return a @code{struct
3896 frame_id} that uniquely identifies an inferior function call's dummy
3897 frame. The value returned must match the dummy frame stack value
3898 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
3899 @xref{SAVE_DUMMY_FRAME_TOS}.
3901 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
3902 @findex DEPRECATED_USE_STRUCT_CONVENTION
3903 If defined, this must be an expression that is nonzero if a value of the
3904 given @var{type} being returned from a function must have space
3905 allocated for it on the stack. @var{gcc_p} is true if the function
3906 being considered is known to have been compiled by GCC; this is helpful
3907 for systems where GCC is known to use different calling convention than
3910 This method has been deprecated in favour of @code{gdbarch_return_value}
3911 (@pxref{gdbarch_return_value}).
3913 @item VALUE_TO_REGISTER(@var{type}, @var{regnum}, @var{from}, @var{to})
3914 @findex VALUE_TO_REGISTER
3915 Convert a value of type @var{type} into the raw contents of register
3917 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3919 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3920 @findex VARIABLES_INSIDE_BLOCK
3921 For dbx-style debugging information, if the compiler puts variable
3922 declarations inside LBRAC/RBRAC blocks, this should be defined to be
3923 nonzero. @var{desc} is the value of @code{n_desc} from the
3924 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
3925 presence of either the @code{GCC_COMPILED_SYMBOL} or the
3926 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
3928 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3929 @findex OS9K_VARIABLES_INSIDE_BLOCK
3930 Similarly, for OS/9000. Defaults to 1.
3933 Motorola M68K target conditionals.
3937 Define this to be the 4-bit location of the breakpoint trap vector. If
3938 not defined, it will default to @code{0xf}.
3940 @item REMOTE_BPT_VECTOR
3941 Defaults to @code{1}.
3943 @item NAME_OF_MALLOC
3944 @findex NAME_OF_MALLOC
3945 A string containing the name of the function to call in order to
3946 allocate some memory in the inferior. The default value is "malloc".
3950 @section Adding a New Target
3952 @cindex adding a target
3953 The following files add a target to @value{GDBN}:
3957 @item gdb/config/@var{arch}/@var{ttt}.mt
3958 Contains a Makefile fragment specific to this target. Specifies what
3959 object files are needed for target @var{ttt}, by defining
3960 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
3961 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
3964 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
3965 but these are now deprecated, replaced by autoconf, and may go away in
3966 future versions of @value{GDBN}.
3968 @item gdb/@var{ttt}-tdep.c
3969 Contains any miscellaneous code required for this target machine. On
3970 some machines it doesn't exist at all. Sometimes the macros in
3971 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
3972 as functions here instead, and the macro is simply defined to call the
3973 function. This is vastly preferable, since it is easier to understand
3976 @item gdb/@var{arch}-tdep.c
3977 @itemx gdb/@var{arch}-tdep.h
3978 This often exists to describe the basic layout of the target machine's
3979 processor chip (registers, stack, etc.). If used, it is included by
3980 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
3983 @item gdb/config/@var{arch}/tm-@var{ttt}.h
3984 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
3985 macro definitions about the target machine's registers, stack frame
3986 format and instructions.
3988 New targets do not need this file and should not create it.
3990 @item gdb/config/@var{arch}/tm-@var{arch}.h
3991 This often exists to describe the basic layout of the target machine's
3992 processor chip (registers, stack, etc.). If used, it is included by
3993 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
3996 New targets do not need this file and should not create it.
4000 If you are adding a new operating system for an existing CPU chip, add a
4001 @file{config/tm-@var{os}.h} file that describes the operating system
4002 facilities that are unusual (extra symbol table info; the breakpoint
4003 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
4004 that just @code{#include}s @file{tm-@var{arch}.h} and
4005 @file{config/tm-@var{os}.h}.
4008 @section Converting an existing Target Architecture to Multi-arch
4009 @cindex converting targets to multi-arch
4011 This section describes the current accepted best practice for converting
4012 an existing target architecture to the multi-arch framework.
4014 The process consists of generating, testing, posting and committing a
4015 sequence of patches. Each patch must contain a single change, for
4021 Directly convert a group of functions into macros (the conversion does
4022 not change the behavior of any of the functions).
4025 Replace a non-multi-arch with a multi-arch mechanism (e.g.,
4029 Enable multi-arch level one.
4032 Delete one or more files.
4037 There isn't a size limit on a patch, however, a developer is strongly
4038 encouraged to keep the patch size down.
4040 Since each patch is well defined, and since each change has been tested
4041 and shows no regressions, the patches are considered @emph{fairly}
4042 obvious. Such patches, when submitted by developers listed in the
4043 @file{MAINTAINERS} file, do not need approval. Occasional steps in the
4044 process may be more complicated and less clear. The developer is
4045 expected to use their judgment and is encouraged to seek advice as
4048 @subsection Preparation
4050 The first step is to establish control. Build (with @option{-Werror}
4051 enabled) and test the target so that there is a baseline against which
4052 the debugger can be compared.
4054 At no stage can the test results regress or @value{GDBN} stop compiling
4055 with @option{-Werror}.
4057 @subsection Add the multi-arch initialization code
4059 The objective of this step is to establish the basic multi-arch
4060 framework. It involves
4065 The addition of a @code{@var{arch}_gdbarch_init} function@footnote{The
4066 above is from the original example and uses K&R C. @value{GDBN}
4067 has since converted to ISO C but lets ignore that.} that creates
4070 static struct gdbarch *
4071 d10v_gdbarch_init (info, arches)
4072 struct gdbarch_info info;
4073 struct gdbarch_list *arches;
4075 struct gdbarch *gdbarch;
4076 /* there is only one d10v architecture */
4078 return arches->gdbarch;
4079 gdbarch = gdbarch_alloc (&info, NULL);
4087 A per-architecture dump function to print any architecture specific
4091 mips_dump_tdep (struct gdbarch *current_gdbarch,
4092 struct ui_file *file)
4094 @dots{} code to print architecture specific info @dots{}
4099 A change to @code{_initialize_@var{arch}_tdep} to register this new
4103 _initialize_mips_tdep (void)
4105 gdbarch_register (bfd_arch_mips, mips_gdbarch_init,
4110 Add the macro @code{GDB_MULTI_ARCH}, defined as 0 (zero), to the file@*
4111 @file{config/@var{arch}/tm-@var{arch}.h}.
4115 @subsection Update multi-arch incompatible mechanisms
4117 Some mechanisms do not work with multi-arch. They include:
4120 @item FRAME_FIND_SAVED_REGS
4121 Replaced with @code{DEPRECATED_FRAME_INIT_SAVED_REGS}
4125 At this stage you could also consider converting the macros into
4128 @subsection Prepare for multi-arch level to one
4130 Temporally set @code{GDB_MULTI_ARCH} to @code{GDB_MULTI_ARCH_PARTIAL}
4131 and then build and start @value{GDBN} (the change should not be
4132 committed). @value{GDBN} may not build, and once built, it may die with
4133 an internal error listing the architecture methods that must be
4136 Fix any build problems (patch(es)).
4138 Convert all the architecture methods listed, which are only macros, into
4139 functions (patch(es)).
4141 Update @code{@var{arch}_gdbarch_init} to set all the missing
4142 architecture methods and wrap the corresponding macros in @code{#if
4143 !GDB_MULTI_ARCH} (patch(es)).
4145 @subsection Set multi-arch level one
4147 Change the value of @code{GDB_MULTI_ARCH} to GDB_MULTI_ARCH_PARTIAL (a
4150 Any problems with throwing ``the switch'' should have been fixed
4153 @subsection Convert remaining macros
4155 Suggest converting macros into functions (and setting the corresponding
4156 architecture method) in small batches.
4158 @subsection Set multi-arch level to two
4160 This should go smoothly.
4162 @subsection Delete the TM file
4164 The @file{tm-@var{arch}.h} can be deleted. @file{@var{arch}.mt} and
4165 @file{configure.in} updated.
4168 @node Target Vector Definition
4170 @chapter Target Vector Definition
4171 @cindex target vector
4173 The target vector defines the interface between @value{GDBN}'s
4174 abstract handling of target systems, and the nitty-gritty code that
4175 actually exercises control over a process or a serial port.
4176 @value{GDBN} includes some 30-40 different target vectors; however,
4177 each configuration of @value{GDBN} includes only a few of them.
4179 @section File Targets
4181 Both executables and core files have target vectors.
4183 @section Standard Protocol and Remote Stubs
4185 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4186 that runs in the target system. @value{GDBN} provides several sample
4187 @dfn{stubs} that can be integrated into target programs or operating
4188 systems for this purpose; they are named @file{*-stub.c}.
4190 The @value{GDBN} user's manual describes how to put such a stub into
4191 your target code. What follows is a discussion of integrating the
4192 SPARC stub into a complicated operating system (rather than a simple
4193 program), by Stu Grossman, the author of this stub.
4195 The trap handling code in the stub assumes the following upon entry to
4200 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4206 you are in the correct trap window.
4209 As long as your trap handler can guarantee those conditions, then there
4210 is no reason why you shouldn't be able to ``share'' traps with the stub.
4211 The stub has no requirement that it be jumped to directly from the
4212 hardware trap vector. That is why it calls @code{exceptionHandler()},
4213 which is provided by the external environment. For instance, this could
4214 set up the hardware traps to actually execute code which calls the stub
4215 first, and then transfers to its own trap handler.
4217 For the most point, there probably won't be much of an issue with
4218 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4219 and often indicate unrecoverable error conditions. Anyway, this is all
4220 controlled by a table, and is trivial to modify. The most important
4221 trap for us is for @code{ta 1}. Without that, we can't single step or
4222 do breakpoints. Everything else is unnecessary for the proper operation
4223 of the debugger/stub.
4225 From reading the stub, it's probably not obvious how breakpoints work.
4226 They are simply done by deposit/examine operations from @value{GDBN}.
4228 @section ROM Monitor Interface
4230 @section Custom Protocols
4232 @section Transport Layer
4234 @section Builtin Simulator
4237 @node Native Debugging
4239 @chapter Native Debugging
4240 @cindex native debugging
4242 Several files control @value{GDBN}'s configuration for native support:
4246 @item gdb/config/@var{arch}/@var{xyz}.mh
4247 Specifies Makefile fragments needed by a @emph{native} configuration on
4248 machine @var{xyz}. In particular, this lists the required
4249 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4250 Also specifies the header file which describes native support on
4251 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4252 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4253 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4255 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4256 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4257 on machine @var{xyz}. While the file is no longer used for this
4258 purpose, the @file{.mh} suffix remains. Perhaps someone will
4259 eventually rename these fragments so that they have a @file{.mn}
4262 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4263 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4264 macro definitions describing the native system environment, such as
4265 child process control and core file support.
4267 @item gdb/@var{xyz}-nat.c
4268 Contains any miscellaneous C code required for this native support of
4269 this machine. On some machines it doesn't exist at all.
4272 There are some ``generic'' versions of routines that can be used by
4273 various systems. These can be customized in various ways by macros
4274 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4275 the @var{xyz} host, you can just include the generic file's name (with
4276 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4278 Otherwise, if your machine needs custom support routines, you will need
4279 to write routines that perform the same functions as the generic file.
4280 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4281 into @code{NATDEPFILES}.
4285 This contains the @emph{target_ops vector} that supports Unix child
4286 processes on systems which use ptrace and wait to control the child.
4289 This contains the @emph{target_ops vector} that supports Unix child
4290 processes on systems which use /proc to control the child.
4293 This does the low-level grunge that uses Unix system calls to do a ``fork
4294 and exec'' to start up a child process.
4297 This is the low level interface to inferior processes for systems using
4298 the Unix @code{ptrace} call in a vanilla way.
4301 @section Native core file Support
4302 @cindex native core files
4305 @findex fetch_core_registers
4306 @item core-aout.c::fetch_core_registers()
4307 Support for reading registers out of a core file. This routine calls
4308 @code{register_addr()}, see below. Now that BFD is used to read core
4309 files, virtually all machines should use @code{core-aout.c}, and should
4310 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4311 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4313 @item core-aout.c::register_addr()
4314 If your @code{nm-@var{xyz}.h} file defines the macro
4315 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4316 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4317 register number @code{regno}. @code{blockend} is the offset within the
4318 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4319 @file{core-aout.c} will define the @code{register_addr()} function and
4320 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4321 you are using the standard @code{fetch_core_registers()}, you will need
4322 to define your own version of @code{register_addr()}, put it into your
4323 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4324 the @code{NATDEPFILES} list. If you have your own
4325 @code{fetch_core_registers()}, you may not need a separate
4326 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4327 implementations simply locate the registers themselves.@refill
4330 When making @value{GDBN} run native on a new operating system, to make it
4331 possible to debug core files, you will need to either write specific
4332 code for parsing your OS's core files, or customize
4333 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4334 machine uses to define the struct of registers that is accessible
4335 (possibly in the u-area) in a core file (rather than
4336 @file{machine/reg.h}), and an include file that defines whatever header
4337 exists on a core file (e.g. the u-area or a @code{struct core}). Then
4338 modify @code{trad_unix_core_file_p} to use these values to set up the
4339 section information for the data segment, stack segment, any other
4340 segments in the core file (perhaps shared library contents or control
4341 information), ``registers'' segment, and if there are two discontiguous
4342 sets of registers (e.g. integer and float), the ``reg2'' segment. This
4343 section information basically delimits areas in the core file in a
4344 standard way, which the section-reading routines in BFD know how to seek
4347 Then back in @value{GDBN}, you need a matching routine called
4348 @code{fetch_core_registers}. If you can use the generic one, it's in
4349 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4350 It will be passed a char pointer to the entire ``registers'' segment,
4351 its length, and a zero; or a char pointer to the entire ``regs2''
4352 segment, its length, and a 2. The routine should suck out the supplied
4353 register values and install them into @value{GDBN}'s ``registers'' array.
4355 If your system uses @file{/proc} to control processes, and uses ELF
4356 format core files, then you may be able to use the same routines for
4357 reading the registers out of processes and out of core files.
4365 @section shared libraries
4367 @section Native Conditionals
4368 @cindex native conditionals
4370 When @value{GDBN} is configured and compiled, various macros are
4371 defined or left undefined, to control compilation when the host and
4372 target systems are the same. These macros should be defined (or left
4373 undefined) in @file{nm-@var{system}.h}.
4377 @item CHILD_PREPARE_TO_STORE
4378 @findex CHILD_PREPARE_TO_STORE
4379 If the machine stores all registers at once in the child process, then
4380 define this to ensure that all values are correct. This usually entails
4381 a read from the child.
4383 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4386 @item FETCH_INFERIOR_REGISTERS
4387 @findex FETCH_INFERIOR_REGISTERS
4388 Define this if the native-dependent code will provide its own routines
4389 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4390 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4391 @file{infptrace.c} is included in this configuration, the default
4392 routines in @file{infptrace.c} are used for these functions.
4396 This macro is normally defined to be the number of the first floating
4397 point register, if the machine has such registers. As such, it would
4398 appear only in target-specific code. However, @file{/proc} support uses this
4399 to decide whether floats are in use on this target.
4401 @item GET_LONGJMP_TARGET
4402 @findex GET_LONGJMP_TARGET
4403 For most machines, this is a target-dependent parameter. On the
4404 DECstation and the Iris, this is a native-dependent parameter, since
4405 @file{setjmp.h} is needed to define it.
4407 This macro determines the target PC address that @code{longjmp} will jump to,
4408 assuming that we have just stopped at a longjmp breakpoint. It takes a
4409 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4410 pointer. It examines the current state of the machine as needed.
4412 @item I386_USE_GENERIC_WATCHPOINTS
4413 An x86-based machine can define this to use the generic x86 watchpoint
4414 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4417 @findex KERNEL_U_ADDR
4418 Define this to the address of the @code{u} structure (the ``user
4419 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
4420 needs to know this so that it can subtract this address from absolute
4421 addresses in the upage, that are obtained via ptrace or from core files.
4422 On systems that don't need this value, set it to zero.
4424 @item KERNEL_U_ADDR_HPUX
4425 @findex KERNEL_U_ADDR_HPUX
4426 Define this to cause @value{GDBN} to determine the address of @code{u} at
4427 runtime, by using HP-style @code{nlist} on the kernel's image in the
4430 @item ONE_PROCESS_WRITETEXT
4431 @findex ONE_PROCESS_WRITETEXT
4432 Define this to be able to, when a breakpoint insertion fails, warn the
4433 user that another process may be running with the same executable.
4436 @findex PROC_NAME_FMT
4437 Defines the format for the name of a @file{/proc} device. Should be
4438 defined in @file{nm.h} @emph{only} in order to override the default
4439 definition in @file{procfs.c}.
4441 @item PTRACE_ARG3_TYPE
4442 @findex PTRACE_ARG3_TYPE
4443 The type of the third argument to the @code{ptrace} system call, if it
4444 exists and is different from @code{int}.
4446 @item REGISTER_U_ADDR
4447 @findex REGISTER_U_ADDR
4448 Defines the offset of the registers in the ``u area''.
4450 @item SHELL_COMMAND_CONCAT
4451 @findex SHELL_COMMAND_CONCAT
4452 If defined, is a string to prefix on the shell command used to start the
4457 If defined, this is the name of the shell to use to run the inferior.
4458 Defaults to @code{"/bin/sh"}.
4460 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4462 Define this to expand into an expression that will cause the symbols in
4463 @var{filename} to be added to @value{GDBN}'s symbol table. If
4464 @var{readsyms} is zero symbols are not read but any necessary low level
4465 processing for @var{filename} is still done.
4467 @item SOLIB_CREATE_INFERIOR_HOOK
4468 @findex SOLIB_CREATE_INFERIOR_HOOK
4469 Define this to expand into any shared-library-relocation code that you
4470 want to be run just after the child process has been forked.
4472 @item START_INFERIOR_TRAPS_EXPECTED
4473 @findex START_INFERIOR_TRAPS_EXPECTED
4474 When starting an inferior, @value{GDBN} normally expects to trap
4476 the shell execs, and once when the program itself execs. If the actual
4477 number of traps is something other than 2, then define this macro to
4478 expand into the number expected.
4482 This determines whether small routines in @file{*-tdep.c}, which
4483 translate register values between @value{GDBN}'s internal
4484 representation and the @file{/proc} representation, are compiled.
4487 @findex U_REGS_OFFSET
4488 This is the offset of the registers in the upage. It need only be
4489 defined if the generic ptrace register access routines in
4490 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4491 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4492 the default value from @file{infptrace.c} is good enough, leave it
4495 The default value means that u.u_ar0 @emph{points to} the location of
4496 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4497 that @code{u.u_ar0} @emph{is} the location of the registers.
4501 See @file{objfiles.c}.
4504 @findex DEBUG_PTRACE
4505 Define this to debug @code{ptrace} calls.
4509 @node Support Libraries
4511 @chapter Support Libraries
4516 BFD provides support for @value{GDBN} in several ways:
4519 @item identifying executable and core files
4520 BFD will identify a variety of file types, including a.out, coff, and
4521 several variants thereof, as well as several kinds of core files.
4523 @item access to sections of files
4524 BFD parses the file headers to determine the names, virtual addresses,
4525 sizes, and file locations of all the various named sections in files
4526 (such as the text section or the data section). @value{GDBN} simply
4527 calls BFD to read or write section @var{x} at byte offset @var{y} for
4530 @item specialized core file support
4531 BFD provides routines to determine the failing command name stored in a
4532 core file, the signal with which the program failed, and whether a core
4533 file matches (i.e.@: could be a core dump of) a particular executable
4536 @item locating the symbol information
4537 @value{GDBN} uses an internal interface of BFD to determine where to find the
4538 symbol information in an executable file or symbol-file. @value{GDBN} itself
4539 handles the reading of symbols, since BFD does not ``understand'' debug
4540 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4545 @cindex opcodes library
4547 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4548 library because it's also used in binutils, for @file{objdump}).
4555 @cindex @code{libiberty} library
4557 The @code{libiberty} library provides a set of functions and features
4558 that integrate and improve on functionality found in modern operating
4559 systems. Broadly speaking, such features can be divided into three
4560 groups: supplemental functions (functions that may be missing in some
4561 environments and operating systems), replacement functions (providing
4562 a uniform and easier to use interface for commonly used standard
4563 functions), and extensions (which provide additional functionality
4564 beyond standard functions).
4566 @value{GDBN} uses various features provided by the @code{libiberty}
4567 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4568 floating format support functions, the input options parser
4569 @samp{getopt}, the @samp{obstack} extension, and other functions.
4571 @subsection @code{obstacks} in @value{GDBN}
4572 @cindex @code{obstacks}
4574 The obstack mechanism provides a convenient way to allocate and free
4575 chunks of memory. Each obstack is a pool of memory that is managed
4576 like a stack. Objects (of any nature, size and alignment) are
4577 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4578 @code{libiberty}'s documenatation for a more detailed explanation of
4581 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4582 object files. There is an obstack associated with each internal
4583 representation of an object file. Lots of things get allocated on
4584 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4585 symbols, minimal symbols, types, vectors of fundamental types, class
4586 fields of types, object files section lists, object files section
4587 offets lists, line tables, symbol tables, partial symbol tables,
4588 string tables, symbol table private data, macros tables, debug
4589 information sections and entries, import and export lists (som),
4590 unwind information (hppa), dwarf2 location expressions data. Plus
4591 various strings such as directory names strings, debug format strings,
4594 An essential and convenient property of all data on @code{obstacks} is
4595 that memory for it gets allocated (with @code{obstack_alloc}) at
4596 various times during a debugging sesssion, but it is released all at
4597 once using the @code{obstack_free} function. The @code{obstack_free}
4598 function takes a pointer to where in the stack it must start the
4599 deletion from (much like the cleanup chains have a pointer to where to
4600 start the cleanups). Because of the stack like structure of the
4601 @code{obstacks}, this allows to free only a top portion of the
4602 obstack. There are a few instances in @value{GDBN} where such thing
4603 happens. Calls to @code{obstack_free} are done after some local data
4604 is allocated to the obstack. Only the local data is deleted from the
4605 obstack. Of course this assumes that nothing between the
4606 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4607 else on the same obstack. For this reason it is best and safest to
4608 use temporary @code{obstacks}.
4610 Releasing the whole obstack is also not safe per se. It is safe only
4611 under the condition that we know the @code{obstacks} memory is no
4612 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4613 when we get rid of the whole objfile(s), for instance upon reading a
4617 @cindex regular expressions library
4628 @item SIGN_EXTEND_CHAR
4630 @item SWITCH_ENUM_BUG
4645 This chapter covers topics that are lower-level than the major
4646 algorithms of @value{GDBN}.
4651 Cleanups are a structured way to deal with things that need to be done
4654 When your code does something (e.g., @code{xmalloc} some memory, or
4655 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4656 the memory or @code{close} the file), it can make a cleanup. The
4657 cleanup will be done at some future point: when the command is finished
4658 and control returns to the top level; when an error occurs and the stack
4659 is unwound; or when your code decides it's time to explicitly perform
4660 cleanups. Alternatively you can elect to discard the cleanups you
4666 @item struct cleanup *@var{old_chain};
4667 Declare a variable which will hold a cleanup chain handle.
4669 @findex make_cleanup
4670 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
4671 Make a cleanup which will cause @var{function} to be called with
4672 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
4673 handle that can later be passed to @code{do_cleanups} or
4674 @code{discard_cleanups}. Unless you are going to call
4675 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
4676 from @code{make_cleanup}.
4679 @item do_cleanups (@var{old_chain});
4680 Do all cleanups added to the chain since the corresponding
4681 @code{make_cleanup} call was made.
4683 @findex discard_cleanups
4684 @item discard_cleanups (@var{old_chain});
4685 Same as @code{do_cleanups} except that it just removes the cleanups from
4686 the chain and does not call the specified functions.
4689 Cleanups are implemented as a chain. The handle returned by
4690 @code{make_cleanups} includes the cleanup passed to the call and any
4691 later cleanups appended to the chain (but not yet discarded or
4695 make_cleanup (a, 0);
4697 struct cleanup *old = make_cleanup (b, 0);
4705 will call @code{c()} and @code{b()} but will not call @code{a()}. The
4706 cleanup that calls @code{a()} will remain in the cleanup chain, and will
4707 be done later unless otherwise discarded.@refill
4709 Your function should explicitly do or discard the cleanups it creates.
4710 Failing to do this leads to non-deterministic behavior since the caller
4711 will arbitrarily do or discard your functions cleanups. This need leads
4712 to two common cleanup styles.
4714 The first style is try/finally. Before it exits, your code-block calls
4715 @code{do_cleanups} with the old cleanup chain and thus ensures that your
4716 code-block's cleanups are always performed. For instance, the following
4717 code-segment avoids a memory leak problem (even when @code{error} is
4718 called and a forced stack unwind occurs) by ensuring that the
4719 @code{xfree} will always be called:
4722 struct cleanup *old = make_cleanup (null_cleanup, 0);
4723 data = xmalloc (sizeof blah);
4724 make_cleanup (xfree, data);
4729 The second style is try/except. Before it exits, your code-block calls
4730 @code{discard_cleanups} with the old cleanup chain and thus ensures that
4731 any created cleanups are not performed. For instance, the following
4732 code segment, ensures that the file will be closed but only if there is
4736 FILE *file = fopen ("afile", "r");
4737 struct cleanup *old = make_cleanup (close_file, file);
4739 discard_cleanups (old);
4743 Some functions, e.g. @code{fputs_filtered()} or @code{error()}, specify
4744 that they ``should not be called when cleanups are not in place''. This
4745 means that any actions you need to reverse in the case of an error or
4746 interruption must be on the cleanup chain before you call these
4747 functions, since they might never return to your code (they
4748 @samp{longjmp} instead).
4750 @section Per-architecture module data
4751 @cindex per-architecture module data
4752 @cindex multi-arch data
4753 @cindex data-pointer, per-architecture/per-module
4755 The multi-arch framework includes a mechanism for adding module
4756 specific per-architecture data-pointers to the @code{struct gdbarch}
4757 architecture object.
4759 A module registers one or more per-architecture data-pointers using:
4761 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
4762 @var{pre_init} is used to, on-demand, allocate an initial value for a
4763 per-architecture data-pointer using the architecture's obstack (passed
4764 in as a parameter). Since @var{pre_init} can be called during
4765 architecture creation, it is not parameterized with the architecture.
4766 and must not call modules that use per-architecture data.
4769 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
4770 @var{post_init} is used to obtain an initial value for a
4771 per-architecture data-pointer @emph{after}. Since @var{post_init} is
4772 always called after architecture creation, it both receives the fully
4773 initialized architecture and is free to call modules that use
4774 per-architecture data (care needs to be taken to ensure that those
4775 other modules do not try to call back to this module as that will
4776 create in cycles in the initialization call graph).
4779 These functions return a @code{struct gdbarch_data} that is used to
4780 identify the per-architecture data-pointer added for that module.
4782 The per-architecture data-pointer is accessed using the function:
4784 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
4785 Given the architecture @var{arch} and module data handle
4786 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
4787 or @code{gdbarch_data_register_post_init}), this function returns the
4788 current value of the per-architecture data-pointer. If the data
4789 pointer is @code{NULL}, it is first initialized by calling the
4790 corresponding @var{pre_init} or @var{post_init} method.
4793 The examples below assume the following definitions:
4796 struct nozel @{ int total; @};
4797 static struct gdbarch_data *nozel_handle;
4800 A module can extend the architecture vector, adding additional
4801 per-architecture data, using the @var{pre_init} method. The module's
4802 per-architecture data is then initialized during architecture
4805 In the below, the module's per-architecture @emph{nozel} is added. An
4806 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
4807 from @code{gdbarch_init}.
4811 nozel_pre_init (struct obstack *obstack)
4813 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
4820 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
4822 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
4823 data->total = nozel;
4827 A module can on-demand create architecture dependant data structures
4828 using @code{post_init}.
4830 In the below, the nozel's total is computed on-demand by
4831 @code{nozel_post_init} using information obtained from the
4836 nozel_post_init (struct gdbarch *gdbarch)
4838 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
4839 nozel->total = gdbarch@dots{} (gdbarch);
4846 nozel_total (struct gdbarch *gdbarch)
4848 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
4853 @section Wrapping Output Lines
4854 @cindex line wrap in output
4857 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
4858 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
4859 added in places that would be good breaking points. The utility
4860 routines will take care of actually wrapping if the line width is
4863 The argument to @code{wrap_here} is an indentation string which is
4864 printed @emph{only} if the line breaks there. This argument is saved
4865 away and used later. It must remain valid until the next call to
4866 @code{wrap_here} or until a newline has been printed through the
4867 @code{*_filtered} functions. Don't pass in a local variable and then
4870 It is usually best to call @code{wrap_here} after printing a comma or
4871 space. If you call it before printing a space, make sure that your
4872 indentation properly accounts for the leading space that will print if
4873 the line wraps there.
4875 Any function or set of functions that produce filtered output must
4876 finish by printing a newline, to flush the wrap buffer, before switching
4877 to unfiltered (@code{printf}) output. Symbol reading routines that
4878 print warnings are a good example.
4880 @section @value{GDBN} Coding Standards
4881 @cindex coding standards
4883 @value{GDBN} follows the GNU coding standards, as described in
4884 @file{etc/standards.texi}. This file is also available for anonymous
4885 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
4886 of the standard; in general, when the GNU standard recommends a practice
4887 but does not require it, @value{GDBN} requires it.
4889 @value{GDBN} follows an additional set of coding standards specific to
4890 @value{GDBN}, as described in the following sections.
4895 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
4898 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
4901 @subsection Memory Management
4903 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
4904 @code{calloc}, @code{free} and @code{asprintf}.
4906 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
4907 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
4908 these functions do not return when the memory pool is empty. Instead,
4909 they unwind the stack using cleanups. These functions return
4910 @code{NULL} when requested to allocate a chunk of memory of size zero.
4912 @emph{Pragmatics: By using these functions, the need to check every
4913 memory allocation is removed. These functions provide portable
4916 @value{GDBN} does not use the function @code{free}.
4918 @value{GDBN} uses the function @code{xfree} to return memory to the
4919 memory pool. Consistent with ISO-C, this function ignores a request to
4920 free a @code{NULL} pointer.
4922 @emph{Pragmatics: On some systems @code{free} fails when passed a
4923 @code{NULL} pointer.}
4925 @value{GDBN} can use the non-portable function @code{alloca} for the
4926 allocation of small temporary values (such as strings).
4928 @emph{Pragmatics: This function is very non-portable. Some systems
4929 restrict the memory being allocated to no more than a few kilobytes.}
4931 @value{GDBN} uses the string function @code{xstrdup} and the print
4932 function @code{xstrprintf}.
4934 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
4935 functions such as @code{sprintf} are very prone to buffer overflow
4939 @subsection Compiler Warnings
4940 @cindex compiler warnings
4942 With few exceptions, developers should include the configuration option
4943 @samp{--enable-gdb-build-warnings=,-Werror} when building @value{GDBN}.
4944 The exceptions are listed in the file @file{gdb/MAINTAINERS}.
4946 This option causes @value{GDBN} (when built using GCC) to be compiled
4947 with a carefully selected list of compiler warning flags. Any warnings
4948 from those flags being treated as errors.
4950 The current list of warning flags includes:
4954 Since @value{GDBN} coding standard requires all functions to be declared
4955 using a prototype, the flag has the side effect of ensuring that
4956 prototyped functions are always visible with out resorting to
4957 @samp{-Wstrict-prototypes}.
4960 Such code often appears to work except on instruction set architectures
4961 that use register windows.
4968 @itemx -Wformat-nonliteral
4969 Since @value{GDBN} uses the @code{format printf} attribute on all
4970 @code{printf} like functions these check not just @code{printf} calls
4971 but also calls to functions such as @code{fprintf_unfiltered}.
4974 This warning includes uses of the assignment operator within an
4975 @code{if} statement.
4977 @item -Wpointer-arith
4979 @item -Wuninitialized
4981 @item -Wunused-label
4982 This warning has the additional benefit of detecting the absence of the
4983 @code{case} reserved word in a switch statement:
4985 enum @{ FD_SCHEDULED, NOTHING_SCHEDULED @} sched;
4998 @item -Wunused-function
5001 @emph{Pragmatics: Due to the way that @value{GDBN} is implemented most
5002 functions have unused parameters. Consequently the warning
5003 @samp{-Wunused-parameter} is precluded from the list. The macro
5004 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
5005 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
5006 is being used. The options @samp{-Wall} and @samp{-Wunused} are also
5007 precluded because they both include @samp{-Wunused-parameter}.}
5009 @emph{Pragmatics: @value{GDBN} has not simply accepted the warnings
5010 enabled by @samp{-Wall -Werror -W...}. Instead it is selecting warnings
5011 when and where their benefits can be demonstrated.}
5013 @subsection Formatting
5015 @cindex source code formatting
5016 The standard GNU recommendations for formatting must be followed
5019 A function declaration should not have its name in column zero. A
5020 function definition should have its name in column zero.
5024 static void foo (void);
5032 @emph{Pragmatics: This simplifies scripting. Function definitions can
5033 be found using @samp{^function-name}.}
5035 There must be a space between a function or macro name and the opening
5036 parenthesis of its argument list (except for macro definitions, as
5037 required by C). There must not be a space after an open paren/bracket
5038 or before a close paren/bracket.
5040 While additional whitespace is generally helpful for reading, do not use
5041 more than one blank line to separate blocks, and avoid adding whitespace
5042 after the end of a program line (as of 1/99, some 600 lines had
5043 whitespace after the semicolon). Excess whitespace causes difficulties
5044 for @code{diff} and @code{patch} utilities.
5046 Pointers are declared using the traditional K&R C style:
5060 @subsection Comments
5062 @cindex comment formatting
5063 The standard GNU requirements on comments must be followed strictly.
5065 Block comments must appear in the following form, with no @code{/*}- or
5066 @code{*/}-only lines, and no leading @code{*}:
5069 /* Wait for control to return from inferior to debugger. If inferior
5070 gets a signal, we may decide to start it up again instead of
5071 returning. That is why there is a loop in this function. When
5072 this function actually returns it means the inferior should be left
5073 stopped and @value{GDBN} should read more commands. */
5076 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5077 comment works correctly, and @kbd{M-q} fills the block consistently.)
5079 Put a blank line between the block comments preceding function or
5080 variable definitions, and the definition itself.
5082 In general, put function-body comments on lines by themselves, rather
5083 than trying to fit them into the 20 characters left at the end of a
5084 line, since either the comment or the code will inevitably get longer
5085 than will fit, and then somebody will have to move it anyhow.
5089 @cindex C data types
5090 Code must not depend on the sizes of C data types, the format of the
5091 host's floating point numbers, the alignment of anything, or the order
5092 of evaluation of expressions.
5094 @cindex function usage
5095 Use functions freely. There are only a handful of compute-bound areas
5096 in @value{GDBN} that might be affected by the overhead of a function
5097 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5098 limited by the target interface (whether serial line or system call).
5100 However, use functions with moderation. A thousand one-line functions
5101 are just as hard to understand as a single thousand-line function.
5103 @emph{Macros are bad, M'kay.}
5104 (But if you have to use a macro, make sure that the macro arguments are
5105 protected with parentheses.)
5109 Declarations like @samp{struct foo *} should be used in preference to
5110 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5113 @subsection Function Prototypes
5114 @cindex function prototypes
5116 Prototypes must be used when both @emph{declaring} and @emph{defining}
5117 a function. Prototypes for @value{GDBN} functions must include both the
5118 argument type and name, with the name matching that used in the actual
5119 function definition.
5121 All external functions should have a declaration in a header file that
5122 callers include, except for @code{_initialize_*} functions, which must
5123 be external so that @file{init.c} construction works, but shouldn't be
5124 visible to random source files.
5126 Where a source file needs a forward declaration of a static function,
5127 that declaration must appear in a block near the top of the source file.
5130 @subsection Internal Error Recovery
5132 During its execution, @value{GDBN} can encounter two types of errors.
5133 User errors and internal errors. User errors include not only a user
5134 entering an incorrect command but also problems arising from corrupt
5135 object files and system errors when interacting with the target.
5136 Internal errors include situations where @value{GDBN} has detected, at
5137 run time, a corrupt or erroneous situation.
5139 When reporting an internal error, @value{GDBN} uses
5140 @code{internal_error} and @code{gdb_assert}.
5142 @value{GDBN} must not call @code{abort} or @code{assert}.
5144 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5145 the code detected a user error, recovered from it and issued a
5146 @code{warning} or the code failed to correctly recover from the user
5147 error and issued an @code{internal_error}.}
5149 @subsection File Names
5151 Any file used when building the core of @value{GDBN} must be in lower
5152 case. Any file used when building the core of @value{GDBN} must be 8.3
5153 unique. These requirements apply to both source and generated files.
5155 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5156 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5157 is introduced to the build process both @file{Makefile.in} and
5158 @file{configure.in} need to be modified accordingly. Compare the
5159 convoluted conversion process needed to transform @file{COPYING} into
5160 @file{copying.c} with the conversion needed to transform
5161 @file{version.in} into @file{version.c}.}
5163 Any file non 8.3 compliant file (that is not used when building the core
5164 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5166 @emph{Pragmatics: This is clearly a compromise.}
5168 When @value{GDBN} has a local version of a system header file (ex
5169 @file{string.h}) the file name based on the POSIX header prefixed with
5170 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5171 independent: they should use only macros defined by @file{configure},
5172 the compiler, or the host; they should include only system headers; they
5173 should refer only to system types. They may be shared between multiple
5174 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5176 For other files @samp{-} is used as the separator.
5179 @subsection Include Files
5181 A @file{.c} file should include @file{defs.h} first.
5183 A @file{.c} file should directly include the @code{.h} file of every
5184 declaration and/or definition it directly refers to. It cannot rely on
5187 A @file{.h} file should directly include the @code{.h} file of every
5188 declaration and/or definition it directly refers to. It cannot rely on
5189 indirect inclusion. Exception: The file @file{defs.h} does not need to
5190 be directly included.
5192 An external declaration should only appear in one include file.
5194 An external declaration should never appear in a @code{.c} file.
5195 Exception: a declaration for the @code{_initialize} function that
5196 pacifies @option{-Wmissing-declaration}.
5198 A @code{typedef} definition should only appear in one include file.
5200 An opaque @code{struct} declaration can appear in multiple @file{.h}
5201 files. Where possible, a @file{.h} file should use an opaque
5202 @code{struct} declaration instead of an include.
5204 All @file{.h} files should be wrapped in:
5207 #ifndef INCLUDE_FILE_NAME_H
5208 #define INCLUDE_FILE_NAME_H
5214 @subsection Clean Design and Portable Implementation
5217 In addition to getting the syntax right, there's the little question of
5218 semantics. Some things are done in certain ways in @value{GDBN} because long
5219 experience has shown that the more obvious ways caused various kinds of
5222 @cindex assumptions about targets
5223 You can't assume the byte order of anything that comes from a target
5224 (including @var{value}s, object files, and instructions). Such things
5225 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5226 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5227 such as @code{bfd_get_32}.
5229 You can't assume that you know what interface is being used to talk to
5230 the target system. All references to the target must go through the
5231 current @code{target_ops} vector.
5233 You can't assume that the host and target machines are the same machine
5234 (except in the ``native'' support modules). In particular, you can't
5235 assume that the target machine's header files will be available on the
5236 host machine. Target code must bring along its own header files --
5237 written from scratch or explicitly donated by their owner, to avoid
5241 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5242 to write the code portably than to conditionalize it for various
5245 @cindex system dependencies
5246 New @code{#ifdef}'s which test for specific compilers or manufacturers
5247 or operating systems are unacceptable. All @code{#ifdef}'s should test
5248 for features. The information about which configurations contain which
5249 features should be segregated into the configuration files. Experience
5250 has proven far too often that a feature unique to one particular system
5251 often creeps into other systems; and that a conditional based on some
5252 predefined macro for your current system will become worthless over
5253 time, as new versions of your system come out that behave differently
5254 with regard to this feature.
5256 Adding code that handles specific architectures, operating systems,
5257 target interfaces, or hosts, is not acceptable in generic code.
5259 @cindex portable file name handling
5260 @cindex file names, portability
5261 One particularly notorious area where system dependencies tend to
5262 creep in is handling of file names. The mainline @value{GDBN} code
5263 assumes Posix semantics of file names: absolute file names begin with
5264 a forward slash @file{/}, slashes are used to separate leading
5265 directories, case-sensitive file names. These assumptions are not
5266 necessarily true on non-Posix systems such as MS-Windows. To avoid
5267 system-dependent code where you need to take apart or construct a file
5268 name, use the following portable macros:
5271 @findex HAVE_DOS_BASED_FILE_SYSTEM
5272 @item HAVE_DOS_BASED_FILE_SYSTEM
5273 This preprocessing symbol is defined to a non-zero value on hosts
5274 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5275 symbol to write conditional code which should only be compiled for
5278 @findex IS_DIR_SEPARATOR
5279 @item IS_DIR_SEPARATOR (@var{c})
5280 Evaluates to a non-zero value if @var{c} is a directory separator
5281 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5282 such a character, but on Windows, both @file{/} and @file{\} will
5285 @findex IS_ABSOLUTE_PATH
5286 @item IS_ABSOLUTE_PATH (@var{file})
5287 Evaluates to a non-zero value if @var{file} is an absolute file name.
5288 For Unix and GNU/Linux hosts, a name which begins with a slash
5289 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5290 @file{x:\bar} are also absolute file names.
5292 @findex FILENAME_CMP
5293 @item FILENAME_CMP (@var{f1}, @var{f2})
5294 Calls a function which compares file names @var{f1} and @var{f2} as
5295 appropriate for the underlying host filesystem. For Posix systems,
5296 this simply calls @code{strcmp}; on case-insensitive filesystems it
5297 will call @code{strcasecmp} instead.
5299 @findex DIRNAME_SEPARATOR
5300 @item DIRNAME_SEPARATOR
5301 Evaluates to a character which separates directories in
5302 @code{PATH}-style lists, typically held in environment variables.
5303 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5305 @findex SLASH_STRING
5307 This evaluates to a constant string you should use to produce an
5308 absolute filename from leading directories and the file's basename.
5309 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5310 @code{"\\"} for some Windows-based ports.
5313 In addition to using these macros, be sure to use portable library
5314 functions whenever possible. For example, to extract a directory or a
5315 basename part from a file name, use the @code{dirname} and
5316 @code{basename} library functions (available in @code{libiberty} for
5317 platforms which don't provide them), instead of searching for a slash
5318 with @code{strrchr}.
5320 Another way to generalize @value{GDBN} along a particular interface is with an
5321 attribute struct. For example, @value{GDBN} has been generalized to handle
5322 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5323 by defining the @code{target_ops} structure and having a current target (as
5324 well as a stack of targets below it, for memory references). Whenever
5325 something needs to be done that depends on which remote interface we are
5326 using, a flag in the current target_ops structure is tested (e.g.,
5327 @code{target_has_stack}), or a function is called through a pointer in the
5328 current target_ops structure. In this way, when a new remote interface
5329 is added, only one module needs to be touched---the one that actually
5330 implements the new remote interface. Other examples of
5331 attribute-structs are BFD access to multiple kinds of object file
5332 formats, or @value{GDBN}'s access to multiple source languages.
5334 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5335 the code interfacing between @code{ptrace} and the rest of
5336 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5337 something was very painful. In @value{GDBN} 4.x, these have all been
5338 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5339 with variations between systems the same way any system-independent
5340 file would (hooks, @code{#if defined}, etc.), and machines which are
5341 radically different don't need to use @file{infptrace.c} at all.
5343 All debugging code must be controllable using the @samp{set debug
5344 @var{module}} command. Do not use @code{printf} to print trace
5345 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5346 @code{#ifdef DEBUG}.
5351 @chapter Porting @value{GDBN}
5352 @cindex porting to new machines
5354 Most of the work in making @value{GDBN} compile on a new machine is in
5355 specifying the configuration of the machine. This is done in a
5356 dizzying variety of header files and configuration scripts, which we
5357 hope to make more sensible soon. Let's say your new host is called an
5358 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5359 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5360 @samp{sparc-sun-sunos4}). In particular:
5364 In the top level directory, edit @file{config.sub} and add @var{arch},
5365 @var{xvend}, and @var{xos} to the lists of supported architectures,
5366 vendors, and operating systems near the bottom of the file. Also, add
5367 @var{xyz} as an alias that maps to
5368 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5372 ./config.sub @var{xyz}
5379 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5383 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5384 and no error messages.
5387 You need to port BFD, if that hasn't been done already. Porting BFD is
5388 beyond the scope of this manual.
5391 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5392 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5393 desired target is already available) also edit @file{gdb/configure.tgt},
5394 setting @code{gdb_target} to something appropriate (for instance,
5397 @emph{Maintainer's note: Work in progress. The file
5398 @file{gdb/configure.host} originally needed to be modified when either a
5399 new native target or a new host machine was being added to @value{GDBN}.
5400 Recent changes have removed this requirement. The file now only needs
5401 to be modified when adding a new native configuration. This will likely
5402 changed again in the future.}
5405 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5406 target-dependent @file{.h} and @file{.c} files used for your
5410 @node Versions and Branches
5411 @chapter Versions and Branches
5415 @value{GDBN}'s version is determined by the file
5416 @file{gdb/version.in} and takes one of the following forms:
5419 @item @var{major}.@var{minor}
5420 @itemx @var{major}.@var{minor}.@var{patchlevel}
5421 an official release (e.g., 6.0 or 6.0.1)
5422 @item @var{major}.@var{minor}.@var{patchlevel}_@var{YYYY}@var{MM}@var{DD}
5423 a snapshot (e.g., 6.0.50_20020630)
5424 @item @var{major}.@var{minor}.@var{patchlevel}_@var{YYYY}-@var{MM}-@var{DD}-cvs
5425 a @sc{cvs} check out (e.g., 6.0.90_2004-02-30-cvs)
5426 @item @var{major}.@var{minor}.@var{patchlevel}_@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5427 a vendor specific release of @value{GDBN}, that while based on@*
5428 @var{major}.@var{minor}.@var{patchlevel}_@var{YYYY}@var{MM}@var{DD},
5429 may contain additional changes
5432 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5433 numbers from the most recent release branch, with a @var{patchlevel}
5434 of 50. As each new release branch is created, the mainline
5435 @var{major} and @var{minor} version numbers are accordingly updated.
5437 @value{GDBN}'s release branch uses a similar, but slightly more
5438 complicated scheme. When the branch is first cut, the mainline's
5439 @var{patchlevel} is changed to .90. As draft releases are drawn from
5440 the branch, the @var{patchlevel} is incremented. Once the first
5441 release (@var{major}.@var{minor}) has been made, the version prefix is
5442 updated to @var{major}.@var{minor}.0.90. Follow on releases have an
5443 incremented @var{patchlevel}.
5445 If the previous @value{GDBN} version is 6.1 and the current version is
5446 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5447 here's an illustration of a typical sequence:
5452 6.1.50_2002-03-02-cvs
5454 +---------------------------.
5457 6.2.50_2002-03-03-cvs 6.1.90 (draft #1)
5459 6.2.50_2002-03-04-cvs 6.1.90_2002-03-04-cvs
5461 6.2.50_2002-03-05-cvs 6.1.91 (draft #2)
5463 6.2.50_2002-03-06-cvs 6.1.91_2002-03-06-cvs
5465 6.2.50_2002-03-07-cvs 6.2 (release)
5467 6.2.50_2002-03-08-cvs 6.2.0.90_2002-03-08-cvs
5469 6.2.50_2002-03-09-cvs 6.2.1 (update)
5471 6.2.50_2002-03-10-cvs <branch closed>
5473 6.2.50_2002-03-11-cvs
5475 +---------------------------.
5478 6.3.50_2002-03-12-cvs 6.2.90 (draft #1)
5482 @section Release Branches
5483 @cindex Release Branches
5485 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
5486 single release branch, and identifies that branch using the @sc{cvs}
5490 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
5491 gdb_@var{major}_@var{minor}-branch
5492 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
5495 @emph{Pragmatics: To help identify the date at which a branch or
5496 release is made, both the branchpoint and release tags include the
5497 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
5498 branch tag, denoting the head of the branch, does not need this.}
5500 @section Vendor Branches
5501 @cindex vendor branches
5503 To avoid version conflicts, vendors are expected to modify the file
5504 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5505 (an official @value{GDBN} release never uses alphabetic characters in
5506 its version identifer). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
5509 @section Experimental Branches
5510 @cindex experimental branches
5512 @subsection Guidelines
5514 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
5515 repository, for experimental development. Branches make it possible
5516 for developers to share preliminary work, and maintainers to examine
5517 significant new developments.
5519 The following are a set of guidelines for creating such branches:
5523 @item a branch has an owner
5524 The owner can set further policy for a branch, but may not change the
5525 ground rules. In particular, they can set a policy for commits (be it
5526 adding more reviewers or deciding who can commit).
5528 @item all commits are posted
5529 All changes committed to a branch shall also be posted to
5530 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
5531 mailing list}. While commentary on such changes are encouraged, people
5532 should remember that the changes only apply to a branch.
5534 @item all commits are covered by an assignment
5535 This ensures that all changes belong to the Free Software Foundation,
5536 and avoids the possibility that the branch may become contaminated.
5538 @item a branch is focused
5539 A focused branch has a single objective or goal, and does not contain
5540 unnecessary or irrelevant changes. Cleanups, where identified, being
5541 be pushed into the mainline as soon as possible.
5543 @item a branch tracks mainline
5544 This keeps the level of divergence under control. It also keeps the
5545 pressure on developers to push cleanups and other stuff into the
5548 @item a branch shall contain the entire @value{GDBN} module
5549 The @value{GDBN} module @code{gdb} should be specified when creating a
5550 branch (branches of individual files should be avoided). @xref{Tags}.
5552 @item a branch shall be branded using @file{version.in}
5553 The file @file{gdb/version.in} shall be modified so that it identifies
5554 the branch @var{owner} and branch @var{name}, e.g.,
5555 @samp{6.2.50_20030303_owner_name} or @samp{6.2 (Owner Name)}.
5562 To simplify the identification of @value{GDBN} branches, the following
5563 branch tagging convention is strongly recommended:
5567 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5568 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
5569 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
5570 date that the branch was created. A branch is created using the
5571 sequence: @anchor{experimental branch tags}
5573 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
5574 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
5575 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
5578 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5579 The tagged point, on the mainline, that was used when merging the branch
5580 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
5581 use a command sequence like:
5583 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
5585 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5586 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5589 Similar sequences can be used to just merge in changes since the last
5595 For further information on @sc{cvs}, see
5596 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
5600 @chapter Releasing @value{GDBN}
5601 @cindex making a new release of gdb
5603 @section Branch Commit Policy
5605 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
5606 5.1 and 5.2 all used the below:
5610 The @file{gdb/MAINTAINERS} file still holds.
5612 Don't fix something on the branch unless/until it is also fixed in the
5613 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
5614 file is better than committing a hack.
5616 When considering a patch for the branch, suggested criteria include:
5617 Does it fix a build? Does it fix the sequence @kbd{break main; run}
5618 when debugging a static binary?
5620 The further a change is from the core of @value{GDBN}, the less likely
5621 the change will worry anyone (e.g., target specific code).
5623 Only post a proposal to change the core of @value{GDBN} after you've
5624 sent individual bribes to all the people listed in the
5625 @file{MAINTAINERS} file @t{;-)}
5628 @emph{Pragmatics: Provided updates are restricted to non-core
5629 functionality there is little chance that a broken change will be fatal.
5630 This means that changes such as adding a new architectures or (within
5631 reason) support for a new host are considered acceptable.}
5634 @section Obsoleting code
5636 Before anything else, poke the other developers (and around the source
5637 code) to see if there is anything that can be removed from @value{GDBN}
5638 (an old target, an unused file).
5640 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
5641 line. Doing this means that it is easy to identify something that has
5642 been obsoleted when greping through the sources.
5644 The process is done in stages --- this is mainly to ensure that the
5645 wider @value{GDBN} community has a reasonable opportunity to respond.
5646 Remember, everything on the Internet takes a week.
5650 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
5651 list} Creating a bug report to track the task's state, is also highly
5656 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
5657 Announcement mailing list}.
5661 Go through and edit all relevant files and lines so that they are
5662 prefixed with the word @code{OBSOLETE}.
5664 Wait until the next GDB version, containing this obsolete code, has been
5667 Remove the obsolete code.
5671 @emph{Maintainer note: While removing old code is regrettable it is
5672 hopefully better for @value{GDBN}'s long term development. Firstly it
5673 helps the developers by removing code that is either no longer relevant
5674 or simply wrong. Secondly since it removes any history associated with
5675 the file (effectively clearing the slate) the developer has a much freer
5676 hand when it comes to fixing broken files.}
5680 @section Before the Branch
5682 The most important objective at this stage is to find and fix simple
5683 changes that become a pain to track once the branch is created. For
5684 instance, configuration problems that stop @value{GDBN} from even
5685 building. If you can't get the problem fixed, document it in the
5686 @file{gdb/PROBLEMS} file.
5688 @subheading Prompt for @file{gdb/NEWS}
5690 People always forget. Send a post reminding them but also if you know
5691 something interesting happened add it yourself. The @code{schedule}
5692 script will mention this in its e-mail.
5694 @subheading Review @file{gdb/README}
5696 Grab one of the nightly snapshots and then walk through the
5697 @file{gdb/README} looking for anything that can be improved. The
5698 @code{schedule} script will mention this in its e-mail.
5700 @subheading Refresh any imported files.
5702 A number of files are taken from external repositories. They include:
5706 @file{texinfo/texinfo.tex}
5708 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
5711 @file{etc/standards.texi}, @file{etc/make-stds.texi}
5714 @subheading Check the ARI
5716 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
5717 (Awk Regression Index ;-) that checks for a number of errors and coding
5718 conventions. The checks include things like using @code{malloc} instead
5719 of @code{xmalloc} and file naming problems. There shouldn't be any
5722 @subsection Review the bug data base
5724 Close anything obviously fixed.
5726 @subsection Check all cross targets build
5728 The targets are listed in @file{gdb/MAINTAINERS}.
5731 @section Cut the Branch
5733 @subheading Create the branch
5738 $ V=`echo $v | sed 's/\./_/g'`
5739 $ D=`date -u +%Y-%m-%d`
5742 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5743 -D $D-gmt gdb_$V-$D-branchpoint insight+dejagnu
5744 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
5745 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight+dejagnu
5748 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5749 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight+dejagnu
5750 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5751 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight+dejagnu
5759 by using @kbd{-D YYYY-MM-DD-gmt} the branch is forced to an exact
5762 the trunk is first taged so that the branch point can easily be found
5764 Insight (which includes GDB) and dejagnu are all tagged at the same time
5766 @file{version.in} gets bumped to avoid version number conflicts
5768 the reading of @file{.cvsrc} is disabled using @file{-f}
5771 @subheading Update @file{version.in}
5776 $ V=`echo $v | sed 's/\./_/g'`
5780 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
5781 -r gdb_$V-branch src/gdb/version.in
5782 cvs -f -d :ext:sources.redhat.com:/cvs/src co
5783 -r gdb_5_2-branch src/gdb/version.in
5785 U src/gdb/version.in
5787 $ echo $u.90-0000-00-00-cvs > version.in
5789 5.1.90-0000-00-00-cvs
5790 $ cvs -f commit version.in
5795 @file{0000-00-00} is used as a date to pump prime the version.in update
5798 @file{.90} and the previous branch version are used as fairly arbitrary
5799 initial branch version number
5803 @subheading Update the web and news pages
5807 @subheading Tweak cron to track the new branch
5809 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
5810 This file needs to be updated so that:
5814 a daily timestamp is added to the file @file{version.in}
5816 the new branch is included in the snapshot process
5820 See the file @file{gdbadmin/cron/README} for how to install the updated
5823 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
5824 any changes. That file is copied to both the branch/ and current/
5825 snapshot directories.
5828 @subheading Update the NEWS and README files
5830 The @file{NEWS} file needs to be updated so that on the branch it refers
5831 to @emph{changes in the current release} while on the trunk it also
5832 refers to @emph{changes since the current release}.
5834 The @file{README} file needs to be updated so that it refers to the
5837 @subheading Post the branch info
5839 Send an announcement to the mailing lists:
5843 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
5845 @email{gdb@@sources.redhat.com, GDB Discsussion mailing list} and
5846 @email{gdb-testers@@sources.redhat.com, GDB Discsussion mailing list}
5849 @emph{Pragmatics: The branch creation is sent to the announce list to
5850 ensure that people people not subscribed to the higher volume discussion
5853 The announcement should include:
5859 how to check out the branch using CVS
5861 the date/number of weeks until the release
5863 the branch commit policy
5867 @section Stabilize the branch
5869 Something goes here.
5871 @section Create a Release
5873 The process of creating and then making available a release is broken
5874 down into a number of stages. The first part addresses the technical
5875 process of creating a releasable tar ball. The later stages address the
5876 process of releasing that tar ball.
5878 When making a release candidate just the first section is needed.
5880 @subsection Create a release candidate
5882 The objective at this stage is to create a set of tar balls that can be
5883 made available as a formal release (or as a less formal release
5886 @subsubheading Freeze the branch
5888 Send out an e-mail notifying everyone that the branch is frozen to
5889 @email{gdb-patches@@sources.redhat.com}.
5891 @subsubheading Establish a few defaults.
5896 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
5898 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
5902 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
5904 /home/gdbadmin/bin/autoconf
5913 Check the @code{autoconf} version carefully. You want to be using the
5914 version taken from the @file{binutils} snapshot directory, which can be
5915 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
5916 unlikely that a system installed version of @code{autoconf} (e.g.,
5917 @file{/usr/bin/autoconf}) is correct.
5920 @subsubheading Check out the relevant modules:
5923 $ for m in gdb insight dejagnu
5925 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
5935 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
5936 any confusion between what is written here and what your local
5937 @code{cvs} really does.
5940 @subsubheading Update relevant files.
5946 Major releases get their comments added as part of the mainline. Minor
5947 releases should probably mention any significant bugs that were fixed.
5949 Don't forget to include the @file{ChangeLog} entry.
5952 $ emacs gdb/src/gdb/NEWS
5957 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
5958 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5963 You'll need to update:
5975 $ emacs gdb/src/gdb/README
5980 $ cp gdb/src/gdb/README insight/src/gdb/README
5981 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5984 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
5985 before the initial branch was cut so just a simple substitute is needed
5988 @emph{Maintainer note: Other projects generate @file{README} and
5989 @file{INSTALL} from the core documentation. This might be worth
5992 @item gdb/version.in
5995 $ echo $v > gdb/src/gdb/version.in
5996 $ cat gdb/src/gdb/version.in
5998 $ emacs gdb/src/gdb/version.in
6001 ... Bump to version ...
6003 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6004 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6007 @item dejagnu/src/dejagnu/configure.in
6009 Dejagnu is more complicated. The version number is a parameter to
6010 @code{AM_INIT_AUTOMAKE}. Tweak it to read something like gdb-5.1.91.
6012 Don't forget to re-generate @file{configure}.
6014 Don't forget to include a @file{ChangeLog} entry.
6017 $ emacs dejagnu/src/dejagnu/configure.in
6022 $ ( cd dejagnu/src/dejagnu && autoconf )
6027 @subsubheading Do the dirty work
6029 This is identical to the process used to create the daily snapshot.
6032 $ for m in gdb insight
6034 ( cd $m/src && gmake -f src-release $m.tar )
6036 $ ( m=dejagnu; cd $m/src && gmake -f src-release $m.tar.bz2 )
6039 If the top level source directory does not have @file{src-release}
6040 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6043 $ for m in gdb insight
6045 ( cd $m/src && gmake -f Makefile.in $m.tar )
6047 $ ( m=dejagnu; cd $m/src && gmake -f Makefile.in $m.tar.bz2 )
6050 @subsubheading Check the source files
6052 You're looking for files that have mysteriously disappeared.
6053 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6054 for the @file{version.in} update @kbd{cronjob}.
6057 $ ( cd gdb/src && cvs -f -q -n update )
6061 @dots{} lots of generated files @dots{}
6066 @dots{} lots of generated files @dots{}
6071 @emph{Don't worry about the @file{gdb.info-??} or
6072 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6073 was also generated only something strange with CVS means that they
6074 didn't get supressed). Fixing it would be nice though.}
6076 @subsubheading Create compressed versions of the release
6082 dejagnu/ dejagnu-gdb-5.2.tar.bz2 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6083 $ for m in gdb insight
6085 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6086 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6096 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6097 in that mode, @code{gzip} does not know the name of the file and, hence,
6098 can not include it in the compressed file. This is also why the release
6099 process runs @code{tar} and @code{bzip2} as separate passes.
6102 @subsection Sanity check the tar ball
6104 Pick a popular machine (Solaris/PPC?) and try the build on that.
6107 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6112 $ ./gdb/gdb ./gdb/gdb
6116 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6118 Starting program: /tmp/gdb-5.2/gdb/gdb
6120 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6121 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6123 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6127 @subsection Make a release candidate available
6129 If this is a release candidate then the only remaining steps are:
6133 Commit @file{version.in} and @file{ChangeLog}
6135 Tweak @file{version.in} (and @file{ChangeLog} to read
6136 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6137 process can restart.
6139 Make the release candidate available in
6140 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6142 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6143 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6146 @subsection Make a formal release available
6148 (And you thought all that was required was to post an e-mail.)
6150 @subsubheading Install on sware
6152 Copy the new files to both the release and the old release directory:
6155 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6156 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6160 Clean up the releases directory so that only the most recent releases
6161 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6164 $ cd ~ftp/pub/gdb/releases
6169 Update the file @file{README} and @file{.message} in the releases
6176 $ ln README .message
6179 @subsubheading Update the web pages.
6183 @item htdocs/download/ANNOUNCEMENT
6184 This file, which is posted as the official announcement, includes:
6187 General announcement
6189 News. If making an @var{M}.@var{N}.1 release, retain the news from
6190 earlier @var{M}.@var{N} release.
6195 @item htdocs/index.html
6196 @itemx htdocs/news/index.html
6197 @itemx htdocs/download/index.html
6198 These files include:
6201 announcement of the most recent release
6203 news entry (remember to update both the top level and the news directory).
6205 These pages also need to be regenerate using @code{index.sh}.
6207 @item download/onlinedocs/
6208 You need to find the magic command that is used to generate the online
6209 docs from the @file{.tar.bz2}. The best way is to look in the output
6210 from one of the nightly @code{cron} jobs and then just edit accordingly.
6214 $ ~/ss/update-web-docs \
6215 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6217 /www/sourceware/htdocs/gdb/download/onlinedocs \
6222 Just like the online documentation. Something like:
6225 $ /bin/sh ~/ss/update-web-ari \
6226 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6228 /www/sourceware/htdocs/gdb/download/ari \
6234 @subsubheading Shadow the pages onto gnu
6236 Something goes here.
6239 @subsubheading Install the @value{GDBN} tar ball on GNU
6241 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6242 @file{~ftp/gnu/gdb}.
6244 @subsubheading Make the @file{ANNOUNCEMENT}
6246 Post the @file{ANNOUNCEMENT} file you created above to:
6250 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6252 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6253 day or so to let things get out)
6255 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6260 The release is out but you're still not finished.
6262 @subsubheading Commit outstanding changes
6264 In particular you'll need to commit any changes to:
6268 @file{gdb/ChangeLog}
6270 @file{gdb/version.in}
6277 @subsubheading Tag the release
6282 $ d=`date -u +%Y-%m-%d`
6285 $ ( cd insight/src/gdb && cvs -f -q update )
6286 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6289 Insight is used since that contains more of the release than
6290 @value{GDBN} (@code{dejagnu} doesn't get tagged but I think we can live
6293 @subsubheading Mention the release on the trunk
6295 Just put something in the @file{ChangeLog} so that the trunk also
6296 indicates when the release was made.
6298 @subsubheading Restart @file{gdb/version.in}
6300 If @file{gdb/version.in} does not contain an ISO date such as
6301 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6302 committed all the release changes it can be set to
6303 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6304 is important - it affects the snapshot process).
6306 Don't forget the @file{ChangeLog}.
6308 @subsubheading Merge into trunk
6310 The files committed to the branch may also need changes merged into the
6313 @subsubheading Revise the release schedule
6315 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6316 Discussion List} with an updated announcement. The schedule can be
6317 generated by running:
6320 $ ~/ss/schedule `date +%s` schedule
6324 The first parameter is approximate date/time in seconds (from the epoch)
6325 of the most recent release.
6327 Also update the schedule @code{cronjob}.
6329 @section Post release
6331 Remove any @code{OBSOLETE} code.
6338 The testsuite is an important component of the @value{GDBN} package.
6339 While it is always worthwhile to encourage user testing, in practice
6340 this is rarely sufficient; users typically use only a small subset of
6341 the available commands, and it has proven all too common for a change
6342 to cause a significant regression that went unnoticed for some time.
6344 The @value{GDBN} testsuite uses the DejaGNU testing framework.
6345 DejaGNU is built using @code{Tcl} and @code{expect}. The tests
6346 themselves are calls to various @code{Tcl} procs; the framework runs all the
6347 procs and summarizes the passes and fails.
6349 @section Using the Testsuite
6351 @cindex running the test suite
6352 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6353 testsuite's objdir) and type @code{make check}. This just sets up some
6354 environment variables and invokes DejaGNU's @code{runtest} script. While
6355 the testsuite is running, you'll get mentions of which test file is in use,
6356 and a mention of any unexpected passes or fails. When the testsuite is
6357 finished, you'll get a summary that looks like this:
6362 # of expected passes 6016
6363 # of unexpected failures 58
6364 # of unexpected successes 5
6365 # of expected failures 183
6366 # of unresolved testcases 3
6367 # of untested testcases 5
6370 The ideal test run consists of expected passes only; however, reality
6371 conspires to keep us from this ideal. Unexpected failures indicate
6372 real problems, whether in @value{GDBN} or in the testsuite. Expected
6373 failures are still failures, but ones which have been decided are too
6374 hard to deal with at the time; for instance, a test case might work
6375 everywhere except on AIX, and there is no prospect of the AIX case
6376 being fixed in the near future. Expected failures should not be added
6377 lightly, since you may be masking serious bugs in @value{GDBN}.
6378 Unexpected successes are expected fails that are passing for some
6379 reason, while unresolved and untested cases often indicate some minor
6380 catastrophe, such as the compiler being unable to deal with a test
6383 When making any significant change to @value{GDBN}, you should run the
6384 testsuite before and after the change, to confirm that there are no
6385 regressions. Note that truly complete testing would require that you
6386 run the testsuite with all supported configurations and a variety of
6387 compilers; however this is more than really necessary. In many cases
6388 testing with a single configuration is sufficient. Other useful
6389 options are to test one big-endian (Sparc) and one little-endian (x86)
6390 host, a cross config with a builtin simulator (powerpc-eabi,
6391 mips-elf), or a 64-bit host (Alpha).
6393 If you add new functionality to @value{GDBN}, please consider adding
6394 tests for it as well; this way future @value{GDBN} hackers can detect
6395 and fix their changes that break the functionality you added.
6396 Similarly, if you fix a bug that was not previously reported as a test
6397 failure, please add a test case for it. Some cases are extremely
6398 difficult to test, such as code that handles host OS failures or bugs
6399 in particular versions of compilers, and it's OK not to try to write
6400 tests for all of those.
6402 DejaGNU supports separate build, host, and target machines. However,
6403 some @value{GDBN} test scripts do not work if the build machine and
6404 the host machine are not the same. In such an environment, these scripts
6405 will give a result of ``UNRESOLVED'', like this:
6408 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6411 @section Testsuite Organization
6413 @cindex test suite organization
6414 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6415 testsuite includes some makefiles and configury, these are very minimal,
6416 and used for little besides cleaning up, since the tests themselves
6417 handle the compilation of the programs that @value{GDBN} will run. The file
6418 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6419 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6420 configuration-specific files, typically used for special-purpose
6421 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6423 The tests themselves are to be found in @file{testsuite/gdb.*} and
6424 subdirectories of those. The names of the test files must always end
6425 with @file{.exp}. DejaGNU collects the test files by wildcarding
6426 in the test directories, so both subdirectories and individual files
6427 get chosen and run in alphabetical order.
6429 The following table lists the main types of subdirectories and what they
6430 are for. Since DejaGNU finds test files no matter where they are
6431 located, and since each test file sets up its own compilation and
6432 execution environment, this organization is simply for convenience and
6437 This is the base testsuite. The tests in it should apply to all
6438 configurations of @value{GDBN} (but generic native-only tests may live here).
6439 The test programs should be in the subset of C that is valid K&R,
6440 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6443 @item gdb.@var{lang}
6444 Language-specific tests for any language @var{lang} besides C. Examples are
6445 @file{gdb.cp} and @file{gdb.java}.
6447 @item gdb.@var{platform}
6448 Non-portable tests. The tests are specific to a specific configuration
6449 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6452 @item gdb.@var{compiler}
6453 Tests specific to a particular compiler. As of this writing (June
6454 1999), there aren't currently any groups of tests in this category that
6455 couldn't just as sensibly be made platform-specific, but one could
6456 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6459 @item gdb.@var{subsystem}
6460 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6461 instance, @file{gdb.disasm} exercises various disassemblers, while
6462 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6465 @section Writing Tests
6466 @cindex writing tests
6468 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6469 should be able to copy existing tests to handle new cases.
6471 You should try to use @code{gdb_test} whenever possible, since it
6472 includes cases to handle all the unexpected errors that might happen.
6473 However, it doesn't cost anything to add new test procedures; for
6474 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6475 calls @code{gdb_test} multiple times.
6477 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6478 necessary, such as when @value{GDBN} has several valid responses to a command.
6480 The source language programs do @emph{not} need to be in a consistent
6481 style. Since @value{GDBN} is used to debug programs written in many different
6482 styles, it's worth having a mix of styles in the testsuite; for
6483 instance, some @value{GDBN} bugs involving the display of source lines would
6484 never manifest themselves if the programs used GNU coding style
6491 Check the @file{README} file, it often has useful information that does not
6492 appear anywhere else in the directory.
6495 * Getting Started:: Getting started working on @value{GDBN}
6496 * Debugging GDB:: Debugging @value{GDBN} with itself
6499 @node Getting Started,,, Hints
6501 @section Getting Started
6503 @value{GDBN} is a large and complicated program, and if you first starting to
6504 work on it, it can be hard to know where to start. Fortunately, if you
6505 know how to go about it, there are ways to figure out what is going on.
6507 This manual, the @value{GDBN} Internals manual, has information which applies
6508 generally to many parts of @value{GDBN}.
6510 Information about particular functions or data structures are located in
6511 comments with those functions or data structures. If you run across a
6512 function or a global variable which does not have a comment correctly
6513 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6514 free to submit a bug report, with a suggested comment if you can figure
6515 out what the comment should say. If you find a comment which is
6516 actually wrong, be especially sure to report that.
6518 Comments explaining the function of macros defined in host, target, or
6519 native dependent files can be in several places. Sometimes they are
6520 repeated every place the macro is defined. Sometimes they are where the
6521 macro is used. Sometimes there is a header file which supplies a
6522 default definition of the macro, and the comment is there. This manual
6523 also documents all the available macros.
6524 @c (@pxref{Host Conditionals}, @pxref{Target
6525 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
6528 Start with the header files. Once you have some idea of how
6529 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
6530 @file{gdbtypes.h}), you will find it much easier to understand the
6531 code which uses and creates those symbol tables.
6533 You may wish to process the information you are getting somehow, to
6534 enhance your understanding of it. Summarize it, translate it to another
6535 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
6536 the code to predict what a test case would do and write the test case
6537 and verify your prediction, etc. If you are reading code and your eyes
6538 are starting to glaze over, this is a sign you need to use a more active
6541 Once you have a part of @value{GDBN} to start with, you can find more
6542 specifically the part you are looking for by stepping through each
6543 function with the @code{next} command. Do not use @code{step} or you
6544 will quickly get distracted; when the function you are stepping through
6545 calls another function try only to get a big-picture understanding
6546 (perhaps using the comment at the beginning of the function being
6547 called) of what it does. This way you can identify which of the
6548 functions being called by the function you are stepping through is the
6549 one which you are interested in. You may need to examine the data
6550 structures generated at each stage, with reference to the comments in
6551 the header files explaining what the data structures are supposed to
6554 Of course, this same technique can be used if you are just reading the
6555 code, rather than actually stepping through it. The same general
6556 principle applies---when the code you are looking at calls something
6557 else, just try to understand generally what the code being called does,
6558 rather than worrying about all its details.
6560 @cindex command implementation
6561 A good place to start when tracking down some particular area is with
6562 a command which invokes that feature. Suppose you want to know how
6563 single-stepping works. As a @value{GDBN} user, you know that the
6564 @code{step} command invokes single-stepping. The command is invoked
6565 via command tables (see @file{command.h}); by convention the function
6566 which actually performs the command is formed by taking the name of
6567 the command and adding @samp{_command}, or in the case of an
6568 @code{info} subcommand, @samp{_info}. For example, the @code{step}
6569 command invokes the @code{step_command} function and the @code{info
6570 display} command invokes @code{display_info}. When this convention is
6571 not followed, you might have to use @code{grep} or @kbd{M-x
6572 tags-search} in emacs, or run @value{GDBN} on itself and set a
6573 breakpoint in @code{execute_command}.
6575 @cindex @code{bug-gdb} mailing list
6576 If all of the above fail, it may be appropriate to ask for information
6577 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
6578 wondering if anyone could give me some tips about understanding
6579 @value{GDBN}''---if we had some magic secret we would put it in this manual.
6580 Suggestions for improving the manual are always welcome, of course.
6582 @node Debugging GDB,,,Hints
6584 @section Debugging @value{GDBN} with itself
6585 @cindex debugging @value{GDBN}
6587 If @value{GDBN} is limping on your machine, this is the preferred way to get it
6588 fully functional. Be warned that in some ancient Unix systems, like
6589 Ultrix 4.2, a program can't be running in one process while it is being
6590 debugged in another. Rather than typing the command @kbd{@w{./gdb
6591 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
6592 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
6594 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
6595 @file{.gdbinit} file that sets up some simple things to make debugging
6596 gdb easier. The @code{info} command, when executed without a subcommand
6597 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
6598 gdb. See @file{.gdbinit} for details.
6600 If you use emacs, you will probably want to do a @code{make TAGS} after
6601 you configure your distribution; this will put the machine dependent
6602 routines for your local machine where they will be accessed first by
6605 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
6606 have run @code{fixincludes} if you are compiling with gcc.
6608 @section Submitting Patches
6610 @cindex submitting patches
6611 Thanks for thinking of offering your changes back to the community of
6612 @value{GDBN} users. In general we like to get well designed enhancements.
6613 Thanks also for checking in advance about the best way to transfer the
6616 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
6617 This manual summarizes what we believe to be clean design for @value{GDBN}.
6619 If the maintainers don't have time to put the patch in when it arrives,
6620 or if there is any question about a patch, it goes into a large queue
6621 with everyone else's patches and bug reports.
6623 @cindex legal papers for code contributions
6624 The legal issue is that to incorporate substantial changes requires a
6625 copyright assignment from you and/or your employer, granting ownership
6626 of the changes to the Free Software Foundation. You can get the
6627 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
6628 and asking for it. We recommend that people write in "All programs
6629 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
6630 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
6632 contributed with only one piece of legalese pushed through the
6633 bureaucracy and filed with the FSF. We can't start merging changes until
6634 this paperwork is received by the FSF (their rules, which we follow
6635 since we maintain it for them).
6637 Technically, the easiest way to receive changes is to receive each
6638 feature as a small context diff or unidiff, suitable for @code{patch}.
6639 Each message sent to me should include the changes to C code and
6640 header files for a single feature, plus @file{ChangeLog} entries for
6641 each directory where files were modified, and diffs for any changes
6642 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
6643 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
6644 single feature, they can be split down into multiple messages.
6646 In this way, if we read and like the feature, we can add it to the
6647 sources with a single patch command, do some testing, and check it in.
6648 If you leave out the @file{ChangeLog}, we have to write one. If you leave
6649 out the doc, we have to puzzle out what needs documenting. Etc., etc.
6651 The reason to send each change in a separate message is that we will not
6652 install some of the changes. They'll be returned to you with questions
6653 or comments. If we're doing our job correctly, the message back to you
6654 will say what you have to fix in order to make the change acceptable.
6655 The reason to have separate messages for separate features is so that
6656 the acceptable changes can be installed while one or more changes are
6657 being reworked. If multiple features are sent in a single message, we
6658 tend to not put in the effort to sort out the acceptable changes from
6659 the unacceptable, so none of the features get installed until all are
6662 If this sounds painful or authoritarian, well, it is. But we get a lot
6663 of bug reports and a lot of patches, and many of them don't get
6664 installed because we don't have the time to finish the job that the bug
6665 reporter or the contributor could have done. Patches that arrive
6666 complete, working, and well designed, tend to get installed on the day
6667 they arrive. The others go into a queue and get installed as time
6668 permits, which, since the maintainers have many demands to meet, may not
6669 be for quite some time.
6671 Please send patches directly to
6672 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
6674 @section Obsolete Conditionals
6675 @cindex obsolete code
6677 Fragments of old code in @value{GDBN} sometimes reference or set the following
6678 configuration macros. They should not be used by new code, and old uses
6679 should be removed as those parts of the debugger are otherwise touched.
6682 @item STACK_END_ADDR
6683 This macro used to define where the end of the stack appeared, for use
6684 in interpreting core file formats that don't record this address in the
6685 core file itself. This information is now configured in BFD, and @value{GDBN}
6686 gets the info portably from there. The values in @value{GDBN}'s configuration
6687 files should be moved into BFD configuration files (if needed there),
6688 and deleted from all of @value{GDBN}'s config files.
6690 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
6691 is so old that it has never been converted to use BFD. Now that's old!
6695 @include observer.texi