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 (C) 1990, 1991, 1992, 1993, 1994, 1996, 1998, 1999, 2000, 2001,
12 2002, 2003, 2004, 2005, 2006
13 Free Software Foundation, Inc.
14 Contributed by Cygnus Solutions. Written by John Gilmore.
15 Second Edition by Stan Shebs.
17 Permission is granted to copy, distribute and/or modify this document
18 under the terms of the GNU Free Documentation License, Version 1.1 or
19 any later version published by the Free Software Foundation; with no
20 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
21 Texts. A copy of the license is included in the section entitled ``GNU
22 Free Documentation License''.
25 @setchapternewpage off
26 @settitle @value{GDBN} Internals
32 @title @value{GDBN} Internals
33 @subtitle{A guide to the internals of the GNU debugger}
35 @author Cygnus Solutions
36 @author Second Edition:
38 @author Cygnus Solutions
41 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
42 \xdef\manvers{\$Revision$} % For use in headers, footers too
44 \hfill Cygnus Solutions\par
46 \hfill \TeX{}info \texinfoversion\par
50 @vskip 0pt plus 1filll
51 Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001,
52 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
54 Permission is granted to copy, distribute and/or modify this document
55 under the terms of the GNU Free Documentation License, Version 1.1 or
56 any later version published by the Free Software Foundation; with no
57 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
58 Texts. A copy of the license is included in the section entitled ``GNU
59 Free Documentation License''.
65 @c Perhaps this should be the title of the document (but only for info,
66 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
67 @top Scope of this Document
69 This document documents the internals of the GNU debugger, @value{GDBN}. It
70 includes description of @value{GDBN}'s key algorithms and operations, as well
71 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
82 * Target Architecture Definition::
83 * Target Vector Definition::
88 * Versions and Branches::
89 * Start of New Year Procedure::
94 * GDB Observers:: @value{GDBN} Currently available observers
95 * GNU Free Documentation License:: The license for this documentation
101 @chapter Requirements
102 @cindex requirements for @value{GDBN}
104 Before diving into the internals, you should understand the formal
105 requirements and other expectations for @value{GDBN}. Although some
106 of these may seem obvious, there have been proposals for @value{GDBN}
107 that have run counter to these requirements.
109 First of all, @value{GDBN} is a debugger. It's not designed to be a
110 front panel for embedded systems. It's not a text editor. It's not a
111 shell. It's not a programming environment.
113 @value{GDBN} is an interactive tool. Although a batch mode is
114 available, @value{GDBN}'s primary role is to interact with a human
117 @value{GDBN} should be responsive to the user. A programmer hot on
118 the trail of a nasty bug, and operating under a looming deadline, is
119 going to be very impatient of everything, including the response time
120 to debugger commands.
122 @value{GDBN} should be relatively permissive, such as for expressions.
123 While the compiler should be picky (or have the option to be made
124 picky), since source code lives for a long time usually, the
125 programmer doing debugging shouldn't be spending time figuring out to
126 mollify the debugger.
128 @value{GDBN} will be called upon to deal with really large programs.
129 Executable sizes of 50 to 100 megabytes occur regularly, and we've
130 heard reports of programs approaching 1 gigabyte in size.
132 @value{GDBN} should be able to run everywhere. No other debugger is
133 available for even half as many configurations as @value{GDBN}
137 @node Overall Structure
139 @chapter Overall Structure
141 @value{GDBN} consists of three major subsystems: user interface,
142 symbol handling (the @dfn{symbol side}), and target system handling (the
145 The user interface consists of several actual interfaces, plus
148 The symbol side consists of object file readers, debugging info
149 interpreters, symbol table management, source language expression
150 parsing, type and value printing.
152 The target side consists of execution control, stack frame analysis, and
153 physical target manipulation.
155 The target side/symbol side division is not formal, and there are a
156 number of exceptions. For instance, core file support involves symbolic
157 elements (the basic core file reader is in BFD) and target elements (it
158 supplies the contents of memory and the values of registers). Instead,
159 this division is useful for understanding how the minor subsystems
162 @section The Symbol Side
164 The symbolic side of @value{GDBN} can be thought of as ``everything
165 you can do in @value{GDBN} without having a live program running''.
166 For instance, you can look at the types of variables, and evaluate
167 many kinds of expressions.
169 @section The Target Side
171 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
172 Although it may make reference to symbolic info here and there, most
173 of the target side will run with only a stripped executable
174 available---or even no executable at all, in remote debugging cases.
176 Operations such as disassembly, stack frame crawls, and register
177 display, are able to work with no symbolic info at all. In some cases,
178 such as disassembly, @value{GDBN} will use symbolic info to present addresses
179 relative to symbols rather than as raw numbers, but it will work either
182 @section Configurations
186 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
187 @dfn{Target} refers to the system where the program being debugged
188 executes. In most cases they are the same machine, in which case a
189 third type of @dfn{Native} attributes come into play.
191 Defines and include files needed to build on the host are host support.
192 Examples are tty support, system defined types, host byte order, host
195 Defines and information needed to handle the target format are target
196 dependent. Examples are the stack frame format, instruction set,
197 breakpoint instruction, registers, and how to set up and tear down the stack
200 Information that is only needed when the host and target are the same,
201 is native dependent. One example is Unix child process support; if the
202 host and target are not the same, doing a fork to start the target
203 process is a bad idea. The various macros needed for finding the
204 registers in the @code{upage}, running @code{ptrace}, and such are all
205 in the native-dependent files.
207 Another example of native-dependent code is support for features that
208 are really part of the target environment, but which require
209 @code{#include} files that are only available on the host system. Core
210 file handling and @code{setjmp} handling are two common cases.
212 When you want to make @value{GDBN} work ``native'' on a particular machine, you
213 have to include all three kinds of information.
221 @value{GDBN} uses a number of debugging-specific algorithms. They are
222 often not very complicated, but get lost in the thicket of special
223 cases and real-world issues. This chapter describes the basic
224 algorithms and mentions some of the specific target definitions that
230 @cindex call stack frame
231 A frame is a construct that @value{GDBN} uses to keep track of calling
232 and called functions.
234 @cindex frame, unwind
235 @value{GDBN}'s current frame model is the result of an incremental
236 cleanup of working code, not a fresh design, so it's a little weird.
238 The natural model would have a frame object, with methods that read
239 and write that frame's registers. Reading or writing the youngest
240 frame's registers would simply read or write the processor's current
241 registers, since the youngest frame is running directly on the
242 processor. Older frames might have some registers saved on the stack
243 by younger frames, so accessing the older frames' registers would do a
244 mix of memory accesses and register accesses, as appropriate.
246 @findex frame_register_unwind
247 Instead, @value{GDBN}'s model is that you find a frame's registers by
248 ``unwinding'' them from the next younger frame. That is, to access
249 the registers of frame #1 (the next-to-youngest frame), you actually
250 apply @code{frame_register_unwind} to frame #0 (the youngest frame).
251 But then the obvious question is: how do you access the registers of
252 the youngest frame itself? How do you ``unwind'' them when they're
255 @cindex sentinel frame
256 @findex get_frame_type
257 @vindex SENTINEL_FRAME
258 To answer this question, GDB has the @dfn{sentinel} frame, the
259 ``-1st'' frame. Unwinding registers from the sentinel frame gives you
260 the current values of the youngest real frame's registers. If @var{f}
261 is a sentinel frame, then @code{get_frame_type (@var{f}) ==
264 @findex create_new_frame
266 @code{FRAME_FP} in the machine description has no meaning to the
267 machine-independent part of @value{GDBN}, except that it is used when
268 setting up a new frame from scratch, as follows:
271 create_new_frame (read_register (DEPRECATED_FP_REGNUM), read_pc ()));
274 @cindex frame pointer register
275 Other than that, all the meaning imparted to @code{DEPRECATED_FP_REGNUM}
276 is imparted by the machine-dependent code. So,
277 @code{DEPRECATED_FP_REGNUM} can have any value that is convenient for
278 the code that creates new frames. (@code{create_new_frame} calls
279 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} if it is defined; that is where
280 you should use the @code{DEPRECATED_FP_REGNUM} value, if your frames are
284 Given a @value{GDBN} frame, define @code{DEPRECATED_FRAME_CHAIN} to
285 determine the address of the calling function's frame. This will be
286 used to create a new @value{GDBN} frame struct, and then
287 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} and
288 @code{DEPRECATED_INIT_FRAME_PC} will be called for the new frame.
290 @section Prologue Analysis
292 @cindex prologue analysis
293 @cindex call frame information
294 @cindex CFI (call frame information)
295 To produce a backtrace and allow the user to manipulate older frames'
296 variables and arguments, @value{GDBN} needs to find the base addresses
297 of older frames, and discover where those frames' registers have been
298 saved. Since a frame's ``callee-saves'' registers get saved by
299 younger frames if and when they're reused, a frame's registers may be
300 scattered unpredictably across younger frames. This means that
301 changing the value of a register-allocated variable in an older frame
302 may actually entail writing to a save slot in some younger frame.
304 Modern versions of GCC emit Dwarf call frame information (``CFI''),
305 which describes how to find frame base addresses and saved registers.
306 But CFI is not always available, so as a fallback @value{GDBN} uses a
307 technique called @dfn{prologue analysis} to find frame sizes and saved
308 registers. A prologue analyzer disassembles the function's machine
309 code starting from its entry point, and looks for instructions that
310 allocate frame space, save the stack pointer in a frame pointer
311 register, save registers, and so on. Obviously, this can't be done
312 accurately in general, but it's tractible to do well enough to be very
313 helpful. Prologue analysis predates the GNU toolchain's support for
314 CFI; at one time, prologue analysis was the only mechanism
315 @value{GDBN} used for stack unwinding at all, when the function
316 calling conventions didn't specify a fixed frame layout.
318 In the olden days, function prologues were generated by hand-written,
319 target-specific code in GCC, and treated as opaque and untouchable by
320 optimizers. Looking at this code, it was usually straightforward to
321 write a prologue analyzer for @value{GDBN} that would accurately
322 understand all the prologues GCC would generate. However, over time
323 GCC became more aggressive about instruction scheduling, and began to
324 understand more about the semantics of the prologue instructions
325 themselves; in response, @value{GDBN}'s analyzers became more complex
326 and fragile. Keeping the prologue analyzers working as GCC (and the
327 instruction sets themselves) evolved became a substantial task.
329 @cindex @file{prologue-value.c}
330 @cindex abstract interpretation of function prologues
331 @cindex pseudo-evaluation of function prologues
332 To try to address this problem, the code in @file{prologue-value.h}
333 and @file{prologue-value.c} provides a general framework for writing
334 prologue analyzers that are simpler and more robust than ad-hoc
335 analyzers. When we analyze a prologue using the prologue-value
336 framework, we're really doing ``abstract interpretation'' or
337 ``pseudo-evaluation'': running the function's code in simulation, but
338 using conservative approximations of the values registers and memory
339 would hold when the code actually runs. For example, if our function
340 starts with the instruction:
343 addi r1, 42 # add 42 to r1
346 we don't know exactly what value will be in @code{r1} after executing
347 this instruction, but we do know it'll be 42 greater than its original
350 If we then see an instruction like:
353 addi r1, 22 # add 22 to r1
356 we still don't know what @code{r1's} value is, but again, we can say
357 it is now 64 greater than its original value.
359 If the next instruction were:
362 mov r2, r1 # set r2 to r1's value
365 then we can say that @code{r2's} value is now the original value of
368 It's common for prologues to save registers on the stack, so we'll
369 need to track the values of stack frame slots, as well as the
370 registers. So after an instruction like this:
376 then we'd know that the stack slot four bytes above the frame pointer
377 holds the original value of @code{r1} plus 64.
381 Of course, this can only go so far before it gets unreasonable. If we
382 wanted to be able to say anything about the value of @code{r1} after
386 xor r1, r3 # exclusive-or r1 and r3, place result in r1
389 then things would get pretty complex. But remember, we're just doing
390 a conservative approximation; if exclusive-or instructions aren't
391 relevant to prologues, we can just say @code{r1}'s value is now
392 ``unknown''. We can ignore things that are too complex, if that loss of
393 information is acceptable for our application.
395 So when we say ``conservative approximation'' here, what we mean is an
396 approximation that is either accurate, or marked ``unknown'', but
399 Using this framework, a prologue analyzer is simply an interpreter for
400 machine code, but one that uses conservative approximations for the
401 contents of registers and memory instead of actual values. Starting
402 from the function's entry point, you simulate instructions up to the
403 current PC, or an instruction that you don't know how to simulate.
404 Now you can examine the state of the registers and stack slots you've
410 To see how large your stack frame is, just check the value of the
411 stack pointer register; if it's the original value of the SP
412 minus a constant, then that constant is the stack frame's size.
413 If the SP's value has been marked as ``unknown'', then that means
414 the prologue has done something too complex for us to track, and
415 we don't know the frame size.
418 To see where we've saved the previous frame's registers, we just
419 search the values we've tracked --- stack slots, usually, but
420 registers, too, if you want --- for something equal to the register's
421 original value. If the calling conventions suggest a standard place
422 to save a given register, then we can check there first, but really,
423 anything that will get us back the original value will probably work.
426 This does take some work. But prologue analyzers aren't
427 quick-and-simple pattern patching to recognize a few fixed prologue
428 forms any more; they're big, hairy functions. Along with inferior
429 function calls, prologue analysis accounts for a substantial portion
430 of the time needed to stabilize a @value{GDBN} port. So it's
431 worthwhile to look for an approach that will be easier to understand
432 and maintain. In the approach described above:
437 It's easier to see that the analyzer is correct: you just see
438 whether the analyzer properly (albiet conservatively) simulates
439 the effect of each instruction.
442 It's easier to extend the analyzer: you can add support for new
443 instructions, and know that you haven't broken anything that
444 wasn't already broken before.
447 It's orthogonal: to gather new information, you don't need to
448 complicate the code for each instruction. As long as your domain
449 of conservative values is already detailed enough to tell you
450 what you need, then all the existing instruction simulations are
451 already gathering the right data for you.
455 The file @file{prologue-value.h} contains detailed comments explaining
456 the framework and how to use it.
459 @section Breakpoint Handling
462 In general, a breakpoint is a user-designated location in the program
463 where the user wants to regain control if program execution ever reaches
466 There are two main ways to implement breakpoints; either as ``hardware''
467 breakpoints or as ``software'' breakpoints.
469 @cindex hardware breakpoints
470 @cindex program counter
471 Hardware breakpoints are sometimes available as a builtin debugging
472 features with some chips. Typically these work by having dedicated
473 register into which the breakpoint address may be stored. If the PC
474 (shorthand for @dfn{program counter})
475 ever matches a value in a breakpoint registers, the CPU raises an
476 exception and reports it to @value{GDBN}.
478 Another possibility is when an emulator is in use; many emulators
479 include circuitry that watches the address lines coming out from the
480 processor, and force it to stop if the address matches a breakpoint's
483 A third possibility is that the target already has the ability to do
484 breakpoints somehow; for instance, a ROM monitor may do its own
485 software breakpoints. So although these are not literally ``hardware
486 breakpoints'', from @value{GDBN}'s point of view they work the same;
487 @value{GDBN} need not do anything more than set the breakpoint and wait
488 for something to happen.
490 Since they depend on hardware resources, hardware breakpoints may be
491 limited in number; when the user asks for more, @value{GDBN} will
492 start trying to set software breakpoints. (On some architectures,
493 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
494 whether there's enough hardware resources to insert all the hardware
495 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
496 an error message only when the program being debugged is continued.)
498 @cindex software breakpoints
499 Software breakpoints require @value{GDBN} to do somewhat more work.
500 The basic theory is that @value{GDBN} will replace a program
501 instruction with a trap, illegal divide, or some other instruction
502 that will cause an exception, and then when it's encountered,
503 @value{GDBN} will take the exception and stop the program. When the
504 user says to continue, @value{GDBN} will restore the original
505 instruction, single-step, re-insert the trap, and continue on.
507 Since it literally overwrites the program being tested, the program area
508 must be writable, so this technique won't work on programs in ROM. It
509 can also distort the behavior of programs that examine themselves,
510 although such a situation would be highly unusual.
512 Also, the software breakpoint instruction should be the smallest size of
513 instruction, so it doesn't overwrite an instruction that might be a jump
514 target, and cause disaster when the program jumps into the middle of the
515 breakpoint instruction. (Strictly speaking, the breakpoint must be no
516 larger than the smallest interval between instructions that may be jump
517 targets; perhaps there is an architecture where only even-numbered
518 instructions may jumped to.) Note that it's possible for an instruction
519 set not to have any instructions usable for a software breakpoint,
520 although in practice only the ARC has failed to define such an
524 The basic definition of the software breakpoint is the macro
527 Basic breakpoint object handling is in @file{breakpoint.c}. However,
528 much of the interesting breakpoint action is in @file{infrun.c}.
531 @cindex insert or remove software breakpoint
532 @findex target_remove_breakpoint
533 @findex target_insert_breakpoint
534 @item target_remove_breakpoint (@var{bp_tgt})
535 @itemx target_insert_breakpoint (@var{bp_tgt})
536 Insert or remove a software breakpoint at address
537 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
538 non-zero for failure. On input, @var{bp_tgt} contains the address of the
539 breakpoint, and is otherwise initialized to zero. The fields of the
540 @code{struct bp_target_info} pointed to by @var{bp_tgt} are updated
541 to contain other information about the breakpoint on output. The field
542 @code{placed_address} may be updated if the breakpoint was placed at a
543 related address; the field @code{shadow_contents} contains the real
544 contents of the bytes where the breakpoint has been inserted,
545 if reading memory would return the breakpoint instead of the
546 underlying memory; the field @code{shadow_len} is the length of
547 memory cached in @code{shadow_contents}, if any; and the field
548 @code{placed_size} is optionally set and used by the target, if
549 it could differ from @code{shadow_len}.
551 For example, the remote target @samp{Z0} packet does not require
552 shadowing memory, so @code{shadow_len} is left at zero. However,
553 the length reported by @code{BREAKPOINT_FROM_PC} is cached in
554 @code{placed_size}, so that a matching @samp{z0} packet can be
555 used to remove the breakpoint.
557 @cindex insert or remove hardware breakpoint
558 @findex target_remove_hw_breakpoint
559 @findex target_insert_hw_breakpoint
560 @item target_remove_hw_breakpoint (@var{bp_tgt})
561 @itemx target_insert_hw_breakpoint (@var{bp_tgt})
562 Insert or remove a hardware-assisted breakpoint at address
563 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
564 non-zero for failure. See @code{target_insert_breakpoint} for
565 a description of the @code{struct bp_target_info} pointed to by
566 @var{bp_tgt}; the @code{shadow_contents} and
567 @code{shadow_len} members are not used for hardware breakpoints,
568 but @code{placed_size} may be.
571 @section Single Stepping
573 @section Signal Handling
575 @section Thread Handling
577 @section Inferior Function Calls
579 @section Longjmp Support
581 @cindex @code{longjmp} debugging
582 @value{GDBN} has support for figuring out that the target is doing a
583 @code{longjmp} and for stopping at the target of the jump, if we are
584 stepping. This is done with a few specialized internal breakpoints,
585 which are visible in the output of the @samp{maint info breakpoint}
588 @findex GET_LONGJMP_TARGET
589 To make this work, you need to define a macro called
590 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
591 structure and extract the longjmp target address. Since @code{jmp_buf}
592 is target specific, you will need to define it in the appropriate
593 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
594 @file{sparc-tdep.c} for examples of how to do this.
599 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
600 breakpoints}) which break when data is accessed rather than when some
601 instruction is executed. When you have data which changes without
602 your knowing what code does that, watchpoints are the silver bullet to
603 hunt down and kill such bugs.
605 @cindex hardware watchpoints
606 @cindex software watchpoints
607 Watchpoints can be either hardware-assisted or not; the latter type is
608 known as ``software watchpoints.'' @value{GDBN} always uses
609 hardware-assisted watchpoints if they are available, and falls back on
610 software watchpoints otherwise. Typical situations where @value{GDBN}
611 will use software watchpoints are:
615 The watched memory region is too large for the underlying hardware
616 watchpoint support. For example, each x86 debug register can watch up
617 to 4 bytes of memory, so trying to watch data structures whose size is
618 more than 16 bytes will cause @value{GDBN} to use software
622 The value of the expression to be watched depends on data held in
623 registers (as opposed to memory).
626 Too many different watchpoints requested. (On some architectures,
627 this situation is impossible to detect until the debugged program is
628 resumed.) Note that x86 debug registers are used both for hardware
629 breakpoints and for watchpoints, so setting too many hardware
630 breakpoints might cause watchpoint insertion to fail.
633 No hardware-assisted watchpoints provided by the target
637 Software watchpoints are very slow, since @value{GDBN} needs to
638 single-step the program being debugged and test the value of the
639 watched expression(s) after each instruction. The rest of this
640 section is mostly irrelevant for software watchpoints.
642 When the inferior stops, @value{GDBN} tries to establish, among other
643 possible reasons, whether it stopped due to a watchpoint being hit.
644 For a data-write watchpoint, it does so by evaluating, for each
645 watchpoint, the expression whose value is being watched, and testing
646 whether the watched value has changed. For data-read and data-access
647 watchpoints, @value{GDBN} needs the target to supply a primitive that
648 returns the address of the data that was accessed or read (see the
649 description of @code{target_stopped_data_address} below): if this
650 primitive returns a valid address, @value{GDBN} infers that a
651 watchpoint triggered if it watches an expression whose evaluation uses
654 @value{GDBN} uses several macros and primitives to support hardware
658 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
659 @item TARGET_HAS_HARDWARE_WATCHPOINTS
660 If defined, the target supports hardware watchpoints.
662 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
663 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
664 Return the number of hardware watchpoints of type @var{type} that are
665 possible to be set. The value is positive if @var{count} watchpoints
666 of this type can be set, zero if setting watchpoints of this type is
667 not supported, and negative if @var{count} is more than the maximum
668 number of watchpoints of type @var{type} that can be set. @var{other}
669 is non-zero if other types of watchpoints are currently enabled (there
670 are architectures which cannot set watchpoints of different types at
673 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
674 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
675 Return non-zero if hardware watchpoints can be used to watch a region
676 whose address is @var{addr} and whose length in bytes is @var{len}.
678 @cindex insert or remove hardware watchpoint
679 @findex target_insert_watchpoint
680 @findex target_remove_watchpoint
681 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
682 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
683 Insert or remove a hardware watchpoint starting at @var{addr}, for
684 @var{len} bytes. @var{type} is the watchpoint type, one of the
685 possible values of the enumerated data type @code{target_hw_bp_type},
686 defined by @file{breakpoint.h} as follows:
689 enum target_hw_bp_type
691 hw_write = 0, /* Common (write) HW watchpoint */
692 hw_read = 1, /* Read HW watchpoint */
693 hw_access = 2, /* Access (read or write) HW watchpoint */
694 hw_execute = 3 /* Execute HW breakpoint */
699 These two macros should return 0 for success, non-zero for failure.
701 @findex target_stopped_data_address
702 @item target_stopped_data_address (@var{addr_p})
703 If the inferior has some watchpoint that triggered, place the address
704 associated with the watchpoint at the location pointed to by
705 @var{addr_p} and return non-zero. Otherwise, return zero. Note that
706 this primitive is used by @value{GDBN} only on targets that support
707 data-read or data-access type watchpoints, so targets that have
708 support only for data-write watchpoints need not implement these
711 @findex HAVE_STEPPABLE_WATCHPOINT
712 @item HAVE_STEPPABLE_WATCHPOINT
713 If defined to a non-zero value, it is not necessary to disable a
714 watchpoint to step over it.
716 @findex HAVE_NONSTEPPABLE_WATCHPOINT
717 @item HAVE_NONSTEPPABLE_WATCHPOINT
718 If defined to a non-zero value, @value{GDBN} should disable a
719 watchpoint to step the inferior over it.
721 @findex HAVE_CONTINUABLE_WATCHPOINT
722 @item HAVE_CONTINUABLE_WATCHPOINT
723 If defined to a non-zero value, it is possible to continue the
724 inferior after a watchpoint has been hit.
726 @findex CANNOT_STEP_HW_WATCHPOINTS
727 @item CANNOT_STEP_HW_WATCHPOINTS
728 If this is defined to a non-zero value, @value{GDBN} will remove all
729 watchpoints before stepping the inferior.
731 @findex STOPPED_BY_WATCHPOINT
732 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
733 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
734 the type @code{struct target_waitstatus}, defined by @file{target.h}.
735 Normally, this macro is defined to invoke the function pointed to by
736 the @code{to_stopped_by_watchpoint} member of the structure (of the
737 type @code{target_ops}, defined on @file{target.h}) that describes the
738 target-specific operations; @code{to_stopped_by_watchpoint} ignores
739 the @var{wait_status} argument.
741 @value{GDBN} does not require the non-zero value returned by
742 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
743 determine for sure whether the inferior stopped due to a watchpoint,
744 it could return non-zero ``just in case''.
747 @subsection x86 Watchpoints
748 @cindex x86 debug registers
749 @cindex watchpoints, on x86
751 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
752 registers designed to facilitate debugging. @value{GDBN} provides a
753 generic library of functions that x86-based ports can use to implement
754 support for watchpoints and hardware-assisted breakpoints. This
755 subsection documents the x86 watchpoint facilities in @value{GDBN}.
757 To use the generic x86 watchpoint support, a port should do the
761 @findex I386_USE_GENERIC_WATCHPOINTS
763 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
764 target-dependent headers.
767 Include the @file{config/i386/nm-i386.h} header file @emph{after}
768 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
771 Add @file{i386-nat.o} to the value of the Make variable
772 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
773 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
776 Provide implementations for the @code{I386_DR_LOW_*} macros described
777 below. Typically, each macro should call a target-specific function
778 which does the real work.
781 The x86 watchpoint support works by maintaining mirror images of the
782 debug registers. Values are copied between the mirror images and the
783 real debug registers via a set of macros which each target needs to
787 @findex I386_DR_LOW_SET_CONTROL
788 @item I386_DR_LOW_SET_CONTROL (@var{val})
789 Set the Debug Control (DR7) register to the value @var{val}.
791 @findex I386_DR_LOW_SET_ADDR
792 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
793 Put the address @var{addr} into the debug register number @var{idx}.
795 @findex I386_DR_LOW_RESET_ADDR
796 @item I386_DR_LOW_RESET_ADDR (@var{idx})
797 Reset (i.e.@: zero out) the address stored in the debug register
800 @findex I386_DR_LOW_GET_STATUS
801 @item I386_DR_LOW_GET_STATUS
802 Return the value of the Debug Status (DR6) register. This value is
803 used immediately after it is returned by
804 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
808 For each one of the 4 debug registers (whose indices are from 0 to 3)
809 that store addresses, a reference count is maintained by @value{GDBN},
810 to allow sharing of debug registers by several watchpoints. This
811 allows users to define several watchpoints that watch the same
812 expression, but with different conditions and/or commands, without
813 wasting debug registers which are in short supply. @value{GDBN}
814 maintains the reference counts internally, targets don't have to do
815 anything to use this feature.
817 The x86 debug registers can each watch a region that is 1, 2, or 4
818 bytes long. The ia32 architecture requires that each watched region
819 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
820 region on 4-byte boundary. However, the x86 watchpoint support in
821 @value{GDBN} can watch unaligned regions and regions larger than 4
822 bytes (up to 16 bytes) by allocating several debug registers to watch
823 a single region. This allocation of several registers per a watched
824 region is also done automatically without target code intervention.
826 The generic x86 watchpoint support provides the following API for the
827 @value{GDBN}'s application code:
830 @findex i386_region_ok_for_watchpoint
831 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
832 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
833 this function. It counts the number of debug registers required to
834 watch a given region, and returns a non-zero value if that number is
835 less than 4, the number of debug registers available to x86
838 @findex i386_stopped_data_address
839 @item i386_stopped_data_address (@var{addr_p})
841 @code{target_stopped_data_address} is set to call this function.
843 function examines the breakpoint condition bits in the DR6 Debug
844 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
845 macro, and returns the address associated with the first bit that is
848 @findex i386_stopped_by_watchpoint
849 @item i386_stopped_by_watchpoint (void)
850 The macro @code{STOPPED_BY_WATCHPOINT}
851 is set to call this function. The
852 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
853 function examines the breakpoint condition bits in the DR6 Debug
854 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
855 macro, and returns true if any bit is set. Otherwise, false is
858 @findex i386_insert_watchpoint
859 @findex i386_remove_watchpoint
860 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
861 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
862 Insert or remove a watchpoint. The macros
863 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
864 are set to call these functions. @code{i386_insert_watchpoint} first
865 looks for a debug register which is already set to watch the same
866 region for the same access types; if found, it just increments the
867 reference count of that debug register, thus implementing debug
868 register sharing between watchpoints. If no such register is found,
869 the function looks for a vacant debug register, sets its mirrored
870 value to @var{addr}, sets the mirrored value of DR7 Debug Control
871 register as appropriate for the @var{len} and @var{type} parameters,
872 and then passes the new values of the debug register and DR7 to the
873 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
874 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
875 required to cover the given region, the above process is repeated for
878 @code{i386_remove_watchpoint} does the opposite: it resets the address
879 in the mirrored value of the debug register and its read/write and
880 length bits in the mirrored value of DR7, then passes these new
881 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
882 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
883 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
884 decrements the reference count, and only calls
885 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
886 the count goes to zero.
888 @findex i386_insert_hw_breakpoint
889 @findex i386_remove_hw_breakpoint
890 @item i386_insert_hw_breakpoint (@var{bp_tgt})
891 @itemx i386_remove_hw_breakpoint (@var{bp_tgt})
892 These functions insert and remove hardware-assisted breakpoints. The
893 macros @code{target_insert_hw_breakpoint} and
894 @code{target_remove_hw_breakpoint} are set to call these functions.
895 The argument is a @code{struct bp_target_info *}, as described in
896 the documentation for @code{target_insert_breakpoint}.
897 These functions work like @code{i386_insert_watchpoint} and
898 @code{i386_remove_watchpoint}, respectively, except that they set up
899 the debug registers to watch instruction execution, and each
900 hardware-assisted breakpoint always requires exactly one debug
903 @findex i386_stopped_by_hwbp
904 @item i386_stopped_by_hwbp (void)
905 This function returns non-zero if the inferior has some watchpoint or
906 hardware breakpoint that triggered. It works like
907 @code{i386_stopped_data_address}, except that it doesn't record the
908 address whose watchpoint triggered.
910 @findex i386_cleanup_dregs
911 @item i386_cleanup_dregs (void)
912 This function clears all the reference counts, addresses, and control
913 bits in the mirror images of the debug registers. It doesn't affect
914 the actual debug registers in the inferior process.
921 x86 processors support setting watchpoints on I/O reads or writes.
922 However, since no target supports this (as of March 2001), and since
923 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
924 watchpoints, this feature is not yet available to @value{GDBN} running
928 x86 processors can enable watchpoints locally, for the current task
929 only, or globally, for all the tasks. For each debug register,
930 there's a bit in the DR7 Debug Control register that determines
931 whether the associated address is watched locally or globally. The
932 current implementation of x86 watchpoint support in @value{GDBN}
933 always sets watchpoints to be locally enabled, since global
934 watchpoints might interfere with the underlying OS and are probably
935 unavailable in many platforms.
941 In the abstract, a checkpoint is a point in the execution history of
942 the program, which the user may wish to return to at some later time.
944 Internally, a checkpoint is a saved copy of the program state, including
945 whatever information is required in order to restore the program to that
946 state at a later time. This can be expected to include the state of
947 registers and memory, and may include external state such as the state
948 of open files and devices.
950 There are a number of ways in which checkpoints may be implemented
951 in gdb, eg. as corefiles, as forked processes, and as some opaque
952 method implemented on the target side.
954 A corefile can be used to save an image of target memory and register
955 state, which can in principle be restored later --- but corefiles do
956 not typically include information about external entities such as
957 open files. Currently this method is not implemented in gdb.
959 A forked process can save the state of user memory and registers,
960 as well as some subset of external (kernel) state. This method
961 is used to implement checkpoints on Linux, and in principle might
962 be used on other systems.
964 Some targets, eg.@: simulators, might have their own built-in
965 method for saving checkpoints, and gdb might be able to take
966 advantage of that capability without necessarily knowing any
967 details of how it is done.
970 @section Observing changes in @value{GDBN} internals
971 @cindex observer pattern interface
972 @cindex notifications about changes in internals
974 In order to function properly, several modules need to be notified when
975 some changes occur in the @value{GDBN} internals. Traditionally, these
976 modules have relied on several paradigms, the most common ones being
977 hooks and gdb-events. Unfortunately, none of these paradigms was
978 versatile enough to become the standard notification mechanism in
979 @value{GDBN}. The fact that they only supported one ``client'' was also
982 A new paradigm, based on the Observer pattern of the @cite{Design
983 Patterns} book, has therefore been implemented. The goal was to provide
984 a new interface overcoming the issues with the notification mechanisms
985 previously available. This new interface needed to be strongly typed,
986 easy to extend, and versatile enough to be used as the standard
987 interface when adding new notifications.
989 See @ref{GDB Observers} for a brief description of the observers
990 currently implemented in GDB. The rationale for the current
991 implementation is also briefly discussed.
995 @chapter User Interface
997 @value{GDBN} has several user interfaces. Although the command-line interface
998 is the most common and most familiar, there are others.
1000 @section Command Interpreter
1002 @cindex command interpreter
1004 The command interpreter in @value{GDBN} is fairly simple. It is designed to
1005 allow for the set of commands to be augmented dynamically, and also
1006 has a recursive subcommand capability, where the first argument to
1007 a command may itself direct a lookup on a different command list.
1009 For instance, the @samp{set} command just starts a lookup on the
1010 @code{setlist} command list, while @samp{set thread} recurses
1011 to the @code{set_thread_cmd_list}.
1015 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
1016 the main command list, and should be used for those commands. The usual
1017 place to add commands is in the @code{_initialize_@var{xyz}} routines at
1018 the ends of most source files.
1020 @findex add_setshow_cmd
1021 @findex add_setshow_cmd_full
1022 To add paired @samp{set} and @samp{show} commands, use
1023 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
1024 a slightly simpler interface which is useful when you don't need to
1025 further modify the new command structures, while the latter returns
1026 the new command structures for manipulation.
1028 @cindex deprecating commands
1029 @findex deprecate_cmd
1030 Before removing commands from the command set it is a good idea to
1031 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1032 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1033 @code{struct cmd_list_element} as it's first argument. You can use the
1034 return value from @code{add_com} or @code{add_cmd} to deprecate the
1035 command immediately after it is created.
1037 The first time a command is used the user will be warned and offered a
1038 replacement (if one exists). Note that the replacement string passed to
1039 @code{deprecate_cmd} should be the full name of the command, i.e. the
1040 entire string the user should type at the command line.
1042 @section UI-Independent Output---the @code{ui_out} Functions
1043 @c This section is based on the documentation written by Fernando
1044 @c Nasser <fnasser@redhat.com>.
1046 @cindex @code{ui_out} functions
1047 The @code{ui_out} functions present an abstraction level for the
1048 @value{GDBN} output code. They hide the specifics of different user
1049 interfaces supported by @value{GDBN}, and thus free the programmer
1050 from the need to write several versions of the same code, one each for
1051 every UI, to produce output.
1053 @subsection Overview and Terminology
1055 In general, execution of each @value{GDBN} command produces some sort
1056 of output, and can even generate an input request.
1058 Output can be generated for the following purposes:
1062 to display a @emph{result} of an operation;
1065 to convey @emph{info} or produce side-effects of a requested
1069 to provide a @emph{notification} of an asynchronous event (including
1070 progress indication of a prolonged asynchronous operation);
1073 to display @emph{error messages} (including warnings);
1076 to show @emph{debug data};
1079 to @emph{query} or prompt a user for input (a special case).
1083 This section mainly concentrates on how to build result output,
1084 although some of it also applies to other kinds of output.
1086 Generation of output that displays the results of an operation
1087 involves one or more of the following:
1091 output of the actual data
1094 formatting the output as appropriate for console output, to make it
1095 easily readable by humans
1098 machine oriented formatting--a more terse formatting to allow for easy
1099 parsing by programs which read @value{GDBN}'s output
1102 annotation, whose purpose is to help legacy GUIs to identify interesting
1106 The @code{ui_out} routines take care of the first three aspects.
1107 Annotations are provided by separate annotation routines. Note that use
1108 of annotations for an interface between a GUI and @value{GDBN} is
1111 Output can be in the form of a single item, which we call a @dfn{field};
1112 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1113 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1114 header and a body. In a BNF-like form:
1117 @item <table> @expansion{}
1118 @code{<header> <body>}
1119 @item <header> @expansion{}
1120 @code{@{ <column> @}}
1121 @item <column> @expansion{}
1122 @code{<width> <alignment> <title>}
1123 @item <body> @expansion{}
1128 @subsection General Conventions
1130 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1131 @code{ui_out_stream_new} (which returns a pointer to the newly created
1132 object) and the @code{make_cleanup} routines.
1134 The first parameter is always the @code{ui_out} vector object, a pointer
1135 to a @code{struct ui_out}.
1137 The @var{format} parameter is like in @code{printf} family of functions.
1138 When it is present, there must also be a variable list of arguments
1139 sufficient used to satisfy the @code{%} specifiers in the supplied
1142 When a character string argument is not used in a @code{ui_out} function
1143 call, a @code{NULL} pointer has to be supplied instead.
1146 @subsection Table, Tuple and List Functions
1148 @cindex list output functions
1149 @cindex table output functions
1150 @cindex tuple output functions
1151 This section introduces @code{ui_out} routines for building lists,
1152 tuples and tables. The routines to output the actual data items
1153 (fields) are presented in the next section.
1155 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1156 containing information about an object; a @dfn{list} is a sequence of
1157 fields where each field describes an identical object.
1159 Use the @dfn{table} functions when your output consists of a list of
1160 rows (tuples) and the console output should include a heading. Use this
1161 even when you are listing just one object but you still want the header.
1163 @cindex nesting level in @code{ui_out} functions
1164 Tables can not be nested. Tuples and lists can be nested up to a
1165 maximum of five levels.
1167 The overall structure of the table output code is something like this:
1182 Here is the description of table-, tuple- and list-related @code{ui_out}
1185 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1186 The function @code{ui_out_table_begin} marks the beginning of the output
1187 of a table. It should always be called before any other @code{ui_out}
1188 function for a given table. @var{nbrofcols} is the number of columns in
1189 the table. @var{nr_rows} is the number of rows in the table.
1190 @var{tblid} is an optional string identifying the table. The string
1191 pointed to by @var{tblid} is copied by the implementation of
1192 @code{ui_out_table_begin}, so the application can free the string if it
1193 was @code{malloc}ed.
1195 The companion function @code{ui_out_table_end}, described below, marks
1196 the end of the table's output.
1199 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1200 @code{ui_out_table_header} provides the header information for a single
1201 table column. You call this function several times, one each for every
1202 column of the table, after @code{ui_out_table_begin}, but before
1203 @code{ui_out_table_body}.
1205 The value of @var{width} gives the column width in characters. The
1206 value of @var{alignment} is one of @code{left}, @code{center}, and
1207 @code{right}, and it specifies how to align the header: left-justify,
1208 center, or right-justify it. @var{colhdr} points to a string that
1209 specifies the column header; the implementation copies that string, so
1210 column header strings in @code{malloc}ed storage can be freed after the
1214 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1215 This function delimits the table header from the table body.
1218 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1219 This function signals the end of a table's output. It should be called
1220 after the table body has been produced by the list and field output
1223 There should be exactly one call to @code{ui_out_table_end} for each
1224 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1225 will signal an internal error.
1228 The output of the tuples that represent the table rows must follow the
1229 call to @code{ui_out_table_body} and precede the call to
1230 @code{ui_out_table_end}. You build a tuple by calling
1231 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1232 calls to functions which actually output fields between them.
1234 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1235 This function marks the beginning of a tuple output. @var{id} points
1236 to an optional string that identifies the tuple; it is copied by the
1237 implementation, and so strings in @code{malloc}ed storage can be freed
1241 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1242 This function signals an end of a tuple output. There should be exactly
1243 one call to @code{ui_out_tuple_end} for each call to
1244 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1248 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1249 This function first opens the tuple and then establishes a cleanup
1250 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1251 and correct implementation of the non-portable@footnote{The function
1252 cast is not portable ISO C.} code sequence:
1254 struct cleanup *old_cleanup;
1255 ui_out_tuple_begin (uiout, "...");
1256 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1261 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1262 This function marks the beginning of a list output. @var{id} points to
1263 an optional string that identifies the list; it is copied by the
1264 implementation, and so strings in @code{malloc}ed storage can be freed
1268 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1269 This function signals an end of a list output. There should be exactly
1270 one call to @code{ui_out_list_end} for each call to
1271 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1275 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1276 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1277 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1278 that will close the list.list.
1281 @subsection Item Output Functions
1283 @cindex item output functions
1284 @cindex field output functions
1286 The functions described below produce output for the actual data
1287 items, or fields, which contain information about the object.
1289 Choose the appropriate function accordingly to your particular needs.
1291 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1292 This is the most general output function. It produces the
1293 representation of the data in the variable-length argument list
1294 according to formatting specifications in @var{format}, a
1295 @code{printf}-like format string. The optional argument @var{fldname}
1296 supplies the name of the field. The data items themselves are
1297 supplied as additional arguments after @var{format}.
1299 This generic function should be used only when it is not possible to
1300 use one of the specialized versions (see below).
1303 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1304 This function outputs a value of an @code{int} variable. It uses the
1305 @code{"%d"} output conversion specification. @var{fldname} specifies
1306 the name of the field.
1309 @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})
1310 This function outputs a value of an @code{int} variable. It differs from
1311 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1312 @var{fldname} specifies
1313 the name of the field.
1316 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1317 This function outputs an address.
1320 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1321 This function outputs a string using the @code{"%s"} conversion
1325 Sometimes, there's a need to compose your output piece by piece using
1326 functions that operate on a stream, such as @code{value_print} or
1327 @code{fprintf_symbol_filtered}. These functions accept an argument of
1328 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1329 used to store the data stream used for the output. When you use one
1330 of these functions, you need a way to pass their results stored in a
1331 @code{ui_file} object to the @code{ui_out} functions. To this end,
1332 you first create a @code{ui_stream} object by calling
1333 @code{ui_out_stream_new}, pass the @code{stream} member of that
1334 @code{ui_stream} object to @code{value_print} and similar functions,
1335 and finally call @code{ui_out_field_stream} to output the field you
1336 constructed. When the @code{ui_stream} object is no longer needed,
1337 you should destroy it and free its memory by calling
1338 @code{ui_out_stream_delete}.
1340 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1341 This function creates a new @code{ui_stream} object which uses the
1342 same output methods as the @code{ui_out} object whose pointer is
1343 passed in @var{uiout}. It returns a pointer to the newly created
1344 @code{ui_stream} object.
1347 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1348 This functions destroys a @code{ui_stream} object specified by
1352 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1353 This function consumes all the data accumulated in
1354 @code{streambuf->stream} and outputs it like
1355 @code{ui_out_field_string} does. After a call to
1356 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1357 the stream is still valid and may be used for producing more fields.
1360 @strong{Important:} If there is any chance that your code could bail
1361 out before completing output generation and reaching the point where
1362 @code{ui_out_stream_delete} is called, it is necessary to set up a
1363 cleanup, to avoid leaking memory and other resources. Here's a
1364 skeleton code to do that:
1367 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1368 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1373 If the function already has the old cleanup chain set (for other kinds
1374 of cleanups), you just have to add your cleanup to it:
1377 mybuf = ui_out_stream_new (uiout);
1378 make_cleanup (ui_out_stream_delete, mybuf);
1381 Note that with cleanups in place, you should not call
1382 @code{ui_out_stream_delete} directly, or you would attempt to free the
1385 @subsection Utility Output Functions
1387 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1388 This function skips a field in a table. Use it if you have to leave
1389 an empty field without disrupting the table alignment. The argument
1390 @var{fldname} specifies a name for the (missing) filed.
1393 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1394 This function outputs the text in @var{string} in a way that makes it
1395 easy to be read by humans. For example, the console implementation of
1396 this method filters the text through a built-in pager, to prevent it
1397 from scrolling off the visible portion of the screen.
1399 Use this function for printing relatively long chunks of text around
1400 the actual field data: the text it produces is not aligned according
1401 to the table's format. Use @code{ui_out_field_string} to output a
1402 string field, and use @code{ui_out_message}, described below, to
1403 output short messages.
1406 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1407 This function outputs @var{nspaces} spaces. It is handy to align the
1408 text produced by @code{ui_out_text} with the rest of the table or
1412 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1413 This function produces a formatted message, provided that the current
1414 verbosity level is at least as large as given by @var{verbosity}. The
1415 current verbosity level is specified by the user with the @samp{set
1416 verbositylevel} command.@footnote{As of this writing (April 2001),
1417 setting verbosity level is not yet implemented, and is always returned
1418 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1419 argument more than zero will cause the message to never be printed.}
1422 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1423 This function gives the console output filter (a paging filter) a hint
1424 of where to break lines which are too long. Ignored for all other
1425 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1426 be printed to indent the wrapped text on the next line; it must remain
1427 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1428 explicit newline is produced by one of the other functions. If
1429 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1432 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1433 This function flushes whatever output has been accumulated so far, if
1434 the UI buffers output.
1438 @subsection Examples of Use of @code{ui_out} functions
1440 @cindex using @code{ui_out} functions
1441 @cindex @code{ui_out} functions, usage examples
1442 This section gives some practical examples of using the @code{ui_out}
1443 functions to generalize the old console-oriented code in
1444 @value{GDBN}. The examples all come from functions defined on the
1445 @file{breakpoints.c} file.
1447 This example, from the @code{breakpoint_1} function, shows how to
1450 The original code was:
1453 if (!found_a_breakpoint++)
1455 annotate_breakpoints_headers ();
1458 printf_filtered ("Num ");
1460 printf_filtered ("Type ");
1462 printf_filtered ("Disp ");
1464 printf_filtered ("Enb ");
1468 printf_filtered ("Address ");
1471 printf_filtered ("What\n");
1473 annotate_breakpoints_table ();
1477 Here's the new version:
1480 nr_printable_breakpoints = @dots{};
1483 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1485 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1487 if (nr_printable_breakpoints > 0)
1488 annotate_breakpoints_headers ();
1489 if (nr_printable_breakpoints > 0)
1491 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1492 if (nr_printable_breakpoints > 0)
1494 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1495 if (nr_printable_breakpoints > 0)
1497 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1498 if (nr_printable_breakpoints > 0)
1500 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1503 if (nr_printable_breakpoints > 0)
1505 if (TARGET_ADDR_BIT <= 32)
1506 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1508 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1510 if (nr_printable_breakpoints > 0)
1512 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1513 ui_out_table_body (uiout);
1514 if (nr_printable_breakpoints > 0)
1515 annotate_breakpoints_table ();
1518 This example, from the @code{print_one_breakpoint} function, shows how
1519 to produce the actual data for the table whose structure was defined
1520 in the above example. The original code was:
1525 printf_filtered ("%-3d ", b->number);
1527 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1528 || ((int) b->type != bptypes[(int) b->type].type))
1529 internal_error ("bptypes table does not describe type #%d.",
1531 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1533 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1535 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1539 This is the new version:
1543 ui_out_tuple_begin (uiout, "bkpt");
1545 ui_out_field_int (uiout, "number", b->number);
1547 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1548 || ((int) b->type != bptypes[(int) b->type].type))
1549 internal_error ("bptypes table does not describe type #%d.",
1551 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1553 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1555 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1559 This example, also from @code{print_one_breakpoint}, shows how to
1560 produce a complicated output field using the @code{print_expression}
1561 functions which requires a stream to be passed. It also shows how to
1562 automate stream destruction with cleanups. The original code was:
1566 print_expression (b->exp, gdb_stdout);
1572 struct ui_stream *stb = ui_out_stream_new (uiout);
1573 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1576 print_expression (b->exp, stb->stream);
1577 ui_out_field_stream (uiout, "what", local_stream);
1580 This example, also from @code{print_one_breakpoint}, shows how to use
1581 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1586 if (b->dll_pathname == NULL)
1587 printf_filtered ("<any library> ");
1589 printf_filtered ("library \"%s\" ", b->dll_pathname);
1596 if (b->dll_pathname == NULL)
1598 ui_out_field_string (uiout, "what", "<any library>");
1599 ui_out_spaces (uiout, 1);
1603 ui_out_text (uiout, "library \"");
1604 ui_out_field_string (uiout, "what", b->dll_pathname);
1605 ui_out_text (uiout, "\" ");
1609 The following example from @code{print_one_breakpoint} shows how to
1610 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1615 if (b->forked_inferior_pid != 0)
1616 printf_filtered ("process %d ", b->forked_inferior_pid);
1623 if (b->forked_inferior_pid != 0)
1625 ui_out_text (uiout, "process ");
1626 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1627 ui_out_spaces (uiout, 1);
1631 Here's an example of using @code{ui_out_field_string}. The original
1636 if (b->exec_pathname != NULL)
1637 printf_filtered ("program \"%s\" ", b->exec_pathname);
1644 if (b->exec_pathname != NULL)
1646 ui_out_text (uiout, "program \"");
1647 ui_out_field_string (uiout, "what", b->exec_pathname);
1648 ui_out_text (uiout, "\" ");
1652 Finally, here's an example of printing an address. The original code:
1656 printf_filtered ("%s ",
1657 hex_string_custom ((unsigned long) b->address, 8));
1664 ui_out_field_core_addr (uiout, "Address", b->address);
1668 @section Console Printing
1677 @cindex @code{libgdb}
1678 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1679 to provide an API to @value{GDBN}'s functionality.
1682 @cindex @code{libgdb}
1683 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1684 better able to support graphical and other environments.
1686 Since @code{libgdb} development is on-going, its architecture is still
1687 evolving. The following components have so far been identified:
1691 Observer - @file{gdb-events.h}.
1693 Builder - @file{ui-out.h}
1695 Event Loop - @file{event-loop.h}
1697 Library - @file{gdb.h}
1700 The model that ties these components together is described below.
1702 @section The @code{libgdb} Model
1704 A client of @code{libgdb} interacts with the library in two ways.
1708 As an observer (using @file{gdb-events}) receiving notifications from
1709 @code{libgdb} of any internal state changes (break point changes, run
1712 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1713 obtain various status values from @value{GDBN}.
1716 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1717 the existing @value{GDBN} CLI), those clients must co-operate when
1718 controlling @code{libgdb}. In particular, a client must ensure that
1719 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1720 before responding to a @file{gdb-event} by making a query.
1722 @section CLI support
1724 At present @value{GDBN}'s CLI is very much entangled in with the core of
1725 @code{libgdb}. Consequently, a client wishing to include the CLI in
1726 their interface needs to carefully co-ordinate its own and the CLI's
1729 It is suggested that the client set @code{libgdb} up to be bi-modal
1730 (alternate between CLI and client query modes). The notes below sketch
1735 The client registers itself as an observer of @code{libgdb}.
1737 The client create and install @code{cli-out} builder using its own
1738 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1739 @code{gdb_stdout} streams.
1741 The client creates a separate custom @code{ui-out} builder that is only
1742 used while making direct queries to @code{libgdb}.
1745 When the client receives input intended for the CLI, it simply passes it
1746 along. Since the @code{cli-out} builder is installed by default, all
1747 the CLI output in response to that command is routed (pronounced rooted)
1748 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1749 At the same time, the client is kept abreast of internal changes by
1750 virtue of being a @code{libgdb} observer.
1752 The only restriction on the client is that it must wait until
1753 @code{libgdb} becomes idle before initiating any queries (using the
1754 client's custom builder).
1756 @section @code{libgdb} components
1758 @subheading Observer - @file{gdb-events.h}
1759 @file{gdb-events} provides the client with a very raw mechanism that can
1760 be used to implement an observer. At present it only allows for one
1761 observer and that observer must, internally, handle the need to delay
1762 the processing of any event notifications until after @code{libgdb} has
1763 finished the current command.
1765 @subheading Builder - @file{ui-out.h}
1766 @file{ui-out} provides the infrastructure necessary for a client to
1767 create a builder. That builder is then passed down to @code{libgdb}
1768 when doing any queries.
1770 @subheading Event Loop - @file{event-loop.h}
1771 @c There could be an entire section on the event-loop
1772 @file{event-loop}, currently non-re-entrant, provides a simple event
1773 loop. A client would need to either plug its self into this loop or,
1774 implement a new event-loop that GDB would use.
1776 The event-loop will eventually be made re-entrant. This is so that
1777 @value{GDBN} can better handle the problem of some commands blocking
1778 instead of returning.
1780 @subheading Library - @file{gdb.h}
1781 @file{libgdb} is the most obvious component of this system. It provides
1782 the query interface. Each function is parameterized by a @code{ui-out}
1783 builder. The result of the query is constructed using that builder
1784 before the query function returns.
1786 @node Symbol Handling
1788 @chapter Symbol Handling
1790 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1791 functions, and types.
1793 @section Symbol Reading
1795 @cindex symbol reading
1796 @cindex reading of symbols
1797 @cindex symbol files
1798 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1799 file is the file containing the program which @value{GDBN} is
1800 debugging. @value{GDBN} can be directed to use a different file for
1801 symbols (with the @samp{symbol-file} command), and it can also read
1802 more symbols via the @samp{add-file} and @samp{load} commands, or while
1803 reading symbols from shared libraries.
1805 @findex find_sym_fns
1806 Symbol files are initially opened by code in @file{symfile.c} using
1807 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1808 of the file by examining its header. @code{find_sym_fns} then uses
1809 this identification to locate a set of symbol-reading functions.
1811 @findex add_symtab_fns
1812 @cindex @code{sym_fns} structure
1813 @cindex adding a symbol-reading module
1814 Symbol-reading modules identify themselves to @value{GDBN} by calling
1815 @code{add_symtab_fns} during their module initialization. The argument
1816 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1817 name (or name prefix) of the symbol format, the length of the prefix,
1818 and pointers to four functions. These functions are called at various
1819 times to process symbol files whose identification matches the specified
1822 The functions supplied by each module are:
1825 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1827 @cindex secondary symbol file
1828 Called from @code{symbol_file_add} when we are about to read a new
1829 symbol file. This function should clean up any internal state (possibly
1830 resulting from half-read previous files, for example) and prepare to
1831 read a new symbol file. Note that the symbol file which we are reading
1832 might be a new ``main'' symbol file, or might be a secondary symbol file
1833 whose symbols are being added to the existing symbol table.
1835 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1836 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1837 new symbol file being read. Its @code{private} field has been zeroed,
1838 and can be modified as desired. Typically, a struct of private
1839 information will be @code{malloc}'d, and a pointer to it will be placed
1840 in the @code{private} field.
1842 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1843 @code{error} if it detects an unavoidable problem.
1845 @item @var{xyz}_new_init()
1847 Called from @code{symbol_file_add} when discarding existing symbols.
1848 This function needs only handle the symbol-reading module's internal
1849 state; the symbol table data structures visible to the rest of
1850 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1851 arguments and no result. It may be called after
1852 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1853 may be called alone if all symbols are simply being discarded.
1855 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1857 Called from @code{symbol_file_add} to actually read the symbols from a
1858 symbol-file into a set of psymtabs or symtabs.
1860 @code{sf} points to the @code{struct sym_fns} originally passed to
1861 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1862 the offset between the file's specified start address and its true
1863 address in memory. @code{mainline} is 1 if this is the main symbol
1864 table being read, and 0 if a secondary symbol file (e.g., shared library
1865 or dynamically loaded file) is being read.@refill
1868 In addition, if a symbol-reading module creates psymtabs when
1869 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1870 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1871 from any point in the @value{GDBN} symbol-handling code.
1874 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1876 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1877 the psymtab has not already been read in and had its @code{pst->symtab}
1878 pointer set. The argument is the psymtab to be fleshed-out into a
1879 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1880 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1881 zero if there were no symbols in that part of the symbol file.
1884 @section Partial Symbol Tables
1886 @value{GDBN} has three types of symbol tables:
1889 @cindex full symbol table
1892 Full symbol tables (@dfn{symtabs}). These contain the main
1893 information about symbols and addresses.
1897 Partial symbol tables (@dfn{psymtabs}). These contain enough
1898 information to know when to read the corresponding part of the full
1901 @cindex minimal symbol table
1904 Minimal symbol tables (@dfn{msymtabs}). These contain information
1905 gleaned from non-debugging symbols.
1908 @cindex partial symbol table
1909 This section describes partial symbol tables.
1911 A psymtab is constructed by doing a very quick pass over an executable
1912 file's debugging information. Small amounts of information are
1913 extracted---enough to identify which parts of the symbol table will
1914 need to be re-read and fully digested later, when the user needs the
1915 information. The speed of this pass causes @value{GDBN} to start up very
1916 quickly. Later, as the detailed rereading occurs, it occurs in small
1917 pieces, at various times, and the delay therefrom is mostly invisible to
1919 @c (@xref{Symbol Reading}.)
1921 The symbols that show up in a file's psymtab should be, roughly, those
1922 visible to the debugger's user when the program is not running code from
1923 that file. These include external symbols and types, static symbols and
1924 types, and @code{enum} values declared at file scope.
1926 The psymtab also contains the range of instruction addresses that the
1927 full symbol table would represent.
1929 @cindex finding a symbol
1930 @cindex symbol lookup
1931 The idea is that there are only two ways for the user (or much of the
1932 code in the debugger) to reference a symbol:
1935 @findex find_pc_function
1936 @findex find_pc_line
1938 By its address (e.g., execution stops at some address which is inside a
1939 function in this file). The address will be noticed to be in the
1940 range of this psymtab, and the full symtab will be read in.
1941 @code{find_pc_function}, @code{find_pc_line}, and other
1942 @code{find_pc_@dots{}} functions handle this.
1944 @cindex lookup_symbol
1947 (e.g., the user asks to print a variable, or set a breakpoint on a
1948 function). Global names and file-scope names will be found in the
1949 psymtab, which will cause the symtab to be pulled in. Local names will
1950 have to be qualified by a global name, or a file-scope name, in which
1951 case we will have already read in the symtab as we evaluated the
1952 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1953 local scope, in which case the first case applies. @code{lookup_symbol}
1954 does most of the work here.
1957 The only reason that psymtabs exist is to cause a symtab to be read in
1958 at the right moment. Any symbol that can be elided from a psymtab,
1959 while still causing that to happen, should not appear in it. Since
1960 psymtabs don't have the idea of scope, you can't put local symbols in
1961 them anyway. Psymtabs don't have the idea of the type of a symbol,
1962 either, so types need not appear, unless they will be referenced by
1965 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1966 been read, and another way if the corresponding symtab has been read
1967 in. Such bugs are typically caused by a psymtab that does not contain
1968 all the visible symbols, or which has the wrong instruction address
1971 The psymtab for a particular section of a symbol file (objfile) could be
1972 thrown away after the symtab has been read in. The symtab should always
1973 be searched before the psymtab, so the psymtab will never be used (in a
1974 bug-free environment). Currently, psymtabs are allocated on an obstack,
1975 and all the psymbols themselves are allocated in a pair of large arrays
1976 on an obstack, so there is little to be gained by trying to free them
1977 unless you want to do a lot more work.
1981 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
1983 @cindex fundamental types
1984 These are the fundamental types that @value{GDBN} uses internally. Fundamental
1985 types from the various debugging formats (stabs, ELF, etc) are mapped
1986 into one of these. They are basically a union of all fundamental types
1987 that @value{GDBN} knows about for all the languages that @value{GDBN}
1990 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
1993 Each time @value{GDBN} builds an internal type, it marks it with one
1994 of these types. The type may be a fundamental type, such as
1995 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
1996 which is a pointer to another type. Typically, several @code{FT_*}
1997 types map to one @code{TYPE_CODE_*} type, and are distinguished by
1998 other members of the type struct, such as whether the type is signed
1999 or unsigned, and how many bits it uses.
2001 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
2003 These are instances of type structs that roughly correspond to
2004 fundamental types and are created as global types for @value{GDBN} to
2005 use for various ugly historical reasons. We eventually want to
2006 eliminate these. Note for example that @code{builtin_type_int}
2007 initialized in @file{gdbtypes.c} is basically the same as a
2008 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
2009 an @code{FT_INTEGER} fundamental type. The difference is that the
2010 @code{builtin_type} is not associated with any particular objfile, and
2011 only one instance exists, while @file{c-lang.c} builds as many
2012 @code{TYPE_CODE_INT} types as needed, with each one associated with
2013 some particular objfile.
2015 @section Object File Formats
2016 @cindex object file formats
2020 @cindex @code{a.out} format
2021 The @code{a.out} format is the original file format for Unix. It
2022 consists of three sections: @code{text}, @code{data}, and @code{bss},
2023 which are for program code, initialized data, and uninitialized data,
2026 The @code{a.out} format is so simple that it doesn't have any reserved
2027 place for debugging information. (Hey, the original Unix hackers used
2028 @samp{adb}, which is a machine-language debugger!) The only debugging
2029 format for @code{a.out} is stabs, which is encoded as a set of normal
2030 symbols with distinctive attributes.
2032 The basic @code{a.out} reader is in @file{dbxread.c}.
2037 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2038 COFF files may have multiple sections, each prefixed by a header. The
2039 number of sections is limited.
2041 The COFF specification includes support for debugging. Although this
2042 was a step forward, the debugging information was woefully limited. For
2043 instance, it was not possible to represent code that came from an
2046 The COFF reader is in @file{coffread.c}.
2050 @cindex ECOFF format
2051 ECOFF is an extended COFF originally introduced for Mips and Alpha
2054 The basic ECOFF reader is in @file{mipsread.c}.
2058 @cindex XCOFF format
2059 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2060 The COFF sections, symbols, and line numbers are used, but debugging
2061 symbols are @code{dbx}-style stabs whose strings are located in the
2062 @code{.debug} section (rather than the string table). For more
2063 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2065 The shared library scheme has a clean interface for figuring out what
2066 shared libraries are in use, but the catch is that everything which
2067 refers to addresses (symbol tables and breakpoints at least) needs to be
2068 relocated for both shared libraries and the main executable. At least
2069 using the standard mechanism this can only be done once the program has
2070 been run (or the core file has been read).
2074 @cindex PE-COFF format
2075 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2076 executables. PE is basically COFF with additional headers.
2078 While BFD includes special PE support, @value{GDBN} needs only the basic
2084 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
2085 to COFF in being organized into a number of sections, but it removes
2086 many of COFF's limitations.
2088 The basic ELF reader is in @file{elfread.c}.
2093 SOM is HP's object file and debug format (not to be confused with IBM's
2094 SOM, which is a cross-language ABI).
2096 The SOM reader is in @file{hpread.c}.
2098 @subsection Other File Formats
2100 @cindex Netware Loadable Module format
2101 Other file formats that have been supported by @value{GDBN} include Netware
2102 Loadable Modules (@file{nlmread.c}).
2104 @section Debugging File Formats
2106 This section describes characteristics of debugging information that
2107 are independent of the object file format.
2111 @cindex stabs debugging info
2112 @code{stabs} started out as special symbols within the @code{a.out}
2113 format. Since then, it has been encapsulated into other file
2114 formats, such as COFF and ELF.
2116 While @file{dbxread.c} does some of the basic stab processing,
2117 including for encapsulated versions, @file{stabsread.c} does
2122 @cindex COFF debugging info
2123 The basic COFF definition includes debugging information. The level
2124 of support is minimal and non-extensible, and is not often used.
2126 @subsection Mips debug (Third Eye)
2128 @cindex ECOFF debugging info
2129 ECOFF includes a definition of a special debug format.
2131 The file @file{mdebugread.c} implements reading for this format.
2135 @cindex DWARF 1 debugging info
2136 DWARF 1 is a debugging format that was originally designed to be
2137 used with ELF in SVR4 systems.
2142 @c If defined, these are the producer strings in a DWARF 1 file. All of
2143 @c these have reasonable defaults already.
2145 The DWARF 1 reader is in @file{dwarfread.c}.
2149 @cindex DWARF 2 debugging info
2150 DWARF 2 is an improved but incompatible version of DWARF 1.
2152 The DWARF 2 reader is in @file{dwarf2read.c}.
2156 @cindex SOM debugging info
2157 Like COFF, the SOM definition includes debugging information.
2159 @section Adding a New Symbol Reader to @value{GDBN}
2161 @cindex adding debugging info reader
2162 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2163 there is probably little to be done.
2165 If you need to add a new object file format, you must first add it to
2166 BFD. This is beyond the scope of this document.
2168 You must then arrange for the BFD code to provide access to the
2169 debugging symbols. Generally @value{GDBN} will have to call swapping routines
2170 from BFD and a few other BFD internal routines to locate the debugging
2171 information. As much as possible, @value{GDBN} should not depend on the BFD
2172 internal data structures.
2174 For some targets (e.g., COFF), there is a special transfer vector used
2175 to call swapping routines, since the external data structures on various
2176 platforms have different sizes and layouts. Specialized routines that
2177 will only ever be implemented by one object file format may be called
2178 directly. This interface should be described in a file
2179 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2181 @section Memory Management for Symbol Files
2183 Most memory associated with a loaded symbol file is stored on
2184 its @code{objfile_obstack}. This includes symbols, types,
2185 namespace data, and other information produced by the symbol readers.
2187 Because this data lives on the objfile's obstack, it is automatically
2188 released when the objfile is unloaded or reloaded. Therefore one
2189 objfile must not reference symbol or type data from another objfile;
2190 they could be unloaded at different times.
2192 User convenience variables, et cetera, have associated types. Normally
2193 these types live in the associated objfile. However, when the objfile
2194 is unloaded, those types are deep copied to global memory, so that
2195 the values of the user variables and history items are not lost.
2198 @node Language Support
2200 @chapter Language Support
2202 @cindex language support
2203 @value{GDBN}'s language support is mainly driven by the symbol reader,
2204 although it is possible for the user to set the source language
2207 @value{GDBN} chooses the source language by looking at the extension
2208 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2209 means Fortran, etc. It may also use a special-purpose language
2210 identifier if the debug format supports it, like with DWARF.
2212 @section Adding a Source Language to @value{GDBN}
2214 @cindex adding source language
2215 To add other languages to @value{GDBN}'s expression parser, follow the
2219 @item Create the expression parser.
2221 @cindex expression parser
2222 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2223 building parsed expressions into a @code{union exp_element} list are in
2226 @cindex language parser
2227 Since we can't depend upon everyone having Bison, and YACC produces
2228 parsers that define a bunch of global names, the following lines
2229 @strong{must} be included at the top of the YACC parser, to prevent the
2230 various parsers from defining the same global names:
2233 #define yyparse @var{lang}_parse
2234 #define yylex @var{lang}_lex
2235 #define yyerror @var{lang}_error
2236 #define yylval @var{lang}_lval
2237 #define yychar @var{lang}_char
2238 #define yydebug @var{lang}_debug
2239 #define yypact @var{lang}_pact
2240 #define yyr1 @var{lang}_r1
2241 #define yyr2 @var{lang}_r2
2242 #define yydef @var{lang}_def
2243 #define yychk @var{lang}_chk
2244 #define yypgo @var{lang}_pgo
2245 #define yyact @var{lang}_act
2246 #define yyexca @var{lang}_exca
2247 #define yyerrflag @var{lang}_errflag
2248 #define yynerrs @var{lang}_nerrs
2251 At the bottom of your parser, define a @code{struct language_defn} and
2252 initialize it with the right values for your language. Define an
2253 @code{initialize_@var{lang}} routine and have it call
2254 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2255 that your language exists. You'll need some other supporting variables
2256 and functions, which will be used via pointers from your
2257 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2258 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2259 for more information.
2261 @item Add any evaluation routines, if necessary
2263 @cindex expression evaluation routines
2264 @findex evaluate_subexp
2265 @findex prefixify_subexp
2266 @findex length_of_subexp
2267 If you need new opcodes (that represent the operations of the language),
2268 add them to the enumerated type in @file{expression.h}. Add support
2269 code for these operations in the @code{evaluate_subexp} function
2270 defined in the file @file{eval.c}. Add cases
2271 for new opcodes in two functions from @file{parse.c}:
2272 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2273 the number of @code{exp_element}s that a given operation takes up.
2275 @item Update some existing code
2277 Add an enumerated identifier for your language to the enumerated type
2278 @code{enum language} in @file{defs.h}.
2280 Update the routines in @file{language.c} so your language is included.
2281 These routines include type predicates and such, which (in some cases)
2282 are language dependent. If your language does not appear in the switch
2283 statement, an error is reported.
2285 @vindex current_language
2286 Also included in @file{language.c} is the code that updates the variable
2287 @code{current_language}, and the routines that translate the
2288 @code{language_@var{lang}} enumerated identifier into a printable
2291 @findex _initialize_language
2292 Update the function @code{_initialize_language} to include your
2293 language. This function picks the default language upon startup, so is
2294 dependent upon which languages that @value{GDBN} is built for.
2296 @findex allocate_symtab
2297 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2298 code so that the language of each symtab (source file) is set properly.
2299 This is used to determine the language to use at each stack frame level.
2300 Currently, the language is set based upon the extension of the source
2301 file. If the language can be better inferred from the symbol
2302 information, please set the language of the symtab in the symbol-reading
2305 @findex print_subexp
2306 @findex op_print_tab
2307 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2308 expression opcodes you have added to @file{expression.h}. Also, add the
2309 printed representations of your operators to @code{op_print_tab}.
2311 @item Add a place of call
2314 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2315 @code{parse_exp_1} (defined in @file{parse.c}).
2317 @item Use macros to trim code
2319 @cindex trimming language-dependent code
2320 The user has the option of building @value{GDBN} for some or all of the
2321 languages. If the user decides to build @value{GDBN} for the language
2322 @var{lang}, then every file dependent on @file{language.h} will have the
2323 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2324 leave out large routines that the user won't need if he or she is not
2325 using your language.
2327 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2328 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2329 compiled form of your parser) is not linked into @value{GDBN} at all.
2331 See the file @file{configure.in} for how @value{GDBN} is configured
2332 for different languages.
2334 @item Edit @file{Makefile.in}
2336 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2337 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2338 not get linked in, or, worse yet, it may not get @code{tar}red into the
2343 @node Host Definition
2345 @chapter Host Definition
2347 With the advent of Autoconf, it's rarely necessary to have host
2348 definition machinery anymore. The following information is provided,
2349 mainly, as an historical reference.
2351 @section Adding a New Host
2353 @cindex adding a new host
2354 @cindex host, adding
2355 @value{GDBN}'s host configuration support normally happens via Autoconf.
2356 New host-specific definitions should not be needed. Older hosts
2357 @value{GDBN} still use the host-specific definitions and files listed
2358 below, but these mostly exist for historical reasons, and will
2359 eventually disappear.
2362 @item gdb/config/@var{arch}/@var{xyz}.mh
2363 This file once contained both host and native configuration information
2364 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2365 configuration information is now handed by Autoconf.
2367 Host configuration information included a definition of
2368 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2369 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2370 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2372 New host only configurations do not need this file.
2374 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2375 This file once contained definitions and includes required when hosting
2376 gdb on machine @var{xyz}. Those definitions and includes are now
2377 handled by Autoconf.
2379 New host and native configurations do not need this file.
2381 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2382 file to define the macros @var{HOST_FLOAT_FORMAT},
2383 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2384 also needs to be replaced with either an Autoconf or run-time test.}
2388 @subheading Generic Host Support Files
2390 @cindex generic host support
2391 There are some ``generic'' versions of routines that can be used by
2392 various systems. These can be customized in various ways by macros
2393 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2394 the @var{xyz} host, you can just include the generic file's name (with
2395 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2397 Otherwise, if your machine needs custom support routines, you will need
2398 to write routines that perform the same functions as the generic file.
2399 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2400 into @code{XDEPFILES}.
2403 @cindex remote debugging support
2404 @cindex serial line support
2406 This contains serial line support for Unix systems. This is always
2407 included, via the makefile variable @code{SER_HARDWIRE}; override this
2408 variable in the @file{.mh} file to avoid it.
2411 This contains serial line support for 32-bit programs running under DOS,
2412 using the DJGPP (a.k.a.@: GO32) execution environment.
2414 @cindex TCP remote support
2416 This contains generic TCP support using sockets.
2419 @section Host Conditionals
2421 When @value{GDBN} is configured and compiled, various macros are
2422 defined or left undefined, to control compilation based on the
2423 attributes of the host system. These macros and their meanings (or if
2424 the meaning is not documented here, then one of the source files where
2425 they are used is indicated) are:
2428 @item @value{GDBN}INIT_FILENAME
2429 The default name of @value{GDBN}'s initialization file (normally
2433 This macro is deprecated.
2435 @item SIGWINCH_HANDLER
2436 If your host defines @code{SIGWINCH}, you can define this to be the name
2437 of a function to be called if @code{SIGWINCH} is received.
2439 @item SIGWINCH_HANDLER_BODY
2440 Define this to expand into code that will define the function named by
2441 the expansion of @code{SIGWINCH_HANDLER}.
2443 @item ALIGN_STACK_ON_STARTUP
2444 @cindex stack alignment
2445 Define this if your system is of a sort that will crash in
2446 @code{tgetent} if the stack happens not to be longword-aligned when
2447 @code{main} is called. This is a rare situation, but is known to occur
2448 on several different types of systems.
2450 @item CRLF_SOURCE_FILES
2451 @cindex DOS text files
2452 Define this if host files use @code{\r\n} rather than @code{\n} as a
2453 line terminator. This will cause source file listings to omit @code{\r}
2454 characters when printing and it will allow @code{\r\n} line endings of files
2455 which are ``sourced'' by gdb. It must be possible to open files in binary
2456 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2458 @item DEFAULT_PROMPT
2460 The default value of the prompt string (normally @code{"(gdb) "}).
2463 @cindex terminal device
2464 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2467 Define this if binary files are opened the same way as text files.
2471 In some cases, use the system call @code{mmap} for reading symbol
2472 tables. For some machines this allows for sharing and quick updates.
2475 Define this if the host system has @code{termio.h}.
2482 Values for host-side constants.
2485 Substitute for isatty, if not available.
2488 This is the longest integer type available on the host. If not defined,
2489 it will default to @code{long long} or @code{long}, depending on
2490 @code{CC_HAS_LONG_LONG}.
2492 @item CC_HAS_LONG_LONG
2493 @cindex @code{long long} data type
2494 Define this if the host C compiler supports @code{long long}. This is set
2495 by the @code{configure} script.
2497 @item PRINTF_HAS_LONG_LONG
2498 Define this if the host can handle printing of long long integers via
2499 the printf format conversion specifier @code{ll}. This is set by the
2500 @code{configure} script.
2502 @item HAVE_LONG_DOUBLE
2503 Define this if the host C compiler supports @code{long double}. This is
2504 set by the @code{configure} script.
2506 @item PRINTF_HAS_LONG_DOUBLE
2507 Define this if the host can handle printing of long double float-point
2508 numbers via the printf format conversion specifier @code{Lg}. This is
2509 set by the @code{configure} script.
2511 @item SCANF_HAS_LONG_DOUBLE
2512 Define this if the host can handle the parsing of long double
2513 float-point numbers via the scanf format conversion specifier
2514 @code{Lg}. This is set by the @code{configure} script.
2516 @item LSEEK_NOT_LINEAR
2517 Define this if @code{lseek (n)} does not necessarily move to byte number
2518 @code{n} in the file. This is only used when reading source files. It
2519 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2522 This macro is used as the argument to @code{lseek} (or, most commonly,
2523 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2524 which is the POSIX equivalent.
2527 If defined, this should be one or more tokens, such as @code{volatile},
2528 that can be used in both the declaration and definition of functions to
2529 indicate that they never return. The default is already set correctly
2530 if compiling with GCC. This will almost never need to be defined.
2533 If defined, this should be one or more tokens, such as
2534 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2535 of functions to indicate that they never return. The default is already
2536 set correctly if compiling with GCC. This will almost never need to be
2541 Define these to appropriate value for the system @code{lseek}, if not already
2545 This is the signal for stopping @value{GDBN}. Defaults to
2546 @code{SIGTSTP}. (Only redefined for the Convex.)
2549 Means that System V (prior to SVR4) include files are in use. (FIXME:
2550 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2551 @file{utils.c} for other things, at the moment.)
2554 Define this to help placate @code{lint} in some situations.
2557 Define this to override the defaults of @code{__volatile__} or
2562 @node Target Architecture Definition
2564 @chapter Target Architecture Definition
2566 @cindex target architecture definition
2567 @value{GDBN}'s target architecture defines what sort of
2568 machine-language programs @value{GDBN} can work with, and how it works
2571 The target architecture object is implemented as the C structure
2572 @code{struct gdbarch *}. The structure, and its methods, are generated
2573 using the Bourne shell script @file{gdbarch.sh}.
2575 @section Operating System ABI Variant Handling
2576 @cindex OS ABI variants
2578 @value{GDBN} provides a mechanism for handling variations in OS
2579 ABIs. An OS ABI variant may have influence over any number of
2580 variables in the target architecture definition. There are two major
2581 components in the OS ABI mechanism: sniffers and handlers.
2583 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2584 (the architecture may be wildcarded) in an attempt to determine the
2585 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2586 to be @dfn{generic}, while sniffers for a specific architecture are
2587 considered to be @dfn{specific}. A match from a specific sniffer
2588 overrides a match from a generic sniffer. Multiple sniffers for an
2589 architecture/flavour may exist, in order to differentiate between two
2590 different operating systems which use the same basic file format. The
2591 OS ABI framework provides a generic sniffer for ELF-format files which
2592 examines the @code{EI_OSABI} field of the ELF header, as well as note
2593 sections known to be used by several operating systems.
2595 @cindex fine-tuning @code{gdbarch} structure
2596 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2597 selected OS ABI. There may be only one handler for a given OS ABI
2598 for each BFD architecture.
2600 The following OS ABI variants are defined in @file{osabi.h}:
2604 @findex GDB_OSABI_UNKNOWN
2605 @item GDB_OSABI_UNKNOWN
2606 The ABI of the inferior is unknown. The default @code{gdbarch}
2607 settings for the architecture will be used.
2609 @findex GDB_OSABI_SVR4
2610 @item GDB_OSABI_SVR4
2611 UNIX System V Release 4
2613 @findex GDB_OSABI_HURD
2614 @item GDB_OSABI_HURD
2615 GNU using the Hurd kernel
2617 @findex GDB_OSABI_SOLARIS
2618 @item GDB_OSABI_SOLARIS
2621 @findex GDB_OSABI_OSF1
2622 @item GDB_OSABI_OSF1
2623 OSF/1, including Digital UNIX and Compaq Tru64 UNIX
2625 @findex GDB_OSABI_LINUX
2626 @item GDB_OSABI_LINUX
2627 GNU using the Linux kernel
2629 @findex GDB_OSABI_FREEBSD_AOUT
2630 @item GDB_OSABI_FREEBSD_AOUT
2631 FreeBSD using the a.out executable format
2633 @findex GDB_OSABI_FREEBSD_ELF
2634 @item GDB_OSABI_FREEBSD_ELF
2635 FreeBSD using the ELF executable format
2637 @findex GDB_OSABI_NETBSD_AOUT
2638 @item GDB_OSABI_NETBSD_AOUT
2639 NetBSD using the a.out executable format
2641 @findex GDB_OSABI_NETBSD_ELF
2642 @item GDB_OSABI_NETBSD_ELF
2643 NetBSD using the ELF executable format
2645 @findex GDB_OSABI_WINCE
2646 @item GDB_OSABI_WINCE
2649 @findex GDB_OSABI_GO32
2650 @item GDB_OSABI_GO32
2653 @findex GDB_OSABI_NETWARE
2654 @item GDB_OSABI_NETWARE
2657 @findex GDB_OSABI_ARM_EABI_V1
2658 @item GDB_OSABI_ARM_EABI_V1
2659 ARM Embedded ABI version 1
2661 @findex GDB_OSABI_ARM_EABI_V2
2662 @item GDB_OSABI_ARM_EABI_V2
2663 ARM Embedded ABI version 2
2665 @findex GDB_OSABI_ARM_APCS
2666 @item GDB_OSABI_ARM_APCS
2667 Generic ARM Procedure Call Standard
2671 Here are the functions that make up the OS ABI framework:
2673 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2674 Return the name of the OS ABI corresponding to @var{osabi}.
2677 @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}))
2678 Register the OS ABI handler specified by @var{init_osabi} for the
2679 architecture, machine type and OS ABI specified by @var{arch},
2680 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2681 machine type, which implies the architecture's default machine type,
2685 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2686 Register the OS ABI file sniffer specified by @var{sniffer} for the
2687 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2688 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2689 be generic, and is allowed to examine @var{flavour}-flavoured files for
2693 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2694 Examine the file described by @var{abfd} to determine its OS ABI.
2695 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2699 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2700 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2701 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2702 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2703 architecture, a warning will be issued and the debugging session will continue
2704 with the defaults already established for @var{gdbarch}.
2707 @section Registers and Memory
2709 @value{GDBN}'s model of the target machine is rather simple.
2710 @value{GDBN} assumes the machine includes a bank of registers and a
2711 block of memory. Each register may have a different size.
2713 @value{GDBN} does not have a magical way to match up with the
2714 compiler's idea of which registers are which; however, it is critical
2715 that they do match up accurately. The only way to make this work is
2716 to get accurate information about the order that the compiler uses,
2717 and to reflect that in the @code{REGISTER_NAME} and related macros.
2719 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2721 @section Pointers Are Not Always Addresses
2722 @cindex pointer representation
2723 @cindex address representation
2724 @cindex word-addressed machines
2725 @cindex separate data and code address spaces
2726 @cindex spaces, separate data and code address
2727 @cindex address spaces, separate data and code
2728 @cindex code pointers, word-addressed
2729 @cindex converting between pointers and addresses
2730 @cindex D10V addresses
2732 On almost all 32-bit architectures, the representation of a pointer is
2733 indistinguishable from the representation of some fixed-length number
2734 whose value is the byte address of the object pointed to. On such
2735 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2736 However, architectures with smaller word sizes are often cramped for
2737 address space, so they may choose a pointer representation that breaks this
2738 identity, and allows a larger code address space.
2740 For example, the Renesas D10V is a 16-bit VLIW processor whose
2741 instructions are 32 bits long@footnote{Some D10V instructions are
2742 actually pairs of 16-bit sub-instructions. However, since you can't
2743 jump into the middle of such a pair, code addresses can only refer to
2744 full 32 bit instructions, which is what matters in this explanation.}.
2745 If the D10V used ordinary byte addresses to refer to code locations,
2746 then the processor would only be able to address 64kb of instructions.
2747 However, since instructions must be aligned on four-byte boundaries, the
2748 low two bits of any valid instruction's byte address are always
2749 zero---byte addresses waste two bits. So instead of byte addresses,
2750 the D10V uses word addresses---byte addresses shifted right two bits---to
2751 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2754 However, this means that code pointers and data pointers have different
2755 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2756 @code{0xC020} when used as a data address, but refers to byte address
2757 @code{0x30080} when used as a code address.
2759 (The D10V also uses separate code and data address spaces, which also
2760 affects the correspondence between pointers and addresses, but we're
2761 going to ignore that here; this example is already too long.)
2763 To cope with architectures like this---the D10V is not the only
2764 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2765 byte numbers, and @dfn{pointers}, which are the target's representation
2766 of an address of a particular type of data. In the example above,
2767 @code{0xC020} is the pointer, which refers to one of the addresses
2768 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2769 @value{GDBN} provides functions for turning a pointer into an address
2770 and vice versa, in the appropriate way for the current architecture.
2772 Unfortunately, since addresses and pointers are identical on almost all
2773 processors, this distinction tends to bit-rot pretty quickly. Thus,
2774 each time you port @value{GDBN} to an architecture which does
2775 distinguish between pointers and addresses, you'll probably need to
2776 clean up some architecture-independent code.
2778 Here are functions which convert between pointers and addresses:
2780 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2781 Treat the bytes at @var{buf} as a pointer or reference of type
2782 @var{type}, and return the address it represents, in a manner
2783 appropriate for the current architecture. This yields an address
2784 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2785 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2788 For example, if the current architecture is the Intel x86, this function
2789 extracts a little-endian integer of the appropriate length from
2790 @var{buf} and returns it. However, if the current architecture is the
2791 D10V, this function will return a 16-bit integer extracted from
2792 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2794 If @var{type} is not a pointer or reference type, then this function
2795 will signal an internal error.
2798 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2799 Store the address @var{addr} in @var{buf}, in the proper format for a
2800 pointer of type @var{type} in the current architecture. Note that
2801 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2804 For example, if the current architecture is the Intel x86, this function
2805 stores @var{addr} unmodified as a little-endian integer of the
2806 appropriate length in @var{buf}. However, if the current architecture
2807 is the D10V, this function divides @var{addr} by four if @var{type} is
2808 a pointer to a function, and then stores it in @var{buf}.
2810 If @var{type} is not a pointer or reference type, then this function
2811 will signal an internal error.
2814 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2815 Assuming that @var{val} is a pointer, return the address it represents,
2816 as appropriate for the current architecture.
2818 This function actually works on integral values, as well as pointers.
2819 For pointers, it performs architecture-specific conversions as
2820 described above for @code{extract_typed_address}.
2823 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2824 Create and return a value representing a pointer of type @var{type} to
2825 the address @var{addr}, as appropriate for the current architecture.
2826 This function performs architecture-specific conversions as described
2827 above for @code{store_typed_address}.
2830 Here are some macros which architectures can define to indicate the
2831 relationship between pointers and addresses. These have default
2832 definitions, appropriate for architectures on which all pointers are
2833 simple unsigned byte addresses.
2835 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2836 Assume that @var{buf} holds a pointer of type @var{type}, in the
2837 appropriate format for the current architecture. Return the byte
2838 address the pointer refers to.
2840 This function may safely assume that @var{type} is either a pointer or a
2841 C@t{++} reference type.
2844 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2845 Store in @var{buf} a pointer of type @var{type} representing the address
2846 @var{addr}, in the appropriate format for the current architecture.
2848 This function may safely assume that @var{type} is either a pointer or a
2849 C@t{++} reference type.
2852 @section Address Classes
2853 @cindex address classes
2854 @cindex DW_AT_byte_size
2855 @cindex DW_AT_address_class
2857 Sometimes information about different kinds of addresses is available
2858 via the debug information. For example, some programming environments
2859 define addresses of several different sizes. If the debug information
2860 distinguishes these kinds of address classes through either the size
2861 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2862 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2863 following macros should be defined in order to disambiguate these
2864 types within @value{GDBN} as well as provide the added information to
2865 a @value{GDBN} user when printing type expressions.
2867 @deftypefn {Target Macro} int ADDRESS_CLASS_TYPE_FLAGS (int @var{byte_size}, int @var{dwarf2_addr_class})
2868 Returns the type flags needed to construct a pointer type whose size
2869 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2870 This function is normally called from within a symbol reader. See
2871 @file{dwarf2read.c}.
2874 @deftypefn {Target Macro} char *ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (int @var{type_flags})
2875 Given the type flags representing an address class qualifier, return
2878 @deftypefn {Target Macro} int ADDRESS_CLASS_NAME_to_TYPE_FLAGS (int @var{name}, int *var{type_flags_ptr})
2879 Given an address qualifier name, set the @code{int} refererenced by @var{type_flags_ptr} to the type flags
2880 for that address class qualifier.
2883 Since the need for address classes is rather rare, none of
2884 the address class macros defined by default. Predicate
2885 macros are provided to detect when they are defined.
2887 Consider a hypothetical architecture in which addresses are normally
2888 32-bits wide, but 16-bit addresses are also supported. Furthermore,
2889 suppose that the @w{DWARF 2} information for this architecture simply
2890 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
2891 of these "short" pointers. The following functions could be defined
2892 to implement the address class macros:
2895 somearch_address_class_type_flags (int byte_size,
2896 int dwarf2_addr_class)
2899 return TYPE_FLAG_ADDRESS_CLASS_1;
2905 somearch_address_class_type_flags_to_name (int type_flags)
2907 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
2914 somearch_address_class_name_to_type_flags (char *name,
2915 int *type_flags_ptr)
2917 if (strcmp (name, "short") == 0)
2919 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
2927 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
2928 to indicate the presence of one of these "short" pointers. E.g, if
2929 the debug information indicates that @code{short_ptr_var} is one of these
2930 short pointers, @value{GDBN} might show the following behavior:
2933 (gdb) ptype short_ptr_var
2934 type = int * @@short
2938 @section Raw and Virtual Register Representations
2939 @cindex raw register representation
2940 @cindex virtual register representation
2941 @cindex representations, raw and virtual registers
2943 @emph{Maintainer note: This section is pretty much obsolete. The
2944 functionality described here has largely been replaced by
2945 pseudo-registers and the mechanisms described in @ref{Target
2946 Architecture Definition, , Using Different Register and Memory Data
2947 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
2948 Bug Tracking Database} and
2949 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
2950 up-to-date information.}
2952 Some architectures use one representation for a value when it lives in a
2953 register, but use a different representation when it lives in memory.
2954 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
2955 the target registers, and the @dfn{virtual} representation is the one
2956 used in memory, and within @value{GDBN} @code{struct value} objects.
2958 @emph{Maintainer note: Notice that the same mechanism is being used to
2959 both convert a register to a @code{struct value} and alternative
2962 For almost all data types on almost all architectures, the virtual and
2963 raw representations are identical, and no special handling is needed.
2964 However, they do occasionally differ. For example:
2968 The x86 architecture supports an 80-bit @code{long double} type. However, when
2969 we store those values in memory, they occupy twelve bytes: the
2970 floating-point number occupies the first ten, and the final two bytes
2971 are unused. This keeps the values aligned on four-byte boundaries,
2972 allowing more efficient access. Thus, the x86 80-bit floating-point
2973 type is the raw representation, and the twelve-byte loosely-packed
2974 arrangement is the virtual representation.
2977 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
2978 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
2979 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
2980 raw representation, and the trimmed 32-bit representation is the
2981 virtual representation.
2984 In general, the raw representation is determined by the architecture, or
2985 @value{GDBN}'s interface to the architecture, while the virtual representation
2986 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
2987 @code{registers}, holds the register contents in raw format, and the
2988 @value{GDBN} remote protocol transmits register values in raw format.
2990 Your architecture may define the following macros to request
2991 conversions between the raw and virtual format:
2993 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
2994 Return non-zero if register number @var{reg}'s value needs different raw
2995 and virtual formats.
2997 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
2998 unless this macro returns a non-zero value for that register.
3001 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
3002 The size of register number @var{reg}'s raw value. This is the number
3003 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
3004 remote protocol packet.
3007 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
3008 The size of register number @var{reg}'s value, in its virtual format.
3009 This is the size a @code{struct value}'s buffer will have, holding that
3013 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
3014 This is the type of the virtual representation of register number
3015 @var{reg}. Note that there is no need for a macro giving a type for the
3016 register's raw form; once the register's value has been obtained, @value{GDBN}
3017 always uses the virtual form.
3020 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3021 Convert the value of register number @var{reg} to @var{type}, which
3022 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3023 at @var{from} holds the register's value in raw format; the macro should
3024 convert the value to virtual format, and place it at @var{to}.
3026 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
3027 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
3028 arguments in different orders.
3030 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
3031 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
3035 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3036 Convert the value of register number @var{reg} to @var{type}, which
3037 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3038 at @var{from} holds the register's value in raw format; the macro should
3039 convert the value to virtual format, and place it at @var{to}.
3041 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
3042 their @var{reg} and @var{type} arguments in different orders.
3046 @section Using Different Register and Memory Data Representations
3047 @cindex register representation
3048 @cindex memory representation
3049 @cindex representations, register and memory
3050 @cindex register data formats, converting
3051 @cindex @code{struct value}, converting register contents to
3053 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
3054 significant change. Many of the macros and functions refered to in this
3055 section are likely to be subject to further revision. See
3056 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
3057 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
3058 further information. cagney/2002-05-06.}
3060 Some architectures can represent a data object in a register using a
3061 form that is different to the objects more normal memory representation.
3067 The Alpha architecture can represent 32 bit integer values in
3068 floating-point registers.
3071 The x86 architecture supports 80-bit floating-point registers. The
3072 @code{long double} data type occupies 96 bits in memory but only 80 bits
3073 when stored in a register.
3077 In general, the register representation of a data type is determined by
3078 the architecture, or @value{GDBN}'s interface to the architecture, while
3079 the memory representation is determined by the Application Binary
3082 For almost all data types on almost all architectures, the two
3083 representations are identical, and no special handling is needed.
3084 However, they do occasionally differ. Your architecture may define the
3085 following macros to request conversions between the register and memory
3086 representations of a data type:
3088 @deftypefn {Target Macro} int CONVERT_REGISTER_P (int @var{reg})
3089 Return non-zero if the representation of a data value stored in this
3090 register may be different to the representation of that same data value
3091 when stored in memory.
3093 When non-zero, the macros @code{REGISTER_TO_VALUE} and
3094 @code{VALUE_TO_REGISTER} are used to perform any necessary conversion.
3097 @deftypefn {Target Macro} void REGISTER_TO_VALUE (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3098 Convert the value of register number @var{reg} to a data object of type
3099 @var{type}. The buffer at @var{from} holds the register's value in raw
3100 format; the converted value should be placed in the buffer at @var{to}.
3102 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
3103 their @var{reg} and @var{type} arguments in different orders.
3105 You should only use @code{REGISTER_TO_VALUE} with registers for which
3106 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
3109 @deftypefn {Target Macro} void VALUE_TO_REGISTER (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3110 Convert a data value of type @var{type} to register number @var{reg}'
3113 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
3114 their @var{reg} and @var{type} arguments in different orders.
3116 You should only use @code{VALUE_TO_REGISTER} with registers for which
3117 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
3120 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
3121 See @file{mips-tdep.c}. It does not do what you want.
3125 @section Frame Interpretation
3127 @section Inferior Call Setup
3129 @section Compiler Characteristics
3131 @section Target Conditionals
3133 This section describes the macros that you can use to define the target
3138 @item ADDR_BITS_REMOVE (addr)
3139 @findex ADDR_BITS_REMOVE
3140 If a raw machine instruction address includes any bits that are not
3141 really part of the address, then define this macro to expand into an
3142 expression that zeroes those bits in @var{addr}. This is only used for
3143 addresses of instructions, and even then not in all contexts.
3145 For example, the two low-order bits of the PC on the Hewlett-Packard PA
3146 2.0 architecture contain the privilege level of the corresponding
3147 instruction. Since instructions must always be aligned on four-byte
3148 boundaries, the processor masks out these bits to generate the actual
3149 address of the instruction. ADDR_BITS_REMOVE should filter out these
3150 bits with an expression such as @code{((addr) & ~3)}.
3152 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS (@var{name}, @var{type_flags_ptr})
3153 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS
3154 If @var{name} is a valid address class qualifier name, set the @code{int}
3155 referenced by @var{type_flags_ptr} to the mask representing the qualifier
3156 and return 1. If @var{name} is not a valid address class qualifier name,
3159 The value for @var{type_flags_ptr} should be one of
3160 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
3161 possibly some combination of these values or'd together.
3162 @xref{Target Architecture Definition, , Address Classes}.
3164 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P ()
3165 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P
3166 Predicate which indicates whether @code{ADDRESS_CLASS_NAME_TO_TYPE_FLAGS}
3169 @item ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
3170 @findex ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
3171 Given a pointers byte size (as described by the debug information) and
3172 the possible @code{DW_AT_address_class} value, return the type flags
3173 used by @value{GDBN} to represent this address class. The value
3174 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
3175 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
3176 values or'd together.
3177 @xref{Target Architecture Definition, , Address Classes}.
3179 @item ADDRESS_CLASS_TYPE_FLAGS_P ()
3180 @findex ADDRESS_CLASS_TYPE_FLAGS_P
3181 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS} has
3184 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (@var{type_flags})
3185 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME
3186 Return the name of the address class qualifier associated with the type
3187 flags given by @var{type_flags}.
3189 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P ()
3190 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P
3191 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS_TO_NAME} has
3193 @xref{Target Architecture Definition, , Address Classes}.
3195 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
3196 @findex ADDRESS_TO_POINTER
3197 Store in @var{buf} a pointer of type @var{type} representing the address
3198 @var{addr}, in the appropriate format for the current architecture.
3199 This macro may safely assume that @var{type} is either a pointer or a
3200 C@t{++} reference type.
3201 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3203 @item BELIEVE_PCC_PROMOTION
3204 @findex BELIEVE_PCC_PROMOTION
3205 Define if the compiler promotes a @code{short} or @code{char}
3206 parameter to an @code{int}, but still reports the parameter as its
3207 original type, rather than the promoted type.
3209 @item BITS_BIG_ENDIAN
3210 @findex BITS_BIG_ENDIAN
3211 Define this if the numbering of bits in the targets does @strong{not} match the
3212 endianness of the target byte order. A value of 1 means that the bits
3213 are numbered in a big-endian bit order, 0 means little-endian.
3217 This is the character array initializer for the bit pattern to put into
3218 memory where a breakpoint is set. Although it's common to use a trap
3219 instruction for a breakpoint, it's not required; for instance, the bit
3220 pattern could be an invalid instruction. The breakpoint must be no
3221 longer than the shortest instruction of the architecture.
3223 @code{BREAKPOINT} has been deprecated in favor of
3224 @code{BREAKPOINT_FROM_PC}.
3226 @item BIG_BREAKPOINT
3227 @itemx LITTLE_BREAKPOINT
3228 @findex LITTLE_BREAKPOINT
3229 @findex BIG_BREAKPOINT
3230 Similar to BREAKPOINT, but used for bi-endian targets.
3232 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
3233 favor of @code{BREAKPOINT_FROM_PC}.
3235 @item DEPRECATED_REMOTE_BREAKPOINT
3236 @itemx DEPRECATED_LITTLE_REMOTE_BREAKPOINT
3237 @itemx DEPRECATED_BIG_REMOTE_BREAKPOINT
3238 @findex DEPRECATED_BIG_REMOTE_BREAKPOINT
3239 @findex DEPRECATED_LITTLE_REMOTE_BREAKPOINT
3240 @findex DEPRECATED_REMOTE_BREAKPOINT
3241 Specify the breakpoint instruction sequence for a remote target.
3242 @code{DEPRECATED_REMOTE_BREAKPOINT},
3243 @code{DEPRECATED_BIG_REMOTE_BREAKPOINT} and
3244 @code{DEPRECATED_LITTLE_REMOTE_BREAKPOINT} have been deprecated in
3245 favor of @code{BREAKPOINT_FROM_PC} (@pxref{BREAKPOINT_FROM_PC}).
3247 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
3248 @findex BREAKPOINT_FROM_PC
3249 @anchor{BREAKPOINT_FROM_PC} Use the program counter to determine the
3250 contents and size of a breakpoint instruction. It returns a pointer to
3251 a string of bytes that encode a breakpoint instruction, stores the
3252 length of the string to @code{*@var{lenptr}}, and adjusts the program
3253 counter (if necessary) to point to the actual memory location where the
3254 breakpoint should be inserted.
3256 Although it is common to use a trap instruction for a breakpoint, it's
3257 not required; for instance, the bit pattern could be an invalid
3258 instruction. The breakpoint must be no longer than the shortest
3259 instruction of the architecture.
3261 Replaces all the other @var{BREAKPOINT} macros.
3263 @item MEMORY_INSERT_BREAKPOINT (@var{bp_tgt})
3264 @itemx MEMORY_REMOVE_BREAKPOINT (@var{bp_tgt})
3265 @findex MEMORY_REMOVE_BREAKPOINT
3266 @findex MEMORY_INSERT_BREAKPOINT
3267 Insert or remove memory based breakpoints. Reasonable defaults
3268 (@code{default_memory_insert_breakpoint} and
3269 @code{default_memory_remove_breakpoint} respectively) have been
3270 provided so that it is not necessary to define these for most
3271 architectures. Architectures which may want to define
3272 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
3273 likely have instructions that are oddly sized or are not stored in a
3274 conventional manner.
3276 It may also be desirable (from an efficiency standpoint) to define
3277 custom breakpoint insertion and removal routines if
3278 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
3281 @item ADJUST_BREAKPOINT_ADDRESS (@var{address})
3282 @findex ADJUST_BREAKPOINT_ADDRESS
3283 @cindex breakpoint address adjusted
3284 Given an address at which a breakpoint is desired, return a breakpoint
3285 address adjusted to account for architectural constraints on
3286 breakpoint placement. This method is not needed by most targets.
3288 The FR-V target (see @file{frv-tdep.c}) requires this method.
3289 The FR-V is a VLIW architecture in which a number of RISC-like
3290 instructions are grouped (packed) together into an aggregate
3291 instruction or instruction bundle. When the processor executes
3292 one of these bundles, the component instructions are executed
3295 In the course of optimization, the compiler may group instructions
3296 from distinct source statements into the same bundle. The line number
3297 information associated with one of the latter statements will likely
3298 refer to some instruction other than the first one in the bundle. So,
3299 if the user attempts to place a breakpoint on one of these latter
3300 statements, @value{GDBN} must be careful to @emph{not} place the break
3301 instruction on any instruction other than the first one in the bundle.
3302 (Remember though that the instructions within a bundle execute
3303 in parallel, so the @emph{first} instruction is the instruction
3304 at the lowest address and has nothing to do with execution order.)
3306 The FR-V's @code{ADJUST_BREAKPOINT_ADDRESS} method will adjust a
3307 breakpoint's address by scanning backwards for the beginning of
3308 the bundle, returning the address of the bundle.
3310 Since the adjustment of a breakpoint may significantly alter a user's
3311 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3312 is initially set and each time that that breakpoint is hit.
3314 @item CALL_DUMMY_LOCATION
3315 @findex CALL_DUMMY_LOCATION
3316 See the file @file{inferior.h}.
3318 This method has been replaced by @code{push_dummy_code}
3319 (@pxref{push_dummy_code}).
3321 @item CANNOT_FETCH_REGISTER (@var{regno})
3322 @findex CANNOT_FETCH_REGISTER
3323 A C expression that should be nonzero if @var{regno} cannot be fetched
3324 from an inferior process. This is only relevant if
3325 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3327 @item CANNOT_STORE_REGISTER (@var{regno})
3328 @findex CANNOT_STORE_REGISTER
3329 A C expression that should be nonzero if @var{regno} should not be
3330 written to the target. This is often the case for program counters,
3331 status words, and other special registers. If this is not defined,
3332 @value{GDBN} will assume that all registers may be written.
3334 @item int CONVERT_REGISTER_P(@var{regnum})
3335 @findex CONVERT_REGISTER_P
3336 Return non-zero if register @var{regnum} can represent data values in a
3338 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3340 @item DECR_PC_AFTER_BREAK
3341 @findex DECR_PC_AFTER_BREAK
3342 Define this to be the amount by which to decrement the PC after the
3343 program encounters a breakpoint. This is often the number of bytes in
3344 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3346 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3347 @findex DISABLE_UNSETTABLE_BREAK
3348 If defined, this should evaluate to 1 if @var{addr} is in a shared
3349 library in which breakpoints cannot be set and so should be disabled.
3351 @item PRINT_FLOAT_INFO()
3352 @findex PRINT_FLOAT_INFO
3353 If defined, then the @samp{info float} command will print information about
3354 the processor's floating point unit.
3356 @item print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3357 @findex print_registers_info
3358 If defined, pretty print the value of the register @var{regnum} for the
3359 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3360 either all registers (@var{all} is non zero) or a select subset of
3361 registers (@var{all} is zero).
3363 The default method prints one register per line, and if @var{all} is
3364 zero omits floating-point registers.
3366 @item PRINT_VECTOR_INFO()
3367 @findex PRINT_VECTOR_INFO
3368 If defined, then the @samp{info vector} command will call this function
3369 to print information about the processor's vector unit.
3371 By default, the @samp{info vector} command will print all vector
3372 registers (the register's type having the vector attribute).
3374 @item DWARF_REG_TO_REGNUM
3375 @findex DWARF_REG_TO_REGNUM
3376 Convert DWARF register number into @value{GDBN} regnum. If not defined,
3377 no conversion will be performed.
3379 @item DWARF2_REG_TO_REGNUM
3380 @findex DWARF2_REG_TO_REGNUM
3381 Convert DWARF2 register number into @value{GDBN} regnum. If not
3382 defined, no conversion will be performed.
3384 @item ECOFF_REG_TO_REGNUM
3385 @findex ECOFF_REG_TO_REGNUM
3386 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
3387 no conversion will be performed.
3389 @item END_OF_TEXT_DEFAULT
3390 @findex END_OF_TEXT_DEFAULT
3391 This is an expression that should designate the end of the text section.
3394 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
3395 @findex EXTRACT_RETURN_VALUE
3396 Define this to extract a function's return value of type @var{type} from
3397 the raw register state @var{regbuf} and copy that, in virtual format,
3400 This method has been deprecated in favour of @code{gdbarch_return_value}
3401 (@pxref{gdbarch_return_value}).
3403 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
3404 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS
3405 @anchor{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}
3406 When defined, extract from the array @var{regbuf} (containing the raw
3407 register state) the @code{CORE_ADDR} at which a function should return
3408 its structure value.
3410 @xref{gdbarch_return_value}.
3412 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P()
3413 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P
3414 Predicate for @code{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}.
3416 @item DEPRECATED_FP_REGNUM
3417 @findex DEPRECATED_FP_REGNUM
3418 If the virtual frame pointer is kept in a register, then define this
3419 macro to be the number (greater than or equal to zero) of that register.
3421 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3424 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3425 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3426 Define this to an expression that returns 1 if the function invocation
3427 represented by @var{fi} does not have a stack frame associated with it.
3430 @item frame_align (@var{address})
3431 @anchor{frame_align}
3433 Define this to adjust @var{address} so that it meets the alignment
3434 requirements for the start of a new stack frame. A stack frame's
3435 alignment requirements are typically stronger than a target processors
3436 stack alignment requirements (@pxref{DEPRECATED_STACK_ALIGN}).
3438 This function is used to ensure that, when creating a dummy frame, both
3439 the initial stack pointer and (if needed) the address of the return
3440 value are correctly aligned.
3442 Unlike @code{DEPRECATED_STACK_ALIGN}, this function always adjusts the
3443 address in the direction of stack growth.
3445 By default, no frame based stack alignment is performed.
3447 @item int frame_red_zone_size
3449 The number of bytes, beyond the innermost-stack-address, reserved by the
3450 @sc{abi}. A function is permitted to use this scratch area (instead of
3451 allocating extra stack space).
3453 When performing an inferior function call, to ensure that it does not
3454 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3455 @var{frame_red_zone_size} bytes before pushing parameters onto the
3458 By default, zero bytes are allocated. The value must be aligned
3459 (@pxref{frame_align}).
3461 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3462 @emph{red zone} when describing this scratch area.
3465 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3466 @findex DEPRECATED_FRAME_CHAIN
3467 Given @var{frame}, return a pointer to the calling frame.
3469 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3470 @findex DEPRECATED_FRAME_CHAIN_VALID
3471 Define this to be an expression that returns zero if the given frame is an
3472 outermost frame, with no caller, and nonzero otherwise. Most normal
3473 situations can be handled without defining this macro, including @code{NULL}
3474 chain pointers, dummy frames, and frames whose PC values are inside the
3475 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3478 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3479 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3480 See @file{frame.h}. Determines the address of all registers in the
3481 current stack frame storing each in @code{frame->saved_regs}. Space for
3482 @code{frame->saved_regs} shall be allocated by
3483 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3484 @code{frame_saved_regs_zalloc}.
3486 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3488 @item FRAME_NUM_ARGS (@var{fi})
3489 @findex FRAME_NUM_ARGS
3490 For the frame described by @var{fi} return the number of arguments that
3491 are being passed. If the number of arguments is not known, return
3494 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3495 @findex DEPRECATED_FRAME_SAVED_PC
3496 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3497 saved there. This is the return address.
3499 This method is deprecated. @xref{unwind_pc}.
3501 @item CORE_ADDR unwind_pc (struct frame_info *@var{this_frame})
3503 @anchor{unwind_pc} Return the instruction address, in @var{this_frame}'s
3504 caller, at which execution will resume after @var{this_frame} returns.
3505 This is commonly refered to as the return address.
3507 The implementation, which must be frame agnostic (work with any frame),
3508 is typically no more than:
3512 frame_unwind_unsigned_register (this_frame, D10V_PC_REGNUM, &pc);
3513 return d10v_make_iaddr (pc);
3517 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3519 @item CORE_ADDR unwind_sp (struct frame_info *@var{this_frame})
3521 @anchor{unwind_sp} Return the frame's inner most stack address. This is
3522 commonly refered to as the frame's @dfn{stack pointer}.
3524 The implementation, which must be frame agnostic (work with any frame),
3525 is typically no more than:
3529 frame_unwind_unsigned_register (this_frame, D10V_SP_REGNUM, &sp);
3530 return d10v_make_daddr (sp);
3534 @xref{TARGET_READ_SP}, which this method replaces.
3536 @item FUNCTION_EPILOGUE_SIZE
3537 @findex FUNCTION_EPILOGUE_SIZE
3538 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3539 function end symbol is 0. For such targets, you must define
3540 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3541 function's epilogue.
3543 @item DEPRECATED_FUNCTION_START_OFFSET
3544 @findex DEPRECATED_FUNCTION_START_OFFSET
3545 An integer, giving the offset in bytes from a function's address (as
3546 used in the values of symbols, function pointers, etc.), and the
3547 function's first genuine instruction.
3549 This is zero on almost all machines: the function's address is usually
3550 the address of its first instruction. However, on the VAX, for
3551 example, each function starts with two bytes containing a bitmask
3552 indicating which registers to save upon entry to the function. The
3553 VAX @code{call} instructions check this value, and save the
3554 appropriate registers automatically. Thus, since the offset from the
3555 function's address to its first instruction is two bytes,
3556 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3558 @item GCC_COMPILED_FLAG_SYMBOL
3559 @itemx GCC2_COMPILED_FLAG_SYMBOL
3560 @findex GCC2_COMPILED_FLAG_SYMBOL
3561 @findex GCC_COMPILED_FLAG_SYMBOL
3562 If defined, these are the names of the symbols that @value{GDBN} will
3563 look for to detect that GCC compiled the file. The default symbols
3564 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3565 respectively. (Currently only defined for the Delta 68.)
3567 @item @value{GDBN}_MULTI_ARCH
3568 @findex @value{GDBN}_MULTI_ARCH
3569 If defined and non-zero, enables support for multiple architectures
3570 within @value{GDBN}.
3572 This support can be enabled at two levels. At level one, only
3573 definitions for previously undefined macros are provided; at level two,
3574 a multi-arch definition of all architecture dependent macros will be
3577 @item @value{GDBN}_TARGET_IS_HPPA
3578 @findex @value{GDBN}_TARGET_IS_HPPA
3579 This determines whether horrible kludge code in @file{dbxread.c} and
3580 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3581 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3584 @item GET_LONGJMP_TARGET
3585 @findex GET_LONGJMP_TARGET
3586 For most machines, this is a target-dependent parameter. On the
3587 DECstation and the Iris, this is a native-dependent parameter, since
3588 the header file @file{setjmp.h} is needed to define it.
3590 This macro determines the target PC address that @code{longjmp} will jump to,
3591 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3592 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3593 pointer. It examines the current state of the machine as needed.
3595 @item DEPRECATED_GET_SAVED_REGISTER
3596 @findex DEPRECATED_GET_SAVED_REGISTER
3597 Define this if you need to supply your own definition for the function
3598 @code{DEPRECATED_GET_SAVED_REGISTER}.
3600 @item DEPRECATED_IBM6000_TARGET
3601 @findex DEPRECATED_IBM6000_TARGET
3602 Shows that we are configured for an IBM RS/6000 system. This
3603 conditional should be eliminated (FIXME) and replaced by
3604 feature-specific macros. It was introduced in a haste and we are
3605 repenting at leisure.
3607 @item I386_USE_GENERIC_WATCHPOINTS
3608 An x86-based target can define this to use the generic x86 watchpoint
3609 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3611 @item SYMBOLS_CAN_START_WITH_DOLLAR
3612 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3613 Some systems have routines whose names start with @samp{$}. Giving this
3614 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3615 routines when parsing tokens that begin with @samp{$}.
3617 On HP-UX, certain system routines (millicode) have names beginning with
3618 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3619 routine that handles inter-space procedure calls on PA-RISC.
3621 @item DEPRECATED_INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3622 @findex DEPRECATED_INIT_EXTRA_FRAME_INFO
3623 If additional information about the frame is required this should be
3624 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3625 is allocated using @code{frame_extra_info_zalloc}.
3627 @item DEPRECATED_INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3628 @findex DEPRECATED_INIT_FRAME_PC
3629 This is a C statement that sets the pc of the frame pointed to by
3630 @var{prev}. [By default...]
3632 @item INNER_THAN (@var{lhs}, @var{rhs})
3634 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3635 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3636 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3639 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3640 @findex gdbarch_in_function_epilogue_p
3641 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3642 The epilogue of a function is defined as the part of a function where
3643 the stack frame of the function already has been destroyed up to the
3644 final `return from function call' instruction.
3646 @item DEPRECATED_SIGTRAMP_START (@var{pc})
3647 @findex DEPRECATED_SIGTRAMP_START
3648 @itemx DEPRECATED_SIGTRAMP_END (@var{pc})
3649 @findex DEPRECATED_SIGTRAMP_END
3650 Define these to be the start and end address of the @code{sigtramp} for the
3651 given @var{pc}. On machines where the address is just a compile time
3652 constant, the macro expansion will typically just ignore the supplied
3655 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3656 @findex IN_SOLIB_CALL_TRAMPOLINE
3657 Define this to evaluate to nonzero if the program is stopped in the
3658 trampoline that connects to a shared library.
3660 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3661 @findex IN_SOLIB_RETURN_TRAMPOLINE
3662 Define this to evaluate to nonzero if the program is stopped in the
3663 trampoline that returns from a shared library.
3665 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3666 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3667 Define this to evaluate to nonzero if the program is stopped in the
3670 @item SKIP_SOLIB_RESOLVER (@var{pc})
3671 @findex SKIP_SOLIB_RESOLVER
3672 Define this to evaluate to the (nonzero) address at which execution
3673 should continue to get past the dynamic linker's symbol resolution
3674 function. A zero value indicates that it is not important or necessary
3675 to set a breakpoint to get through the dynamic linker and that single
3676 stepping will suffice.
3678 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3679 @findex INTEGER_TO_ADDRESS
3680 @cindex converting integers to addresses
3681 Define this when the architecture needs to handle non-pointer to address
3682 conversions specially. Converts that value to an address according to
3683 the current architectures conventions.
3685 @emph{Pragmatics: When the user copies a well defined expression from
3686 their source code and passes it, as a parameter, to @value{GDBN}'s
3687 @code{print} command, they should get the same value as would have been
3688 computed by the target program. Any deviation from this rule can cause
3689 major confusion and annoyance, and needs to be justified carefully. In
3690 other words, @value{GDBN} doesn't really have the freedom to do these
3691 conversions in clever and useful ways. It has, however, been pointed
3692 out that users aren't complaining about how @value{GDBN} casts integers
3693 to pointers; they are complaining that they can't take an address from a
3694 disassembly listing and give it to @code{x/i}. Adding an architecture
3695 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3696 @value{GDBN} to ``get it right'' in all circumstances.}
3698 @xref{Target Architecture Definition, , Pointers Are Not Always
3701 @item NO_HIF_SUPPORT
3702 @findex NO_HIF_SUPPORT
3703 (Specific to the a29k.)
3705 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3706 @findex POINTER_TO_ADDRESS
3707 Assume that @var{buf} holds a pointer of type @var{type}, in the
3708 appropriate format for the current architecture. Return the byte
3709 address the pointer refers to.
3710 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3712 @item REGISTER_CONVERTIBLE (@var{reg})
3713 @findex REGISTER_CONVERTIBLE
3714 Return non-zero if @var{reg} uses different raw and virtual formats.
3715 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3717 @item REGISTER_TO_VALUE(@var{regnum}, @var{type}, @var{from}, @var{to})
3718 @findex REGISTER_TO_VALUE
3719 Convert the raw contents of register @var{regnum} into a value of type
3721 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3723 @item DEPRECATED_REGISTER_RAW_SIZE (@var{reg})
3724 @findex DEPRECATED_REGISTER_RAW_SIZE
3725 Return the raw size of @var{reg}; defaults to the size of the register's
3727 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3729 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3730 @findex register_reggroup_p
3731 @cindex register groups
3732 Return non-zero if register @var{regnum} is a member of the register
3733 group @var{reggroup}.
3735 By default, registers are grouped as follows:
3738 @item float_reggroup
3739 Any register with a valid name and a floating-point type.
3740 @item vector_reggroup
3741 Any register with a valid name and a vector type.
3742 @item general_reggroup
3743 Any register with a valid name and a type other than vector or
3744 floating-point. @samp{float_reggroup}.
3746 @itemx restore_reggroup
3748 Any register with a valid name.
3751 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3752 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3753 Return the virtual size of @var{reg}; defaults to the size of the
3754 register's virtual type.
3755 Return the virtual size of @var{reg}.
3756 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3758 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3759 @findex REGISTER_VIRTUAL_TYPE
3760 Return the virtual type of @var{reg}.
3761 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3763 @item struct type *register_type (@var{gdbarch}, @var{reg})
3764 @findex register_type
3765 If defined, return the type of register @var{reg}. This function
3766 superseeds @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3767 Definition, , Raw and Virtual Register Representations}.
3769 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3770 @findex REGISTER_CONVERT_TO_VIRTUAL
3771 Convert the value of register @var{reg} from its raw form to its virtual
3773 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3775 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3776 @findex REGISTER_CONVERT_TO_RAW
3777 Convert the value of register @var{reg} from its virtual form to its raw
3779 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3781 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3782 @findex regset_from_core_section
3783 Return the appropriate register set for a core file section with name
3784 @var{sect_name} and size @var{sect_size}.
3786 @item SOFTWARE_SINGLE_STEP_P()
3787 @findex SOFTWARE_SINGLE_STEP_P
3788 Define this as 1 if the target does not have a hardware single-step
3789 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3791 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3792 @findex SOFTWARE_SINGLE_STEP
3793 A function that inserts or removes (depending on
3794 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3795 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3798 @item SOFUN_ADDRESS_MAYBE_MISSING
3799 @findex SOFUN_ADDRESS_MAYBE_MISSING
3800 Somebody clever observed that, the more actual addresses you have in the
3801 debug information, the more time the linker has to spend relocating
3802 them. So whenever there's some other way the debugger could find the
3803 address it needs, you should omit it from the debug info, to make
3806 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3807 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3808 entries in stabs-format debugging information. @code{N_SO} stabs mark
3809 the beginning and ending addresses of compilation units in the text
3810 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3812 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3816 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3817 addresses where the function starts by taking the function name from
3818 the stab, and then looking that up in the minsyms (the
3819 linker/assembler symbol table). In other words, the stab has the
3820 name, and the linker/assembler symbol table is the only place that carries
3824 @code{N_SO} stabs have an address of zero, too. You just look at the
3825 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3826 and guess the starting and ending addresses of the compilation unit from
3830 @item PC_LOAD_SEGMENT
3831 @findex PC_LOAD_SEGMENT
3832 If defined, print information about the load segment for the program
3833 counter. (Defined only for the RS/6000.)
3837 If the program counter is kept in a register, then define this macro to
3838 be the number (greater than or equal to zero) of that register.
3840 This should only need to be defined if @code{TARGET_READ_PC} and
3841 @code{TARGET_WRITE_PC} are not defined.
3844 @findex PARM_BOUNDARY
3845 If non-zero, round arguments to a boundary of this many bits before
3846 pushing them on the stack.
3848 @item stabs_argument_has_addr (@var{gdbarch}, @var{type})
3849 @findex stabs_argument_has_addr
3850 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3851 @anchor{stabs_argument_has_addr} Define this to return nonzero if a
3852 function argument of type @var{type} is passed by reference instead of
3855 This method replaces @code{DEPRECATED_REG_STRUCT_HAS_ADDR}
3856 (@pxref{DEPRECATED_REG_STRUCT_HAS_ADDR}).
3858 @item PROCESS_LINENUMBER_HOOK
3859 @findex PROCESS_LINENUMBER_HOOK
3860 A hook defined for XCOFF reading.
3862 @item PROLOGUE_FIRSTLINE_OVERLAP
3863 @findex PROLOGUE_FIRSTLINE_OVERLAP
3864 (Only used in unsupported Convex configuration.)
3868 If defined, this is the number of the processor status register. (This
3869 definition is only used in generic code when parsing "$ps".)
3871 @item DEPRECATED_POP_FRAME
3872 @findex DEPRECATED_POP_FRAME
3874 If defined, used by @code{frame_pop} to remove a stack frame. This
3875 method has been superseeded by generic code.
3877 @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})
3878 @findex push_dummy_call
3879 @findex DEPRECATED_PUSH_ARGUMENTS.
3880 @anchor{push_dummy_call} Define this to push the dummy frame's call to
3881 the inferior function onto the stack. In addition to pushing
3882 @var{nargs}, the code should push @var{struct_addr} (when
3883 @var{struct_return}), and the return address (@var{bp_addr}).
3885 @var{function} is a pointer to a @code{struct value}; on architectures that use
3886 function descriptors, this contains the function descriptor value.
3888 Returns the updated top-of-stack pointer.
3890 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3892 @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})
3893 @findex push_dummy_code
3894 @anchor{push_dummy_code} Given a stack based call dummy, push the
3895 instruction sequence (including space for a breakpoint) to which the
3896 called function should return.
3898 Set @var{bp_addr} to the address at which the breakpoint instruction
3899 should be inserted, @var{real_pc} to the resume address when starting
3900 the call sequence, and return the updated inner-most stack address.
3902 By default, the stack is grown sufficient to hold a frame-aligned
3903 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3904 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3906 This method replaces @code{CALL_DUMMY_LOCATION},
3907 @code{DEPRECATED_REGISTER_SIZE}.
3909 @item REGISTER_NAME(@var{i})
3910 @findex REGISTER_NAME
3911 Return the name of register @var{i} as a string. May return @code{NULL}
3912 or @code{NUL} to indicate that register @var{i} is not valid.
3914 @item DEPRECATED_REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3915 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3916 @anchor{DEPRECATED_REG_STRUCT_HAS_ADDR}Define this to return 1 if the
3917 given type will be passed by pointer rather than directly.
3919 This method has been replaced by @code{stabs_argument_has_addr}
3920 (@pxref{stabs_argument_has_addr}).
3922 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3923 @findex SAVE_DUMMY_FRAME_TOS
3924 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3925 notify the target dependent code of the top-of-stack value that will be
3926 passed to the the inferior code. This is the value of the @code{SP}
3927 after both the dummy frame and space for parameters/results have been
3928 allocated on the stack. @xref{unwind_dummy_id}.
3930 @item SDB_REG_TO_REGNUM
3931 @findex SDB_REG_TO_REGNUM
3932 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3933 defined, no conversion will be done.
3935 @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})
3936 @findex gdbarch_return_value
3937 @anchor{gdbarch_return_value} Given a function with a return-value of
3938 type @var{rettype}, return which return-value convention that function
3941 @value{GDBN} currently recognizes two function return-value conventions:
3942 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3943 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3944 value is found in memory and the address of that memory location is
3945 passed in as the function's first parameter.
3947 If the register convention is being used, and @var{writebuf} is
3948 non-@code{NULL}, also copy the return-value in @var{writebuf} into
3951 If the register convention is being used, and @var{readbuf} is
3952 non-@code{NULL}, also copy the return value from @var{regcache} into
3953 @var{readbuf} (@var{regcache} contains a copy of the registers from the
3954 just returned function).
3956 @xref{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}, for a description of how
3957 return-values that use the struct convention are handled.
3959 @emph{Maintainer note: This method replaces separate predicate, extract,
3960 store methods. By having only one method, the logic needed to determine
3961 the return-value convention need only be implemented in one place. If
3962 @value{GDBN} were written in an @sc{oo} language, this method would
3963 instead return an object that knew how to perform the register
3964 return-value extract and store.}
3966 @emph{Maintainer note: This method does not take a @var{gcc_p}
3967 parameter, and such a parameter should not be added. If an architecture
3968 that requires per-compiler or per-function information be identified,
3969 then the replacement of @var{rettype} with @code{struct value}
3970 @var{function} should be persued.}
3972 @emph{Maintainer note: The @var{regcache} parameter limits this methods
3973 to the inner most frame. While replacing @var{regcache} with a
3974 @code{struct frame_info} @var{frame} parameter would remove that
3975 limitation there has yet to be a demonstrated need for such a change.}
3977 @item SKIP_PERMANENT_BREAKPOINT
3978 @findex SKIP_PERMANENT_BREAKPOINT
3979 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3980 steps over a breakpoint by removing it, stepping one instruction, and
3981 re-inserting the breakpoint. However, permanent breakpoints are
3982 hardwired into the inferior, and can't be removed, so this strategy
3983 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
3984 state so that execution will resume just after the breakpoint. This
3985 macro does the right thing even when the breakpoint is in the delay slot
3986 of a branch or jump.
3988 @item SKIP_PROLOGUE (@var{pc})
3989 @findex SKIP_PROLOGUE
3990 A C expression that returns the address of the ``real'' code beyond the
3991 function entry prologue found at @var{pc}.
3993 @item SKIP_TRAMPOLINE_CODE (@var{pc})
3994 @findex SKIP_TRAMPOLINE_CODE
3995 If the target machine has trampoline code that sits between callers and
3996 the functions being called, then define this macro to return a new PC
3997 that is at the start of the real function.
4001 If the stack-pointer is kept in a register, then define this macro to be
4002 the number (greater than or equal to zero) of that register, or -1 if
4003 there is no such register.
4005 @item STAB_REG_TO_REGNUM
4006 @findex STAB_REG_TO_REGNUM
4007 Define this to convert stab register numbers (as gotten from `r'
4008 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
4011 @item DEPRECATED_STACK_ALIGN (@var{addr})
4012 @anchor{DEPRECATED_STACK_ALIGN}
4013 @findex DEPRECATED_STACK_ALIGN
4014 Define this to increase @var{addr} so that it meets the alignment
4015 requirements for the processor's stack.
4017 Unlike @ref{frame_align}, this function always adjusts @var{addr}
4020 By default, no stack alignment is performed.
4022 @item STEP_SKIPS_DELAY (@var{addr})
4023 @findex STEP_SKIPS_DELAY
4024 Define this to return true if the address is of an instruction with a
4025 delay slot. If a breakpoint has been placed in the instruction's delay
4026 slot, @value{GDBN} will single-step over that instruction before resuming
4027 normally. Currently only defined for the Mips.
4029 @item STORE_RETURN_VALUE (@var{type}, @var{regcache}, @var{valbuf})
4030 @findex STORE_RETURN_VALUE
4031 A C expression that writes the function return value, found in
4032 @var{valbuf}, into the @var{regcache}. @var{type} is the type of the
4033 value that is to be returned.
4035 This method has been deprecated in favour of @code{gdbarch_return_value}
4036 (@pxref{gdbarch_return_value}).
4038 @item SYMBOL_RELOADING_DEFAULT
4039 @findex SYMBOL_RELOADING_DEFAULT
4040 The default value of the ``symbol-reloading'' variable. (Never defined in
4043 @item TARGET_CHAR_BIT
4044 @findex TARGET_CHAR_BIT
4045 Number of bits in a char; defaults to 8.
4047 @item TARGET_CHAR_SIGNED
4048 @findex TARGET_CHAR_SIGNED
4049 Non-zero if @code{char} is normally signed on this architecture; zero if
4050 it should be unsigned.
4052 The ISO C standard requires the compiler to treat @code{char} as
4053 equivalent to either @code{signed char} or @code{unsigned char}; any
4054 character in the standard execution set is supposed to be positive.
4055 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4056 on the IBM S/390, RS6000, and PowerPC targets.
4058 @item TARGET_COMPLEX_BIT
4059 @findex TARGET_COMPLEX_BIT
4060 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
4062 At present this macro is not used.
4064 @item TARGET_DOUBLE_BIT
4065 @findex TARGET_DOUBLE_BIT
4066 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
4068 @item TARGET_DOUBLE_COMPLEX_BIT
4069 @findex TARGET_DOUBLE_COMPLEX_BIT
4070 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
4072 At present this macro is not used.
4074 @item TARGET_FLOAT_BIT
4075 @findex TARGET_FLOAT_BIT
4076 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
4078 @item TARGET_INT_BIT
4079 @findex TARGET_INT_BIT
4080 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
4082 @item TARGET_LONG_BIT
4083 @findex TARGET_LONG_BIT
4084 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
4086 @item TARGET_LONG_DOUBLE_BIT
4087 @findex TARGET_LONG_DOUBLE_BIT
4088 Number of bits in a long double float;
4089 defaults to @code{2 * TARGET_DOUBLE_BIT}.
4091 @item TARGET_LONG_LONG_BIT
4092 @findex TARGET_LONG_LONG_BIT
4093 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
4095 @item TARGET_PTR_BIT
4096 @findex TARGET_PTR_BIT
4097 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
4099 @item TARGET_SHORT_BIT
4100 @findex TARGET_SHORT_BIT
4101 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
4103 @item TARGET_READ_PC
4104 @findex TARGET_READ_PC
4105 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
4106 @findex TARGET_WRITE_PC
4107 @anchor{TARGET_WRITE_PC}
4108 @itemx TARGET_READ_SP
4109 @findex TARGET_READ_SP
4110 @itemx TARGET_READ_FP
4111 @findex TARGET_READ_FP
4116 @anchor{TARGET_READ_SP} These change the behavior of @code{read_pc},
4117 @code{write_pc}, and @code{read_sp}. For most targets, these may be
4118 left undefined. @value{GDBN} will call the read and write register
4119 functions with the relevant @code{_REGNUM} argument.
4121 These macros are useful when a target keeps one of these registers in a
4122 hard to get at place; for example, part in a segment register and part
4123 in an ordinary register.
4125 @xref{unwind_sp}, which replaces @code{TARGET_READ_SP}.
4127 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
4128 @findex TARGET_VIRTUAL_FRAME_POINTER
4129 Returns a @code{(register, offset)} pair representing the virtual frame
4130 pointer in use at the code address @var{pc}. If virtual frame pointers
4131 are not used, a default definition simply returns
4132 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
4134 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4135 If non-zero, the target has support for hardware-assisted
4136 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4137 other related macros.
4139 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
4140 @findex TARGET_PRINT_INSN
4141 This is the function used by @value{GDBN} to print an assembly
4142 instruction. It prints the instruction at address @var{addr} in
4143 debugged memory and returns the length of the instruction, in bytes. If
4144 a target doesn't define its own printing routine, it defaults to an
4145 accessor function for the global pointer
4146 @code{deprecated_tm_print_insn}. This usually points to a function in
4147 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
4148 @var{info} is a structure (of type @code{disassemble_info}) defined in
4149 @file{include/dis-asm.h} used to pass information to the instruction
4152 @item struct frame_id unwind_dummy_id (struct frame_info *@var{frame})
4153 @findex unwind_dummy_id
4154 @anchor{unwind_dummy_id} Given @var{frame} return a @code{struct
4155 frame_id} that uniquely identifies an inferior function call's dummy
4156 frame. The value returned must match the dummy frame stack value
4157 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
4158 @xref{SAVE_DUMMY_FRAME_TOS}.
4160 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
4161 @findex DEPRECATED_USE_STRUCT_CONVENTION
4162 If defined, this must be an expression that is nonzero if a value of the
4163 given @var{type} being returned from a function must have space
4164 allocated for it on the stack. @var{gcc_p} is true if the function
4165 being considered is known to have been compiled by GCC; this is helpful
4166 for systems where GCC is known to use different calling convention than
4169 This method has been deprecated in favour of @code{gdbarch_return_value}
4170 (@pxref{gdbarch_return_value}).
4172 @item VALUE_TO_REGISTER(@var{type}, @var{regnum}, @var{from}, @var{to})
4173 @findex VALUE_TO_REGISTER
4174 Convert a value of type @var{type} into the raw contents of register
4176 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4178 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
4179 @findex VARIABLES_INSIDE_BLOCK
4180 For dbx-style debugging information, if the compiler puts variable
4181 declarations inside LBRAC/RBRAC blocks, this should be defined to be
4182 nonzero. @var{desc} is the value of @code{n_desc} from the
4183 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
4184 presence of either the @code{GCC_COMPILED_SYMBOL} or the
4185 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
4187 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
4188 @findex OS9K_VARIABLES_INSIDE_BLOCK
4189 Similarly, for OS/9000. Defaults to 1.
4192 Motorola M68K target conditionals.
4196 Define this to be the 4-bit location of the breakpoint trap vector. If
4197 not defined, it will default to @code{0xf}.
4199 @item REMOTE_BPT_VECTOR
4200 Defaults to @code{1}.
4202 @item NAME_OF_MALLOC
4203 @findex NAME_OF_MALLOC
4204 A string containing the name of the function to call in order to
4205 allocate some memory in the inferior. The default value is "malloc".
4209 @section Adding a New Target
4211 @cindex adding a target
4212 The following files add a target to @value{GDBN}:
4216 @item gdb/config/@var{arch}/@var{ttt}.mt
4217 Contains a Makefile fragment specific to this target. Specifies what
4218 object files are needed for target @var{ttt}, by defining
4219 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
4220 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
4223 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
4224 but these are now deprecated, replaced by autoconf, and may go away in
4225 future versions of @value{GDBN}.
4227 @item gdb/@var{ttt}-tdep.c
4228 Contains any miscellaneous code required for this target machine. On
4229 some machines it doesn't exist at all. Sometimes the macros in
4230 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
4231 as functions here instead, and the macro is simply defined to call the
4232 function. This is vastly preferable, since it is easier to understand
4235 @item gdb/@var{arch}-tdep.c
4236 @itemx gdb/@var{arch}-tdep.h
4237 This often exists to describe the basic layout of the target machine's
4238 processor chip (registers, stack, etc.). If used, it is included by
4239 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
4242 @item gdb/config/@var{arch}/tm-@var{ttt}.h
4243 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
4244 macro definitions about the target machine's registers, stack frame
4245 format and instructions.
4247 New targets do not need this file and should not create it.
4249 @item gdb/config/@var{arch}/tm-@var{arch}.h
4250 This often exists to describe the basic layout of the target machine's
4251 processor chip (registers, stack, etc.). If used, it is included by
4252 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
4255 New targets do not need this file and should not create it.
4259 If you are adding a new operating system for an existing CPU chip, add a
4260 @file{config/tm-@var{os}.h} file that describes the operating system
4261 facilities that are unusual (extra symbol table info; the breakpoint
4262 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
4263 that just @code{#include}s @file{tm-@var{arch}.h} and
4264 @file{config/tm-@var{os}.h}.
4267 @section Converting an existing Target Architecture to Multi-arch
4268 @cindex converting targets to multi-arch
4270 This section describes the current accepted best practice for converting
4271 an existing target architecture to the multi-arch framework.
4273 The process consists of generating, testing, posting and committing a
4274 sequence of patches. Each patch must contain a single change, for
4280 Directly convert a group of functions into macros (the conversion does
4281 not change the behavior of any of the functions).
4284 Replace a non-multi-arch with a multi-arch mechanism (e.g.,
4288 Enable multi-arch level one.
4291 Delete one or more files.
4296 There isn't a size limit on a patch, however, a developer is strongly
4297 encouraged to keep the patch size down.
4299 Since each patch is well defined, and since each change has been tested
4300 and shows no regressions, the patches are considered @emph{fairly}
4301 obvious. Such patches, when submitted by developers listed in the
4302 @file{MAINTAINERS} file, do not need approval. Occasional steps in the
4303 process may be more complicated and less clear. The developer is
4304 expected to use their judgment and is encouraged to seek advice as
4307 @subsection Preparation
4309 The first step is to establish control. Build (with @option{-Werror}
4310 enabled) and test the target so that there is a baseline against which
4311 the debugger can be compared.
4313 At no stage can the test results regress or @value{GDBN} stop compiling
4314 with @option{-Werror}.
4316 @subsection Add the multi-arch initialization code
4318 The objective of this step is to establish the basic multi-arch
4319 framework. It involves
4324 The addition of a @code{@var{arch}_gdbarch_init} function@footnote{The
4325 above is from the original example and uses K&R C. @value{GDBN}
4326 has since converted to ISO C but lets ignore that.} that creates
4329 static struct gdbarch *
4330 d10v_gdbarch_init (info, arches)
4331 struct gdbarch_info info;
4332 struct gdbarch_list *arches;
4334 struct gdbarch *gdbarch;
4335 /* there is only one d10v architecture */
4337 return arches->gdbarch;
4338 gdbarch = gdbarch_alloc (&info, NULL);
4346 A per-architecture dump function to print any architecture specific
4350 mips_dump_tdep (struct gdbarch *current_gdbarch,
4351 struct ui_file *file)
4353 @dots{} code to print architecture specific info @dots{}
4358 A change to @code{_initialize_@var{arch}_tdep} to register this new
4362 _initialize_mips_tdep (void)
4364 gdbarch_register (bfd_arch_mips, mips_gdbarch_init,
4369 Add the macro @code{GDB_MULTI_ARCH}, defined as 0 (zero), to the file@*
4370 @file{config/@var{arch}/tm-@var{arch}.h}.
4374 @subsection Update multi-arch incompatible mechanisms
4376 Some mechanisms do not work with multi-arch. They include:
4379 @item FRAME_FIND_SAVED_REGS
4380 Replaced with @code{DEPRECATED_FRAME_INIT_SAVED_REGS}
4384 At this stage you could also consider converting the macros into
4387 @subsection Prepare for multi-arch level to one
4389 Temporally set @code{GDB_MULTI_ARCH} to @code{GDB_MULTI_ARCH_PARTIAL}
4390 and then build and start @value{GDBN} (the change should not be
4391 committed). @value{GDBN} may not build, and once built, it may die with
4392 an internal error listing the architecture methods that must be
4395 Fix any build problems (patch(es)).
4397 Convert all the architecture methods listed, which are only macros, into
4398 functions (patch(es)).
4400 Update @code{@var{arch}_gdbarch_init} to set all the missing
4401 architecture methods and wrap the corresponding macros in @code{#if
4402 !GDB_MULTI_ARCH} (patch(es)).
4404 @subsection Set multi-arch level one
4406 Change the value of @code{GDB_MULTI_ARCH} to GDB_MULTI_ARCH_PARTIAL (a
4409 Any problems with throwing ``the switch'' should have been fixed
4412 @subsection Convert remaining macros
4414 Suggest converting macros into functions (and setting the corresponding
4415 architecture method) in small batches.
4417 @subsection Set multi-arch level to two
4419 This should go smoothly.
4421 @subsection Delete the TM file
4423 The @file{tm-@var{arch}.h} can be deleted. @file{@var{arch}.mt} and
4424 @file{configure.in} updated.
4427 @node Target Vector Definition
4429 @chapter Target Vector Definition
4430 @cindex target vector
4432 The target vector defines the interface between @value{GDBN}'s
4433 abstract handling of target systems, and the nitty-gritty code that
4434 actually exercises control over a process or a serial port.
4435 @value{GDBN} includes some 30-40 different target vectors; however,
4436 each configuration of @value{GDBN} includes only a few of them.
4438 @section File Targets
4440 Both executables and core files have target vectors.
4442 @section Standard Protocol and Remote Stubs
4444 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4445 that runs in the target system. @value{GDBN} provides several sample
4446 @dfn{stubs} that can be integrated into target programs or operating
4447 systems for this purpose; they are named @file{*-stub.c}.
4449 The @value{GDBN} user's manual describes how to put such a stub into
4450 your target code. What follows is a discussion of integrating the
4451 SPARC stub into a complicated operating system (rather than a simple
4452 program), by Stu Grossman, the author of this stub.
4454 The trap handling code in the stub assumes the following upon entry to
4459 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4465 you are in the correct trap window.
4468 As long as your trap handler can guarantee those conditions, then there
4469 is no reason why you shouldn't be able to ``share'' traps with the stub.
4470 The stub has no requirement that it be jumped to directly from the
4471 hardware trap vector. That is why it calls @code{exceptionHandler()},
4472 which is provided by the external environment. For instance, this could
4473 set up the hardware traps to actually execute code which calls the stub
4474 first, and then transfers to its own trap handler.
4476 For the most point, there probably won't be much of an issue with
4477 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4478 and often indicate unrecoverable error conditions. Anyway, this is all
4479 controlled by a table, and is trivial to modify. The most important
4480 trap for us is for @code{ta 1}. Without that, we can't single step or
4481 do breakpoints. Everything else is unnecessary for the proper operation
4482 of the debugger/stub.
4484 From reading the stub, it's probably not obvious how breakpoints work.
4485 They are simply done by deposit/examine operations from @value{GDBN}.
4487 @section ROM Monitor Interface
4489 @section Custom Protocols
4491 @section Transport Layer
4493 @section Builtin Simulator
4496 @node Native Debugging
4498 @chapter Native Debugging
4499 @cindex native debugging
4501 Several files control @value{GDBN}'s configuration for native support:
4505 @item gdb/config/@var{arch}/@var{xyz}.mh
4506 Specifies Makefile fragments needed by a @emph{native} configuration on
4507 machine @var{xyz}. In particular, this lists the required
4508 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4509 Also specifies the header file which describes native support on
4510 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4511 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4512 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4514 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4515 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4516 on machine @var{xyz}. While the file is no longer used for this
4517 purpose, the @file{.mh} suffix remains. Perhaps someone will
4518 eventually rename these fragments so that they have a @file{.mn}
4521 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4522 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4523 macro definitions describing the native system environment, such as
4524 child process control and core file support.
4526 @item gdb/@var{xyz}-nat.c
4527 Contains any miscellaneous C code required for this native support of
4528 this machine. On some machines it doesn't exist at all.
4531 There are some ``generic'' versions of routines that can be used by
4532 various systems. These can be customized in various ways by macros
4533 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4534 the @var{xyz} host, you can just include the generic file's name (with
4535 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4537 Otherwise, if your machine needs custom support routines, you will need
4538 to write routines that perform the same functions as the generic file.
4539 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4540 into @code{NATDEPFILES}.
4544 This contains the @emph{target_ops vector} that supports Unix child
4545 processes on systems which use ptrace and wait to control the child.
4548 This contains the @emph{target_ops vector} that supports Unix child
4549 processes on systems which use /proc to control the child.
4552 This does the low-level grunge that uses Unix system calls to do a ``fork
4553 and exec'' to start up a child process.
4556 This is the low level interface to inferior processes for systems using
4557 the Unix @code{ptrace} call in a vanilla way.
4560 @section Native core file Support
4561 @cindex native core files
4564 @findex fetch_core_registers
4565 @item core-aout.c::fetch_core_registers()
4566 Support for reading registers out of a core file. This routine calls
4567 @code{register_addr()}, see below. Now that BFD is used to read core
4568 files, virtually all machines should use @code{core-aout.c}, and should
4569 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4570 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4572 @item core-aout.c::register_addr()
4573 If your @code{nm-@var{xyz}.h} file defines the macro
4574 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4575 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4576 register number @code{regno}. @code{blockend} is the offset within the
4577 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4578 @file{core-aout.c} will define the @code{register_addr()} function and
4579 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4580 you are using the standard @code{fetch_core_registers()}, you will need
4581 to define your own version of @code{register_addr()}, put it into your
4582 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4583 the @code{NATDEPFILES} list. If you have your own
4584 @code{fetch_core_registers()}, you may not need a separate
4585 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4586 implementations simply locate the registers themselves.@refill
4589 When making @value{GDBN} run native on a new operating system, to make it
4590 possible to debug core files, you will need to either write specific
4591 code for parsing your OS's core files, or customize
4592 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4593 machine uses to define the struct of registers that is accessible
4594 (possibly in the u-area) in a core file (rather than
4595 @file{machine/reg.h}), and an include file that defines whatever header
4596 exists on a core file (e.g., the u-area or a @code{struct core}). Then
4597 modify @code{trad_unix_core_file_p} to use these values to set up the
4598 section information for the data segment, stack segment, any other
4599 segments in the core file (perhaps shared library contents or control
4600 information), ``registers'' segment, and if there are two discontiguous
4601 sets of registers (e.g., integer and float), the ``reg2'' segment. This
4602 section information basically delimits areas in the core file in a
4603 standard way, which the section-reading routines in BFD know how to seek
4606 Then back in @value{GDBN}, you need a matching routine called
4607 @code{fetch_core_registers}. If you can use the generic one, it's in
4608 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4609 It will be passed a char pointer to the entire ``registers'' segment,
4610 its length, and a zero; or a char pointer to the entire ``regs2''
4611 segment, its length, and a 2. The routine should suck out the supplied
4612 register values and install them into @value{GDBN}'s ``registers'' array.
4614 If your system uses @file{/proc} to control processes, and uses ELF
4615 format core files, then you may be able to use the same routines for
4616 reading the registers out of processes and out of core files.
4624 @section shared libraries
4626 @section Native Conditionals
4627 @cindex native conditionals
4629 When @value{GDBN} is configured and compiled, various macros are
4630 defined or left undefined, to control compilation when the host and
4631 target systems are the same. These macros should be defined (or left
4632 undefined) in @file{nm-@var{system}.h}.
4636 @item CHILD_PREPARE_TO_STORE
4637 @findex CHILD_PREPARE_TO_STORE
4638 If the machine stores all registers at once in the child process, then
4639 define this to ensure that all values are correct. This usually entails
4640 a read from the child.
4642 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4645 @item FETCH_INFERIOR_REGISTERS
4646 @findex FETCH_INFERIOR_REGISTERS
4647 Define this if the native-dependent code will provide its own routines
4648 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4649 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4650 @file{infptrace.c} is included in this configuration, the default
4651 routines in @file{infptrace.c} are used for these functions.
4655 This macro is normally defined to be the number of the first floating
4656 point register, if the machine has such registers. As such, it would
4657 appear only in target-specific code. However, @file{/proc} support uses this
4658 to decide whether floats are in use on this target.
4660 @item GET_LONGJMP_TARGET
4661 @findex GET_LONGJMP_TARGET
4662 For most machines, this is a target-dependent parameter. On the
4663 DECstation and the Iris, this is a native-dependent parameter, since
4664 @file{setjmp.h} is needed to define it.
4666 This macro determines the target PC address that @code{longjmp} will jump to,
4667 assuming that we have just stopped at a longjmp breakpoint. It takes a
4668 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4669 pointer. It examines the current state of the machine as needed.
4671 @item I386_USE_GENERIC_WATCHPOINTS
4672 An x86-based machine can define this to use the generic x86 watchpoint
4673 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4676 @findex KERNEL_U_ADDR
4677 Define this to the address of the @code{u} structure (the ``user
4678 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
4679 needs to know this so that it can subtract this address from absolute
4680 addresses in the upage, that are obtained via ptrace or from core files.
4681 On systems that don't need this value, set it to zero.
4683 @item KERNEL_U_ADDR_HPUX
4684 @findex KERNEL_U_ADDR_HPUX
4685 Define this to cause @value{GDBN} to determine the address of @code{u} at
4686 runtime, by using HP-style @code{nlist} on the kernel's image in the
4689 @item ONE_PROCESS_WRITETEXT
4690 @findex ONE_PROCESS_WRITETEXT
4691 Define this to be able to, when a breakpoint insertion fails, warn the
4692 user that another process may be running with the same executable.
4695 @findex PROC_NAME_FMT
4696 Defines the format for the name of a @file{/proc} device. Should be
4697 defined in @file{nm.h} @emph{only} in order to override the default
4698 definition in @file{procfs.c}.
4700 @item PTRACE_ARG3_TYPE
4701 @findex PTRACE_ARG3_TYPE
4702 The type of the third argument to the @code{ptrace} system call, if it
4703 exists and is different from @code{int}.
4705 @item REGISTER_U_ADDR
4706 @findex REGISTER_U_ADDR
4707 Defines the offset of the registers in the ``u area''.
4709 @item SHELL_COMMAND_CONCAT
4710 @findex SHELL_COMMAND_CONCAT
4711 If defined, is a string to prefix on the shell command used to start the
4716 If defined, this is the name of the shell to use to run the inferior.
4717 Defaults to @code{"/bin/sh"}.
4719 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4721 Define this to expand into an expression that will cause the symbols in
4722 @var{filename} to be added to @value{GDBN}'s symbol table. If
4723 @var{readsyms} is zero symbols are not read but any necessary low level
4724 processing for @var{filename} is still done.
4726 @item SOLIB_CREATE_INFERIOR_HOOK
4727 @findex SOLIB_CREATE_INFERIOR_HOOK
4728 Define this to expand into any shared-library-relocation code that you
4729 want to be run just after the child process has been forked.
4731 @item START_INFERIOR_TRAPS_EXPECTED
4732 @findex START_INFERIOR_TRAPS_EXPECTED
4733 When starting an inferior, @value{GDBN} normally expects to trap
4735 the shell execs, and once when the program itself execs. If the actual
4736 number of traps is something other than 2, then define this macro to
4737 expand into the number expected.
4741 This determines whether small routines in @file{*-tdep.c}, which
4742 translate register values between @value{GDBN}'s internal
4743 representation and the @file{/proc} representation, are compiled.
4746 @findex U_REGS_OFFSET
4747 This is the offset of the registers in the upage. It need only be
4748 defined if the generic ptrace register access routines in
4749 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4750 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4751 the default value from @file{infptrace.c} is good enough, leave it
4754 The default value means that u.u_ar0 @emph{points to} the location of
4755 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4756 that @code{u.u_ar0} @emph{is} the location of the registers.
4760 See @file{objfiles.c}.
4763 @findex DEBUG_PTRACE
4764 Define this to debug @code{ptrace} calls.
4768 @node Support Libraries
4770 @chapter Support Libraries
4775 BFD provides support for @value{GDBN} in several ways:
4778 @item identifying executable and core files
4779 BFD will identify a variety of file types, including a.out, coff, and
4780 several variants thereof, as well as several kinds of core files.
4782 @item access to sections of files
4783 BFD parses the file headers to determine the names, virtual addresses,
4784 sizes, and file locations of all the various named sections in files
4785 (such as the text section or the data section). @value{GDBN} simply
4786 calls BFD to read or write section @var{x} at byte offset @var{y} for
4789 @item specialized core file support
4790 BFD provides routines to determine the failing command name stored in a
4791 core file, the signal with which the program failed, and whether a core
4792 file matches (i.e.@: could be a core dump of) a particular executable
4795 @item locating the symbol information
4796 @value{GDBN} uses an internal interface of BFD to determine where to find the
4797 symbol information in an executable file or symbol-file. @value{GDBN} itself
4798 handles the reading of symbols, since BFD does not ``understand'' debug
4799 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4804 @cindex opcodes library
4806 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4807 library because it's also used in binutils, for @file{objdump}).
4814 @cindex @code{libiberty} library
4816 The @code{libiberty} library provides a set of functions and features
4817 that integrate and improve on functionality found in modern operating
4818 systems. Broadly speaking, such features can be divided into three
4819 groups: supplemental functions (functions that may be missing in some
4820 environments and operating systems), replacement functions (providing
4821 a uniform and easier to use interface for commonly used standard
4822 functions), and extensions (which provide additional functionality
4823 beyond standard functions).
4825 @value{GDBN} uses various features provided by the @code{libiberty}
4826 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4827 floating format support functions, the input options parser
4828 @samp{getopt}, the @samp{obstack} extension, and other functions.
4830 @subsection @code{obstacks} in @value{GDBN}
4831 @cindex @code{obstacks}
4833 The obstack mechanism provides a convenient way to allocate and free
4834 chunks of memory. Each obstack is a pool of memory that is managed
4835 like a stack. Objects (of any nature, size and alignment) are
4836 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4837 @code{libiberty}'s documenatation for a more detailed explanation of
4840 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4841 object files. There is an obstack associated with each internal
4842 representation of an object file. Lots of things get allocated on
4843 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4844 symbols, minimal symbols, types, vectors of fundamental types, class
4845 fields of types, object files section lists, object files section
4846 offets lists, line tables, symbol tables, partial symbol tables,
4847 string tables, symbol table private data, macros tables, debug
4848 information sections and entries, import and export lists (som),
4849 unwind information (hppa), dwarf2 location expressions data. Plus
4850 various strings such as directory names strings, debug format strings,
4853 An essential and convenient property of all data on @code{obstacks} is
4854 that memory for it gets allocated (with @code{obstack_alloc}) at
4855 various times during a debugging sesssion, but it is released all at
4856 once using the @code{obstack_free} function. The @code{obstack_free}
4857 function takes a pointer to where in the stack it must start the
4858 deletion from (much like the cleanup chains have a pointer to where to
4859 start the cleanups). Because of the stack like structure of the
4860 @code{obstacks}, this allows to free only a top portion of the
4861 obstack. There are a few instances in @value{GDBN} where such thing
4862 happens. Calls to @code{obstack_free} are done after some local data
4863 is allocated to the obstack. Only the local data is deleted from the
4864 obstack. Of course this assumes that nothing between the
4865 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4866 else on the same obstack. For this reason it is best and safest to
4867 use temporary @code{obstacks}.
4869 Releasing the whole obstack is also not safe per se. It is safe only
4870 under the condition that we know the @code{obstacks} memory is no
4871 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4872 when we get rid of the whole objfile(s), for instance upon reading a
4876 @cindex regular expressions library
4887 @item SIGN_EXTEND_CHAR
4889 @item SWITCH_ENUM_BUG
4904 This chapter covers topics that are lower-level than the major
4905 algorithms of @value{GDBN}.
4910 Cleanups are a structured way to deal with things that need to be done
4913 When your code does something (e.g., @code{xmalloc} some memory, or
4914 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4915 the memory or @code{close} the file), it can make a cleanup. The
4916 cleanup will be done at some future point: when the command is finished
4917 and control returns to the top level; when an error occurs and the stack
4918 is unwound; or when your code decides it's time to explicitly perform
4919 cleanups. Alternatively you can elect to discard the cleanups you
4925 @item struct cleanup *@var{old_chain};
4926 Declare a variable which will hold a cleanup chain handle.
4928 @findex make_cleanup
4929 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
4930 Make a cleanup which will cause @var{function} to be called with
4931 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
4932 handle that can later be passed to @code{do_cleanups} or
4933 @code{discard_cleanups}. Unless you are going to call
4934 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
4935 from @code{make_cleanup}.
4938 @item do_cleanups (@var{old_chain});
4939 Do all cleanups added to the chain since the corresponding
4940 @code{make_cleanup} call was made.
4942 @findex discard_cleanups
4943 @item discard_cleanups (@var{old_chain});
4944 Same as @code{do_cleanups} except that it just removes the cleanups from
4945 the chain and does not call the specified functions.
4948 Cleanups are implemented as a chain. The handle returned by
4949 @code{make_cleanups} includes the cleanup passed to the call and any
4950 later cleanups appended to the chain (but not yet discarded or
4954 make_cleanup (a, 0);
4956 struct cleanup *old = make_cleanup (b, 0);
4964 will call @code{c()} and @code{b()} but will not call @code{a()}. The
4965 cleanup that calls @code{a()} will remain in the cleanup chain, and will
4966 be done later unless otherwise discarded.@refill
4968 Your function should explicitly do or discard the cleanups it creates.
4969 Failing to do this leads to non-deterministic behavior since the caller
4970 will arbitrarily do or discard your functions cleanups. This need leads
4971 to two common cleanup styles.
4973 The first style is try/finally. Before it exits, your code-block calls
4974 @code{do_cleanups} with the old cleanup chain and thus ensures that your
4975 code-block's cleanups are always performed. For instance, the following
4976 code-segment avoids a memory leak problem (even when @code{error} is
4977 called and a forced stack unwind occurs) by ensuring that the
4978 @code{xfree} will always be called:
4981 struct cleanup *old = make_cleanup (null_cleanup, 0);
4982 data = xmalloc (sizeof blah);
4983 make_cleanup (xfree, data);
4988 The second style is try/except. Before it exits, your code-block calls
4989 @code{discard_cleanups} with the old cleanup chain and thus ensures that
4990 any created cleanups are not performed. For instance, the following
4991 code segment, ensures that the file will be closed but only if there is
4995 FILE *file = fopen ("afile", "r");
4996 struct cleanup *old = make_cleanup (close_file, file);
4998 discard_cleanups (old);
5002 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
5003 that they ``should not be called when cleanups are not in place''. This
5004 means that any actions you need to reverse in the case of an error or
5005 interruption must be on the cleanup chain before you call these
5006 functions, since they might never return to your code (they
5007 @samp{longjmp} instead).
5009 @section Per-architecture module data
5010 @cindex per-architecture module data
5011 @cindex multi-arch data
5012 @cindex data-pointer, per-architecture/per-module
5014 The multi-arch framework includes a mechanism for adding module
5015 specific per-architecture data-pointers to the @code{struct gdbarch}
5016 architecture object.
5018 A module registers one or more per-architecture data-pointers using:
5020 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
5021 @var{pre_init} is used to, on-demand, allocate an initial value for a
5022 per-architecture data-pointer using the architecture's obstack (passed
5023 in as a parameter). Since @var{pre_init} can be called during
5024 architecture creation, it is not parameterized with the architecture.
5025 and must not call modules that use per-architecture data.
5028 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
5029 @var{post_init} is used to obtain an initial value for a
5030 per-architecture data-pointer @emph{after}. Since @var{post_init} is
5031 always called after architecture creation, it both receives the fully
5032 initialized architecture and is free to call modules that use
5033 per-architecture data (care needs to be taken to ensure that those
5034 other modules do not try to call back to this module as that will
5035 create in cycles in the initialization call graph).
5038 These functions return a @code{struct gdbarch_data} that is used to
5039 identify the per-architecture data-pointer added for that module.
5041 The per-architecture data-pointer is accessed using the function:
5043 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
5044 Given the architecture @var{arch} and module data handle
5045 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
5046 or @code{gdbarch_data_register_post_init}), this function returns the
5047 current value of the per-architecture data-pointer. If the data
5048 pointer is @code{NULL}, it is first initialized by calling the
5049 corresponding @var{pre_init} or @var{post_init} method.
5052 The examples below assume the following definitions:
5055 struct nozel @{ int total; @};
5056 static struct gdbarch_data *nozel_handle;
5059 A module can extend the architecture vector, adding additional
5060 per-architecture data, using the @var{pre_init} method. The module's
5061 per-architecture data is then initialized during architecture
5064 In the below, the module's per-architecture @emph{nozel} is added. An
5065 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
5066 from @code{gdbarch_init}.
5070 nozel_pre_init (struct obstack *obstack)
5072 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
5079 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
5081 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5082 data->total = nozel;
5086 A module can on-demand create architecture dependant data structures
5087 using @code{post_init}.
5089 In the below, the nozel's total is computed on-demand by
5090 @code{nozel_post_init} using information obtained from the
5095 nozel_post_init (struct gdbarch *gdbarch)
5097 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
5098 nozel->total = gdbarch@dots{} (gdbarch);
5105 nozel_total (struct gdbarch *gdbarch)
5107 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5112 @section Wrapping Output Lines
5113 @cindex line wrap in output
5116 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
5117 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
5118 added in places that would be good breaking points. The utility
5119 routines will take care of actually wrapping if the line width is
5122 The argument to @code{wrap_here} is an indentation string which is
5123 printed @emph{only} if the line breaks there. This argument is saved
5124 away and used later. It must remain valid until the next call to
5125 @code{wrap_here} or until a newline has been printed through the
5126 @code{*_filtered} functions. Don't pass in a local variable and then
5129 It is usually best to call @code{wrap_here} after printing a comma or
5130 space. If you call it before printing a space, make sure that your
5131 indentation properly accounts for the leading space that will print if
5132 the line wraps there.
5134 Any function or set of functions that produce filtered output must
5135 finish by printing a newline, to flush the wrap buffer, before switching
5136 to unfiltered (@code{printf}) output. Symbol reading routines that
5137 print warnings are a good example.
5139 @section @value{GDBN} Coding Standards
5140 @cindex coding standards
5142 @value{GDBN} follows the GNU coding standards, as described in
5143 @file{etc/standards.texi}. This file is also available for anonymous
5144 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
5145 of the standard; in general, when the GNU standard recommends a practice
5146 but does not require it, @value{GDBN} requires it.
5148 @value{GDBN} follows an additional set of coding standards specific to
5149 @value{GDBN}, as described in the following sections.
5154 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
5157 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
5160 @subsection Memory Management
5162 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
5163 @code{calloc}, @code{free} and @code{asprintf}.
5165 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
5166 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
5167 these functions do not return when the memory pool is empty. Instead,
5168 they unwind the stack using cleanups. These functions return
5169 @code{NULL} when requested to allocate a chunk of memory of size zero.
5171 @emph{Pragmatics: By using these functions, the need to check every
5172 memory allocation is removed. These functions provide portable
5175 @value{GDBN} does not use the function @code{free}.
5177 @value{GDBN} uses the function @code{xfree} to return memory to the
5178 memory pool. Consistent with ISO-C, this function ignores a request to
5179 free a @code{NULL} pointer.
5181 @emph{Pragmatics: On some systems @code{free} fails when passed a
5182 @code{NULL} pointer.}
5184 @value{GDBN} can use the non-portable function @code{alloca} for the
5185 allocation of small temporary values (such as strings).
5187 @emph{Pragmatics: This function is very non-portable. Some systems
5188 restrict the memory being allocated to no more than a few kilobytes.}
5190 @value{GDBN} uses the string function @code{xstrdup} and the print
5191 function @code{xstrprintf}.
5193 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
5194 functions such as @code{sprintf} are very prone to buffer overflow
5198 @subsection Compiler Warnings
5199 @cindex compiler warnings
5201 With few exceptions, developers should include the configuration option
5202 @samp{--enable-gdb-build-warnings=,-Werror} when building @value{GDBN}.
5203 The exceptions are listed in the file @file{gdb/MAINTAINERS}.
5205 This option causes @value{GDBN} (when built using GCC) to be compiled
5206 with a carefully selected list of compiler warning flags. Any warnings
5207 from those flags being treated as errors.
5209 The current list of warning flags includes:
5213 Since @value{GDBN} coding standard requires all functions to be declared
5214 using a prototype, the flag has the side effect of ensuring that
5215 prototyped functions are always visible with out resorting to
5216 @samp{-Wstrict-prototypes}.
5219 Such code often appears to work except on instruction set architectures
5220 that use register windows.
5227 @itemx -Wformat-nonliteral
5228 Since @value{GDBN} uses the @code{format printf} attribute on all
5229 @code{printf} like functions these check not just @code{printf} calls
5230 but also calls to functions such as @code{fprintf_unfiltered}.
5233 This warning includes uses of the assignment operator within an
5234 @code{if} statement.
5236 @item -Wpointer-arith
5238 @item -Wuninitialized
5240 @item -Wunused-label
5241 This warning has the additional benefit of detecting the absence of the
5242 @code{case} reserved word in a switch statement:
5244 enum @{ FD_SCHEDULED, NOTHING_SCHEDULED @} sched;
5257 @item -Wunused-function
5259 @item -Wno-pointer-sign
5260 In version 4.0, GCC began warning about pointer argument passing or
5261 assignment even when the source and destination differed only in
5262 signedness. However, most @value{GDBN} code doesn't distinguish
5263 carefully between @code{char} and @code{unsigned char}. In early 2006
5264 the @value{GDBN} developers decided correcting these warnings wasn't
5265 worth the time it would take.
5269 @emph{Pragmatics: Due to the way that @value{GDBN} is implemented most
5270 functions have unused parameters. Consequently the warning
5271 @samp{-Wunused-parameter} is precluded from the list. The macro
5272 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
5273 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
5274 is being used. The options @samp{-Wall} and @samp{-Wunused} are also
5275 precluded because they both include @samp{-Wunused-parameter}.}
5277 @emph{Pragmatics: @value{GDBN} has not simply accepted the warnings
5278 enabled by @samp{-Wall -Werror -W...}. Instead it is selecting warnings
5279 when and where their benefits can be demonstrated.}
5281 @subsection Formatting
5283 @cindex source code formatting
5284 The standard GNU recommendations for formatting must be followed
5287 A function declaration should not have its name in column zero. A
5288 function definition should have its name in column zero.
5292 static void foo (void);
5300 @emph{Pragmatics: This simplifies scripting. Function definitions can
5301 be found using @samp{^function-name}.}
5303 There must be a space between a function or macro name and the opening
5304 parenthesis of its argument list (except for macro definitions, as
5305 required by C). There must not be a space after an open paren/bracket
5306 or before a close paren/bracket.
5308 While additional whitespace is generally helpful for reading, do not use
5309 more than one blank line to separate blocks, and avoid adding whitespace
5310 after the end of a program line (as of 1/99, some 600 lines had
5311 whitespace after the semicolon). Excess whitespace causes difficulties
5312 for @code{diff} and @code{patch} utilities.
5314 Pointers are declared using the traditional K&R C style:
5328 @subsection Comments
5330 @cindex comment formatting
5331 The standard GNU requirements on comments must be followed strictly.
5333 Block comments must appear in the following form, with no @code{/*}- or
5334 @code{*/}-only lines, and no leading @code{*}:
5337 /* Wait for control to return from inferior to debugger. If inferior
5338 gets a signal, we may decide to start it up again instead of
5339 returning. That is why there is a loop in this function. When
5340 this function actually returns it means the inferior should be left
5341 stopped and @value{GDBN} should read more commands. */
5344 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5345 comment works correctly, and @kbd{M-q} fills the block consistently.)
5347 Put a blank line between the block comments preceding function or
5348 variable definitions, and the definition itself.
5350 In general, put function-body comments on lines by themselves, rather
5351 than trying to fit them into the 20 characters left at the end of a
5352 line, since either the comment or the code will inevitably get longer
5353 than will fit, and then somebody will have to move it anyhow.
5357 @cindex C data types
5358 Code must not depend on the sizes of C data types, the format of the
5359 host's floating point numbers, the alignment of anything, or the order
5360 of evaluation of expressions.
5362 @cindex function usage
5363 Use functions freely. There are only a handful of compute-bound areas
5364 in @value{GDBN} that might be affected by the overhead of a function
5365 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5366 limited by the target interface (whether serial line or system call).
5368 However, use functions with moderation. A thousand one-line functions
5369 are just as hard to understand as a single thousand-line function.
5371 @emph{Macros are bad, M'kay.}
5372 (But if you have to use a macro, make sure that the macro arguments are
5373 protected with parentheses.)
5377 Declarations like @samp{struct foo *} should be used in preference to
5378 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5381 @subsection Function Prototypes
5382 @cindex function prototypes
5384 Prototypes must be used when both @emph{declaring} and @emph{defining}
5385 a function. Prototypes for @value{GDBN} functions must include both the
5386 argument type and name, with the name matching that used in the actual
5387 function definition.
5389 All external functions should have a declaration in a header file that
5390 callers include, except for @code{_initialize_*} functions, which must
5391 be external so that @file{init.c} construction works, but shouldn't be
5392 visible to random source files.
5394 Where a source file needs a forward declaration of a static function,
5395 that declaration must appear in a block near the top of the source file.
5398 @subsection Internal Error Recovery
5400 During its execution, @value{GDBN} can encounter two types of errors.
5401 User errors and internal errors. User errors include not only a user
5402 entering an incorrect command but also problems arising from corrupt
5403 object files and system errors when interacting with the target.
5404 Internal errors include situations where @value{GDBN} has detected, at
5405 run time, a corrupt or erroneous situation.
5407 When reporting an internal error, @value{GDBN} uses
5408 @code{internal_error} and @code{gdb_assert}.
5410 @value{GDBN} must not call @code{abort} or @code{assert}.
5412 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5413 the code detected a user error, recovered from it and issued a
5414 @code{warning} or the code failed to correctly recover from the user
5415 error and issued an @code{internal_error}.}
5417 @subsection File Names
5419 Any file used when building the core of @value{GDBN} must be in lower
5420 case. Any file used when building the core of @value{GDBN} must be 8.3
5421 unique. These requirements apply to both source and generated files.
5423 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5424 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5425 is introduced to the build process both @file{Makefile.in} and
5426 @file{configure.in} need to be modified accordingly. Compare the
5427 convoluted conversion process needed to transform @file{COPYING} into
5428 @file{copying.c} with the conversion needed to transform
5429 @file{version.in} into @file{version.c}.}
5431 Any file non 8.3 compliant file (that is not used when building the core
5432 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5434 @emph{Pragmatics: This is clearly a compromise.}
5436 When @value{GDBN} has a local version of a system header file (ex
5437 @file{string.h}) the file name based on the POSIX header prefixed with
5438 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5439 independent: they should use only macros defined by @file{configure},
5440 the compiler, or the host; they should include only system headers; they
5441 should refer only to system types. They may be shared between multiple
5442 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5444 For other files @samp{-} is used as the separator.
5447 @subsection Include Files
5449 A @file{.c} file should include @file{defs.h} first.
5451 A @file{.c} file should directly include the @code{.h} file of every
5452 declaration and/or definition it directly refers to. It cannot rely on
5455 A @file{.h} file should directly include the @code{.h} file of every
5456 declaration and/or definition it directly refers to. It cannot rely on
5457 indirect inclusion. Exception: The file @file{defs.h} does not need to
5458 be directly included.
5460 An external declaration should only appear in one include file.
5462 An external declaration should never appear in a @code{.c} file.
5463 Exception: a declaration for the @code{_initialize} function that
5464 pacifies @option{-Wmissing-declaration}.
5466 A @code{typedef} definition should only appear in one include file.
5468 An opaque @code{struct} declaration can appear in multiple @file{.h}
5469 files. Where possible, a @file{.h} file should use an opaque
5470 @code{struct} declaration instead of an include.
5472 All @file{.h} files should be wrapped in:
5475 #ifndef INCLUDE_FILE_NAME_H
5476 #define INCLUDE_FILE_NAME_H
5482 @subsection Clean Design and Portable Implementation
5485 In addition to getting the syntax right, there's the little question of
5486 semantics. Some things are done in certain ways in @value{GDBN} because long
5487 experience has shown that the more obvious ways caused various kinds of
5490 @cindex assumptions about targets
5491 You can't assume the byte order of anything that comes from a target
5492 (including @var{value}s, object files, and instructions). Such things
5493 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5494 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5495 such as @code{bfd_get_32}.
5497 You can't assume that you know what interface is being used to talk to
5498 the target system. All references to the target must go through the
5499 current @code{target_ops} vector.
5501 You can't assume that the host and target machines are the same machine
5502 (except in the ``native'' support modules). In particular, you can't
5503 assume that the target machine's header files will be available on the
5504 host machine. Target code must bring along its own header files --
5505 written from scratch or explicitly donated by their owner, to avoid
5509 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5510 to write the code portably than to conditionalize it for various
5513 @cindex system dependencies
5514 New @code{#ifdef}'s which test for specific compilers or manufacturers
5515 or operating systems are unacceptable. All @code{#ifdef}'s should test
5516 for features. The information about which configurations contain which
5517 features should be segregated into the configuration files. Experience
5518 has proven far too often that a feature unique to one particular system
5519 often creeps into other systems; and that a conditional based on some
5520 predefined macro for your current system will become worthless over
5521 time, as new versions of your system come out that behave differently
5522 with regard to this feature.
5524 Adding code that handles specific architectures, operating systems,
5525 target interfaces, or hosts, is not acceptable in generic code.
5527 @cindex portable file name handling
5528 @cindex file names, portability
5529 One particularly notorious area where system dependencies tend to
5530 creep in is handling of file names. The mainline @value{GDBN} code
5531 assumes Posix semantics of file names: absolute file names begin with
5532 a forward slash @file{/}, slashes are used to separate leading
5533 directories, case-sensitive file names. These assumptions are not
5534 necessarily true on non-Posix systems such as MS-Windows. To avoid
5535 system-dependent code where you need to take apart or construct a file
5536 name, use the following portable macros:
5539 @findex HAVE_DOS_BASED_FILE_SYSTEM
5540 @item HAVE_DOS_BASED_FILE_SYSTEM
5541 This preprocessing symbol is defined to a non-zero value on hosts
5542 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5543 symbol to write conditional code which should only be compiled for
5546 @findex IS_DIR_SEPARATOR
5547 @item IS_DIR_SEPARATOR (@var{c})
5548 Evaluates to a non-zero value if @var{c} is a directory separator
5549 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5550 such a character, but on Windows, both @file{/} and @file{\} will
5553 @findex IS_ABSOLUTE_PATH
5554 @item IS_ABSOLUTE_PATH (@var{file})
5555 Evaluates to a non-zero value if @var{file} is an absolute file name.
5556 For Unix and GNU/Linux hosts, a name which begins with a slash
5557 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5558 @file{x:\bar} are also absolute file names.
5560 @findex FILENAME_CMP
5561 @item FILENAME_CMP (@var{f1}, @var{f2})
5562 Calls a function which compares file names @var{f1} and @var{f2} as
5563 appropriate for the underlying host filesystem. For Posix systems,
5564 this simply calls @code{strcmp}; on case-insensitive filesystems it
5565 will call @code{strcasecmp} instead.
5567 @findex DIRNAME_SEPARATOR
5568 @item DIRNAME_SEPARATOR
5569 Evaluates to a character which separates directories in
5570 @code{PATH}-style lists, typically held in environment variables.
5571 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5573 @findex SLASH_STRING
5575 This evaluates to a constant string you should use to produce an
5576 absolute filename from leading directories and the file's basename.
5577 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5578 @code{"\\"} for some Windows-based ports.
5581 In addition to using these macros, be sure to use portable library
5582 functions whenever possible. For example, to extract a directory or a
5583 basename part from a file name, use the @code{dirname} and
5584 @code{basename} library functions (available in @code{libiberty} for
5585 platforms which don't provide them), instead of searching for a slash
5586 with @code{strrchr}.
5588 Another way to generalize @value{GDBN} along a particular interface is with an
5589 attribute struct. For example, @value{GDBN} has been generalized to handle
5590 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5591 by defining the @code{target_ops} structure and having a current target (as
5592 well as a stack of targets below it, for memory references). Whenever
5593 something needs to be done that depends on which remote interface we are
5594 using, a flag in the current target_ops structure is tested (e.g.,
5595 @code{target_has_stack}), or a function is called through a pointer in the
5596 current target_ops structure. In this way, when a new remote interface
5597 is added, only one module needs to be touched---the one that actually
5598 implements the new remote interface. Other examples of
5599 attribute-structs are BFD access to multiple kinds of object file
5600 formats, or @value{GDBN}'s access to multiple source languages.
5602 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5603 the code interfacing between @code{ptrace} and the rest of
5604 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5605 something was very painful. In @value{GDBN} 4.x, these have all been
5606 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5607 with variations between systems the same way any system-independent
5608 file would (hooks, @code{#if defined}, etc.), and machines which are
5609 radically different don't need to use @file{infptrace.c} at all.
5611 All debugging code must be controllable using the @samp{set debug
5612 @var{module}} command. Do not use @code{printf} to print trace
5613 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5614 @code{#ifdef DEBUG}.
5619 @chapter Porting @value{GDBN}
5620 @cindex porting to new machines
5622 Most of the work in making @value{GDBN} compile on a new machine is in
5623 specifying the configuration of the machine. This is done in a
5624 dizzying variety of header files and configuration scripts, which we
5625 hope to make more sensible soon. Let's say your new host is called an
5626 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5627 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5628 @samp{sparc-sun-sunos4}). In particular:
5632 In the top level directory, edit @file{config.sub} and add @var{arch},
5633 @var{xvend}, and @var{xos} to the lists of supported architectures,
5634 vendors, and operating systems near the bottom of the file. Also, add
5635 @var{xyz} as an alias that maps to
5636 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5640 ./config.sub @var{xyz}
5647 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5651 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5652 and no error messages.
5655 You need to port BFD, if that hasn't been done already. Porting BFD is
5656 beyond the scope of this manual.
5659 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5660 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5661 desired target is already available) also edit @file{gdb/configure.tgt},
5662 setting @code{gdb_target} to something appropriate (for instance,
5665 @emph{Maintainer's note: Work in progress. The file
5666 @file{gdb/configure.host} originally needed to be modified when either a
5667 new native target or a new host machine was being added to @value{GDBN}.
5668 Recent changes have removed this requirement. The file now only needs
5669 to be modified when adding a new native configuration. This will likely
5670 changed again in the future.}
5673 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5674 target-dependent @file{.h} and @file{.c} files used for your
5678 @node Versions and Branches
5679 @chapter Versions and Branches
5683 @value{GDBN}'s version is determined by the file
5684 @file{gdb/version.in} and takes one of the following forms:
5687 @item @var{major}.@var{minor}
5688 @itemx @var{major}.@var{minor}.@var{patchlevel}
5689 an official release (e.g., 6.2 or 6.2.1)
5690 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
5691 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
5692 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
5693 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
5694 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
5695 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
5696 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5697 a vendor specific release of @value{GDBN}, that while based on@*
5698 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
5699 may include additional changes
5702 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5703 numbers from the most recent release branch, with a @var{patchlevel}
5704 of 50. At the time each new release branch is created, the mainline's
5705 @var{major} and @var{minor} version numbers are updated.
5707 @value{GDBN}'s release branch is similar. When the branch is cut, the
5708 @var{patchlevel} is changed from 50 to 90. As draft releases are
5709 drawn from the branch, the @var{patchlevel} is incremented. Once the
5710 first release (@var{major}.@var{minor}) has been made, the
5711 @var{patchlevel} is set to 0 and updates have an incremented
5714 For snapshots, and @sc{cvs} check outs, it is also possible to
5715 identify the @sc{cvs} origin:
5718 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
5719 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
5720 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
5721 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
5722 drawn from a release branch prior to the release (e.g.,
5724 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
5725 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
5726 drawn from a release branch after the release (e.g., 6.2.0.20020308)
5729 If the previous @value{GDBN} version is 6.1 and the current version is
5730 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5731 here's an illustration of a typical sequence:
5738 +--------------------------.
5741 6.2.50.20020303-cvs 6.1.90 (draft #1)
5743 6.2.50.20020304-cvs 6.1.90.20020304-cvs
5745 6.2.50.20020305-cvs 6.1.91 (draft #2)
5747 6.2.50.20020306-cvs 6.1.91.20020306-cvs
5749 6.2.50.20020307-cvs 6.2 (release)
5751 6.2.50.20020308-cvs 6.2.0.20020308-cvs
5753 6.2.50.20020309-cvs 6.2.1 (update)
5755 6.2.50.20020310-cvs <branch closed>
5759 +--------------------------.
5762 6.3.50.20020312-cvs 6.2.90 (draft #1)
5766 @section Release Branches
5767 @cindex Release Branches
5769 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
5770 single release branch, and identifies that branch using the @sc{cvs}
5774 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
5775 gdb_@var{major}_@var{minor}-branch
5776 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
5779 @emph{Pragmatics: To help identify the date at which a branch or
5780 release is made, both the branchpoint and release tags include the
5781 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
5782 branch tag, denoting the head of the branch, does not need this.}
5784 @section Vendor Branches
5785 @cindex vendor branches
5787 To avoid version conflicts, vendors are expected to modify the file
5788 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5789 (an official @value{GDBN} release never uses alphabetic characters in
5790 its version identifer). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
5793 @section Experimental Branches
5794 @cindex experimental branches
5796 @subsection Guidelines
5798 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
5799 repository, for experimental development. Branches make it possible
5800 for developers to share preliminary work, and maintainers to examine
5801 significant new developments.
5803 The following are a set of guidelines for creating such branches:
5807 @item a branch has an owner
5808 The owner can set further policy for a branch, but may not change the
5809 ground rules. In particular, they can set a policy for commits (be it
5810 adding more reviewers or deciding who can commit).
5812 @item all commits are posted
5813 All changes committed to a branch shall also be posted to
5814 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
5815 mailing list}. While commentary on such changes are encouraged, people
5816 should remember that the changes only apply to a branch.
5818 @item all commits are covered by an assignment
5819 This ensures that all changes belong to the Free Software Foundation,
5820 and avoids the possibility that the branch may become contaminated.
5822 @item a branch is focused
5823 A focused branch has a single objective or goal, and does not contain
5824 unnecessary or irrelevant changes. Cleanups, where identified, being
5825 be pushed into the mainline as soon as possible.
5827 @item a branch tracks mainline
5828 This keeps the level of divergence under control. It also keeps the
5829 pressure on developers to push cleanups and other stuff into the
5832 @item a branch shall contain the entire @value{GDBN} module
5833 The @value{GDBN} module @code{gdb} should be specified when creating a
5834 branch (branches of individual files should be avoided). @xref{Tags}.
5836 @item a branch shall be branded using @file{version.in}
5837 The file @file{gdb/version.in} shall be modified so that it identifies
5838 the branch @var{owner} and branch @var{name}, e.g.,
5839 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
5846 To simplify the identification of @value{GDBN} branches, the following
5847 branch tagging convention is strongly recommended:
5851 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5852 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
5853 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
5854 date that the branch was created. A branch is created using the
5855 sequence: @anchor{experimental branch tags}
5857 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
5858 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
5859 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
5862 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5863 The tagged point, on the mainline, that was used when merging the branch
5864 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
5865 use a command sequence like:
5867 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
5869 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5870 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5873 Similar sequences can be used to just merge in changes since the last
5879 For further information on @sc{cvs}, see
5880 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
5882 @node Start of New Year Procedure
5883 @chapter Start of New Year Procedure
5884 @cindex new year procedure
5886 At the start of each new year, the following actions should be performed:
5890 Rotate the ChangeLog file
5892 The current @file{ChangeLog} file should be renamed into
5893 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
5894 A new @file{ChangeLog} file should be created, and its contents should
5895 contain a reference to the previous ChangeLog. The following should
5896 also be preserved at the end of the new ChangeLog, in order to provide
5897 the appropriate settings when editing this file with Emacs:
5903 version-control: never
5908 Update the copyright year in the startup message
5910 Update the copyright year in file @file{top.c}, function
5911 @code{print_gdb_version}.
5916 @chapter Releasing @value{GDBN}
5917 @cindex making a new release of gdb
5919 @section Branch Commit Policy
5921 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
5922 5.1 and 5.2 all used the below:
5926 The @file{gdb/MAINTAINERS} file still holds.
5928 Don't fix something on the branch unless/until it is also fixed in the
5929 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
5930 file is better than committing a hack.
5932 When considering a patch for the branch, suggested criteria include:
5933 Does it fix a build? Does it fix the sequence @kbd{break main; run}
5934 when debugging a static binary?
5936 The further a change is from the core of @value{GDBN}, the less likely
5937 the change will worry anyone (e.g., target specific code).
5939 Only post a proposal to change the core of @value{GDBN} after you've
5940 sent individual bribes to all the people listed in the
5941 @file{MAINTAINERS} file @t{;-)}
5944 @emph{Pragmatics: Provided updates are restricted to non-core
5945 functionality there is little chance that a broken change will be fatal.
5946 This means that changes such as adding a new architectures or (within
5947 reason) support for a new host are considered acceptable.}
5950 @section Obsoleting code
5952 Before anything else, poke the other developers (and around the source
5953 code) to see if there is anything that can be removed from @value{GDBN}
5954 (an old target, an unused file).
5956 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
5957 line. Doing this means that it is easy to identify something that has
5958 been obsoleted when greping through the sources.
5960 The process is done in stages --- this is mainly to ensure that the
5961 wider @value{GDBN} community has a reasonable opportunity to respond.
5962 Remember, everything on the Internet takes a week.
5966 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
5967 list} Creating a bug report to track the task's state, is also highly
5972 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
5973 Announcement mailing list}.
5977 Go through and edit all relevant files and lines so that they are
5978 prefixed with the word @code{OBSOLETE}.
5980 Wait until the next GDB version, containing this obsolete code, has been
5983 Remove the obsolete code.
5987 @emph{Maintainer note: While removing old code is regrettable it is
5988 hopefully better for @value{GDBN}'s long term development. Firstly it
5989 helps the developers by removing code that is either no longer relevant
5990 or simply wrong. Secondly since it removes any history associated with
5991 the file (effectively clearing the slate) the developer has a much freer
5992 hand when it comes to fixing broken files.}
5996 @section Before the Branch
5998 The most important objective at this stage is to find and fix simple
5999 changes that become a pain to track once the branch is created. For
6000 instance, configuration problems that stop @value{GDBN} from even
6001 building. If you can't get the problem fixed, document it in the
6002 @file{gdb/PROBLEMS} file.
6004 @subheading Prompt for @file{gdb/NEWS}
6006 People always forget. Send a post reminding them but also if you know
6007 something interesting happened add it yourself. The @code{schedule}
6008 script will mention this in its e-mail.
6010 @subheading Review @file{gdb/README}
6012 Grab one of the nightly snapshots and then walk through the
6013 @file{gdb/README} looking for anything that can be improved. The
6014 @code{schedule} script will mention this in its e-mail.
6016 @subheading Refresh any imported files.
6018 A number of files are taken from external repositories. They include:
6022 @file{texinfo/texinfo.tex}
6024 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
6027 @file{etc/standards.texi}, @file{etc/make-stds.texi}
6030 @subheading Check the ARI
6032 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
6033 (Awk Regression Index ;-) that checks for a number of errors and coding
6034 conventions. The checks include things like using @code{malloc} instead
6035 of @code{xmalloc} and file naming problems. There shouldn't be any
6038 @subsection Review the bug data base
6040 Close anything obviously fixed.
6042 @subsection Check all cross targets build
6044 The targets are listed in @file{gdb/MAINTAINERS}.
6047 @section Cut the Branch
6049 @subheading Create the branch
6054 $ V=`echo $v | sed 's/\./_/g'`
6055 $ D=`date -u +%Y-%m-%d`
6058 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6059 -D $D-gmt gdb_$V-$D-branchpoint insight+dejagnu
6060 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
6061 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight+dejagnu
6064 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6065 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight+dejagnu
6066 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6067 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight+dejagnu
6075 by using @kbd{-D YYYY-MM-DD-gmt} the branch is forced to an exact
6078 the trunk is first taged so that the branch point can easily be found
6080 Insight (which includes GDB) and dejagnu are all tagged at the same time
6082 @file{version.in} gets bumped to avoid version number conflicts
6084 the reading of @file{.cvsrc} is disabled using @file{-f}
6087 @subheading Update @file{version.in}
6092 $ V=`echo $v | sed 's/\./_/g'`
6096 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
6097 -r gdb_$V-branch src/gdb/version.in
6098 cvs -f -d :ext:sources.redhat.com:/cvs/src co
6099 -r gdb_5_2-branch src/gdb/version.in
6101 U src/gdb/version.in
6103 $ echo $u.90-0000-00-00-cvs > version.in
6105 5.1.90-0000-00-00-cvs
6106 $ cvs -f commit version.in
6111 @file{0000-00-00} is used as a date to pump prime the version.in update
6114 @file{.90} and the previous branch version are used as fairly arbitrary
6115 initial branch version number
6119 @subheading Update the web and news pages
6123 @subheading Tweak cron to track the new branch
6125 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
6126 This file needs to be updated so that:
6130 a daily timestamp is added to the file @file{version.in}
6132 the new branch is included in the snapshot process
6136 See the file @file{gdbadmin/cron/README} for how to install the updated
6139 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
6140 any changes. That file is copied to both the branch/ and current/
6141 snapshot directories.
6144 @subheading Update the NEWS and README files
6146 The @file{NEWS} file needs to be updated so that on the branch it refers
6147 to @emph{changes in the current release} while on the trunk it also
6148 refers to @emph{changes since the current release}.
6150 The @file{README} file needs to be updated so that it refers to the
6153 @subheading Post the branch info
6155 Send an announcement to the mailing lists:
6159 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6161 @email{gdb@@sources.redhat.com, GDB Discsussion mailing list} and
6162 @email{gdb-testers@@sources.redhat.com, GDB Discsussion mailing list}
6165 @emph{Pragmatics: The branch creation is sent to the announce list to
6166 ensure that people people not subscribed to the higher volume discussion
6169 The announcement should include:
6175 how to check out the branch using CVS
6177 the date/number of weeks until the release
6179 the branch commit policy
6183 @section Stabilize the branch
6185 Something goes here.
6187 @section Create a Release
6189 The process of creating and then making available a release is broken
6190 down into a number of stages. The first part addresses the technical
6191 process of creating a releasable tar ball. The later stages address the
6192 process of releasing that tar ball.
6194 When making a release candidate just the first section is needed.
6196 @subsection Create a release candidate
6198 The objective at this stage is to create a set of tar balls that can be
6199 made available as a formal release (or as a less formal release
6202 @subsubheading Freeze the branch
6204 Send out an e-mail notifying everyone that the branch is frozen to
6205 @email{gdb-patches@@sources.redhat.com}.
6207 @subsubheading Establish a few defaults.
6212 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
6214 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6218 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6220 /home/gdbadmin/bin/autoconf
6229 Check the @code{autoconf} version carefully. You want to be using the
6230 version taken from the @file{binutils} snapshot directory, which can be
6231 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
6232 unlikely that a system installed version of @code{autoconf} (e.g.,
6233 @file{/usr/bin/autoconf}) is correct.
6236 @subsubheading Check out the relevant modules:
6239 $ for m in gdb insight dejagnu
6241 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
6251 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
6252 any confusion between what is written here and what your local
6253 @code{cvs} really does.
6256 @subsubheading Update relevant files.
6262 Major releases get their comments added as part of the mainline. Minor
6263 releases should probably mention any significant bugs that were fixed.
6265 Don't forget to include the @file{ChangeLog} entry.
6268 $ emacs gdb/src/gdb/NEWS
6273 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
6274 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6279 You'll need to update:
6291 $ emacs gdb/src/gdb/README
6296 $ cp gdb/src/gdb/README insight/src/gdb/README
6297 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6300 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
6301 before the initial branch was cut so just a simple substitute is needed
6304 @emph{Maintainer note: Other projects generate @file{README} and
6305 @file{INSTALL} from the core documentation. This might be worth
6308 @item gdb/version.in
6311 $ echo $v > gdb/src/gdb/version.in
6312 $ cat gdb/src/gdb/version.in
6314 $ emacs gdb/src/gdb/version.in
6317 ... Bump to version ...
6319 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6320 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6323 @item dejagnu/src/dejagnu/configure.in
6325 Dejagnu is more complicated. The version number is a parameter to
6326 @code{AM_INIT_AUTOMAKE}. Tweak it to read something like gdb-5.1.91.
6328 Don't forget to re-generate @file{configure}.
6330 Don't forget to include a @file{ChangeLog} entry.
6333 $ emacs dejagnu/src/dejagnu/configure.in
6338 $ ( cd dejagnu/src/dejagnu && autoconf )
6343 @subsubheading Do the dirty work
6345 This is identical to the process used to create the daily snapshot.
6348 $ for m in gdb insight
6350 ( cd $m/src && gmake -f src-release $m.tar )
6352 $ ( m=dejagnu; cd $m/src && gmake -f src-release $m.tar.bz2 )
6355 If the top level source directory does not have @file{src-release}
6356 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6359 $ for m in gdb insight
6361 ( cd $m/src && gmake -f Makefile.in $m.tar )
6363 $ ( m=dejagnu; cd $m/src && gmake -f Makefile.in $m.tar.bz2 )
6366 @subsubheading Check the source files
6368 You're looking for files that have mysteriously disappeared.
6369 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6370 for the @file{version.in} update @kbd{cronjob}.
6373 $ ( cd gdb/src && cvs -f -q -n update )
6377 @dots{} lots of generated files @dots{}
6382 @dots{} lots of generated files @dots{}
6387 @emph{Don't worry about the @file{gdb.info-??} or
6388 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6389 was also generated only something strange with CVS means that they
6390 didn't get supressed). Fixing it would be nice though.}
6392 @subsubheading Create compressed versions of the release
6398 dejagnu/ dejagnu-gdb-5.2.tar.bz2 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6399 $ for m in gdb insight
6401 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6402 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6412 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6413 in that mode, @code{gzip} does not know the name of the file and, hence,
6414 can not include it in the compressed file. This is also why the release
6415 process runs @code{tar} and @code{bzip2} as separate passes.
6418 @subsection Sanity check the tar ball
6420 Pick a popular machine (Solaris/PPC?) and try the build on that.
6423 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6428 $ ./gdb/gdb ./gdb/gdb
6432 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6434 Starting program: /tmp/gdb-5.2/gdb/gdb
6436 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6437 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6439 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6443 @subsection Make a release candidate available
6445 If this is a release candidate then the only remaining steps are:
6449 Commit @file{version.in} and @file{ChangeLog}
6451 Tweak @file{version.in} (and @file{ChangeLog} to read
6452 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6453 process can restart.
6455 Make the release candidate available in
6456 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6458 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6459 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6462 @subsection Make a formal release available
6464 (And you thought all that was required was to post an e-mail.)
6466 @subsubheading Install on sware
6468 Copy the new files to both the release and the old release directory:
6471 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6472 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6476 Clean up the releases directory so that only the most recent releases
6477 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6480 $ cd ~ftp/pub/gdb/releases
6485 Update the file @file{README} and @file{.message} in the releases
6492 $ ln README .message
6495 @subsubheading Update the web pages.
6499 @item htdocs/download/ANNOUNCEMENT
6500 This file, which is posted as the official announcement, includes:
6503 General announcement
6505 News. If making an @var{M}.@var{N}.1 release, retain the news from
6506 earlier @var{M}.@var{N} release.
6511 @item htdocs/index.html
6512 @itemx htdocs/news/index.html
6513 @itemx htdocs/download/index.html
6514 These files include:
6517 announcement of the most recent release
6519 news entry (remember to update both the top level and the news directory).
6521 These pages also need to be regenerate using @code{index.sh}.
6523 @item download/onlinedocs/
6524 You need to find the magic command that is used to generate the online
6525 docs from the @file{.tar.bz2}. The best way is to look in the output
6526 from one of the nightly @code{cron} jobs and then just edit accordingly.
6530 $ ~/ss/update-web-docs \
6531 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6533 /www/sourceware/htdocs/gdb/download/onlinedocs \
6538 Just like the online documentation. Something like:
6541 $ /bin/sh ~/ss/update-web-ari \
6542 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6544 /www/sourceware/htdocs/gdb/download/ari \
6550 @subsubheading Shadow the pages onto gnu
6552 Something goes here.
6555 @subsubheading Install the @value{GDBN} tar ball on GNU
6557 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6558 @file{~ftp/gnu/gdb}.
6560 @subsubheading Make the @file{ANNOUNCEMENT}
6562 Post the @file{ANNOUNCEMENT} file you created above to:
6566 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6568 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6569 day or so to let things get out)
6571 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6576 The release is out but you're still not finished.
6578 @subsubheading Commit outstanding changes
6580 In particular you'll need to commit any changes to:
6584 @file{gdb/ChangeLog}
6586 @file{gdb/version.in}
6593 @subsubheading Tag the release
6598 $ d=`date -u +%Y-%m-%d`
6601 $ ( cd insight/src/gdb && cvs -f -q update )
6602 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6605 Insight is used since that contains more of the release than
6606 @value{GDBN} (@code{dejagnu} doesn't get tagged but I think we can live
6609 @subsubheading Mention the release on the trunk
6611 Just put something in the @file{ChangeLog} so that the trunk also
6612 indicates when the release was made.
6614 @subsubheading Restart @file{gdb/version.in}
6616 If @file{gdb/version.in} does not contain an ISO date such as
6617 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6618 committed all the release changes it can be set to
6619 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6620 is important - it affects the snapshot process).
6622 Don't forget the @file{ChangeLog}.
6624 @subsubheading Merge into trunk
6626 The files committed to the branch may also need changes merged into the
6629 @subsubheading Revise the release schedule
6631 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6632 Discussion List} with an updated announcement. The schedule can be
6633 generated by running:
6636 $ ~/ss/schedule `date +%s` schedule
6640 The first parameter is approximate date/time in seconds (from the epoch)
6641 of the most recent release.
6643 Also update the schedule @code{cronjob}.
6645 @section Post release
6647 Remove any @code{OBSOLETE} code.
6654 The testsuite is an important component of the @value{GDBN} package.
6655 While it is always worthwhile to encourage user testing, in practice
6656 this is rarely sufficient; users typically use only a small subset of
6657 the available commands, and it has proven all too common for a change
6658 to cause a significant regression that went unnoticed for some time.
6660 The @value{GDBN} testsuite uses the DejaGNU testing framework.
6661 DejaGNU is built using @code{Tcl} and @code{expect}. The tests
6662 themselves are calls to various @code{Tcl} procs; the framework runs all the
6663 procs and summarizes the passes and fails.
6665 @section Using the Testsuite
6667 @cindex running the test suite
6668 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6669 testsuite's objdir) and type @code{make check}. This just sets up some
6670 environment variables and invokes DejaGNU's @code{runtest} script. While
6671 the testsuite is running, you'll get mentions of which test file is in use,
6672 and a mention of any unexpected passes or fails. When the testsuite is
6673 finished, you'll get a summary that looks like this:
6678 # of expected passes 6016
6679 # of unexpected failures 58
6680 # of unexpected successes 5
6681 # of expected failures 183
6682 # of unresolved testcases 3
6683 # of untested testcases 5
6686 To run a specific test script, type:
6688 make check RUNTESTFLAGS='@var{tests}'
6690 where @var{tests} is a list of test script file names, separated by
6693 The ideal test run consists of expected passes only; however, reality
6694 conspires to keep us from this ideal. Unexpected failures indicate
6695 real problems, whether in @value{GDBN} or in the testsuite. Expected
6696 failures are still failures, but ones which have been decided are too
6697 hard to deal with at the time; for instance, a test case might work
6698 everywhere except on AIX, and there is no prospect of the AIX case
6699 being fixed in the near future. Expected failures should not be added
6700 lightly, since you may be masking serious bugs in @value{GDBN}.
6701 Unexpected successes are expected fails that are passing for some
6702 reason, while unresolved and untested cases often indicate some minor
6703 catastrophe, such as the compiler being unable to deal with a test
6706 When making any significant change to @value{GDBN}, you should run the
6707 testsuite before and after the change, to confirm that there are no
6708 regressions. Note that truly complete testing would require that you
6709 run the testsuite with all supported configurations and a variety of
6710 compilers; however this is more than really necessary. In many cases
6711 testing with a single configuration is sufficient. Other useful
6712 options are to test one big-endian (Sparc) and one little-endian (x86)
6713 host, a cross config with a builtin simulator (powerpc-eabi,
6714 mips-elf), or a 64-bit host (Alpha).
6716 If you add new functionality to @value{GDBN}, please consider adding
6717 tests for it as well; this way future @value{GDBN} hackers can detect
6718 and fix their changes that break the functionality you added.
6719 Similarly, if you fix a bug that was not previously reported as a test
6720 failure, please add a test case for it. Some cases are extremely
6721 difficult to test, such as code that handles host OS failures or bugs
6722 in particular versions of compilers, and it's OK not to try to write
6723 tests for all of those.
6725 DejaGNU supports separate build, host, and target machines. However,
6726 some @value{GDBN} test scripts do not work if the build machine and
6727 the host machine are not the same. In such an environment, these scripts
6728 will give a result of ``UNRESOLVED'', like this:
6731 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6734 @section Testsuite Organization
6736 @cindex test suite organization
6737 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6738 testsuite includes some makefiles and configury, these are very minimal,
6739 and used for little besides cleaning up, since the tests themselves
6740 handle the compilation of the programs that @value{GDBN} will run. The file
6741 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6742 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6743 configuration-specific files, typically used for special-purpose
6744 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6746 The tests themselves are to be found in @file{testsuite/gdb.*} and
6747 subdirectories of those. The names of the test files must always end
6748 with @file{.exp}. DejaGNU collects the test files by wildcarding
6749 in the test directories, so both subdirectories and individual files
6750 get chosen and run in alphabetical order.
6752 The following table lists the main types of subdirectories and what they
6753 are for. Since DejaGNU finds test files no matter where they are
6754 located, and since each test file sets up its own compilation and
6755 execution environment, this organization is simply for convenience and
6760 This is the base testsuite. The tests in it should apply to all
6761 configurations of @value{GDBN} (but generic native-only tests may live here).
6762 The test programs should be in the subset of C that is valid K&R,
6763 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6766 @item gdb.@var{lang}
6767 Language-specific tests for any language @var{lang} besides C. Examples are
6768 @file{gdb.cp} and @file{gdb.java}.
6770 @item gdb.@var{platform}
6771 Non-portable tests. The tests are specific to a specific configuration
6772 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6775 @item gdb.@var{compiler}
6776 Tests specific to a particular compiler. As of this writing (June
6777 1999), there aren't currently any groups of tests in this category that
6778 couldn't just as sensibly be made platform-specific, but one could
6779 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6782 @item gdb.@var{subsystem}
6783 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6784 instance, @file{gdb.disasm} exercises various disassemblers, while
6785 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6788 @section Writing Tests
6789 @cindex writing tests
6791 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6792 should be able to copy existing tests to handle new cases.
6794 You should try to use @code{gdb_test} whenever possible, since it
6795 includes cases to handle all the unexpected errors that might happen.
6796 However, it doesn't cost anything to add new test procedures; for
6797 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6798 calls @code{gdb_test} multiple times.
6800 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6801 necessary, such as when @value{GDBN} has several valid responses to a command.
6803 The source language programs do @emph{not} need to be in a consistent
6804 style. Since @value{GDBN} is used to debug programs written in many different
6805 styles, it's worth having a mix of styles in the testsuite; for
6806 instance, some @value{GDBN} bugs involving the display of source lines would
6807 never manifest themselves if the programs used GNU coding style
6814 Check the @file{README} file, it often has useful information that does not
6815 appear anywhere else in the directory.
6818 * Getting Started:: Getting started working on @value{GDBN}
6819 * Debugging GDB:: Debugging @value{GDBN} with itself
6822 @node Getting Started,,, Hints
6824 @section Getting Started
6826 @value{GDBN} is a large and complicated program, and if you first starting to
6827 work on it, it can be hard to know where to start. Fortunately, if you
6828 know how to go about it, there are ways to figure out what is going on.
6830 This manual, the @value{GDBN} Internals manual, has information which applies
6831 generally to many parts of @value{GDBN}.
6833 Information about particular functions or data structures are located in
6834 comments with those functions or data structures. If you run across a
6835 function or a global variable which does not have a comment correctly
6836 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6837 free to submit a bug report, with a suggested comment if you can figure
6838 out what the comment should say. If you find a comment which is
6839 actually wrong, be especially sure to report that.
6841 Comments explaining the function of macros defined in host, target, or
6842 native dependent files can be in several places. Sometimes they are
6843 repeated every place the macro is defined. Sometimes they are where the
6844 macro is used. Sometimes there is a header file which supplies a
6845 default definition of the macro, and the comment is there. This manual
6846 also documents all the available macros.
6847 @c (@pxref{Host Conditionals}, @pxref{Target
6848 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
6851 Start with the header files. Once you have some idea of how
6852 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
6853 @file{gdbtypes.h}), you will find it much easier to understand the
6854 code which uses and creates those symbol tables.
6856 You may wish to process the information you are getting somehow, to
6857 enhance your understanding of it. Summarize it, translate it to another
6858 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
6859 the code to predict what a test case would do and write the test case
6860 and verify your prediction, etc. If you are reading code and your eyes
6861 are starting to glaze over, this is a sign you need to use a more active
6864 Once you have a part of @value{GDBN} to start with, you can find more
6865 specifically the part you are looking for by stepping through each
6866 function with the @code{next} command. Do not use @code{step} or you
6867 will quickly get distracted; when the function you are stepping through
6868 calls another function try only to get a big-picture understanding
6869 (perhaps using the comment at the beginning of the function being
6870 called) of what it does. This way you can identify which of the
6871 functions being called by the function you are stepping through is the
6872 one which you are interested in. You may need to examine the data
6873 structures generated at each stage, with reference to the comments in
6874 the header files explaining what the data structures are supposed to
6877 Of course, this same technique can be used if you are just reading the
6878 code, rather than actually stepping through it. The same general
6879 principle applies---when the code you are looking at calls something
6880 else, just try to understand generally what the code being called does,
6881 rather than worrying about all its details.
6883 @cindex command implementation
6884 A good place to start when tracking down some particular area is with
6885 a command which invokes that feature. Suppose you want to know how
6886 single-stepping works. As a @value{GDBN} user, you know that the
6887 @code{step} command invokes single-stepping. The command is invoked
6888 via command tables (see @file{command.h}); by convention the function
6889 which actually performs the command is formed by taking the name of
6890 the command and adding @samp{_command}, or in the case of an
6891 @code{info} subcommand, @samp{_info}. For example, the @code{step}
6892 command invokes the @code{step_command} function and the @code{info
6893 display} command invokes @code{display_info}. When this convention is
6894 not followed, you might have to use @code{grep} or @kbd{M-x
6895 tags-search} in emacs, or run @value{GDBN} on itself and set a
6896 breakpoint in @code{execute_command}.
6898 @cindex @code{bug-gdb} mailing list
6899 If all of the above fail, it may be appropriate to ask for information
6900 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
6901 wondering if anyone could give me some tips about understanding
6902 @value{GDBN}''---if we had some magic secret we would put it in this manual.
6903 Suggestions for improving the manual are always welcome, of course.
6905 @node Debugging GDB,,,Hints
6907 @section Debugging @value{GDBN} with itself
6908 @cindex debugging @value{GDBN}
6910 If @value{GDBN} is limping on your machine, this is the preferred way to get it
6911 fully functional. Be warned that in some ancient Unix systems, like
6912 Ultrix 4.2, a program can't be running in one process while it is being
6913 debugged in another. Rather than typing the command @kbd{@w{./gdb
6914 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
6915 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
6917 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
6918 @file{.gdbinit} file that sets up some simple things to make debugging
6919 gdb easier. The @code{info} command, when executed without a subcommand
6920 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
6921 gdb. See @file{.gdbinit} for details.
6923 If you use emacs, you will probably want to do a @code{make TAGS} after
6924 you configure your distribution; this will put the machine dependent
6925 routines for your local machine where they will be accessed first by
6928 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
6929 have run @code{fixincludes} if you are compiling with gcc.
6931 @section Submitting Patches
6933 @cindex submitting patches
6934 Thanks for thinking of offering your changes back to the community of
6935 @value{GDBN} users. In general we like to get well designed enhancements.
6936 Thanks also for checking in advance about the best way to transfer the
6939 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
6940 This manual summarizes what we believe to be clean design for @value{GDBN}.
6942 If the maintainers don't have time to put the patch in when it arrives,
6943 or if there is any question about a patch, it goes into a large queue
6944 with everyone else's patches and bug reports.
6946 @cindex legal papers for code contributions
6947 The legal issue is that to incorporate substantial changes requires a
6948 copyright assignment from you and/or your employer, granting ownership
6949 of the changes to the Free Software Foundation. You can get the
6950 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
6951 and asking for it. We recommend that people write in "All programs
6952 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
6953 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
6955 contributed with only one piece of legalese pushed through the
6956 bureaucracy and filed with the FSF. We can't start merging changes until
6957 this paperwork is received by the FSF (their rules, which we follow
6958 since we maintain it for them).
6960 Technically, the easiest way to receive changes is to receive each
6961 feature as a small context diff or unidiff, suitable for @code{patch}.
6962 Each message sent to me should include the changes to C code and
6963 header files for a single feature, plus @file{ChangeLog} entries for
6964 each directory where files were modified, and diffs for any changes
6965 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
6966 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
6967 single feature, they can be split down into multiple messages.
6969 In this way, if we read and like the feature, we can add it to the
6970 sources with a single patch command, do some testing, and check it in.
6971 If you leave out the @file{ChangeLog}, we have to write one. If you leave
6972 out the doc, we have to puzzle out what needs documenting. Etc., etc.
6974 The reason to send each change in a separate message is that we will not
6975 install some of the changes. They'll be returned to you with questions
6976 or comments. If we're doing our job correctly, the message back to you
6977 will say what you have to fix in order to make the change acceptable.
6978 The reason to have separate messages for separate features is so that
6979 the acceptable changes can be installed while one or more changes are
6980 being reworked. If multiple features are sent in a single message, we
6981 tend to not put in the effort to sort out the acceptable changes from
6982 the unacceptable, so none of the features get installed until all are
6985 If this sounds painful or authoritarian, well, it is. But we get a lot
6986 of bug reports and a lot of patches, and many of them don't get
6987 installed because we don't have the time to finish the job that the bug
6988 reporter or the contributor could have done. Patches that arrive
6989 complete, working, and well designed, tend to get installed on the day
6990 they arrive. The others go into a queue and get installed as time
6991 permits, which, since the maintainers have many demands to meet, may not
6992 be for quite some time.
6994 Please send patches directly to
6995 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
6997 @section Obsolete Conditionals
6998 @cindex obsolete code
7000 Fragments of old code in @value{GDBN} sometimes reference or set the following
7001 configuration macros. They should not be used by new code, and old uses
7002 should be removed as those parts of the debugger are otherwise touched.
7005 @item STACK_END_ADDR
7006 This macro used to define where the end of the stack appeared, for use
7007 in interpreting core file formats that don't record this address in the
7008 core file itself. This information is now configured in BFD, and @value{GDBN}
7009 gets the info portably from there. The values in @value{GDBN}'s configuration
7010 files should be moved into BFD configuration files (if needed there),
7011 and deleted from all of @value{GDBN}'s config files.
7013 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
7014 is so old that it has never been converted to use BFD. Now that's old!
7018 @include observer.texi